the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 15 Mar 2023
FASDD: An Open-access 100,000-level Flame and Smoke Detection Dataset for Deep Learning in Fire Detection
Abstract. With the advancement of computer vision, artificial intelligence, and remote sensing technologies, deep learning algorithms are increasingly used in terrestrial, airborne, and spaceborne-based fire detection systems. The performance and generalization of these data-driven fire detection algorithms, however, are restricted by the limited number and source of fire detection datasets. A large-scale fire detection benchmark dataset covering complex and varied fire scenarios is urgently needed. This work constructs a 100,000-level Flame and Smoke Detection Dataset (FASDD) based on multi-source heterogeneous flame and smoke images. It holds rich variations in image size, resolution, illumination (day and night), scenario (indoor and outdoor), image range (far and near), viewing angle (top view and side view), platform (surveillance cameras, drones, and satellites), and data source (Internet, social media, and open-access fire datasets). To the best of our knowledge, FASDD is currently the most versatile and comprehensive dataset for fire detection. It provides a challenging benchmark to drive the continuous evolution of fire detection models. Additionally, we formulate a unified workflow for preprocessing, annotation, and quality control of fire samples. Out-of-the-box annotations are published in four different formats for training deep learning models. Extensive performance evaluations based on classical methods show that most of the object detection models trained on FASDD can achieve satisfactory fire detection results, and especially YOLOv5x achieves nearly 80 % mAP@0.5 accuracy on heterogeneous images spanning two domains of computer vision and remote sensing. And the application in wildfire location demonstrates that deep learning models trained on our dataset can be used in recognizing and monitoring forest fires. Deep learning models trained with FASDD can be simultaneously deployed on satellites, drones, and ground sensors, thus realizing collaborative fire observation and detection in a space-air-ground integrated network environment. The dataset is available from the Science Data Bank website at https://doi.org/10.57760/sciencedb.j00104.00103 (Wang et al., 2022a).
This preprint has been withdrawn.
-
Withdrawal notice
This preprint has been withdrawn.
Please read the editorial note first before accessing the preprint.
-
Preprint
(7002 KB)
-
This preprint has been withdrawn.
Please read the editorial note first before accessing the preprint.
- Preprint
(7002 KB) - Metadata XML
- BibTeX
- EndNote
Interactive discussion
Status: closed
-
CC1: 'Comment on essd-2023-73', w jr, 29 Mar 2023
We very much look forward to your data set being helpful to our research.
Citation: https://doi.org/10.5194/essd-2023-73-CC1 -
AC1: 'Reply on CC1', Ming Wang, 29 Mar 2023
Thank you very much for your interest and comment. As soon as our paper is published, we will make this dataset publicly available.
Hope our dataset could contribute to your research.
Citation: https://doi.org/10.5194/essd-2023-73-AC1
-
AC1: 'Reply on CC1', Ming Wang, 29 Mar 2023
-
RC1: 'Comment on essd-2023-73', Anonymous Referee #1, 17 Apr 2023
Summary
Fire has wide impacts on Earth Systems and human society, and efficient fire detection could promote better understanding, modeling, and preventing fires. Wang et al., synthesized a comprehensive dataset (FASDD) covering images from terrestrial, airborne, and spaceborne sensors. For the previous version, I had nine major concerns about the data generation, data validation, and the usefulness of the data, but the current responses and version can hardly convince me for the majority of my major concerns. Here I reclaim my major concerns unaddressed and why they are critically important for fire detection using spaceborne satellite. I suggest the authors to handle those major concerns, otherwise, I really apologize and have to reject this work since these are critical or fundamental drawbacks for this study.
Major comments
(1) Data generation: it’s well known that near infrared (NIR) and short-wave infrared (SWIR) are two commonly used bands for fire detection while the authors only used the visible bands via visual interpretation. If only based on true colors, it can hardly convince me about the generality of the dataset for large spatial scale fire detection (e.g., some land surface items could show similar colors with fires in remote sensing images and thus mislead machine leaning models). Also, the results in Table 3 showed the really low performance on detecting remote sensing fires even with advanced machine learning models. With such a lower accuracy, how can the data help improve fire detection.
My response to the authors’ responses: I understand some previous works have used optical camera or RGB for smoke detection, but those purely RGB based fire detection works are used for terrestrial or near-surface fire monitoring instead of spaceborne satellite. The FASDD_RS data include sentinel-2 and landsat8, and the NIR or SWIR bands are fundamentally important for monitoring fire-induced vegetation and dryness changes, and are basically used for fire detection. For example, the listed works bellow all used bands like NRI or SWIR. Some basic indexes with NIR and SWIR can achieve substantially higher accuracy (e.g., Castillo et al., 2020) than the reported accuracy by this study. Also the deep learning work in Pereira et al.2021, achieved much higher accuracy than your work. I have never seen a published reliable fire dataset derived from satellite observations only used RGB.
Barboza Castillo, E., Turpo Cayo, E. Y., de Almeida, C. M., Salas López, R., Rojas Briceño, N. B., Silva López, J. O., ... & Espinoza-Villar, R. (2020). Monitoring wildfires in the northeastern peruvian amazon using landsat-8 and sentinel-2 imagery in the GEE platform. ISPRS International Journal of Geo-Information, 9(10), 564.
de Almeida Pereira, G. H., Fusioka, A. M., Nassu, B. T., & Minetto, R. (2021). Active fire detection in Landsat-8 imagery: A large-scale dataset and a deep-learning study. ISPRS Journal of Photogrammetry and Remote Sensing, 178, 171-186.
Hu, X., Ban, Y., & Nascetti, A. (2021). Sentinel-2 MSI data for active fire detection in major fire-prone biomes: A multi-criteria approach. International Journal of Applied Earth Observation and Geoinformation, 101, 102347.
(2) Data generation: for active fire detection, middle infrared and thermal bands are also
important but were ignored.My response to the authors’ responses: Actually, landsat8 has thermal infrared bands. My concern is that why only use RGB for fire detection? The ultimate goal of this dataset is for improving fire detection instead of some technique details like only using one machine learning model with the same-format inputting dataset. If fusing different datasets can improve fire detection, why not use comprehensive models (e.g., different models) to fuse different datasets and finally achieve better fire detection (e.g., based on an ensemble of machine learning models)? If you use this dataset to develop one single model but throwing important bands information, I don’t think it is a good idea.
(3) Data generation: for the fire detection, the authors used the top-of-atmosphere reflectance instead of atmospheric-corrected land surface reflectance which should be a problem. For example, if the smoke is white or grey, how to classify smoke versus clouds only through visual interpretation of true-colors? Meanwhile, the CV fire images should be obtained on land surface with a much higher spatial resolution (can be with sub-meter resolution in Fig. 4). Can such kinds of CV image trained machine learning models be directly used to large-scale remote sensing data without atmospheric-correction and with different spatial resolutions?
My response to the authors’ responses: I understand that FASDD_CV have images from UAVs. But from the results in Table 4, it seems that for FASDD_RS, the model performance even dropped if we used FASDD (including FASDD_RS and FASDD_CV) for training the model. The dropped performance means that the knowledge derived from FASDD_CV can hardly be transferred or used for improving FASDD_RS. If so, why we should integrate those two heterogenous datasets?
(4) Data annotation: the “minimum bounding rectangle” was used to label the images. Commonly, fire detection is to classify whether each pixel is burned or not instead of a rectangle (e.g., the MODIS fire product in Giglio et al., (2018) and Giglio et al., (2016)). Meanwhile, the fire patch perimeter was always not rectangle (Laurent et al., 2018), therefore a bounding rectangle could contain both burned and unburned pixels, right?
My response to the authors’ responses: Please clarify the differences for fire detection in terms of scene classification, object detection, and semantic segmentation; clarify the main focus of this dataset (e.g., objection detection tasks) and its differences with previous fire detection datasets from satellites in the main text. Currently, section 2.1 only mentioned about existing fire detection datasets for terrestrial and airborne but not including existing satellite fire detection datasets.
(5) Data validation: the true-color based visual interpretation could also involve biases, therefore it’s important to validate the generated data against other reliable fire dataset, such as the MTBS data
My response to the authors’ responses: it can hardly convince me by validating the dataset only using one specific fire event in MTBS. Additionally, for Figure 6, why not use metrics to quantitively validate the dataset? Please comprehensively validate your dataset. If the dataset is not reliable, how can we trust it and use it?
(6) Line 245-246, the dataset consists of 95,314 computer vision fire samples but only 5,773 remote sensing samples. Due to the data imbalance, the model performance (Table 2) on FASDD data therefore mainly depends on the model performance on FASDD_CV. For the limited number of remote sensing fire samples, most samples in each region were distributed within ten days of a specific year (Table 1). Can such limited number of wildfires reflect all the seasonal and interannual changes of environmental conditions and fire dynamics so that machine learning models could learn from enough data? To my knowledge, the fire occurrence conditions changed across seasons and therefore the fire detectability could also be affected.
My response to the authors’ responses: a) can CV dataset help improve fire detection using RS (Table 4)? b) I disagree that “the features of fire objects do not have a high seasonal dependence” responded by the authors. The fire features including fire radiative power or fire intensity, fire spread, fire duration, and fire size, are strongly controlled by environmental conditions which could vary across seasons (e.g., dry season versus non-dry seasons) or within a season.
(7) Evaluation in section 4: the evaluation mainly showed the extent to which machine learning models could detect fires of generated FASDD data, therefore whether FASDD is reliable remains unknown. The FASDD data needed to be validated against other reliable fire products.
My response to the authors’ responses: please see my comment in (5).
(8) There are many existing remote sensing fire products (e.g., MTBS, MODIS fire products) and CV fire data sets. I understand that combining the two kinds of dataset was the main difference of FASDD relative to other datasets, however, the authors did not show why combining these two kinds of dataset is important? Can combining these two kinds of datasets improve fire detection relative to existing fire detection algorithms (e.g., the method for MODIS or MTBS fire product)? If not, why people use such complex dataset (with different spatial resolution and without atmospheric-correction)
My response to the authors’ responses: For table 4, the accuracy on FASDD_RS is really lower than existing fire detection algorithms (see my major comment 1), and it seems that integrating those two datasets can achieve very limited benefit or even performance loss for three of total four machine learning methods. If so, why people use such complex dataset? Additionally, the MTBS is based on Landsat, not necessarily having a lower spatial resolution than your RS dataset.
(9) Line 365-370: chaning "epoch" is not ablation experiments. Ablation experiments commonly refer to changing a component of machine learning model (i.e., model structure changes)
Citation: https://doi.org/10.5194/essd-2023-73-RC1 -
AC2: 'Reply on RC1', Ming Wang, 17 Apr 2023
Thank you very much for your comments and suggestions. We insist that our datasets, in particular FASDD_CV, contribute to fire detection research in the field of computer vision. Thus, we decide to withdraw this submission and reconstruct the remote sensing datasets by considering NIR and SWIR bands, with the hope of achieving better fire detection performance.
Thank you again for your comments and help.
Citation: https://doi.org/10.5194/essd-2023-73-AC2 -
AC3: 'Reply on RC1', Ming Wang, 17 Apr 2023
Thank you very much for your comments and suggestions. We insist that our datasets, in particular FASDD_CV, contribute to fire detection research in the field of computer vision. Thus, we decide to withdraw this submission and reconstruct the remote sensing datasets by considering NIR and SWIR bands, with the hope of achieving better fire detection performance.
Thank you again for your comments and help.
Citation: https://doi.org/10.5194/essd-2023-73-AC3 -
AC4: 'Reply on RC1', Ming Wang, 18 Apr 2023
We would like to make a clarification that after careful evaluation, we should be able to add additional experiments before the deadline to address your concerns based on your suggestions. Thank you again for your suggestions.
Citation: https://doi.org/10.5194/essd-2023-73-AC4 - AC5: 'Reply on RC1', Ming Wang, 29 May 2023
-
AC17: 'Reply on RC1', Ming Wang, 24 Sep 2023
We sincerely appreciate your valuable comments and suggestions. Your insights have played a crucial role in improving the quality of both our dataset and the manuscript. In response to all your concerns, we conducted a significant number of experiments and made corresponding revisions to the manuscript.
Please refer to the attached document for specific responses and revisions.
-
AC2: 'Reply on RC1', Ming Wang, 17 Apr 2023
-
CC2: 'Request for Dataset for my Thesis on Fire Detection', Ebrahim Akbari, 26 May 2023
Dear Team,
I am writing to kindly request access to the FASDD dataset for my master's thesis on fire detection. I have come across your dataset paper titled "FASDD: An Open-access 100,000-level Flame and Smoke Detection Dataset for Deep Learning in Fire Detection" and I believe it would greatly contribute to the research I am conducting.
I am particularly interested in utilizing the annotated images from the dataset to train and evaluate deep learning models for fire detection. The comprehensive nature of the FASDD dataset, along with its large-scale and open-access features, align perfectly with the objectives of my thesis.
I understand that the dataset paper is currently under review and open for public discussion in the Earth System Science Data (ESSD) journal. I would be more than willing to provide valuable feedback and actively participate in the discussion to support the publication process and promote the sharing of this dataset with the research community.
I kindly request access to a portion of the annotated dataset, approximately 5000 images, to facilitate the progress of my thesis research. I assure you that the dataset will be used solely for academic purposes and in accordance with any terms and conditions set by your platform.
Thank you very much for considering my request. I greatly appreciate your support and contribution to the field of fire detection. Please let me know if there are any further steps or requirements I need to fulfill in order to obtain access to the FASDD dataset.
I look forward to your positive response.
Sincerely,
Ebrahim Akbari
A master's student in Telecommunications Engineering at Quchan University of Technology
ebrahimakbari118@gmail.com
+989153858085
Citation: https://doi.org/10.5194/essd-2023-73-CC2 -
CC3: 'Reply on CC2', Liangcun Jiang, 27 May 2023
Dear Ebrahim Akbari,
Thank you for your interest in our dataset. We are thrilled that it can contribute to your research.
The FASDD dataset is specifically designed for training and evaluating deep learning models in fire detection, with over 120,000 diverse flame and smoke images captured in various scenarios. Our team has meticulously annotated these images through manual visual annotation and multiple quality checks. The annotations aim to improve the generalization performance of deep learning models and provide a benchmark for advanced fire detection models.We have taken note of your data requirements and will reach out to you via email soon to provide a portion of the dataset for your research purposes.
Once again, we appreciate your interest, support, and acknowledgment of our work. We hope that our dataset proves valuable to your research and contributes to the field of fire detection.
Citation: https://doi.org/10.5194/essd-2023-73-CC3
-
CC3: 'Reply on CC2', Liangcun Jiang, 27 May 2023
-
CC4: 'Comment on essd-2023-73', Rakesh Kumar Mallik, 29 May 2023
Hi This is Rakesh.
I am a AI/Machine learning ensthusiastic person, who has worked in couple of corporates as data sceintist and ML engineer.
I came across this valubale data set of Fire and Somke detection. I beleive this dataset can be used for designing a robust work place safetly system. Not only that based on the variance in the data distribution, it can be used for industrial safety and precaution, can detect roadside fire etc.. I see small small openset dataset on Fire and smoke many places here and there, but None has gobne to this lenght of creating of 1L images with segmenattion mask. Even its just for my pure educational intrest but i would love to get a hange of this dataset and explore more what this can offer.
Thank you
Citation: https://doi.org/10.5194/essd-2023-73-CC4 -
AC6: 'Reply on CC4', Ming Wang, 30 May 2023
Dear Rakesh,
Thank you for expressing your interest in our dataset. The FASDD dataset has been specifically curated as an object detection dataset for training and evaluating deep learning models in fire detection. It consists of over 120,000 distinct images capturing flames and smoke in various scenarios. Our team has meticulously annotated these images through manual visual inspection and multiple quality checks. These annotations aim to improve the generalization performance of deep learning models and provide a benchmark for advanced fire detection systems.
We appreciate your data request, and we will be contacting you shortly via email to provide a portion of the dataset for your research purposes. It's important to note that the current dataset we have prepared primarily focuses on object detection tasks. However, with additional efforts, you can generate segmentation masks based on the existing annotations to suit your specific applications. In fact, we are actively working on creating a separate dataset specifically tailored for semantic segmentation tasks, which will incorporate segmentation masks of flames and smoke.
Once again, we sincerely appreciate your interest, support, and recognition of our work. We hope that our dataset proves valuable for your research endeavors and contributes to the field of fire detection.
Citation: https://doi.org/10.5194/essd-2023-73-AC6
-
AC6: 'Reply on CC4', Ming Wang, 30 May 2023
-
CC5: 'Comment on essd-2023-73', Stephan Sturges, 21 Jun 2023
This dataset would be a very welcome addition to the field. There is a global explosion in capabilities when it comes to overhead sensing whether from satellites, drones, blimps, planes etc... And there a real need for datasets like this to allow detection capabilites to be created for these sensing modalities.
Citation: https://doi.org/10.5194/essd-2023-73-CC5 -
AC7: 'Reply on CC5', Ming Wang, 23 Jun 2023
Dear Stephan Sturges,
Thank you for your valuable feedback on our dataset. We greatly appreciate your recognition of the significance of our contribution to the fire detection field. We completely agree that there is currently a global surge in capabilities for overhead sensing, including data collected from satellites, drones, blimps, planes, and other platforms.
Your comment highlights the real need for comprehensive datasets such as ours, as they play a crucial role in enabling the development of detection capabilities for these various sensing modalities. By providing access to this dataset, we aim to facilitate research and innovation in the fire detection field by offering a valuable resource for researchers and practitioners.
Once again, we sincerely thank you for acknowledging the importance of our dataset, and we hope that it proves to be a valuable asset for future advancements in fire detection.
Best regards,
The authors
Citation: https://doi.org/10.5194/essd-2023-73-AC7
-
AC7: 'Reply on CC5', Ming Wang, 23 Jun 2023
-
CC6: 'Comment on essd-2023-73', Rafik Ghali, 27 Jun 2023
Dear sir,
I hope this email finds you well.
I am writing to you as a researcher interested in your recent publication titled "FASDD: An Open-access 100,000-level Flame And Smoke Detection Dataset for Deep Learning in Fire Detection". I have read your work with great interest and found the dataset used in your study to be of significant relevance to my own research objectives.
Thank you for your attention to this matter. I look forward to your positive response. Please do not hesitate to contact me if you require any further information or have any questions. I truly appreciate your time and consideration.
Best regards,
Citation: https://doi.org/10.5194/essd-2023-73-CC6 -
AC8: 'Reply on CC6', Ming Wang, 03 Jul 2023
Dear Rafik Ghali,
Thank you for reaching out and expressing your interest in our recent publication titled "FASDD: An Open-access 100,000-level Flame And Smoke Detection Dataset for Deep Learning in Fire Detection." We are glad to hear that you found our dataset to be of significant relevance to your own research objectives.
After thorough discussions among the authors, we have decided to share a portion of the dataset with interested researchers like yourself. We believe that collaboration and knowledge sharing are crucial for advancing research in the field of fire detection. We appreciate your enthusiasm and dedication to furthering the scientific community's understanding in this area.
Thank you once again for your interest in our work.
Best regards,
The authors
Citation: https://doi.org/10.5194/essd-2023-73-AC8
-
AC8: 'Reply on CC6', Ming Wang, 03 Jul 2023
-
CC7: 'dataset access', dipesh digwal, 03 Jul 2023
please can you please provide me fire dataset as it is to be used for a fire safety device being constructed by my company
Citation: https://doi.org/10.5194/essd-2023-73-CC7 -
AC9: 'Reply on CC7', Ming Wang, 03 Jul 2023
Dear Dipesh Digwal,
Thank you for your email and your interest in our fire dataset. We appreciate your recognition of the value and contribution of our dataset.
After careful evaluation, the authors have decided to share a portion of the dataset with you to support your project. However, please note that the dataset is subject to certain conditions and restrictions, which we will outline below:
Intended Use: The dataset is intended to be used for fire safety purposes only. It should not be utilized for any other purposes without obtaining proper permissions or rights.
Attribution: It is important to acknowledge the source of the dataset appropriately. Please ensure that you credit the dataset by mentioning the authors and providing a reference to the original publication or source.
Confidentiality and Security: The shared dataset should be treated with utmost confidentiality and stored securely. It must not be shared with any third parties without prior consent from our team.
Compliance with Legal and Ethical Guidelines: The usage of the dataset must comply with all applicable laws, regulations, and ethical guidelines. It should not infringe upon the rights of individuals or organizations, and any potential risks or biases should be carefully considered and addressed.
We appreciate your understanding and cooperation in this matter. We look forward to supporting your efforts in fire safety device construction.
Best regards,
The authors
Citation: https://doi.org/10.5194/essd-2023-73-AC9
-
AC9: 'Reply on CC7', Ming Wang, 03 Jul 2023
-
CC8: 'Comment on essd-2023-73', Rafik Ghali, 05 Jul 2023
Dear Sir
I hope this e-mail finds you well. I am writing to ask you to kindly provide me with a link to the FASDD_RS satellite image dataset.The images available in the FASDD_RS database are renowned for their quality and relevance to my research area.I would be very grateful if you could provide me with access or information on the access procedure. Access to the FASDD_RS database will contribute significantly to the success of my research.Thank you in advance for your help and prompt reply. Please let me know if there are any other requirements or procedures to follow.Best regards,Citation: https://doi.org/10.5194/essd-2023-73-CC8 -
AC10: 'Reply on CC8', Ming Wang, 06 Jul 2023
Dear Rafik Ghali,
I hope this email finds you well. Thank you for expressing your interest in the FASDD_RS satellite image database. Unfortunately, at this stage, our database is still undergoing the peer-review process for the associated paper, and we are unable to publicly release all of the data. However, we understand the importance of access to relevant imagery for your research.
In consideration of your needs, we are able to provide you with a portion of the annotated fire and smoke images from our database. These images can be valuable assets for your research. We believe that this subset will still contribute significantly to your studies while we continue the review process for the full dataset.
Thank you once again for your interest and recognition.
Best wishes,
The authors
Citation: https://doi.org/10.5194/essd-2023-73-AC10 -
AC11: 'Reply on CC8', Ming Wang, 06 Jul 2023
Dear Rafik Ghali,
I hope this email finds you well. Thank you for expressing your interest in the FASDD_RS satellite image database. Unfortunately, at this stage, our database is still undergoing the peer-review process for the associated paper, and we are unable to publicly release all of the data. However, we understand the importance of access to relevant imagery for your research.
In consideration of your needs, we are able to provide you with a portion of the annotated fire and smoke images from our database. These images can be valuable assets for your research. We believe that this subset will still contribute significantly to your studies while we continue the review process for the full dataset.
Thank you once again for your interest and recognition.
Best wishes,
The authors
Citation: https://doi.org/10.5194/essd-2023-73-AC11
-
AC10: 'Reply on CC8', Ming Wang, 06 Jul 2023
-
CC9: 'Comment on essd-2023-73', lin jing, 08 Jul 2023
Hello, I am training a model to detect fire smoke, I believe your dataset will help me a lot, looking forward to your reply!
Citation: https://doi.org/10.5194/essd-2023-73-CC9 -
AC12: 'Reply on CC9', Ming Wang, 08 Jul 2023
Dear Lin,
Thank you for your interest in our dataset (FASDD). Having reliable data is crucial for improving the accuracy and effectiveness of deep learning models. We are delighted to hear that you believe our dataset will be beneficial for your model training.
However, our dataset is currently not available in its entirety due to ongoing research paper review process. Nevertheless, we are willing to share a partial subset of the dataset with you. This will enable you to make progress in your training process while respecting the limitations imposed by the review process. We appreciate your understanding regarding the partial sharing of the dataset. We believe that even with this subset, you will be able to make valuable advancements in your work. We will reach out to you shortly to discuss the details of sharing the subset of the dataset.
Once again, we sincerely appreciate your interest and recognition of FASDD dataset. We wish you every success in your model training.
Best regards,
The authors
Citation: https://doi.org/10.5194/essd-2023-73-AC12
-
AC12: 'Reply on CC9', Ming Wang, 08 Jul 2023
-
CC10: 'Comment on essd-2023-73', sumanth reddy, 18 Jul 2023
Dear Ming Wang,
I would like to extend my heartfelt appreciation to the creators and contributors of the FASSD dataset. As a researcher in the field of [mention your research area], I have had the privilege of utilizing this dataset in my recent studies, and I am truly impressed by its quality and scope.
The FASSD dataset has been an invaluable asset to the research community, providing a diverse range of data that is essential for advancing our understanding of [mention the specific research area or application]. The attention to detail and the meticulous curation of the dataset have been evident in every aspect, enabling researchers like myself to derive meaningful insights and produce reliable results.
Furthermore, the seamless accessibility and proper documentation of the FASSD dataset have streamlined my research process significantly. I commend the team behind this initiative for their dedication to promoting transparency and reproducibility in the scientific community.
The impact of the FASSD dataset on the research landscape cannot be overstated. It has not only facilitated my own research but has also inspired new research directions and collaborations within the community. The dataset's versatility and comprehensive coverage have sparked innovative ideas that promise to push the boundaries of our field even further.
In conclusion, I want to express my deep gratitude to the individuals and organizations responsible for the creation and maintenance of the FASSD dataset. Your contribution has empowered researchers worldwide and accelerated the pace of scientific discovery. I look forward to witnessing the continued growth and impact of this remarkable resource.
Thank you once again for this outstanding dataset, and I eagerly await further advancements in the field, driven by the remarkable work of Ming Wang.
Citation: https://doi.org/10.5194/essd-2023-73-CC10 -
AC13: 'Reply on CC10', Ming Wang, 18 Jul 2023
Dear Sumanth Reddy,
Thank you for your comment expressing your heartfelt appreciation for the creators and contributors of the FASSD dataset. We are delighted to hear that you have found immense value in utilizing the dataset in your recent studies.
As a fellow researcher in the field of fire detection, we share your admiration for the quality and scope of the FASSD dataset. Its diverse range of data has undoubtedly been instrumental in advancing our understanding of fire. The attention to detail and meticulous curation evident in every aspect of the dataset have allowed researchers like us to derive meaningful insights and achieve reliable results.
The impact of the FASSD dataset on the research landscape is indeed profound. Not only has it facilitated your own research, but it has also inspired new research directions and collaborations within the community. The dataset's versatility and comprehensive coverage have sparked innovative ideas that hold great potential to push the boundaries of our field even further.
Thank you once again for recognizing the outstanding value of the dataset and for your kind words. We join you in eagerly awaiting further advancements in the field, driven by some remarkable work.
Best regards,
The authors
Citation: https://doi.org/10.5194/essd-2023-73-AC13
-
AC13: 'Reply on CC10', Ming Wang, 18 Jul 2023
-
CC11: 'Comment on essd-2023-73', Shokhruz Kak, 30 Jul 2023
Dear Authors,
I hope this email finds you well. I am writing to request access to the Flame And Smoke Detection Dataset (FASDD) mentioned in your paper titled "FASDD: An Open-access 100,000-level Flame And Smoke Detection Dataset for Deep Learning in Fire Detection".
he FASDD: An Open-access 100,000-level Flame And Smoke Detection Dataset for Deep Learning in Fire Detection is a remarkable contribution to the field of fire detection in several ways. Firstly, it is accessible dataset that is necessary for training deep learning models to accurately detect fire in different situations. This can significantly improve fire management and save numerous lives. Secondly, this dataset opens up new possibilities for research in the field of deep learning, which could lead to the development of innovative approaches and solutions for fire detection. Overall, the FASDD is an excellent resource that offers a variety of benefits and opportunities to researchers, practitioners, and firefighters, and its availability will undoubtedly bring about significant progress in the field of fire detection and management.
I am particularly interested in your dataset for my research paper in machine learning to detect wildfires. Access to the FASDD dataset would greatly contribute to my research by enabling a comparative analysis of different fire detection methodologies and enhancing the accuracy and robustness of my proposed model.
I kindly request your assistance in providing me with access to the FASDD dataset, or any alternative means to obtain the necessary data. I assure you that the dataset will be used solely for research purposes and in accordance with the terms and conditions set by the dataset's creators. Any data shared with me will be treated with the utmost confidentiality and will not be shared with any third parties.Should you require any further information or have any concerns, please do not hesitate to contact me. Thank you for your time and attention.SincerelyCitation: https://doi.org/10.5194/essd-2023-73-CC11 -
AC14: 'Reply on CC11', Ming Wang, 30 Jul 2023
Dear Shokhruz Kak,
I hope this email finds you well. Thank you for expressing your interest in our Flame And Smoke Detection Dataset (FASDD) and for your kind words about its potential contributions to the field of fire detection.
However, at the moment, our dataset is not fully available for external use due to an ongoing research paper review process. We have received a significant number of requests from researchers like yourself, which has contributed to the evaluation of the dataset's potential impact and robustness.
Nonetheless, we understand the importance of collaboration and knowledge sharing, and we are more than willing to support your research efforts. As such, we can offer you access to a partial subset of the FASDD. While it may not encompass the entire dataset, we believe it can still provide valuable insights and contribute to the comparative analysis you intend to conduct. We will reach out to you shortly to discuss the details of sharing the subset of the dataset.
Thank you for your understanding and cooperation. We look forward to assisting you in your research endeavors.
Best regards,
The authors
Citation: https://doi.org/10.5194/essd-2023-73-AC14
-
AC14: 'Reply on CC11', Ming Wang, 30 Jul 2023
-
CC12: 'Comment on essd-2023-73', hossein rajoli, 12 Aug 2023
Being a part of an engaged research team focused on wildfire detection and management, obtaining access to the dataset would be immensely appreciated. Such access has the potential to significantly advance my research efforts in this domain.
Bests.
Citation: https://doi.org/10.5194/essd-2023-73-CC12 -
AC15: 'Reply on CC12', Ming Wang, 12 Aug 2023
Dear hossein rajoli,
Thank you for expressing your interest in our Flame And Smoke Detection Dataset (FASDD).
We will reach out to you shortly to discuss the details of sharing the dataset.
Thank you once again for your interest and recognition of FASDD dataset. We hope that the dataset will be helpful to you in your research.
Best regards,
The authors
Citation: https://doi.org/10.5194/essd-2023-73-AC15
-
AC15: 'Reply on CC12', Ming Wang, 12 Aug 2023
-
RC2: 'Comment on essd-2023-73', Anonymous Referee #2, 23 Aug 2023
The comment was uploaded in the form of a supplement: https://essd.copernicus.org/preprints/essd-2023-73/essd-2023-73-RC2-supplement.pdf
-
AC16: 'Reply on RC2', Ming Wang, 24 Sep 2023
We sincerely appreciate your valuable comments and suggestions. In response to your concerns, we have conducted extensive experiments and made corresponding revisions to the manuscript.
Please refer to the attached document for specific responses and revisions.
-
AC16: 'Reply on RC2', Ming Wang, 24 Sep 2023
Interactive discussion
Status: closed
-
CC1: 'Comment on essd-2023-73', w jr, 29 Mar 2023
We very much look forward to your data set being helpful to our research.
Citation: https://doi.org/10.5194/essd-2023-73-CC1 -
AC1: 'Reply on CC1', Ming Wang, 29 Mar 2023
Thank you very much for your interest and comment. As soon as our paper is published, we will make this dataset publicly available.
Hope our dataset could contribute to your research.
Citation: https://doi.org/10.5194/essd-2023-73-AC1
-
AC1: 'Reply on CC1', Ming Wang, 29 Mar 2023
-
RC1: 'Comment on essd-2023-73', Anonymous Referee #1, 17 Apr 2023
Summary
Fire has wide impacts on Earth Systems and human society, and efficient fire detection could promote better understanding, modeling, and preventing fires. Wang et al., synthesized a comprehensive dataset (FASDD) covering images from terrestrial, airborne, and spaceborne sensors. For the previous version, I had nine major concerns about the data generation, data validation, and the usefulness of the data, but the current responses and version can hardly convince me for the majority of my major concerns. Here I reclaim my major concerns unaddressed and why they are critically important for fire detection using spaceborne satellite. I suggest the authors to handle those major concerns, otherwise, I really apologize and have to reject this work since these are critical or fundamental drawbacks for this study.
Major comments
(1) Data generation: it’s well known that near infrared (NIR) and short-wave infrared (SWIR) are two commonly used bands for fire detection while the authors only used the visible bands via visual interpretation. If only based on true colors, it can hardly convince me about the generality of the dataset for large spatial scale fire detection (e.g., some land surface items could show similar colors with fires in remote sensing images and thus mislead machine leaning models). Also, the results in Table 3 showed the really low performance on detecting remote sensing fires even with advanced machine learning models. With such a lower accuracy, how can the data help improve fire detection.
My response to the authors’ responses: I understand some previous works have used optical camera or RGB for smoke detection, but those purely RGB based fire detection works are used for terrestrial or near-surface fire monitoring instead of spaceborne satellite. The FASDD_RS data include sentinel-2 and landsat8, and the NIR or SWIR bands are fundamentally important for monitoring fire-induced vegetation and dryness changes, and are basically used for fire detection. For example, the listed works bellow all used bands like NRI or SWIR. Some basic indexes with NIR and SWIR can achieve substantially higher accuracy (e.g., Castillo et al., 2020) than the reported accuracy by this study. Also the deep learning work in Pereira et al.2021, achieved much higher accuracy than your work. I have never seen a published reliable fire dataset derived from satellite observations only used RGB.
Barboza Castillo, E., Turpo Cayo, E. Y., de Almeida, C. M., Salas López, R., Rojas Briceño, N. B., Silva López, J. O., ... & Espinoza-Villar, R. (2020). Monitoring wildfires in the northeastern peruvian amazon using landsat-8 and sentinel-2 imagery in the GEE platform. ISPRS International Journal of Geo-Information, 9(10), 564.
de Almeida Pereira, G. H., Fusioka, A. M., Nassu, B. T., & Minetto, R. (2021). Active fire detection in Landsat-8 imagery: A large-scale dataset and a deep-learning study. ISPRS Journal of Photogrammetry and Remote Sensing, 178, 171-186.
Hu, X., Ban, Y., & Nascetti, A. (2021). Sentinel-2 MSI data for active fire detection in major fire-prone biomes: A multi-criteria approach. International Journal of Applied Earth Observation and Geoinformation, 101, 102347.
(2) Data generation: for active fire detection, middle infrared and thermal bands are also
important but were ignored.My response to the authors’ responses: Actually, landsat8 has thermal infrared bands. My concern is that why only use RGB for fire detection? The ultimate goal of this dataset is for improving fire detection instead of some technique details like only using one machine learning model with the same-format inputting dataset. If fusing different datasets can improve fire detection, why not use comprehensive models (e.g., different models) to fuse different datasets and finally achieve better fire detection (e.g., based on an ensemble of machine learning models)? If you use this dataset to develop one single model but throwing important bands information, I don’t think it is a good idea.
(3) Data generation: for the fire detection, the authors used the top-of-atmosphere reflectance instead of atmospheric-corrected land surface reflectance which should be a problem. For example, if the smoke is white or grey, how to classify smoke versus clouds only through visual interpretation of true-colors? Meanwhile, the CV fire images should be obtained on land surface with a much higher spatial resolution (can be with sub-meter resolution in Fig. 4). Can such kinds of CV image trained machine learning models be directly used to large-scale remote sensing data without atmospheric-correction and with different spatial resolutions?
My response to the authors’ responses: I understand that FASDD_CV have images from UAVs. But from the results in Table 4, it seems that for FASDD_RS, the model performance even dropped if we used FASDD (including FASDD_RS and FASDD_CV) for training the model. The dropped performance means that the knowledge derived from FASDD_CV can hardly be transferred or used for improving FASDD_RS. If so, why we should integrate those two heterogenous datasets?
(4) Data annotation: the “minimum bounding rectangle” was used to label the images. Commonly, fire detection is to classify whether each pixel is burned or not instead of a rectangle (e.g., the MODIS fire product in Giglio et al., (2018) and Giglio et al., (2016)). Meanwhile, the fire patch perimeter was always not rectangle (Laurent et al., 2018), therefore a bounding rectangle could contain both burned and unburned pixels, right?
My response to the authors’ responses: Please clarify the differences for fire detection in terms of scene classification, object detection, and semantic segmentation; clarify the main focus of this dataset (e.g., objection detection tasks) and its differences with previous fire detection datasets from satellites in the main text. Currently, section 2.1 only mentioned about existing fire detection datasets for terrestrial and airborne but not including existing satellite fire detection datasets.
(5) Data validation: the true-color based visual interpretation could also involve biases, therefore it’s important to validate the generated data against other reliable fire dataset, such as the MTBS data
My response to the authors’ responses: it can hardly convince me by validating the dataset only using one specific fire event in MTBS. Additionally, for Figure 6, why not use metrics to quantitively validate the dataset? Please comprehensively validate your dataset. If the dataset is not reliable, how can we trust it and use it?
(6) Line 245-246, the dataset consists of 95,314 computer vision fire samples but only 5,773 remote sensing samples. Due to the data imbalance, the model performance (Table 2) on FASDD data therefore mainly depends on the model performance on FASDD_CV. For the limited number of remote sensing fire samples, most samples in each region were distributed within ten days of a specific year (Table 1). Can such limited number of wildfires reflect all the seasonal and interannual changes of environmental conditions and fire dynamics so that machine learning models could learn from enough data? To my knowledge, the fire occurrence conditions changed across seasons and therefore the fire detectability could also be affected.
My response to the authors’ responses: a) can CV dataset help improve fire detection using RS (Table 4)? b) I disagree that “the features of fire objects do not have a high seasonal dependence” responded by the authors. The fire features including fire radiative power or fire intensity, fire spread, fire duration, and fire size, are strongly controlled by environmental conditions which could vary across seasons (e.g., dry season versus non-dry seasons) or within a season.
(7) Evaluation in section 4: the evaluation mainly showed the extent to which machine learning models could detect fires of generated FASDD data, therefore whether FASDD is reliable remains unknown. The FASDD data needed to be validated against other reliable fire products.
My response to the authors’ responses: please see my comment in (5).
(8) There are many existing remote sensing fire products (e.g., MTBS, MODIS fire products) and CV fire data sets. I understand that combining the two kinds of dataset was the main difference of FASDD relative to other datasets, however, the authors did not show why combining these two kinds of dataset is important? Can combining these two kinds of datasets improve fire detection relative to existing fire detection algorithms (e.g., the method for MODIS or MTBS fire product)? If not, why people use such complex dataset (with different spatial resolution and without atmospheric-correction)
My response to the authors’ responses: For table 4, the accuracy on FASDD_RS is really lower than existing fire detection algorithms (see my major comment 1), and it seems that integrating those two datasets can achieve very limited benefit or even performance loss for three of total four machine learning methods. If so, why people use such complex dataset? Additionally, the MTBS is based on Landsat, not necessarily having a lower spatial resolution than your RS dataset.
(9) Line 365-370: chaning "epoch" is not ablation experiments. Ablation experiments commonly refer to changing a component of machine learning model (i.e., model structure changes)
Citation: https://doi.org/10.5194/essd-2023-73-RC1 -
AC2: 'Reply on RC1', Ming Wang, 17 Apr 2023
Thank you very much for your comments and suggestions. We insist that our datasets, in particular FASDD_CV, contribute to fire detection research in the field of computer vision. Thus, we decide to withdraw this submission and reconstruct the remote sensing datasets by considering NIR and SWIR bands, with the hope of achieving better fire detection performance.
Thank you again for your comments and help.
Citation: https://doi.org/10.5194/essd-2023-73-AC2 -
AC3: 'Reply on RC1', Ming Wang, 17 Apr 2023
Thank you very much for your comments and suggestions. We insist that our datasets, in particular FASDD_CV, contribute to fire detection research in the field of computer vision. Thus, we decide to withdraw this submission and reconstruct the remote sensing datasets by considering NIR and SWIR bands, with the hope of achieving better fire detection performance.
Thank you again for your comments and help.
Citation: https://doi.org/10.5194/essd-2023-73-AC3 -
AC4: 'Reply on RC1', Ming Wang, 18 Apr 2023
We would like to make a clarification that after careful evaluation, we should be able to add additional experiments before the deadline to address your concerns based on your suggestions. Thank you again for your suggestions.
Citation: https://doi.org/10.5194/essd-2023-73-AC4 - AC5: 'Reply on RC1', Ming Wang, 29 May 2023
-
AC17: 'Reply on RC1', Ming Wang, 24 Sep 2023
We sincerely appreciate your valuable comments and suggestions. Your insights have played a crucial role in improving the quality of both our dataset and the manuscript. In response to all your concerns, we conducted a significant number of experiments and made corresponding revisions to the manuscript.
Please refer to the attached document for specific responses and revisions.
-
AC2: 'Reply on RC1', Ming Wang, 17 Apr 2023
-
CC2: 'Request for Dataset for my Thesis on Fire Detection', Ebrahim Akbari, 26 May 2023
Dear Team,
I am writing to kindly request access to the FASDD dataset for my master's thesis on fire detection. I have come across your dataset paper titled "FASDD: An Open-access 100,000-level Flame and Smoke Detection Dataset for Deep Learning in Fire Detection" and I believe it would greatly contribute to the research I am conducting.
I am particularly interested in utilizing the annotated images from the dataset to train and evaluate deep learning models for fire detection. The comprehensive nature of the FASDD dataset, along with its large-scale and open-access features, align perfectly with the objectives of my thesis.
I understand that the dataset paper is currently under review and open for public discussion in the Earth System Science Data (ESSD) journal. I would be more than willing to provide valuable feedback and actively participate in the discussion to support the publication process and promote the sharing of this dataset with the research community.
I kindly request access to a portion of the annotated dataset, approximately 5000 images, to facilitate the progress of my thesis research. I assure you that the dataset will be used solely for academic purposes and in accordance with any terms and conditions set by your platform.
Thank you very much for considering my request. I greatly appreciate your support and contribution to the field of fire detection. Please let me know if there are any further steps or requirements I need to fulfill in order to obtain access to the FASDD dataset.
I look forward to your positive response.
Sincerely,
Ebrahim Akbari
A master's student in Telecommunications Engineering at Quchan University of Technology
ebrahimakbari118@gmail.com
+989153858085
Citation: https://doi.org/10.5194/essd-2023-73-CC2 -
CC3: 'Reply on CC2', Liangcun Jiang, 27 May 2023
Dear Ebrahim Akbari,
Thank you for your interest in our dataset. We are thrilled that it can contribute to your research.
The FASDD dataset is specifically designed for training and evaluating deep learning models in fire detection, with over 120,000 diverse flame and smoke images captured in various scenarios. Our team has meticulously annotated these images through manual visual annotation and multiple quality checks. The annotations aim to improve the generalization performance of deep learning models and provide a benchmark for advanced fire detection models.We have taken note of your data requirements and will reach out to you via email soon to provide a portion of the dataset for your research purposes.
Once again, we appreciate your interest, support, and acknowledgment of our work. We hope that our dataset proves valuable to your research and contributes to the field of fire detection.
Citation: https://doi.org/10.5194/essd-2023-73-CC3
-
CC3: 'Reply on CC2', Liangcun Jiang, 27 May 2023
-
CC4: 'Comment on essd-2023-73', Rakesh Kumar Mallik, 29 May 2023
Hi This is Rakesh.
I am a AI/Machine learning ensthusiastic person, who has worked in couple of corporates as data sceintist and ML engineer.
I came across this valubale data set of Fire and Somke detection. I beleive this dataset can be used for designing a robust work place safetly system. Not only that based on the variance in the data distribution, it can be used for industrial safety and precaution, can detect roadside fire etc.. I see small small openset dataset on Fire and smoke many places here and there, but None has gobne to this lenght of creating of 1L images with segmenattion mask. Even its just for my pure educational intrest but i would love to get a hange of this dataset and explore more what this can offer.
Thank you
Citation: https://doi.org/10.5194/essd-2023-73-CC4 -
AC6: 'Reply on CC4', Ming Wang, 30 May 2023
Dear Rakesh,
Thank you for expressing your interest in our dataset. The FASDD dataset has been specifically curated as an object detection dataset for training and evaluating deep learning models in fire detection. It consists of over 120,000 distinct images capturing flames and smoke in various scenarios. Our team has meticulously annotated these images through manual visual inspection and multiple quality checks. These annotations aim to improve the generalization performance of deep learning models and provide a benchmark for advanced fire detection systems.
We appreciate your data request, and we will be contacting you shortly via email to provide a portion of the dataset for your research purposes. It's important to note that the current dataset we have prepared primarily focuses on object detection tasks. However, with additional efforts, you can generate segmentation masks based on the existing annotations to suit your specific applications. In fact, we are actively working on creating a separate dataset specifically tailored for semantic segmentation tasks, which will incorporate segmentation masks of flames and smoke.
Once again, we sincerely appreciate your interest, support, and recognition of our work. We hope that our dataset proves valuable for your research endeavors and contributes to the field of fire detection.
Citation: https://doi.org/10.5194/essd-2023-73-AC6
-
AC6: 'Reply on CC4', Ming Wang, 30 May 2023
-
CC5: 'Comment on essd-2023-73', Stephan Sturges, 21 Jun 2023
This dataset would be a very welcome addition to the field. There is a global explosion in capabilities when it comes to overhead sensing whether from satellites, drones, blimps, planes etc... And there a real need for datasets like this to allow detection capabilites to be created for these sensing modalities.
Citation: https://doi.org/10.5194/essd-2023-73-CC5 -
AC7: 'Reply on CC5', Ming Wang, 23 Jun 2023
Dear Stephan Sturges,
Thank you for your valuable feedback on our dataset. We greatly appreciate your recognition of the significance of our contribution to the fire detection field. We completely agree that there is currently a global surge in capabilities for overhead sensing, including data collected from satellites, drones, blimps, planes, and other platforms.
Your comment highlights the real need for comprehensive datasets such as ours, as they play a crucial role in enabling the development of detection capabilities for these various sensing modalities. By providing access to this dataset, we aim to facilitate research and innovation in the fire detection field by offering a valuable resource for researchers and practitioners.
Once again, we sincerely thank you for acknowledging the importance of our dataset, and we hope that it proves to be a valuable asset for future advancements in fire detection.
Best regards,
The authors
Citation: https://doi.org/10.5194/essd-2023-73-AC7
-
AC7: 'Reply on CC5', Ming Wang, 23 Jun 2023
-
CC6: 'Comment on essd-2023-73', Rafik Ghali, 27 Jun 2023
Dear sir,
I hope this email finds you well.
I am writing to you as a researcher interested in your recent publication titled "FASDD: An Open-access 100,000-level Flame And Smoke Detection Dataset for Deep Learning in Fire Detection". I have read your work with great interest and found the dataset used in your study to be of significant relevance to my own research objectives.
Thank you for your attention to this matter. I look forward to your positive response. Please do not hesitate to contact me if you require any further information or have any questions. I truly appreciate your time and consideration.
Best regards,
Citation: https://doi.org/10.5194/essd-2023-73-CC6 -
AC8: 'Reply on CC6', Ming Wang, 03 Jul 2023
Dear Rafik Ghali,
Thank you for reaching out and expressing your interest in our recent publication titled "FASDD: An Open-access 100,000-level Flame And Smoke Detection Dataset for Deep Learning in Fire Detection." We are glad to hear that you found our dataset to be of significant relevance to your own research objectives.
After thorough discussions among the authors, we have decided to share a portion of the dataset with interested researchers like yourself. We believe that collaboration and knowledge sharing are crucial for advancing research in the field of fire detection. We appreciate your enthusiasm and dedication to furthering the scientific community's understanding in this area.
Thank you once again for your interest in our work.
Best regards,
The authors
Citation: https://doi.org/10.5194/essd-2023-73-AC8
-
AC8: 'Reply on CC6', Ming Wang, 03 Jul 2023
-
CC7: 'dataset access', dipesh digwal, 03 Jul 2023
please can you please provide me fire dataset as it is to be used for a fire safety device being constructed by my company
Citation: https://doi.org/10.5194/essd-2023-73-CC7 -
AC9: 'Reply on CC7', Ming Wang, 03 Jul 2023
Dear Dipesh Digwal,
Thank you for your email and your interest in our fire dataset. We appreciate your recognition of the value and contribution of our dataset.
After careful evaluation, the authors have decided to share a portion of the dataset with you to support your project. However, please note that the dataset is subject to certain conditions and restrictions, which we will outline below:
Intended Use: The dataset is intended to be used for fire safety purposes only. It should not be utilized for any other purposes without obtaining proper permissions or rights.
Attribution: It is important to acknowledge the source of the dataset appropriately. Please ensure that you credit the dataset by mentioning the authors and providing a reference to the original publication or source.
Confidentiality and Security: The shared dataset should be treated with utmost confidentiality and stored securely. It must not be shared with any third parties without prior consent from our team.
Compliance with Legal and Ethical Guidelines: The usage of the dataset must comply with all applicable laws, regulations, and ethical guidelines. It should not infringe upon the rights of individuals or organizations, and any potential risks or biases should be carefully considered and addressed.
We appreciate your understanding and cooperation in this matter. We look forward to supporting your efforts in fire safety device construction.
Best regards,
The authors
Citation: https://doi.org/10.5194/essd-2023-73-AC9
-
AC9: 'Reply on CC7', Ming Wang, 03 Jul 2023
-
CC8: 'Comment on essd-2023-73', Rafik Ghali, 05 Jul 2023
Dear Sir
I hope this e-mail finds you well. I am writing to ask you to kindly provide me with a link to the FASDD_RS satellite image dataset.The images available in the FASDD_RS database are renowned for their quality and relevance to my research area.I would be very grateful if you could provide me with access or information on the access procedure. Access to the FASDD_RS database will contribute significantly to the success of my research.Thank you in advance for your help and prompt reply. Please let me know if there are any other requirements or procedures to follow.Best regards,Citation: https://doi.org/10.5194/essd-2023-73-CC8 -
AC10: 'Reply on CC8', Ming Wang, 06 Jul 2023
Dear Rafik Ghali,
I hope this email finds you well. Thank you for expressing your interest in the FASDD_RS satellite image database. Unfortunately, at this stage, our database is still undergoing the peer-review process for the associated paper, and we are unable to publicly release all of the data. However, we understand the importance of access to relevant imagery for your research.
In consideration of your needs, we are able to provide you with a portion of the annotated fire and smoke images from our database. These images can be valuable assets for your research. We believe that this subset will still contribute significantly to your studies while we continue the review process for the full dataset.
Thank you once again for your interest and recognition.
Best wishes,
The authors
Citation: https://doi.org/10.5194/essd-2023-73-AC10 -
AC11: 'Reply on CC8', Ming Wang, 06 Jul 2023
Dear Rafik Ghali,
I hope this email finds you well. Thank you for expressing your interest in the FASDD_RS satellite image database. Unfortunately, at this stage, our database is still undergoing the peer-review process for the associated paper, and we are unable to publicly release all of the data. However, we understand the importance of access to relevant imagery for your research.
In consideration of your needs, we are able to provide you with a portion of the annotated fire and smoke images from our database. These images can be valuable assets for your research. We believe that this subset will still contribute significantly to your studies while we continue the review process for the full dataset.
Thank you once again for your interest and recognition.
Best wishes,
The authors
Citation: https://doi.org/10.5194/essd-2023-73-AC11
-
AC10: 'Reply on CC8', Ming Wang, 06 Jul 2023
-
CC9: 'Comment on essd-2023-73', lin jing, 08 Jul 2023
Hello, I am training a model to detect fire smoke, I believe your dataset will help me a lot, looking forward to your reply!
Citation: https://doi.org/10.5194/essd-2023-73-CC9 -
AC12: 'Reply on CC9', Ming Wang, 08 Jul 2023
Dear Lin,
Thank you for your interest in our dataset (FASDD). Having reliable data is crucial for improving the accuracy and effectiveness of deep learning models. We are delighted to hear that you believe our dataset will be beneficial for your model training.
However, our dataset is currently not available in its entirety due to ongoing research paper review process. Nevertheless, we are willing to share a partial subset of the dataset with you. This will enable you to make progress in your training process while respecting the limitations imposed by the review process. We appreciate your understanding regarding the partial sharing of the dataset. We believe that even with this subset, you will be able to make valuable advancements in your work. We will reach out to you shortly to discuss the details of sharing the subset of the dataset.
Once again, we sincerely appreciate your interest and recognition of FASDD dataset. We wish you every success in your model training.
Best regards,
The authors
Citation: https://doi.org/10.5194/essd-2023-73-AC12
-
AC12: 'Reply on CC9', Ming Wang, 08 Jul 2023
-
CC10: 'Comment on essd-2023-73', sumanth reddy, 18 Jul 2023
Dear Ming Wang,
I would like to extend my heartfelt appreciation to the creators and contributors of the FASSD dataset. As a researcher in the field of [mention your research area], I have had the privilege of utilizing this dataset in my recent studies, and I am truly impressed by its quality and scope.
The FASSD dataset has been an invaluable asset to the research community, providing a diverse range of data that is essential for advancing our understanding of [mention the specific research area or application]. The attention to detail and the meticulous curation of the dataset have been evident in every aspect, enabling researchers like myself to derive meaningful insights and produce reliable results.
Furthermore, the seamless accessibility and proper documentation of the FASSD dataset have streamlined my research process significantly. I commend the team behind this initiative for their dedication to promoting transparency and reproducibility in the scientific community.
The impact of the FASSD dataset on the research landscape cannot be overstated. It has not only facilitated my own research but has also inspired new research directions and collaborations within the community. The dataset's versatility and comprehensive coverage have sparked innovative ideas that promise to push the boundaries of our field even further.
In conclusion, I want to express my deep gratitude to the individuals and organizations responsible for the creation and maintenance of the FASSD dataset. Your contribution has empowered researchers worldwide and accelerated the pace of scientific discovery. I look forward to witnessing the continued growth and impact of this remarkable resource.
Thank you once again for this outstanding dataset, and I eagerly await further advancements in the field, driven by the remarkable work of Ming Wang.
Citation: https://doi.org/10.5194/essd-2023-73-CC10 -
AC13: 'Reply on CC10', Ming Wang, 18 Jul 2023
Dear Sumanth Reddy,
Thank you for your comment expressing your heartfelt appreciation for the creators and contributors of the FASSD dataset. We are delighted to hear that you have found immense value in utilizing the dataset in your recent studies.
As a fellow researcher in the field of fire detection, we share your admiration for the quality and scope of the FASSD dataset. Its diverse range of data has undoubtedly been instrumental in advancing our understanding of fire. The attention to detail and meticulous curation evident in every aspect of the dataset have allowed researchers like us to derive meaningful insights and achieve reliable results.
The impact of the FASSD dataset on the research landscape is indeed profound. Not only has it facilitated your own research, but it has also inspired new research directions and collaborations within the community. The dataset's versatility and comprehensive coverage have sparked innovative ideas that hold great potential to push the boundaries of our field even further.
Thank you once again for recognizing the outstanding value of the dataset and for your kind words. We join you in eagerly awaiting further advancements in the field, driven by some remarkable work.
Best regards,
The authors
Citation: https://doi.org/10.5194/essd-2023-73-AC13
-
AC13: 'Reply on CC10', Ming Wang, 18 Jul 2023
-
CC11: 'Comment on essd-2023-73', Shokhruz Kak, 30 Jul 2023
Dear Authors,
I hope this email finds you well. I am writing to request access to the Flame And Smoke Detection Dataset (FASDD) mentioned in your paper titled "FASDD: An Open-access 100,000-level Flame And Smoke Detection Dataset for Deep Learning in Fire Detection".
he FASDD: An Open-access 100,000-level Flame And Smoke Detection Dataset for Deep Learning in Fire Detection is a remarkable contribution to the field of fire detection in several ways. Firstly, it is accessible dataset that is necessary for training deep learning models to accurately detect fire in different situations. This can significantly improve fire management and save numerous lives. Secondly, this dataset opens up new possibilities for research in the field of deep learning, which could lead to the development of innovative approaches and solutions for fire detection. Overall, the FASDD is an excellent resource that offers a variety of benefits and opportunities to researchers, practitioners, and firefighters, and its availability will undoubtedly bring about significant progress in the field of fire detection and management.
I am particularly interested in your dataset for my research paper in machine learning to detect wildfires. Access to the FASDD dataset would greatly contribute to my research by enabling a comparative analysis of different fire detection methodologies and enhancing the accuracy and robustness of my proposed model.
I kindly request your assistance in providing me with access to the FASDD dataset, or any alternative means to obtain the necessary data. I assure you that the dataset will be used solely for research purposes and in accordance with the terms and conditions set by the dataset's creators. Any data shared with me will be treated with the utmost confidentiality and will not be shared with any third parties.Should you require any further information or have any concerns, please do not hesitate to contact me. Thank you for your time and attention.SincerelyCitation: https://doi.org/10.5194/essd-2023-73-CC11 -
AC14: 'Reply on CC11', Ming Wang, 30 Jul 2023
Dear Shokhruz Kak,
I hope this email finds you well. Thank you for expressing your interest in our Flame And Smoke Detection Dataset (FASDD) and for your kind words about its potential contributions to the field of fire detection.
However, at the moment, our dataset is not fully available for external use due to an ongoing research paper review process. We have received a significant number of requests from researchers like yourself, which has contributed to the evaluation of the dataset's potential impact and robustness.
Nonetheless, we understand the importance of collaboration and knowledge sharing, and we are more than willing to support your research efforts. As such, we can offer you access to a partial subset of the FASDD. While it may not encompass the entire dataset, we believe it can still provide valuable insights and contribute to the comparative analysis you intend to conduct. We will reach out to you shortly to discuss the details of sharing the subset of the dataset.
Thank you for your understanding and cooperation. We look forward to assisting you in your research endeavors.
Best regards,
The authors
Citation: https://doi.org/10.5194/essd-2023-73-AC14
-
AC14: 'Reply on CC11', Ming Wang, 30 Jul 2023
-
CC12: 'Comment on essd-2023-73', hossein rajoli, 12 Aug 2023
Being a part of an engaged research team focused on wildfire detection and management, obtaining access to the dataset would be immensely appreciated. Such access has the potential to significantly advance my research efforts in this domain.
Bests.
Citation: https://doi.org/10.5194/essd-2023-73-CC12 -
AC15: 'Reply on CC12', Ming Wang, 12 Aug 2023
Dear hossein rajoli,
Thank you for expressing your interest in our Flame And Smoke Detection Dataset (FASDD).
We will reach out to you shortly to discuss the details of sharing the dataset.
Thank you once again for your interest and recognition of FASDD dataset. We hope that the dataset will be helpful to you in your research.
Best regards,
The authors
Citation: https://doi.org/10.5194/essd-2023-73-AC15
-
AC15: 'Reply on CC12', Ming Wang, 12 Aug 2023
-
RC2: 'Comment on essd-2023-73', Anonymous Referee #2, 23 Aug 2023
The comment was uploaded in the form of a supplement: https://essd.copernicus.org/preprints/essd-2023-73/essd-2023-73-RC2-supplement.pdf
-
AC16: 'Reply on RC2', Ming Wang, 24 Sep 2023
We sincerely appreciate your valuable comments and suggestions. In response to your concerns, we have conducted extensive experiments and made corresponding revisions to the manuscript.
Please refer to the attached document for specific responses and revisions.
-
AC16: 'Reply on RC2', Ming Wang, 24 Sep 2023
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 3,127 | 912 | 152 | 4,191 | 85 | 70 |
- HTML: 3,127
- PDF: 912
- XML: 152
- Total: 4,191
- BibTeX: 85
- EndNote: 70
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Cited
5 citations as recorded by crossref.
- Fire Detection in Urban Areas Using Multimodal Data and Federated Learning A. Sharma et al. 10.3390/fire7040104
- Video Fire Detection Methods Based on Deep Learning: Datasets, Methods, and Future Directions C. Jin et al. 10.3390/fire6080315
- Enhancing Fire Detection Technology: A UV-Based System Utilizing Fourier Spectrum Analysis for Reliable and Accurate Fire Detection C. Truong et al. 10.3390/app13137845
- Recent Advances and Emerging Directions in Fire Detection Systems Based on Machine Learning Algorithms B. Diaconu 10.3390/fire6110441
- Review and perspectives of digital twin systems for wildland fire management Y. Li et al. 10.1007/s11676-024-01810-x
Ming Wang
Liangcun Jiang
Peng Yue
Tianyu Tuo
This preprint has been withdrawn.
Please read the editorial note first before accessing the preprint.
- Preprint
(7002 KB) - Metadata XML