UGS-1m: Fine-grained urban green space mapping of 34 major cities in China based on the deep learning framework
- 1School of Geography and Planning, Sun Yat-sen University, Guangzhou 510275, China
- 2Guangdong Key Laboratory for Urbanization and Geo-simulation, Guangzhou 510275, China
- 3Dept. of Physics and Technology, UiT the Arctic University of Norway, 9019 Tromsø, Norway
- 4Dept. of Engineering, University of Cambridge, Cambridge CB2 1PZ, UK
- 1School of Geography and Planning, Sun Yat-sen University, Guangzhou 510275, China
- 2Guangdong Key Laboratory for Urbanization and Geo-simulation, Guangzhou 510275, China
- 3Dept. of Physics and Technology, UiT the Arctic University of Norway, 9019 Tromsø, Norway
- 4Dept. of Engineering, University of Cambridge, Cambridge CB2 1PZ, UK
Abstract. Urban green space (UGS) is an important component in the urban ecosystem and has great significance to the urban ecological environment. Although the development of remote sensing platforms and deep learning technologies have provided opportunities for UGS mapping from high-resolution images (HRIs), challenges still exist in its large-scale and fine-grained application, due to insufficient annotated datasets and specially designed methods for UGS. Moreover, the domain shift between images from different regions is also a problem that must be solved. To address these issues, a general deep learning (DL) framework is proposed for UGS mapping in the large scale, and the fine-grained UGS maps of 34 major cities/areas in China are generated (UGS-1m). The DL framework consists of a generator and a discriminator. The generator is a fully convolutional network designed for UGS extraction (UGSNet), which integrates attention mechanisms to improve the discrimination to UGS, and employs a point rending strategy for edge recovery. The discriminator is a fully connected network aiming to deal with the domain shift between images. To support the model training, an urban green space dataset (UGSet) with a total number of 4,454 samples of size 512×512 is provided. The main steps to obtain UGS-1m can be summarized as follows: a) Firstly, the UGSNet will be pre-trained on the UGSet in order to get a good starting training point for the generator; b) After pre-training on the UGSet, the discriminator is responsible to adapt the pre-trained UGSNet to different cities/areas through adversarial training; c) Finally, the UGS results of the 34 major cities/areas in China (UGS-1m) are obtained using 2,343 Google Earth images with a data frame of 7'30" in longitude and 5'00" in latitude, and a spatial resolution of nearly 1.1 meters. Evaluating the performance of the proposed approach on samples from Guangzhou city shows the validity of the UGS-1m products, with an overall accuracy of 87.4 % and an F1 score of 81.14 %. Furthermore, experiments on UGSet with the existing state-of-the-art (SOTA) DL models proves the effectiveness of UGSNet, with the highest F1 of 77.30 %. Finally, the comparisons with existing products further shows the feasibility of the UGS-1m and the effectiveness and great potential of the proposed DL framework. The UGS-1m can be downloaded from https://doi.org/10.5281/zenodo.6155516 (Shi et al., 2022).
Qian Shi et al.
Status: closed
-
RC1: 'Comment on essd-2022-75', Anonymous Referee #1, 18 Jul 2022
This paper provides a high-resolution urban green space (UGS) maps for 34 megacities in China, as well as a UGS dataset for the deep learning models. The paper is complete in structure and the results the results have shown the effectiveness of the proposed deep learning framework. I think the paper can contribute to UGS research in terms of both algorithm, dataset and products. However, some revisions still should be made to the manuscript before considering publication:
- The caption of Figure 1 should be expanded in combination with its content to improve understanding.
- The descriptions of the reminders of the paper (Line 100) should be revised, while the Conclusions are arranged in Section 7.
- In the deep learning framework, the author use “four enhanced Coordinate attention (ECA) modules (Hou et al., 2021) to enhance feature representations” (Line 164). The descriptions of the ECA is simple with only a Figure 7. Is there any difference between the ECA and that in the reference (Hou et al., 2021)? Please elaborate.
- The Point Head in 3.1.2 sounds interesting. However, some details are not clarified. For example, how does the N sampling points obtain? And what is the rationale behind the Point Head to improve accuracies? Besides, the value of N is missed in the paper.
- What does “D“ denote in (4) and (5)?
- Some training details and the network configuration are missed. For example, the details of the data augmentation strategies.
- The Section 4 should be elaborated to upgrade the readability and approval of the results, which is brief relatively at present.
- The introductions of Figure 12 (Line 260) should also be extended.
- There are several typos and statements without any references or evidence. For example, the references of the "Guidance of the General Office of the State Council on Scientific Greening" (Line 108) should be provided.
- For writing, the manuscript needs proofreading.
-
AC1: 'Reply on RC1', Mengxi Liu, 20 Jul 2022
We feel great thanks for your affirmation and careful review work on our article. Your professional suggestions are very helpful to make our study clearer and more comprehensive. In the next, we will carefully consider your suggestions and revise the manuscript accordingly, and give point-by-point responses to your suggestions.
-
RC2: 'Comment on essd-2022-75', Robbe Neyns, 02 Sep 2022
The paper makes a clear contribution to the field of urban green space mapping by providing a new annotated dataset of urban green spaces, a new deep learning framework and a detailed green space map of 34 cities in China. However, I still have several question/comments that need to be addressed before considering publication.
General remarks
- One of the key contributions of the paper is the use of adversarial training is the use of adversarial training to increase the generalisation capacity of the deep learning model. Yet the actual added value of the adversarial part of the model is only addressed to a limited extent. If I understand correctly (this should also be clarified in the paper), the other semantic segmentation models were not finetuned using adversarial training. This means that one could deduce the effectiveness of the adversarial training approach from the difference in accuracy with the other models. But this will always leave the question: is the improvement related to the model structure of the generator or due to the inclusion of a discriminator? In short, I would have liked to see the difference in performance with/without the discriminator.
- Related to the previous remark I would have liked a description of the spatial differences in accuracy in the discussion section. Do some cities show a lower accuracy? Could we explain this by the type of green space that is dominant? Difference in phenological phase when the imagery was taken?...
- The discussion of the type of errors in the resulting green space map is very limited. Do you mainly notice problems at the edges of green spaces? For high/low or dense/sparse vegetation?
- Are there any remarks in relation to the dataset? For example, in figure 12 image 3 (starting from above) I noticed that the residential area has been indicated as non-vegetated while there is clearly vegetation next to the buildings.
Specific remarks
- Which value did you use for N in section 3.1.2?
- Figure 8: I believe the final 8x8x1 image is flattened before it is fed through the softmax function?
- How was the parameter optimization performed?
- A batch size of 8 image pairs is rather small, was this decided through parameter optimization experiments or because of memory limitations?
- The same learning rate, batch size and number of epochs was used for all models? Was this done to facilitate the experiments?
-
AC2: 'Reply on RC2', Mengxi Liu, 05 Sep 2022
We really appreciate your professional review work on our article. As you are concerned, there are several problems that need to be addressed. According to your useful suggestions, we will make extensive corrections to our draft as soon as possible, and give point-by-point responses to you.
Status: closed
-
RC1: 'Comment on essd-2022-75', Anonymous Referee #1, 18 Jul 2022
This paper provides a high-resolution urban green space (UGS) maps for 34 megacities in China, as well as a UGS dataset for the deep learning models. The paper is complete in structure and the results the results have shown the effectiveness of the proposed deep learning framework. I think the paper can contribute to UGS research in terms of both algorithm, dataset and products. However, some revisions still should be made to the manuscript before considering publication:
- The caption of Figure 1 should be expanded in combination with its content to improve understanding.
- The descriptions of the reminders of the paper (Line 100) should be revised, while the Conclusions are arranged in Section 7.
- In the deep learning framework, the author use “four enhanced Coordinate attention (ECA) modules (Hou et al., 2021) to enhance feature representations” (Line 164). The descriptions of the ECA is simple with only a Figure 7. Is there any difference between the ECA and that in the reference (Hou et al., 2021)? Please elaborate.
- The Point Head in 3.1.2 sounds interesting. However, some details are not clarified. For example, how does the N sampling points obtain? And what is the rationale behind the Point Head to improve accuracies? Besides, the value of N is missed in the paper.
- What does “D“ denote in (4) and (5)?
- Some training details and the network configuration are missed. For example, the details of the data augmentation strategies.
- The Section 4 should be elaborated to upgrade the readability and approval of the results, which is brief relatively at present.
- The introductions of Figure 12 (Line 260) should also be extended.
- There are several typos and statements without any references or evidence. For example, the references of the "Guidance of the General Office of the State Council on Scientific Greening" (Line 108) should be provided.
- For writing, the manuscript needs proofreading.
-
AC1: 'Reply on RC1', Mengxi Liu, 20 Jul 2022
We feel great thanks for your affirmation and careful review work on our article. Your professional suggestions are very helpful to make our study clearer and more comprehensive. In the next, we will carefully consider your suggestions and revise the manuscript accordingly, and give point-by-point responses to your suggestions.
-
RC2: 'Comment on essd-2022-75', Robbe Neyns, 02 Sep 2022
The paper makes a clear contribution to the field of urban green space mapping by providing a new annotated dataset of urban green spaces, a new deep learning framework and a detailed green space map of 34 cities in China. However, I still have several question/comments that need to be addressed before considering publication.
General remarks
- One of the key contributions of the paper is the use of adversarial training is the use of adversarial training to increase the generalisation capacity of the deep learning model. Yet the actual added value of the adversarial part of the model is only addressed to a limited extent. If I understand correctly (this should also be clarified in the paper), the other semantic segmentation models were not finetuned using adversarial training. This means that one could deduce the effectiveness of the adversarial training approach from the difference in accuracy with the other models. But this will always leave the question: is the improvement related to the model structure of the generator or due to the inclusion of a discriminator? In short, I would have liked to see the difference in performance with/without the discriminator.
- Related to the previous remark I would have liked a description of the spatial differences in accuracy in the discussion section. Do some cities show a lower accuracy? Could we explain this by the type of green space that is dominant? Difference in phenological phase when the imagery was taken?...
- The discussion of the type of errors in the resulting green space map is very limited. Do you mainly notice problems at the edges of green spaces? For high/low or dense/sparse vegetation?
- Are there any remarks in relation to the dataset? For example, in figure 12 image 3 (starting from above) I noticed that the residential area has been indicated as non-vegetated while there is clearly vegetation next to the buildings.
Specific remarks
- Which value did you use for N in section 3.1.2?
- Figure 8: I believe the final 8x8x1 image is flattened before it is fed through the softmax function?
- How was the parameter optimization performed?
- A batch size of 8 image pairs is rather small, was this decided through parameter optimization experiments or because of memory limitations?
- The same learning rate, batch size and number of epochs was used for all models? Was this done to facilitate the experiments?
-
AC2: 'Reply on RC2', Mengxi Liu, 05 Sep 2022
We really appreciate your professional review work on our article. As you are concerned, there are several problems that need to be addressed. According to your useful suggestions, we will make extensive corrections to our draft as soon as possible, and give point-by-point responses to you.
Qian Shi et al.
Data sets
UGS-1m: Fine-grained urban green space mapping of 34 major cities in China based on the deep learning framework Qian Shi, Mengxi Liu and Andrea Marinoni https://doi.org/10.5281/zenodo.6155516
Qian Shi et al.
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
444 | 132 | 21 | 597 | 10 | 9 |
- HTML: 444
- PDF: 132
- XML: 21
- Total: 597
- BibTeX: 10
- EndNote: 9
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1