Received: 11 Jan 2022 – Discussion started: 11 Apr 2022
Abstract. Roadside noise barriers (RNBs) are important urban infrastructures to develop a liveable city. However, the absence of accurate and large-scale geospatial data on RNBs has impeded the increasing progress of rational urban planning, sustainable cities, and healthy environments. To address this problem, this study proposes a geospatial artificial intelligence framework to create a vectorized RNB dataset in China using street view imagery. To begin, intensive sampling is performed on the road network of each city based on OpenStreetMap, which is used as the geo-reference to download 5.6 million Baidu Street View (BSV) images. Furthermore, considering the prior geographic knowledge contained in street view images, convolutional neural networks incorporating image context information (IC-CNNs) based on an ensemble learning strategy are developed to detect RNBs from the BSV images. Subsequently, the RNB dataset presented by polylines is generated based on the identified RNB locations, with a total length of 2,227 km in 215 cities. At last, the quality of the RNB dataset is evaluated from two perspectives: first, the detection accuracy; second, the completeness and positional accuracy. Specifically, based on a set of randomly selected samples containing 10,000 BSV images, four quantitative metrics are calculated, with an overall accuracy of 98.61 %, recall of 87.14 %, precision of 76.44 %, and F1-score of 81.44 %. Moreover, a total length of 254 km of roads in different cities are manually surveyed using BSV images to evaluate the mileage deviation and overlap level between the generated and surveyed RNBs. The root-mean-squared error for mileage deviation is 0.08 km, and the intersection over union for overlay level is 88.08 % ± 2.95 %. The evaluation results suggest that the generated RNB dataset is of high quality and can be applied as an accurate and reliable dataset for a variety of large-scale urban studies. The generated vectorized RNB dataset and the labelled BSV image benchmark dataset are publicly available at https://doi.org/10.11888/Others.tpdc.271914 (Chen, 2021).
This paper is well written and inspires me by providing two kinds of datasets (i.e., RNB dataset and street view imagery benchmark dataset) as well as developing a deep learning method incorporating image context information. This work is important as the reuslts fill a gap in the shortage of national infrastructure datasets (i.e., RNBs).
This paper creates a vectorized RNB dataset, which is an impressive work of good quality. This dataset will be beneficial for further studies as it is not always easy to create and find such data. Also, the street view image benchmark dataset is provided, which can be used as a training dataset for further work. There are some comments for the authors to consider: 1. Do the authors see any possibility to extent the application of this approach to other regions outside of China in the future? 2. I hope this dataset can be updated regularly to follow the frequency of Baidu Maps adaptation, and it is very beneficial, although the workload is enormous. 3. Line 247-248: As said in this article, "where blank areas indicate no RNBs or lack of BSV images", from my perspective, it will be exciting and crucial to know exact information from blank areas, such as which cities lack RNBs or BSV images.
NoiseBarrierIdentificationZhen Qian, Min Chen, Yue Yang, Teng Zhong, Fan Zhang, Rui Zhu, Kai Zhang, Zhixin Zhang, Zhuo Sun, Peilong Ma, Guonian Lü, Yu Ye, Jinyue Yan http://dx.doi.org/10.11888/Others.tpdc.271914
Roadside Noise Barrier IdentificationZhen Qian, Min Chen, Yue Yang, Kai Zhang, Teng Zhong, Fan Zhang, Rui Zhu, Zhixin Zhang, Zhuo Sun, Peilong Ma, Guonian Lü, and Yu Ye https://github.com/ChanceQZ/NoiseBarrierIdentification
Zhen Qian et al.
Viewed
Total article views: 328 (including HTML, PDF, and XML)
HTML
PDF
XML
Total
BibTeX
EndNote
262
52
14
328
3
5
HTML: 262
PDF: 52
XML: 14
Total: 328
BibTeX: 3
EndNote: 5
Views and downloads (calculated since 11 Apr 2022)
Cumulative views and downloads
(calculated since 11 Apr 2022)
Viewed (geographical distribution)
Total article views: 311 (including HTML, PDF, and XML)
Thereof 311 with geography defined
and 0 with unknown origin.
Roadside noise barriers (RNBs) are important urban infrastructures to develop a liveable city. This study provides the first reliable and nationwide vectorized RNB dataset by street view imagery in China. The generated RNB dataset is evaluated in two aspects, i.e., the detection accuracy, as well as the completeness and positional accuracy. The method is based on a developed geospatial artificial intelligence framework.
Roadside noise barriers (RNBs) are important urban infrastructures to develop a liveable city....