Refined Monitoring of Water Bodies in Built-Up Areas and Analysis and Evaluation of Water Color Characteristics in Complex Urban Backgrounds
With the rapid construction and development of cities in China, urban water environment changes and water quality monitoring face increasingly complex challenges. Water body extraction in complex urban settings is susceptible to environmental disturbances, reducing monitoring accuracy and efficiency...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/11021300/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849433294747205632 |
|---|---|
| author | Bin Li Yong Xie Zui Tao Wen Shao Gan Qin |
| author_facet | Bin Li Yong Xie Zui Tao Wen Shao Gan Qin |
| author_sort | Bin Li |
| collection | DOAJ |
| description | With the rapid construction and development of cities in China, urban water environment changes and water quality monitoring face increasingly complex challenges. Water body extraction in complex urban settings is susceptible to environmental disturbances, reducing monitoring accuracy and efficiency. Additionally, traditional water quality detection methods, relying on laboratory sampling and analysis, are hindered by high costs and long cycles. To address these issues, this study constructs a high-resolution satellite imagery dataset for urban contexts and trains four deep learning models (UNet, DeepLab, PSPNet, Segformer) for water body extraction. A method combining the water color index and geographic detector for rapid water quality assessment is also proposed. The research shows that Segformer performs best in handling complex urban background issues (e.g., shadow problems), achieving a water body extraction accuracy of 98.64%. The self-constructed dataset demonstrates good accuracy across all four models, effectively addressing urban environment diversity and interference factors. For water quality assessment, the study combines the water color index and geographic detector, establishing a rapid evaluation method. Using the water color index for preliminary assessment and the geographic detector to analyze the impact of environmental factors on water quality changes, the results for Shenzhen’s built-up area show that primary and secondary water colors represent 39.17% and 52.81%, respectively. Comparison with measured data reveals a root mean square error of 3.9° and an average absolute percentage error of 1.5%, validating the proposed method’s accuracy. Industrial location, population density, soil erosion, and other factors significantly affect water color, particularly as industrial activities and soil erosion alter water composition and suspended solids, directly impacting water color. The interaction analysis using the geographic detector highlights how the combined influence of factors such as industrial activities and soil erosion affects water color changes, providing a clearer understanding of the spatial factors driving water quality variations in urban environments. |
| format | Article |
| id | doaj-art-c6523178fa5d405e9bf78ef86617fc29 |
| institution | Kabale University |
| issn | 1939-1404 2151-1535 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj-art-c6523178fa5d405e9bf78ef86617fc292025-08-20T03:27:06ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-0118148971491110.1109/JSTARS.2025.357584711021300Refined Monitoring of Water Bodies in Built-Up Areas and Analysis and Evaluation of Water Color Characteristics in Complex Urban BackgroundsBin Li0https://orcid.org/0009-0003-2907-0360Yong Xie1https://orcid.org/0000-0002-7863-7170Zui Tao2https://orcid.org/0000-0002-8369-4452Wen Shao3https://orcid.org/0000-0003-1203-0883Gan Qin4School of Geographical Sciences, Nanjing University of Information Science and Technology, Nanjing, ChinaSchool of Geographical Sciences, Nanjing University of Information Science and Technology, Nanjing, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing, ChinaSchool of Geographical Sciences, Nanjing University of Information Science and Technology, Nanjing, ChinaSchool of Geographical Sciences, Nanjing University of Information Science and Technology, Nanjing, ChinaWith the rapid construction and development of cities in China, urban water environment changes and water quality monitoring face increasingly complex challenges. Water body extraction in complex urban settings is susceptible to environmental disturbances, reducing monitoring accuracy and efficiency. Additionally, traditional water quality detection methods, relying on laboratory sampling and analysis, are hindered by high costs and long cycles. To address these issues, this study constructs a high-resolution satellite imagery dataset for urban contexts and trains four deep learning models (UNet, DeepLab, PSPNet, Segformer) for water body extraction. A method combining the water color index and geographic detector for rapid water quality assessment is also proposed. The research shows that Segformer performs best in handling complex urban background issues (e.g., shadow problems), achieving a water body extraction accuracy of 98.64%. The self-constructed dataset demonstrates good accuracy across all four models, effectively addressing urban environment diversity and interference factors. For water quality assessment, the study combines the water color index and geographic detector, establishing a rapid evaluation method. Using the water color index for preliminary assessment and the geographic detector to analyze the impact of environmental factors on water quality changes, the results for Shenzhen’s built-up area show that primary and secondary water colors represent 39.17% and 52.81%, respectively. Comparison with measured data reveals a root mean square error of 3.9° and an average absolute percentage error of 1.5%, validating the proposed method’s accuracy. Industrial location, population density, soil erosion, and other factors significantly affect water color, particularly as industrial activities and soil erosion alter water composition and suspended solids, directly impacting water color. The interaction analysis using the geographic detector highlights how the combined influence of factors such as industrial activities and soil erosion affects water color changes, providing a clearer understanding of the spatial factors driving water quality variations in urban environments.https://ieeexplore.ieee.org/document/11021300/Deep learningfactor detectionurbanized areawater body extractionwater body sample setwater color analysis |
| spellingShingle | Bin Li Yong Xie Zui Tao Wen Shao Gan Qin Refined Monitoring of Water Bodies in Built-Up Areas and Analysis and Evaluation of Water Color Characteristics in Complex Urban Backgrounds IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Deep learning factor detection urbanized area water body extraction water body sample set water color analysis |
| title | Refined Monitoring of Water Bodies in Built-Up Areas and Analysis and Evaluation of Water Color Characteristics in Complex Urban Backgrounds |
| title_full | Refined Monitoring of Water Bodies in Built-Up Areas and Analysis and Evaluation of Water Color Characteristics in Complex Urban Backgrounds |
| title_fullStr | Refined Monitoring of Water Bodies in Built-Up Areas and Analysis and Evaluation of Water Color Characteristics in Complex Urban Backgrounds |
| title_full_unstemmed | Refined Monitoring of Water Bodies in Built-Up Areas and Analysis and Evaluation of Water Color Characteristics in Complex Urban Backgrounds |
| title_short | Refined Monitoring of Water Bodies in Built-Up Areas and Analysis and Evaluation of Water Color Characteristics in Complex Urban Backgrounds |
| title_sort | refined monitoring of water bodies in built up areas and analysis and evaluation of water color characteristics in complex urban backgrounds |
| topic | Deep learning factor detection urbanized area water body extraction water body sample set water color analysis |
| url | https://ieeexplore.ieee.org/document/11021300/ |
| work_keys_str_mv | AT binli refinedmonitoringofwaterbodiesinbuiltupareasandanalysisandevaluationofwatercolorcharacteristicsincomplexurbanbackgrounds AT yongxie refinedmonitoringofwaterbodiesinbuiltupareasandanalysisandevaluationofwatercolorcharacteristicsincomplexurbanbackgrounds AT zuitao refinedmonitoringofwaterbodiesinbuiltupareasandanalysisandevaluationofwatercolorcharacteristicsincomplexurbanbackgrounds AT wenshao refinedmonitoringofwaterbodiesinbuiltupareasandanalysisandevaluationofwatercolorcharacteristicsincomplexurbanbackgrounds AT ganqin refinedmonitoringofwaterbodiesinbuiltupareasandanalysisandevaluationofwatercolorcharacteristicsincomplexurbanbackgrounds |