Remote Sensing Extraction Method of Urban Roof and Pavement and Analysis in Pollutant Load

Dekun CHEN,Huapeng QIN,Hongliang XU,Xiangli ZHANG

DOI: https://doi.org/10.3969/j.issn.1003-6504.2017.02.015

IF: 11.4

2017-01-01

Environmental Science & Technology

Abstract:Urban non-point source pollution is one of the main reasons to cause water quality deterioration.Load and flushing effects of pollutants from roofs and pavements are significantly different,and effective extraction of roof and pavement information from remote sensing images is the basis of urban non-point source evaluation and management.For three regions of different building densities in a Shenzhen' s high-resolution remote sensing image,the strategy based on the combination of pixel-based classification and object-oriented method was adopted to extract the roof and pavement information,and the results were then compared to those derived from using visual interpretation,pixel-based classification and object-oriented method techniques;pollutant load per unit area in typical samples from roofs and pavements were estimated based on wet sampling method and then used together with the roof and pavement areas calculated from the remote sensing images to estimate accumulative pollutant load.The study showed this method improved the visual effect and classification accuracy of roof and pavement information extraction,and the overall accuracies were 83.82％,93.02％ and 88.31％ for high,medium and low building density regions,respectively.COD load per unit area of roof was 97.32 mg/m2,and COD loads per unit area were 196.02 mg/m2,104.50 mg/m2 and 50.65 mg/m2 for the pavements in high,medium and low building density regions,respectively.For all the COD loads from impervious areas in high,medium and low building density regions,total roof COD loads accounted for 30.05％,34.09％ and 71.74％,respectively,while the percentages of the total pavement COD loads were 69.95％,65.91％,and 28.26％,respectively.