基于夜光遥感影像的城市建成区面积提取研究

Research on Square Measure Extraction of Urban Built up Area Based on Nightlight Remote Sensing Images

夜光遥感影像通过城市灯光分布情况及强弱程度来分析城市用地信息,从而反映城市化情况。使用两类夜光遥感影像,首先采用遥感指数对DMSP/OLS数据进行饱和性纠正;同时采用幂指数函数模型对两类夜光遥感影像进行连续性纠正,完成图像的预处理工作;其次通过改进的二分阈值迭代法提取2009年,2014年,2019年鞍山市城市建成区面积,并完成精度验证;最后计算城市扩张速度及强度定量分析鞍山市过去十年城市化进程。结果表明,(1)图像预处理方面,遥感指数法有效改善DMSS/OLS数据的饱和效应,幂指数函数模型实现了不同夜光遥感数据间的连续性纠正,该结果有利于后续工作开展;(2)建成区面积提取,改进的二分阈值迭代法有利于选取合适的初始阈值,减少迭代次数、提高提取速度,三个时期建成区面积提取总体精度都在94%以上,F1分数在65%以上;(3)城市化进程分析,速度指数由2009—2014年的4.2提升到2014—2019年的13.0,强度指数由2009—2014年的0.06提升到2014—2019年的0.20,反映出鞍山市过去十年间城市化进程不断加快。

Nightlight remote sensing images analyze urban land information through the distribution and intensity of urban lighting,thereby reflecting the urbanization situation.This article uses two types of night light remote sensing images,first using remote sensing indices to correct the saturation of DMSP/OLS data;Simultaneously,a power exponential function model is used to perform continuity correction on two types of night light remote sensing images,completing image preprocessing work;Secondly,the improved binary threshold iteration method was used to extract the urban built-up area of Anshan prefecture in 2009,2014,and 2019,and accuracy verification was completed;Finally,calculate the speed and intensity of urban expansion to quantitatively analyze the urbanization process of Anshan City in the past decade.The results indicate that①in terms of image preprocessing,the remote sensing index method effectively improves the saturation effect of DMSS/OLS data,and the power exponential function model achieves continuity correction between different night light remote sensing data,which is beneficial for subsequent work;②The improved binary threshold iteration method for extracting built-up area is conducive to selecting appropriate initial thresholds,reducing iteration times,and improving extraction speed.The overall accuracy of built-up area extraction in all three periods is above 94%,and the F1 score is above 65%;③The analysis of urbanization process shows that the speed index increased from 4.2 in 2009-2014 to 13.0 in 2014-2019,and the intensity index increased from 0.06 in 2009-2014 to 0.20 in 2014-2019,reflecting the continuous acceleration of urbanization process in Anshan prefecture over the past decade.