基于遥感技术的北京市平原区砂石坑识别探索

The Identification of Deserted Pits in Plain Area of Beijing Based on Remote Sensing Image

为满足城市建设对建材的需求,一般会在城市周边开挖砂石,从而遗留了很多砂石坑。遗留下的砂石坑是一种退化的生态系统,也是导致植被丧失、景观破坏并引起沙尘暴的主要原因之一,同时由于对其无序的管理和垃圾、废水的排放,它也成为影响城市环境质量的重要因素之一。由于遗留的砂石坑数量大,分布广,通常需要花费很长的时间和人力进行实际普查,才能掌握其数目、大小和分布。为配合北京市的平原区砂石坑综合整治规划研究,尝试利用ASTER遥感数据,识别北京市平原区的砂石坑分布。本研究针对砂石坑这一特定地物信息,采用监督分类的方法对北京市平原区的砂石坑信息进行识别和提取,并采用最大多数聚焦分析对分类图像进行后处理,优化分类结果。通过与前期调查数据的比对,分类的生产者精度和用户精度分别为85.88%和60.33%。随着核实调查的开展,分类精度有望继续提高。通过遥感影像分类所提取的砂石坑信息,对快速了解北京市平原区砂石坑的总体分布情况,从而进一步深入展开砂石坑全面调查研究,并为北京市政府进行砂石坑的生态恢复提供合理、可行的规划方案奠定了基础,具有十分重要的指导意义。

The pits left from the urban development and construction are degraded ecosystems which will cause a loss of vegetation, landscape destruction and sandstorm, and thus have a harmful influence on the urban environmental quality. Based on remotely sensed data ASTER, using supervised classification method and focal majority analysis, the information of pit＇s location in the plain area of Beijing is recognized, extracted and the results are optimized. The produce accuracy and user accuracy of the classification is 85.88 % and 60.33% respectively according to the data from a previous research. The classification accuracy has the potential to rise after the verification works. By classification, the information extracted from remotely sensed image will help us in knowing quickly the general distribution of the pits, and on the basis of this it will be possible to support the study on ecosystem restoration of the pits.