融合流程优化思维的遥感影像变化检测自动化研究

目前土地卫片执法中对影像变化图斑的提取主要依靠作业人员对两时相高分辨率卫星遥感影像的目视解译，以确定土地利用变化发生的空间位置。图斑的正确与否完全依靠解译人员的目视判读经验，容易产生错误。通用的检测流程针对特定遥感影像数据可以得到较好的检测结果，但是面对大面积、特征多样、分辨率较高的城市遥感影像时，应用效果可能不佳。融合流程优化思维，整合现有成熟的遥感影像变化检测相关技术，利用数字城市建设中积累的大量高精度GIS数据，并结合地物形状特征指数和检测人员的作业经验，进行变化检测自动化研究，包括人机交互检测和批量自动检测两个主要流程，并应用于深圳市土地卫片执法中的土地利用变化图斑提取环节，可提高其自动化程度，有效降低时间和人力成本，及时发现并阻止土地违法利用行为。

Automatic Change Detection of Remote Sensing Image Integrated with Process Optimization

At present, the method for extract change spot with remote imagery, in the law enforcement of land use, is mainly making full use of the operators' visual interpretation of the two high-resolution remote sensing imageries, to get the spatial position of the land use change. Therefore, whether the spots are right or not, it is easy to generate errors, depends by the experience of the interpreters. With general detection process, there is a good result for the particular remote sensing data. But handling the city's remote sensing image on a large scale, with various characteristics, and higher resolution, this detection method seems to be limited and hard to be applied. This paper presents a remote sensing imagechange detection process which includes two main parts:remote sensing image human-computer interactive change detection and batch automatic detection. The system has integrated thought of process optimization and the existing remote sensing image change detection methods, and used a large number of high-precision GIS data accumulated during the development of digital city, which is fully combined with the features' shape index and people's working experience. It has been applied to the extraction of land use change spot in law enforcement inspection of satellite land image in Shenzhen, improved the level of automation, and reduced the time and labor costs effectively, which is helpful for the related government to find andprevent the occurrence of illegal land use in time.