论文部分内容阅读
以天绘一号和资源一号两期遥感影像为研究对象,采用面向对象多尺度分割和基于特征值阈值的方法提取采伐迹地,将研究区分成采伐迹地和非采伐迹地两类,使用分类后比较法检测森林覆盖变化。结果表明:国产的天绘一号和资源一号遥感影像有较好的内部几何一致性,正射校正后的两期影像可以达到1个像元内的校准精度。遥感变化检测的面积精度和重合率与实际变化相比较,分别为93.1%和95.0%,该方法能较好地检测出森林覆盖变化,适合于县域森林资源年度更新。
Taking the two remote sensing images of Tianyu-1 and Huayuan-1 as research object, the object-oriented multi-scale segmentation and eigenvalue-based method were used to extract the deforested land. The study was divided into two types of deforested land and non-deforested land. Comparative method to detect changes in forest cover. The results show that there is a good internal geometric consistency between domestically produced Tianyu One and Resource One remote sensing images, and orthorectified two images can achieve the calibration accuracy within one pixel. Compared with the actual change, the precision and coincidence rate of remote sensing change detection are 93.1% and 95.0% respectively. This method can detect the change of forest cover well and is suitable for the annual update of forest resources in the county.