面向地理国情监测的植被区域变化发现研究--以重庆市璧山区为例

VEGETATION REGIONAL CHANGE DISCOVERY FOR NATIONAL GEOGRAPHICAL CONDITIONS MONITORING—TAKING BISHAN DISTRICT IN CHONGQING AS AN EXAMPLE

植被变化发现是地理国情监测中的重点工作之一,以重庆市璧山区为例,设计一种顾及前期成果的变化发现技术方案。根据地类编码,对前期矢量成果进行归类合并,并提取出前期植被区域矢量;在植被区域矢量图的约束下,对后期影像进行分割,并以该分割线为基础获得前/后时期影像的相应图斑对;提取图斑对的两种植被指数特征,采用欧式距离作为特征的相似性度量;自适应变化阈值取特征向量距离的3倍中误差。实验结果表明,该技术方案针对植被的变化发现精度较高,可应用于生产。

Vegetation change discovery is one of the key tasks in national geographical conditions monitoring.Taking Bishan district of Chongqing as an example,a change discovery technology scheme is designed,which takes into account the previous results.According to the terrestrial coding,the early-time vector results were classified and merged,and the pre-vegetation region vector was extracted;under the constraint of the vegetation area vector,the late image was segmented,and we obtained the corresponding map pairs of the images in the pre/post period based on the dividing line;two vegetation index features of map pairs were extracted,and Euclidean distance was used as similarity measure;the adaptive change threshold took 3 times of root mean square error of the feature vector distance.The experimental results show that our scheme is highly accurate for vegetation change discovery,and can be applied to production.