论文部分内容阅读
以顺昌县为研究区域,探讨利用遥感影像提取我国南方地区竹资源专题信息的技术。通过分析竹资源与其它林地在TM影像上的光谱特征差异,采用最大似然法和叠加、掩膜等处理方法,实现提取竹资源专题信息的光谱片层分割。在此基础上,基于灰度共生矩阵,分析竹资源与其它林地的纹理特征差异,提出“基于光谱片层—面向纹理类”的竹资源专题信息提取技术。研究结果表明,采用该技术,竹资源专题信息提取精度可达84.8%,在我国南方地区具有较好的可行性和优越性。
Taking Shunchang as a research area, this paper discusses the technology of using remote sensing images to extract thematic information of bamboo resources in southern China. By analyzing the differences of spectral characteristics between TM images and other bamboo resources, the method of maximum likelihood, superposition and masking was used to achieve the spectral slice segmentation of extracting thematic information of bamboo resources. On this basis, based on the gray level co-occurrence matrix, the difference of texture features between bamboo resources and other woodland is analyzed, and the bamboo information extraction technology based on “slice layer-oriented texture class” is put forward. The results show that using this technique, the extraction accuracy of thematic information of bamboo resources can reach 84.8%, which has good feasibility and superiority in southern China.