文献类型：期刊文献

中文题名：基于SPOT5遥感影像的小班区划技术

英文题名：Division of Forest Resources Based on SPOT5 Remote Sensing Images

作者：吴春争[1] 冯益明[2] 舒清态[1] 李增元[3] 武红敢[3] 车腾腾[3]

第一作者：吴春争

机构：[1]西南林业大学;[2]中国林业科学研究院荒漠化研究所;[3]中国林业科学研究院资源信息研究所

年份：2011

卷号：39

期号：7

起止页码：120-122

中文期刊名：东北林业大学学报

外文期刊名：Journal of Northeast Forestry University

收录：CSTPCD;;北大核心:【北大核心2008】；CSCD:【CSCD2011_2012】；

基金：林业公益科研专项(200704019);“863”项目(2007AA12Z181);云南省自然科学基金(2008ZC094M)

语种：中文

中文关键词：SPOT5影像;小班区划;多尺度分割;规则分类

外文关键词：SPOT5 images ; Sub-compartment division; Muhi-resolution segmentation; Classification rules

分类号：S771.8

摘要：采用自主开发设计的森林资源规划调查系统,在大的地性线与已有林班线为边界控制条件下,对常用于森林资源调查的SPOT5遥感影像进行了多尺度、多层次逐级分割,获取了影像对象。同时,综合运用分割对象的光谱、空间特征和纹理特征形成分类规则,在此基础上进行小班区划,试验结果表明,该方法得到了较好的小班区划精度。
SPOT5 remote sensing images commonly used in forest inventory were segmented in multi-level and muhi-scale. The image objects were extracted under controls of large terrain feature lines and compartment lines using the self-developed forest resource planning and survey system. Then, the classification rules were constructed based on the spectra of the objects and their geometric and textural features. Based on this, sub-compartment was divided. High accuracy of sub-compartment division was obtained.