基于C5_0算法森林资源变化检测方法的研究_以山东省徂徕山林区为例.pdfVIP

2011, 26( 5) : 185~ 191 Journal of Northwest Forestry University 5. 0 以山东省徂徕山林区为例 1, 2 1* 2 王志慧 , 李世明 , 张艺伟 ( 1. , 100091; 2. ( ) , 100083) : 以山东省徂徕山林场为试验区, 利用两时相的T M 与ET M+ 遥感数据对该地区的针叶 林阔叶林等 林资源的变化进行研究将基于 5.0算法的决策树分类方法应用于 林变化检 测, 并对3 种检测方案进行试验比较: ( 1) 以单一时相图像作为数据源并各自分类,分类后作比较提 取变化信息; (2) 以两时相图像的原始波段数据作为数据源训练规则, 并生成变化检测图; (3) 以两 时相图像加上邻近相关分析图像作为数据源训练规则, 生成变化检测图试验结果表明, 基于 5.0算法的决策树分类可以有效的进行 林变化检测, 并且加入邻近相关分析图像后的变化检测 精度达到最高 : 变化检测; 5.0; 决策树; 邻近相关分析 : S757.24 : A :2011) Methodsological Study on the Detection of the Variations of Forest Resources Based on 5. 0 Algorithm- A ase of ulai Forest in Shandong 1, 2 1* 2 WANG Zhihui , LI Shiming , ZHANG Yiwei (1. Institute of Forest Resources Inf ormation Techniques, ChineseA cademy of Forestry, Beij ing 100091, China; 2. China University of ining Technology (Beij ing), Beij ing 100083) Abstract:With doubletemporal T M and ET M+ remote sensing data, the information of the variation of forest resources of ulai Mountain in Shandong Province, hina was explored. Decision tree classification based on 5.0 algorithm to forest change detection was applied. T hree different detection methods were compared: 1) to classify singletemporal data by 5. 0 respectively, and extract change information after comparing classification results; 2) to create 5.0 train rules through doubletemporal raw data, then gen erate change detection map; 3) in addition to doubletemporal remote sensing data, neighborhood correla tion analysis images are also added as one of the data sources of 5.0, and generate change detection map of variation. T he experimental result shows that decisio