基于TM影像的林地变化检测方法对比分析——以伊春市为例

Comparative Analysis with the Method of Detecting the Changes of Forests Resources Based on TM Image——A Case of Yichun

以伊春市为试验区,利用两期TM遥感影像数据对该地区的林地面积的变化进行了检测。实验使用了归一化植被指数差值法、主成分分析差异法和基于C5.0决策树的分类后比较法,并对三种结果做了对比。结果表明:1)1999年至2010年,林地总面积是增加的,阔叶林显著减少,混交林显著增加并发展成为主要林地类型。2)三种方法的精度分别为79.33%、82.54%、87.24%,说明分类后比较法能够更有效地对森林覆盖的变化进行检测,并且能够定性定量的对变化区域进行分析。

Taking the Yichun in Heilongjiang Province as a case,the changes of wood were detected in this study by utilizing double-temporal TM remote sensing data.Three methods,including Normalized vegetation index difference,principal component analysis difference,and classification based on C5.0 decision tree were applied and compared in this paper.The results shows the following：1）The area of forests had the overall increased tendency in Yichun from1999 to 2010.The area of broad-leaved forest decreased obviously,while the area of theropencedrymion increased obviously and became the principal type.2）The accuracy for three methods were 79.33%,82.54%,and87.24%,respectively,suggesting that the method of classification can significantly outperform the other two methods in detecting the changes of forests resources.Meanwhile,it can be employed to analyze the change areas qualitatively and quantitatively.