HJ-1星多光谱影像火烧迹地信息提取研究--以百色市为例

Study on Information Extraction of Fire Scars Based on Multi-spectral Image of HJ-1 Satellite:A Case Study of Baise City

为识别火烧迹地等地类,以广西百色市为研究区,采用HJ-1星多光谱影像数据近红外波段光谱值、林火发生前后两时相各自NDVI值以及NDVI变化值,基于先验知识和统计分析构建决策树分类模型,通过与传统最大似然分类提取结果的比较分析,表明基于多特征的决策树模型能够有效地对HJ-1星多光谱遥感数据进行火烧迹地等地类提取,在研究区并具有良好的推广性。

In order to extract the fire scars information from HJ-1 satellite multispectral image data,with Baise city as the study area,the four features,i.e.the spectral value of near-infrared band,NDVI of bi-temporal HJ-1 image before and after forest fire,and the changed values both of them,were extracted and used for the classification of the fire scars.On the basis of experiments,the decision tree model was designed based on prior knowledge and statistical analysis.A comparison of the classification results with the results of Maximum Likelihood Classifier shows that the decision tree,which uses multiple features,is higher and more effective in forest fire classification from HJ-1 satellite image.