基于决策树分类的大屯矿区地物信息提取及矿区污染分析

本文在对国内外遥感图像分类方法充分研究分析的基础上,选择决策树分类法对大屯矿区的Landsat 8遥感图像进行分类研究。选取样本提取并分析研究区典型地类光谱特征曲线,依据光谱曲线特征和归一化植被指数建立了土地利用分类决策树模型,通过反复试验和修正,筛选出适宜大屯矿区地物分类的决策树最优阈值,对研究区进行分类和精度评价,最后通过分类结果对研究区的水体污染状况进行简要分析。

Extraction of Features Based on the Decision Tree Classification and Analysis of Plo lution in Datun County

This article selected and used decision tree classification to classify and study Landsat 8 images of Datun Mining Area , on the basis of fully researching and analyzing Domestic and foreign remote sensing image classification method .Samples are chosen and subtracted to analyze the typical spectrum characteristic curve of the researched area and establish the land usage decision tree classifi -cation model according to the typical spectrum curve characteristic and NDVI .After repeated classifications and experiments , the opti-mal threshold value is picked out to analyze the preciously evaluate the researched area .Finally, brief analysis is conduct to water pol-lution in the researched area through the classification results .