论文部分内容阅读
综合面向对象和CART决策树方法,对浙江省安吉县山川乡毛竹林分布信息及胸径、树高、郁闭度等调查因子和地上部分碳储量进行遥感定量估算.结果表明:综合基于多尺度分割的对象特征及决策树,能够充分利用不同尺度层次信息关联的优势,实现毛竹林专题信息高精度提取,其用户精度达到89.1%;基于对象特征构建的毛竹林调查因子回归树估算模型,其估算结果均能达到正常或较好水平,其中,郁闭度回归树模型的精度最高为67.9%,估算效果较好;而平均胸径和树高估算的总精度相对较低,这与采用光学遥感数据进行森林树高、胸径估算达不到理想结果的结论一致;毛竹林地上部分碳储量回归树模型的估算结果较好,高值区域估算精度达到80%以上.
Based on the comprehensive object-oriented and CART decision tree method, the distribution information and the investigation factors such as DBH, tree height, canopy density and aboveground carbon stocks of Phyllostachys pubescens forest in Shanchuan Township, Anji County, Zhejiang Province were quantitatively evaluated by remote sensing. The results showed that: Which can make full use of the advantages of information association at different scales and achieve high accuracy extraction of thematic information of Maozhulin, and the accuracy of the user reaches 89.1%. Based on the object characteristics, the estimation model of the regression tree of the surveyed factors of Moso bamboo forest is obtained, which is estimated The results can reach a normal or good level, of which, the highest accuracy of the regression model of canopy regression tree 67.9%, the estimated effect is better; and the average accuracy of the average diameter at breast height and tree height estimation is relatively low, which is consistent with the use of optical remote sensing data The results of tree height and DBH estimation did not reach the ideal results. The estimation results of the aboveground carbon storage regression tree model of Phyllostachys pubescens forest were better, and the estimation accuracy of high value area reached more than 80%.