摘要
以河北省秦皇岛市山海关公益林为研究对象,结合Landsat TM数据和森林资源二类调查数据,运用灰度共生矩阵分析法提取纹理信息,采用逐步回归法建立多元线性回归模型,进行森林蓄积量的估算。结果表明:选取纹理因子参与建模,建立的线性回归方程的拟合效果较好,估测模型的R^2值达0.766,估计值的标准误差最小,标准误差最小值为28.036,说明纹理因子对提高森林蓄积量的估测精度有重要影响。
With Shanhaiguan Public Welfare Forest of Qinhuangdao City in Hebei Province as the research area,we extracted texture features by gray level co-occurrence matrix( GLCM) based on the forest resource inventory data for management and Landsat TM images,and the multiple linear regression model was established by stepwise regression method to estimate the forest volume. The fitting result was better if selected texture factors to establish model with R^2 of 0.766,the smallest standard error of the estimated value,and the minimum error of standard error of 28.036. Therefore,the texture factor has important influence on the estimation precision of forest volume.
出处
《东北林业大学学报》
CAS
CSCD
北大核心
2017年第11期21-25,共5页
Journal of Northeast Forestry University
基金
国家重点研发计划项目(2017YFD0600904)
国家高分专项(30-Y20A37-9003-15/17-3)