Study On Estimation Of Forest Volume Based On TM Image

The forest stock volume is to reflect one of the basic indicators of the total scale and the level of a national or regional forest resources, but also reflects the rich forest resources.It is an important basis to measure the merits of forest ecological environment. Traditional forest volume survey method requires a lot of intensive labor and investment in property. In this study,the study area is Daqingshan in Hohhot of Inner Mongolia.This study mainly used TM images and the forest resources survey data to established remote sensing estimation of forest volume. Geometric correction and atmospheric correction of remote sensing images has been made and extract images average reflectivity as remote sensing factors. Finishing the two types of survey data extracted forest separation of the conditions of factors, the above factors as independent variables of remote sensing estimation of the volume of Pinus tabulaeformis plantation related, the main contents and conclusions of the study are summarized as follows:(1)Correlation analysis and stepwise regression analysis has been done on17factors. The results show that the model which removed out of the larger correlation factors is better.(2)The volume estimation model is:y=-125.656＋188.995x7＋3.814x8-19.122x9＋3.842x10＋102.92x16(3)40samples volume has been used to tested the estimation model, the average estimation accuracy is86.09%.