Estimation And Retrieval Of Forest Volume By Remote Sensing Based On KNN

Since 1970, space remote sensing technology has been used in forest resources survey. How to use '3S' technology to estimate and retrieve forest volume has become a hot spot in the field of forest resources survey.Using remote sensing for the estimation and retrieval of forest volume is the modeling technology which establishes the mathematical model between the survey data and remote sensing image and gives the expression of the number of forest volume(estimation)and distribution(retrieval)with the promotion of the remote sensing,computer technology and the mathematical physics model and with the combination of satellite image data and a small amount of ground survey data(or provisional information on the ground plot). ,Traditional methods of estimating forest volume, including multipal regression, PLS,the optimal regression equation and so on. But all of them contain a serious problem, that is, they are unsuitable for the estimation of forest volume if a multiple collinearity exists in RS and GIS information which has an impact on the forest volume estimation,however,KNN can solve this problem.KNN method, as a non-parametric method, has been widely used in the multi-source surveys and the estimation and retrieval of forest volume. At home,the method of KNN is less used in the estimation and retrieval of forest volume in the boreal and sub-boreal regions, even not researched in the subtropical regions.This paper takes Wengyuan county in Guangdong province as the study field. KNN is firstly applied to the estimation and retrieval of forest volume in subtropical regions of China.The results show that:the accuracy of remote sensing image classification and the estimation and retrieval of forest volume based on KNN parameter optimization is obviously better than the traditional means of classification and the estimation of multipal linear regression. KNN can be successfully used in the estimation and retrieval of forest volume in subtropical zone.