An Application Study Of TM Image In Forest Management Investigation

Forest resource is material foundation of human society sustainable development. It is the renewable resources in the economic construction and ecological environment construction, having the ecological function of water and soil conservation, climate regulation. Forest resources situation and its dynamics change, not only affect the region's sustained economic development, but also affect regional and even global environmental changes. Therefore, it has attracted attention extremely.Remote sensing technology has the macro, dynamic, efficient, low cost and the cycle repeated many advantages. It is very suitable for monitoring the status of forest resources. TM data has extensive application in the survey of forest resources at present.In this research ,we chose the fogang city of Guangdong province as the study area. With the help of remote sensing data of TM and the field survey data of forest department, some key problems are discussed about geometric correction, image processing, band combination selection, the means of classifications for image , vegetation index extraction and forest volume estimate.The main contents and conclusions are summarized as follows:(1)Comparing the study area TM images of the band correlation and Analysis of the spectral characteristics of forest types, Through qualitative analysis and value of the OIF, the result shows : TM 3-4-7 in remote sensing imaging research forest type is a best combination, with the largest amount of information and the least redundant information.(2) The tests of supervised classification have been done to TM image.The methods are maximum-likelihood and SVM. The result shows: with the method of maximum-likelihood Classification the coefficient of KAPPA is 0.5659, the total precision is 63.2058%. with the method of SVM Classification the coefficient of KAPPA is 0.5867, the total precision is 65.0880%. Classification of TM there exists great disparity on the area between the forest type and the data of field survey. Especially in the conifer, the broadleaf and the confuse of conifer with broadleaf. The results can not achieve requirements of application.lt has a good separation of the forest and no-forest ,water, construction, farmland. It can be used for the ecological survey and design in forest management.(3)Six vegetation indexes were extracted in the study area.The relationship was analyzed between the vegetation indexes and Volume. The result shows:Vegetation indexes can generally shown Volume trend, but all kinds of vegetation indices reflect the volume of small differences in ability.After atmospheric correction,the RVI has good response in high vegetation coverage Forest.(4)Estimates volume from a stratified sample with unequal probabilities in the study area The result shows :the samping was feasible and efficient with∑RVI as a auxiliary. In the same sampling accuracy, it can be greatly reduced number of samples.It's a improvement in systemic sampling at present.