The Application Research Of Object-oriented High Resolution Remote Sensing Image Classification In Forest Volume Estimate

High spatial resolution remote sensing images contain abundant structure and semantic information,with high scores on object-oriented remote sensing image classification technology uses image pixel around several homogenous pixel of image objects, can make full use of the structure and semantic features of space which the image object contains, and construct a classifier to classify. Compared with the traditional needle objects classification method, the classified method of object-oriented classification effect is better for high resolution remote sensing image. Using the monitoring area of the existing fixed sample plot survey data to establish estimate model, and the area of land use types of woodland accumulation to estimate reservoir distribution is the method to estimate forest volume. So the remote sensing images of monitoring area need to be classified, in order to distinguish the scope of the woodland. Because of forest land types include: coniferous forest, broad-leaved forest and mixed needle, such as different types, different forest types of volume is put in bigger difference. In the forest volume estimating, forest category of main categories accumulation estimate equation respectively, then the monitoring area by forest category type of remote sensing image classification, the accumulation of various forest category corresponding to estimate equation was used to monitor area on the remote sensing image matching precision forest category distribution area, can improve the forest stock volume estimation accuracy. Remote sensing images with high marks in the survey application in the field of forest resources in our country, using the object-oriented classification technology of monitoring area of remote sensing image classification, on the basis of precision, can improve the accuracy of forest volume estimate, to kind of forest resources in our country and the second class survey precision and efficiency, has the vital significance.In this paper, I realize remote sensing image multi-scale segmentation, object-oriented classification and the main sorts of forest category forest stock volume estimation based on the C # and ArcGIS Engine secondary development component programming, and analyze the use of object-oriented remote sensing image classification method for covering the monitoring area of high resolution remote sensing image classification of forest type effect and points with the main sorts of forest stock volume estimation accuracy. It can effective implementation of small class precision forest and forest land type classification, the high resolution remote sensing image in forest resource survey and application of forest land change survey can play a role.