Small Scale Extraction Based On TM Remote Sensing Image Segmentation

One of the most important contents of forest resource investigation system is the investigation of forest resources planning and design.This is due to the accuracy of the spatial distribution of forest area,the accuracy of the area measurement and the reliability of the final resources,which are directly affected by the division of small classes and the drawing.In the actual investigation,3 methods are used in small class division:the use of aerial photographs,topographic maps,and mapping.A common method of three kinds of drawing small,large labor intensity,small boundary displacement and small area size is not accurate at present,with the development of computer science and technology and geographic information system and its wide application,digital terrain data to replace the traditional paper map has become a reality.Therefore,it is of special significance to study the terrain features extracted from digital terrain data automatically and accurately.A method to extract slope and topographic map of the slope,and some experts and scholars have done some research,mostly concentrated in the algorithm itself,on how to use these methods for forestry survey service,seldom discussed.In this study,the Heilongjiang area of Greater Khingan Range Province,Tahe Forestry Bureau as the research object,using Landset TM images as the basic data,using object-oriented multi-scale segmentation technique of Tahe forestry bureau class boundaries can be extracted,and compared with the data of two small classes of forest resource survey in 2008,the results show that the TM data of this study the use of DEM,and the slope based on multi-scale segmentation to auxiliary extraction method is better,with small extraction can reach 85%.As early as the compartment boundary line of control of image segmentation,remote sensing image of each Linbannei corresponding to the combination of multi-scale image segmentation method,the best extraction effect of subcompartment boundary extraction,using the method of coincidence degree can be further improved.The extraction results in further editing can be used as a working map of subcompartment division.