鍏抽敭璇�: 闈㈠悜瀵硅薄褰卞儚鍒嗘瀽, 鏁版嵁鎸栨帢, 鍒嗗壊鍒嗙被灞傛浣撶郴, 鍒嗗壊浼樺寲, 闅忔満妫灄

Abstract: With the wide application of high-resolution satellite锛孫bject based Image Analysis (OBIA) has gradually become main stream of extracting land cover information.Segmentation optimization is a fundamental step in OBIA.Different land cover types usually have different optimized segmentation parameters.How to make full use of the optimal Multi-Resolution Segmentation (MRS) to establish a segmentation classification hierarchy and to achieve high-precision land cover mapping锛宨s a challenge in object-oriented image analysis.based on the optimal segmentation parameters of different land cover types锛宼his paper explores a method to construct a segmentation optimized hierarchical classification system based on the minimum optimized segmentation unit.Experiments show that this method can effectively reduce the dependence on the operator's personal experience when setting up the classification hierarchy system锛宨mprove the classification accuracy锛宎nd meet the requirements of automatic drawing.

Key words: Object based Image Analysis(OBIA), Data mining, Segmentation classification hierarchical system, Segmentation optimized, Random forest