Multi-feature model analysis-based target identification for remote sensing image

摘要

Recently, target identification has become a key topic due to the richness of nature resources contained in high-resolution remote sensing image. Many algorithms have been proposed in this aspect. In order to reduce algorithmic complexity and shorten computing cost, a multi-feature analysis model is proposed to identify targets in remote sensing images. The originality of the method includes two aspects. (1) City planning diagram, which is a vector file in the term of polygons of indispensable attributes, is used for image partition. (2) Multiple features are modeled, and then an identification rule is developed based on the model and minimum distance classification. Experimental results show that the proposed method is reliable in performing target identification.