A Method for Information Measurement of Classification Maps
Various remote-sensing thematic maps are produced based on primary classification maps. Quality of the thematic maps is closely related with the classification accuracy of the primary classification maps and the information amount contained in them. The existing literature hasnt paid much attention to the information measurement of classification maps. From information theory, this paper has explored the unified mathematical foundation for measuring the spatial uncertainty and property uncertainty, and the emphasis has been put on the method for measuring the information of classification maps. Based on the information theory, this paper has respectively discussed in detail the methods for measuring the information of vector polygons, vector lines and vector points by adding constraints. Finally the information calculation of the vector classification maps is converted into information calculation of key points. In view of the fact that information representation is isomorphic, de-redundancy information measurement of grid classification maps is converted into information measurement of vector key points so that a method for measuring information of de-redundancy grid classification maps is obtained. In order to verify the proposed method, this paper has explained the practical significance of this method from the perspective of encoding amount of image data. In this way, it has provided some reference to the information measurement of maps.
Information Measurement Classification Map Information Entropy
Yan Chen Kaimin Sun
Information Institute HUST WENHUA College wuhan, China LIESMARS Wuhan University Wuhan, China
国际会议
太原
英文
551-556
2011-02-26（万方平台首次上网日期，不代表论文的发表时间）