A New Method of Bridge Target Detection on ASAR Imagery

摘要

The result of linear target detection is advantageous to the content analysis and interpretation of ASAR imagery and can be regarded as the imagery matching reference as well as the GIS platform input to realize the map automatic update. On the foundation of the previous studies, the method of bridge target detection based on Combined wavelet transformation and mathematical morphology analysis is proposed according to the good local characteristic, multi-resolution and effectiveness of wavelet transformation in this article, and the result is revised by combining the fine characteristics of different methods, which is applied to ASAR data of Zhaoqing and Zhuhai of China for carrying on the bridge detection as well as the comparative experiment with classical methods, such as Sobel and Canny etc detection operators. Results indicate that the combined methods can detect the target bridge perfectly on the direction and size, at the same time can restrain the target missing and false warning.