关键词: 特征点检测, 视觉对比度方法, 图像增强算法, 距离比例不变性

Abstract: In the geological disaster monitoring system, the image change detection can be used to monitor the occurred disaster. Image matching is needed to change detection accurately. Aiming at the problem of low matching precision with the image in the fog condition, an image matching method based on image contrast enhancement of SURF is proposed, enhanceing the matching precision of the fog image efficiently. It uses multi-scale Retinex algorithm to enhance the fog image, and then completes the extraction of feature points with SURF algorithm, uses Euclidean distance to measure the similarity of characteristic points, according to the similarity triangular distance ratio invariance, it deletes point pairs of mismatch. Experimental results show that the method significantly improves the precision of image feature points matching, provides a good foundation for subsequent use of image change detection for disaster monitoring.

Key words: feature point detection, visual contrast measure, image enhancement algorithm, invariant-scale distance