：An IR/Visual image matching algorithm based on line features and SIFT points is proposed.Firstly,line segments are extracted from the images by using line segment detection algorithm.Then the most visually prominent lines are reserved after handling the results.Gaussian scale space is introduced and description matrix is set up by collecting gradient column vectors of line support region.The line descriptor is obtained by computing the mean and standard deviation,and line features from IR/visual images are made to match.The cross points of the matched lines are used to make SIFT description and matching.Finally,the mismatching lines are removed.Experimental results show that the proposed algorithm is highly accurate for IR/visual images matching,especially when applied to images containing obvious linear contour.