A asynchronous point matching algorithm is proposed to resolve the problem of IR and visible images matching.Firstly

feature points are extracted by FAST -9 corner detector with Gaussian scale space.The main orientation of each point is assigned according to the neighborhood information.Then

a SIFT -like descriptor is applied to describe control points.Then

the Nearest Neighbor Distance Ratio is used to rough matching.After that

a shape context descriptor is constituted with the point neighborhood information of edge.Finally

the results of rough matching are purified by the corresponding similarity measure algorithm.Experimental results indicate that the proposed algorithm is invariant to image rotation

scale and occlusion changes.Even under special weather condition and different time

it still performs robust and gets a higher average correct matching ratio than SURF.