摘要:
In order to effectively improve the speed and accuracy of the image matching algorithm,this paper proposed an image matching algorithm.First,in order to get the contour of the image which can be used as the feature area,the Log-polar transformation was carried on on the reference image and the target image respectively.Then,project the contour area on the orientations of distance axis and angle axis separately.Finally,compare the projection value to determine whether the images match.In order to reduce the computation time which is brought by calculating the whole image,this paper proposed a method of sampling by step.Experiment results show that,this algorithm executes quickly,and it is robust on resisting the noise or JPEG compression.The images can be matched accurately,though the target image is rotated or zoomed relatively to the reference image.
展开