Robust Method for Panorama Mosaic Based on Block Matching in ROI
Ling Zhong · Jisuanji yingyong yanjiu · 2007
This paper presents a robust method for constructing a panorama from images.Firstly,in order to mitigate the variation of the illumination conditions,the histograms of the overlapped areas are equalized.Secondly,extracting the seed points,the highly textured point in the overlapped area,by using phase correlation.Because ROI is restricted within overlapped areas,more accurate correspondences are obtained.Finally,the weighted Block Matching Algorithm(BMA) is used to minimize image distortion caused by camera rotation.Comparing the performance of the methods which is Peak Signal to Noise Ratio(PSNR) with multi-types images,results demonstrates the robustness of the method.