基于双遍历和膨胀滤波的图像拼接定位算法

A Image Stitching Positioning Algorithm Based on Double Traversal and Expansion Filter

针对图像拼接操作会对边缘像素的连续性造成破坏等问题,本文提出了一种基于双遍历峰值和膨胀滤波的彩色图像拼接定位算法。该算法首先对待检测图像进行行列差分,确定双遍历峰值,同时再进行R,G,B三通道拼接点融合,从而全面获取拼接点信息,并利用数学形态学的膨胀滤波,强化拼接边缘,削弱噪点的影响。仿真结果表明,彩色图像拼接定位算法,时空复杂度小,能够有效地检测图像拼接伪造区域、大小及形状,可以对伪造区域进行准确定位。该研究可应用于图像取证等相关行业。

This paper puts forward a double traverse peak color image stitching positioning algorithm and filtering,for image stitching operation will damage to the continuity of edge pixels. The algorithm firstly treats differential detection image procession and determine double traverse spikes then it treats R, G, B three channel fusion splicing point, comprehensive information for splicing point; Finally, the expansion of the use of mathematical morphology filtering reinforces stitching edge, and weakens the noise effect. The simulation results show that the time and space complexity is small, algorithm can effectively detect the image splicing forgery area, size, and shape, and it can accurate positioning for fake areas. Itcan be used in image forensics.