存在运动目标时的图像镶嵌方法研究*

针对在图像镶嵌过程中，存在运动目标时所导致的虚影现象及明显拼接缝的问题，提出了一种图像无缝镶嵌方法。该方法在结合支持向量机处理运动目标的基础上，通过尺度不变特征点的提取以及变换模型参数的更新，得到更加精确的变换矩阵。另外，在图像融合过程中，最优拼接缝的寻找综合了颜色、纹理及结构等多方面的信息。实验表明，该方法能够很好地消除虚影现象和拼接缝隙，得到高质量图像，从而获得无缝镶嵌的效果。

Research on image mosaic method with moving object *

For the phenomenon of the virtual and obvious problems of the seam in the process of image mosaic, it proposed a seamless mosaic image method. This method based on the combination of SVM to deal with moving target, through the scale invariant feature point extraction and update of transformation model parameters to get more accurate transformation matrix. In addition, during the image fusion, it considered color, texture and structure information into the optimal joint search. Experimental results show that, this method can effectively eliminate the phenomenon of virtual shadow phenomenon and splicing gap, and get high quality images to have a seamless Mosaic effect.