欧阳宁, 翟紫伶, 首照宇, 张彤, 袁华. 基于图切割的多频带融合图像拼接[J]. 微电子学与计算机, 2013, 30(7): 107-110.
引用本文: 欧阳宁, 翟紫伶, 首照宇, 张彤, 袁华. 基于图切割的多频带融合图像拼接[J]. 微电子学与计算机, 2013, 30(7): 107-110.
OUYANG Ning, ZHAI Ziling, SHOU Zhaoyu, ZHANG Tong, YUAN Hua. Iamge Stitching of Multi-band Blending Based on Graph Cut[J]. Microelectronics & Computer, 2013, 30(7): 107-110.
Citation: OUYANG Ning, ZHAI Ziling, SHOU Zhaoyu, ZHANG Tong, YUAN Hua. Iamge Stitching of Multi-band Blending Based on Graph Cut[J]. Microelectronics & Computer, 2013, 30(7): 107-110.

基于图切割的多频带融合图像拼接

Iamge Stitching of Multi-band Blending Based on Graph Cut

  • 摘要: 由于在视频帧间隔不大的图像拼接过程中,重叠区域存在运动物体和配准误差时,简单的融合会产生融合鬼影和拼接缝隙。针对上述两个问题,首先利用SURF提取特征点,经过配准求单应性矩阵得到图像的重叠区域,再通过基于网络最大流的图切割查找到全局最优的缝合线,最后使用多频带融合得到宽视野的全景拼接图。实验结果表明,该方法在消除鬼影和拼接缝隙方面具有很好的效果。

     

    Abstract: As the image stitching process between the small interval video frame,a simple integration will produce fusion ghosting and the stitching gap in the overlapping regions when there is the moving object and the registration error.for resolving these two problems,First,use of the SURF to extract the feature points,and the overlapping areas of the image to be gotten by homography matrix after the feature registration,and then there is the global optimum suture to be found by graph cut based on the maximum flow of network.the last,it is used to the multi -band blending to have a panorama stitching image with wide view.Experimental results shows that this method has a good effect in eliminating ghosting and stitching gap.

     

/

返回文章
返回