摘要
图像拼接技术将存在重叠区域的多幅图像经过配准和融合后得到单幅宽视场图像。由于误差的积累,多幅图像拼接后在重叠区域会有明显的拼接痕迹,所以需要对拼接后的图像进行优化。首先研究了变换矩阵及其参数,然后提出一种图像对齐方法,完成图像拼接,最后用全局优化策略消除累积误差。实验证明,该方法在存在较大光照强度变化,重叠区域小的情况下能够鲁棒地完成多幅图像的拼接。
Multiple images with overlap region are registered and fused into a wide view image by Image stitching technology. Because of the error accumulation, there will be obvious stitching hint in the overlap region after multiple image mosaic, so it calls for optimization to the image after stitching. First, the transformation matrix and its parameters are studied, and then a method of image alignment is proposed to complete image stitching, at last global optimization strategy are used to eliminate the accumulated error. Experiments show that this method can stitch multiple images robustly with the exist of large intensity change, little overlapping region.
出处
《电脑知识与技术(过刊)》
2010年第3X期1966-1968,共3页
Computer Knowledge and Technology
关键词
图像拼接
图像对齐
全局优化
特征匹配
变换矩阵
Image stitching
alignment
optimization
feature matching
transformation matrix