图像拼接的核心是图像的配准.提出一种准确、高效的图像拼接方法:利用Harris角点检测算子提取相邻图

像的特征点,通过Euclid度量改进熵变换的算法,进行特征点匹配,并通过最小均值二乘估计求解相邻图像的变换

矩阵,以此实现宽视角图像的拼接.实验结果表明:所给出的算法能够实现场景图像的精确匹配.

-Image mosaic is the core of image registration. Make an accurate and efficient method of image mosaic: Using Harris corner detection operator extracting feature points of adjacent images, through the improvement of Euclid metric entropy transform algorithm, the feature point matching, and by minimizing the mean estimate for solving二乘adjacent image transformation matrix, as the realization of wide viewing angle image mosaic. The experimental results show that: given algorithm to achieve accurate matching scenes images.