基于特征点的图像拼接方法

提出了一种基于特征点匹配的全景图像拼接方法.该方法首先利用sift算法提取各图像中的特征点并利用Harris算法对图像特征点提取进行了优化,然后采用基于K-d树的BBF算法查找和确定初始匹配点对,完成特征点的粗匹配,再根据图像配准结果使用稳健的RANSAC算法对粗匹配的特征点进行筛选,计算出图像间变换矩阵H,最后采用渐入渐出的加权平均的融合算法对两幅图像进行无缝拼接,形成一幅完整的全景画面.实验结果验证了该方法的有效性,拼接效果较好.

Image Mosaic Technology Based on Feature Point Match

This paper proposes a panoramic image mosaic method based on feature points matching. This method firstly uses the sift algorithm to extract the image feature points and uses Harris algorithm to optimize the image feature extraction. Then the BBF algorithm based on K-d tree is used to find and determine the initial matching points and complete the coarse matching of the feature points. Then according to the result of image registration, robust RANSAC algorithm is utilized to filter coarse matching feature points and calculate the transformation matrix H. Finally, the gradually fading out fusion algorithm of the weighted average is used in the seamless Mosaic of two images, form a complete panoramic view picture. Experimental results verify the effectiveness of the proposed method and the splicing effect is better.