Abstract: In order to expand the viewing angle for robot visual navigation, this paper presents a fast image stitching algorithm based on ORB algorithm and OECF model. Firstly, feature points are extracted by SIFT algorithm after images pre-processing. Secondly, the binary descriptors of the feature points are obtained by ORB algorithm to achieve images matches. The project transformation model matrix between two matching images is calculated. Finally, to solve the color inconsistency problem of images stitching, the Opto-Electronic Conversion Function of the camera is estimated to conduct color adjustment. Experimental results show that the proposed algorithm is effective for image stitching, moreover the stitching quality and stitching speed are much better than other image stitching algorithms.

Key words: ORB(Oriented Brief) algorithm, OECF model, image stitching, color adjustment

摘要： 为了增加移动机器人视觉导航中的成像范围，提出了一种基于ORB算法和OECF模型的图像拼接算法，获取的图像进行预处理后，先利用SIFT算法提取特征点，再用ORB算法获取特征点的二进制码串描述子进行匹配，而后通过投影变换模型获取参数，最后估计摄像机的光电转换函数OECF（Opto-Electronic Conversion Function），利用Laguerre OECF模型参数对拼接图像进行色彩调整，解决了多幅图像拼接中颜色不一致的问题。实验结果表明，该算法能有效地对图像进行拼接，配准速度和拼接视觉效果和优于其他算法。

关键词: ORB算法, OECF模型, 图像拼接, 色彩平衡