摘要
针对室内环境下利用传统加速鲁棒特征(SURF)算法进行图像实时拼接效率不高,以及室内环境图像快速拼接的要求,提出了一种基于改进SURF算法的室内环境图像快速拼接方法。首先,采用SURF算法对获得到的室内环境图像进行初步特征点提取,同时利用双向K最近邻分类算法筛选提取到的SURF特征点,得到粗略的匹配点对;其次,采用随机采样一致性(RANSAC)算法更新匹配点对,剔除被错误提取的匹配点对;最后,采取加权平均融合算法对配准完成后的图像进行融合拼接。通过该方法得到了完整的室内环境拼接图像,并提高了图像匹配准确度和拼接效率。从实验结果来看,提出的改进方法能极大降低室内图像的匹配错误,提高了室内图像拼接效率,同时能够取得较好的拼接效果。
In order to solve the problem of poor effect on indoor environment image mosaic caused by Speed-Up Robust Features （SURF） algorithm and meet the demands of the fast indoor environment images matching,in this paper, the improved algorithm of SURF is proposed. First,the feature points are detected by SURF algorithm,using bidirectional k-nearest neighbor algorithm to filtrate the feature points to obtain the fuzzy matching points pair. Then, the error feature points are eliminated by RANSAC algorithm. Finally, the weighted average method is used to fuse and join the mosaic images. Through the above steps, the final images have been obtained and the accuracy and efficiency of the images matching has been improved. The experimental results show that the algorithm proposed in this paper can reduce the matching error,increasing the images mosaic efficiency and acquiring the good stitching effect.
出处
《计算机技术与发展》
2015年第8期39-42,47,共5页
Computer Technology and Development
基金
国家自然科学基金资助项目(51375477)
关键词
室内环境
图像拼接
加速鲁棒特征
随机采样一致性
最近邻分类
indoor environment
image mosaic
speed-up robust features
random sample consensus
nearest neighbor classification
