基于全局和局部特征的图像快速拼接算法  

Fast Image Mosaic Algorithm Based on Global and Local Features

作　　者：徐奇[1] 张佑春[1] XU Qi;ZHANG Youchun(Anhui Business Vocational College,Hefei 230001,China)

机构地区：[1]安徽工商职业学院,安徽合肥230001

出　　处：《佳木斯大学学报（自然科学版）》2023年第6期73-76,共4页Journal of Jiamusi University：Natural Science Edition

基　　金：安徽省高校自然科学研究重点项目(KJ2020A1095)。

摘　　要：为了提升图像快速拼接效果,提出了基于全局和局部特征的图像快速拼接算法,利用改进形状上下文算法,提取图像全局形状特征,按照轮廓点序列间最小欧式距离完成图像粗匹配。通过关键点提取算法提取原始图像与粗匹配图像的局部特征,利用应用随机抽样一致性方法根据局部特征进行图像精匹配,通过渐入渐出融合算法融合精匹配图像拼接时重叠区域的边缘,完成图像快速无缝拼接。实验证明:该算法在图像模糊与视角变换时,均可有效快速拼接图像,拼接后的图像清晰度佳、细节信息含量多,获得理想的图像拼接效果。In order to improve the effect of fast image mosaic,a fast image mosaic algorithm based on global and local features is proposed.Using the improved shape context algorithm,the global shape features of the image are extracted,and the rough image matching is completed according to the minimum Euclidean distance between contour points.The key point extraction algorithm is used to extract the local features of the original image and the coarse matching image.The random sampling consistency method is used to carry out the fine matching of the image according to the local features.The edge of the overlapping area during the fine matching image stitching is fused by the gradual in and gradual out fusion algorithm to complete the fast and seamless image stitching.Experimental results show that the algorithm can effectively and quickly mosaic images when the image is blurred and the angle of view is changed.The stitched images have good clarity and more detail information,and the ideal image stitching effect is obtained.

关 键 词：全局特征 局部特征 图像快速拼接 梯度角度 抽样一致性 融合算法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]