基于改进ASIFT的可见光图像拼接研究

Visible Image Mosaic Algorithm Based on Improved ASIFT

为了解决目前基于特征点的图像拼接算法在图像重叠度较低情况下,图像拼接效果差以及无法满足实时图像拼接的问题,提出了一种基于改进仿射—尺度不变特征转换(Affine Scale Invariant Feature Transform,ASIFT)的快速图像拼接算法,在特征点匹配过程中引入主分量分析(Principal Component Analysis,PCA)法进行处理,提出了一种PCAASIFT描述符对特征点重新进行描述。实验结果表明,与基于SIFT和SURF的拼接算法相比,该算法实现了高精度拼接,并且比传统ASIFT拼接算法提高了拼接的速度。

In order to solve the problem that image mosaic algorithm based on feature points causes the image stitching difficult andtakes highly computational cost,when the degree of overlap of two pictures is low, a fast image mosaic algorithm based on improvedASIFT is presented.PCA method is introduced into feature matching and a PCA-ASIFT descriptor is presented to describe the featurepoint.Experimental results demonstrate that this algorithm has better performance than the mosaic method based on SIFT or SURF, andreduces image mosaic time than the mosaic method based on traditional ASIFT.