1 引言图像拼接是指将具有部分景物重合的图像进行无缝拼合,生成一张具有较宽视角的高分辨率图像的技术.现有的图像拼接技术主要有两类:基于区域的方法和基于特征的方法.基于区域的方法[1]通常是取参考图像中某一窗口中的点作为特征模板,建立描述
A coarse-fine method is developed to register images, which uses the features of both time domain and frequency domain. Firstly, the coarse relationship between images is estimated by Fourier transformation in frequency domain. Secondly, matching feature points in time domain are found. Finally, accurate transformation is derived and images are stitched together. Lab experiment proves that the method is efficient, which stitches images seamlessly, without any manual intervention.