基于改进SIFT算法的地形测绘图像拼接方法

Terrain mapping image mosaic method based on improved SIFT algorithm

为解决传统方法拼接地形测绘图像容易出现匹配错误的问题,以改进SIFT算法为基础研究一种新的地形测绘图像拼接方法 .该方法通过对尺度不变特征变换(SIFT)描述子生成过程的分析改进,得到关键点的位置坐标;根据特征点检测模板中心点像素灰度的差值,提取地形测绘图像的特征点;计算目标匹配点的方向误差,通过构建直方图剔除错误的匹配点;使用高斯加权窗口对特征梯度值进行加权处理,获取特征描述子,对SIFT算法中的特征描述子进行降维处理,提高计算速度,以完成SIFT算法改进.并通过双向匹配配准测绘图像,根据待拼接地形测绘图像的线性过渡因子,计算重叠区域灰度值,实现图像重叠区域的无缝拼接.实验结果表明,与传统的拼接方法相比,改进后的方法能够准确提取待拼接地形测绘图像的特征点,拼接效果更优,耗时约为0.57 min,信噪比达到了32.61 dB.

In order to solve the problem that the traditional method of stitching topographic mapping images is prone to matching errors,a new method of stitching topographic mapping images is studied based on the improved SIFT algorithm.The generation process of sift descriptor is analyzed to obtain the position coordinates of key points.According to the difference of pixel gray of the central point of the feature point detection template,the feature points of the topographic mapping image are extracted.We calculate the direction error of target matching points,and eliminate the wrong matching points by constructing histograms.The Gaussian weighting window is used to weight the feature gradient value,obtain the feature descriptor,and reduce the dimension of the feature descriptor in SIFT algorithm to improve the calculation speed.After the improvement of SIFT algorithm,the mapping image is registered by two-way matching.According to the linear transition factor of the topographic mapping image to be spliced,the gray value of the overlapping area is calculated to realize the seamless splicing of the image overlapping area.The experimental results show that,compared with the traditional stitching methods,the improved method can accurately extract the feature points of the terrain surveying and mapping images to be spliced,resulting in better splicing performance.The splicing time is about 0.57 minutes,and the signal-to-noise ratio reaches 32.61 dB.