基于灰度特征编码的快速景象匹配方法

Fast Scene Matching Method Based on Gray Feature Coding

为解决复杂条件下快速目标匹配的问题,提出一种基于分形编码思想的景象匹配方法。基本原理是将图像分成小块,通过各个块灰度值之和的比较进行编码,提取图像中与分辨率无关的灰度特征,得到特征编码矩阵,再采用NNPROD算法作为相关性度量函数进行匹配。实验结果表明,该方法使匹配速度大幅提高,对不同分辨率、各种噪声的图像均能快速准确匹配,体现了很好的鲁棒性和适应性。

In order to solve the fast target matching problem in complicated condition, a fast scene matching method based on feature coding is presented. The basic principle is that the image is divided into pieces of the same size, coding is done by comparing the gray value of each piece in image, then the gray feature of image which is not related to resolution is extracted, and the feature coding matrix is getted. Then the NNPROD correlation measure function is used in matching. The experimental results show that this method greatly increase the matching speed, and it can get exact the / matching result with the image of different resolution and all kinds of noise, which shows better robustness and adaptability.