二值化编码和相位相关的鲁棒景像匹配方法

Robust Scene Matching Method based on Binary Coding and Phase Correlation

结合二值化编码和相位相关,提出了一种快速有效的鲁棒景像匹配方法。首先利用2×2窗口上主对角线上像素灰度值的升降对实时图和基准子图进行二值化编码;然后,搜索二值化后的实时图在基准图中的位置,即,根据二值化后的实时图和基准子图的相位相关函数进行定位。多组实验与分析表明,该景像匹配定位算法对部分遮挡、图像灰度与对比度变化以及复杂噪声干扰等都具有比较强的鲁棒性。

By combining binary coding with phase correlation, a robust scene matching method is proposed. The sensed image and the reference sub-image are coded in binary numbers by the use of increment and decrement information of the pixels＇ values in the main diagonal of each 2 × 2 window. The binary-coded sensed image can be located in the reference image by computing the value of the phase correlation functions. Theoretical analysis and simulation experiments show the effectiveness of the algorithm in matching images with partial occlusion, intensity or contrast changes and complex noise corruptions, etc.