基于相位相关算法的研究与实现

相位相关算法是一种应用很广泛的图像匹配方法，它具有对噪声有较高的容忍度、检测结果与照度无关以及受几何失真小等突出优点。它与极坐标变换相结合可以得到旋转不变相位相关算法，两者可以对目标进行判准并精确检测出平移量和旋转量。本文对它们进行了分析和仿真，证明了算法的优越性，并在仿真过程中对旋转不变相位相关算法进行了优化，且达到了提高匹配精度和减少计算量的目的。

Study and implementation of phase correlation algorithm

Phase correlation algorithm as a kind of image matching method has many advantages such as insensitive to noise, unrelated to illumination of target, immune to geometric distortion, and is used widely. There is the rotation-invariant phase correlation algorithm which combines polar coordinate transform with phase correlation, they can recognize the correlated image and detect translation and rotation angle between two correlated images. This paper provides the analysis and simulation for them, and the superiority of the algorithms is testified. Furthermore, it improves the matching precision and reduces the calculation by optimizing the rotation-invariant phase correlation during the simulation.