基于云遗传算法的图像相关匹配

Image Correlation Matching Based on Cloud Genetic Algorithm

针对图像相关匹配计算量大的问题,提出基于云遗传算法的图像相关匹配方法。考虑到图像平均量的存在会增加匹配的难度,对传统归一化相关测度进行修正。为寻找最佳匹配点,将修正后的相关测度作为适应度函数,采用云遗传算法进行寻优。由于云遗传算法具有收敛速度快、局部寻优能力强和不易产生早熟现象等优点,新方法的匹配精度和速度都得到提高,且抗噪声能力强。仿真实验结果表明,新方法对无噪声和有噪声图像都能实现高精度匹配,在匹配精度和速度上优于基于自适应遗传算法的匹配方法。

Considering image matching is a heavy computation task, this paper proposes a novel image correlation matching method based on Cloud Genetic Algorithm（CGA）. To avoid mean image value increases the difficulty of image matching, an improved normalized correlation measure is developed as a fitness function for searching the best matching point. Since CGA can converge fast to a high quality local optimal, the novel method＇s accuracy and speed are high, and it is robust to noise. Simulation results show that the proposed method can match both noise free images and noisy images with higher accuracy and higher speed than the adaptive genetic algorithm based matching approach.