低信噪比图像的模糊增强算法

The fuzzy image enhancement algorithm for iow SNR image

传统的机器视觉方法应用于低信噪比图像的目标增强,增强结果容易受噪声影响,鲁棒性较差。本文提出一种基于最大模糊熵的图象增强算法,对不同复杂背景下的图像具有增强函数强度自动可调和运算速度快的优点。其算法的实现关键是利用遗传算法自动搜寻满足最大熵条件的模糊增强函数的最佳带宽参数。进化过程中运用实数编码,在父代染色体构成的超平面内运用高斯交叉、变异算子创建后代,并采用欧式距离自适应调节交叉、变异概率。通过实验证明,该算法能够有效的增强复杂背景下的微弱目标信号。

Based on the traditional machine visual methods, the enhancement results of low SNR image are easily affected by the noise. In this paper, a novel fuzzy enhancement algorithm is proposed based on the maximum fuzzy entropy principle. For the diverse complicated images, the algorithm has advantage of adjusting the enhancement strength automatically and computing fast. Searching the best band-width parameters of the membership function to meet the maximum en-trepy principle automatically is the key procedure to implement the algorithm. Namely, the real coding-based off-spring chromosomes are created by the Gaus-sian crossover and mutation operator from the hyperplane formed by the parent ones. And the crossover and mutation rates can be self-adapted by the Euclidean distance. The experiments verified that the weak signals submerged in the complex background can be enhanced efficiently by this algorithm.