采用快速混合蛙跳算法的微光图像增强

Enhancement of low-light-level image based on FSFLA algorithm

由于探测器灵敏度的限制,激光雷达、夜视等图像各点的相对灰度较低。基于更快速的混合蛙跳算法(A Fast Shuffled Frog Leaping Algorithm,FSFLA)提出了一种图像自适应快速增强算法。该算法采用了一种具有更大搜索范围、更快的收敛速度的快速混合蛙跳算法,降低了运算时间。它应用于微光图像的处理方面上,较传统的SFLA算法能更快地达到图像增强的效果,更适用于实际应用场合。在给出24个初值的情况下,传统的SFLA算法须迭代平均20次才能达到稳定的效果,而FSFLA仅需6次,因而大幅度地提高了运算时间。

Because of the limit of the sensitivity of detector, the relative gray scale of radar image and night viewing image is usually very low. So it is necessary to improve the relative gray scale of radar image and night viewing image. An adaptive image enhancement algorithm was proposed based on FSFLA algorithm （A Fast Shuffled Frog Leaping Algorithm）. This algorithm was based on FSFLA with larger scan scope, faster convergence rate and shorter computing time. When the algorithm was applied to low light level image, the effect of image enhancement was faster to achieve compared with traditional SFLA algorithm （Shuffled Frog Leaping Algorithm）. As a result, this algorithm was more applicable in actual occasion. Under the condition of twenty four given initial value, traditional SFLA algorithm needs to converge twenty times to achieve stability in average while FSFLA algorithm only needs six times in average, thus greatly improving the computing time.