基于改进混合蛙跳算法的图像阈值分割算法

摘要

针对最大类间方差法在图像分割时存在造成噪声干扰和过分割的缺点,提出一种基于改进混合蛙跳算法的图像阈值分割算法. 算法将苹果图像编码处理,选取图像的类间方差作为改进混合蛙跳算法的适应度值,通过改进的混合蛙跳算法寻找最大的分割阈值,利用该最优阈值使用经典最大类间方差法对花牛苹果图像进行分割. 选取强光、较强光、较弱光和弱光条件下四幅花牛苹果图像进行分割实验,结果表明,采用基于改进混合蛙跳算法的图像阈值分割算法较最大类间方差法和基于混合蛙跳算法的图像阈值分割算法均具有较好的图像阈值寻优能力,可有效改善花牛苹果图像的分割效果.%To solve the defect of noise interference and over-segmentation the OTSU has in image segmentation, we propose an image threshold segmentation algorithm which is based on improved shuffled frog leaping (ISFL) algorithm.By coding the apple images, the algorithm selects be-tween-class variance as the fitness value of ISFL algorithm and searches the maximal segmentation threshold through ISFL, and this optimal threshold is then employed to segment Huaniu apple images with OTSU.By choosing four Huaniu apple images under four conditions of strong light, faintish strong light, faintish weak light and weak light for segmentation experiments, the results indicate that the image threshold segmentation algorithm based on ISFL algorithm has better image threshold optimisation performance compared with both OTSU and the image threshold segmentation algo-rithm based on shuffled frog leaping algorithm, it can effectively improve the segmentation effect of Huaniu apple images.