摘要
针对最大类间方差法在图像分割时产生的阈值会引起过分割的缺点,提出一种基于和声搜索算法的图像阈值寻优算法。算法将苹果图像预处理、计算图像目标和背景两类间方差作为和声搜索算法的目标函数,通过和声搜索算法寻找最优分割阈值,选取6种不同光照条件下的苹果图像进行阈值寻优实验。结果表明,采用基于和声搜索算法的图像阈值寻优算法较最大类间方差法和基于混合蛙跳算法的图像阈值寻优算法具有较好的图像阈值寻优能力。
To solve the over segmentation problem of OTSU threshold in image segmentation, the image threshold optimization algorithm has been proposed based on harmony search algorithm. By preprocessing the apple images, the algorithm computes the class variance between target image and background to get the objective function of harmony search algorithm for obtaining the optimal segmentation threshold. Six apple images under different illumination conditions have been selected for threshold optimization experiments, and the results indicate that image threshold optimization algorithm based on harmony search algorithm has much better image threshold optimization performance than OTSU and image threshold optimization algorithm based on shuffled frog leaping algorithm.
出处
《洛阳理工学院学报（自然科学版）》
2014年第3期48-50,共3页
Journal of Luoyang Institute of Science and Technology：Natural Science Edition
基金
国家自然科学基金(61063028)
中国博士后科学基金资助项目(2013M542398)
甘肃省自然科学研究基金计划(1308RJZA214
1208RJZA133)
甘肃农业大学盛彤笙科技创新基金资助项目(GSAU-STS-1322)
甘肃省高等学校研究生导师科研项目(1202-04
1102-05)
兰州交通大学青年科学基金项目(2013032)资助
