基于细菌觅食优化算法的自适应阈值边缘检测

摘要

Under the threshold high or low, Sobel operator will form the fake edge and the missing of edge. In order to overcome it, this paper proposes a new method which combines Bacterial Foraging Optimization(BFO) algorithm with Otsu algorithm, makes use of global optimization ability of BFO algorithm, uses the Otsu algorithm as the fitness function to search the best threshold, and automatically detects image edges. Experimental result shows that the method improves the retrieval speed of threshold and makes better effect of edge segmentation.%Sobel算子在阈值偏高或偏低的情况下会产生伪边缘或边缘丢失的问题.为此,将细菌觅食优化算法与最大类间方差相结合,利用细菌觅食优化算法的全局寻优能力,以最大类间方差为适应度函数搜索最佳边缘检测阈值,对图像进行自动边缘检测.实验结果证明,该方法能够取得较好的边缘分割效果,提高阈值检索速度.