基于细菌觅食与改进Otsu算法的自适应阈值边缘提取

摘要

针对Canny算子在阈值偏高或偏低的情况下会产生边缘丢失或伪边缘的缺陷，提出了一种基于细菌觅食算法与改进最大类间方差法（Otsu算法）相结合的自适应阈值Canny算子。根据图像的梯度幅值直方图将像素点进行分类，并基于改进的Otsu算法定义描述类间方差的适应度函数，通过细菌觅食算法自动获取使适应度函数最优的高低阈值。实验结果表明，该方法在目标边缘提取的准确性上优于传统算法。%In order to fix the defects of broken edges or fake edges while the thresholds are too high or too low in original Canny algorithm, this paper proposes a new algorithm for Canny edge extraction with adaptive thresholds based on Bacte-rial Foraging Optimization algorithm(BFA)and enhanced Otsu algorithm. According to the value of gradient amplitude histogram, pixels are divided into edge type, non-edge type and pending type;A fitness function is defined to describe the mean square error among the three pixel type based on the enhanced Otsu algorithm;Both of the low and high thresholds are selected with the process of BFA. The experimental results show that this method is more accurate than original methods in edge extraction.