蚁群算法在低对比度图像边缘检测中的应用

Application of Ant Colony Algorithm in Edge Detection for Lower Contrast Image

蚁群算法应用于大多数图像边缘检测均具有抗噪声能力强、提取边缘精细等优点,但在处理含噪声的低对比度图像边缘时会出现边缘部分缺失、边缘不平滑等现象。为了对低对比度图像的边缘检测达到理想效果,文中通过对蚁群算法中信息素矩阵和阈值选取方法进行分析,将传统蚁群算法中四种启发函数得到的信息素矩阵进行叠加,再对其元素进行统计排序选取合适的阈值进行边缘提取。实验结果表明,文中方法能有效提取含一定噪声的低对比度图像边缘。

The image edge detection based on the ant colony algorithm has many advantages, such as strong ability of resisting noise and fine edge extraction. But when it is used in the edge extraction of lower contrast image with noise, several bad phenomenon occur, such as the hiatus of edge portion and unsmooth margin. In order to achieve the desired result for the lower contrast image edge extraction,in this paper, use the method which plus the pheromone matrixes got by four traditional heuristic functions to gain the pheromone matrix contai- ning more rich edge information,and select the appropriate threshold through the ordered matrix elements to produce the edge extraction. After comparing with several traditional methods, the experimental results show that this method can efficiently extract the edge of low contrast images with noise.