蚁群算法在二维Otsu图像分割中的应用

提出了一种基于蚁群算法和二维Otsu的图像分割方法,利用蚁群算法快速寻优的特点,求出二维Otsu图像分割的阈值分割点,对图像进行分割。根据源图像和邻域平滑后图像的灰度,以及灰度频数进行聚类。通过灰度直方图的峰值点设置精确的初始聚类中心,解决了蚁群算法运算次数多、计算量大的问题;针对具体应用,对聚类半径、信息激素和启发引导函数进行了修正。实验表明该算法速度快、划分特性好、抗噪声能力强,可以准确地分割出目标。

Application of Ant Colony Algorithm in Two-Dimensional Otsu Image Segmentation

This pape proposes an image segmentation method based on ant colony algorithm and twodimensional Otsu and figures out threshold segmentation point of two-dimensional Otsu image to segment the image.Clustering is conducted according to source image,image grayscale after neighbor smoothing and grayscale frequency.The problems of ant colony algorithm（i.e.many operation times and large amount of calculation）are solved by using the peak point of gray histogram to set precise initial clustering center.According to the specific application,this paper corrects the cluster radius,pheromones and enlightening guiding function.The experiments show that the algorithm is fast and can accurately segment the target,with excellent segmentation property and strong anti-noise ability.