基于适应性阈值和遗传算法的图像分割

An Image Segmentation Method Using the Adaptability Threshold and Genetic Algorithms

图像分割是自动目标识别的关键和首要步骤,其目的是将目标和背景分离,为计算机视觉的后续处理提供依据。提出了一种基于适应性阈值和遗传算法进行图像分割的方法,对图像的不同区域分别作灰度值的评价标准,把图像分割问题定义为一个多区域的独立优化问题,利用遗传算法的寻优高效性,搜索到能使分割质量达到最优的图像分割阈值。试验表明,使用该方法可使图像分割获得满意的效果。

Image segmentation is a key and initial step of automa tic target partition. Its purpose is to separate target and background,and offer basis for follow-up handle of computer vision. An image segmentation m ethod based on the adaptability threshold and genetic algorithms (GA) is present ed. It applies the segmentation quality criteria to different districts in i mage and formulates the segmentation problem as independent optimization of many districts. The GA efficiently searches the segmentation parameter space and the goal of this method is to obtain the optimal image segmentation threshold f or segmentation quality. Experiment shows that using this method it will obt ain satisfied effect for image segmentation.