图像分割中算法的应用研究

研究图像优化分割问题,最佳阈值选取直接影响到图像分割的清晰度质量。传统采用经验法进行分割,难以获得最佳阈值,导致分割准确率低,易产生图像误分割。为了提高图象分割准确率,提出一种基于遗传算法的Otsu图像分割。首先对图像进行去噪处理并绘制直方图;然后直方图信息选取适当灰度值作为遗传算法中的初始种群,最优阈值作为目标函数,最后通过选择、交叉和变异等遗传操作得到图像分割最优阈值,并进行图像分割。实验结果表明,遗传算法的Otsu图像分割加快了速度,减少了计算量,提高了图像分割准确率,证明适应于图像实时处理。

Application research on Image Segmentation Method

Otsud is a simple image segmentation method with small calculation and stable performance,in which the key problem is to determine the optimal threshold.Traditionally experiences are used to determine the optimal threshold,but the segmentation accuracy is low,the antimonies ability is poor,and it is easy to generate the false division.In order to improve the image segmentation accuracy,an improved Otsud image segmentation method was proposed based on the genetic algorithm.Firstly,the noise of image was dealt with and the histograms was drawn.Then appropriate histogram information grey values were selected as the initial population genetic algorithm,and the optimum threshold was as objective function.Finally,by selecting,crossover and mutation genetic operations,the optimal threshold of image segmentation was obtained.Experimental results show that Otsud image segmentation algorithm based on genetic algorithm speeds up the segmentation and improves the image segmentation accuracy.It is more adapted to the image real-time processing.