一种改进阈值分割算法在镜片缺陷检测中的应用

Application of an Improved Threshold Segmentation Algorithm in Lens Defect Detection

在光学镜片缺陷检测中,为提高光学镜片图像阈值分割的精度和速度,提出一种新的粒子群算法(PSO)+Otsu阈值分割算法。该算法通过改进PSO权重因子更新策略,增加权重因子在迭代初期位于较大值的时间,增强全局搜索能力,计算粒子的最优位置,并把最优位置赋值给Otsu算法,最终实现光学镜片图像的阈值分割。改进的权重因子更新策略能够克服典型线性递减权重因子更新策略由于迭代初期的全局搜索能力不足,导致后期陷入局部极值的缺点。实验结果表明,该算法在提高图像阈值分割精度同时,还提高了阈值分割的速度。

In the optical lens defect detection,in order to improve the accuracy and speed of the optical lens image threshold segmentation,a new particle swarm optimization(PSO)+Otsu threshold segmentation algorithm is proposed.The algorithm improves the PSO weight factor update strategy,increases the time when the weight factor is at a larger value at the beginning of the iteration,enhances the global search ability,calculates the optimal position of the particle,and assigns the optimal position to the Otsu algorithm.Finally realize the threshold segmentation of the optical lens image.The improved weight factor update strategy can overcome the shortcomings of the typical linearly decreasing weight factor update strategy that the global search ability at the initial stage of the iteration is insufficient,which leads to the local extreme value in the later stage.Experimental results show that this algorithm improves the speed of threshold segmentation while improving the accuracy of image threshold segmentation.