最大类间方差耦合分水岭分割的目标定位算法

为了解决当前复杂目标受噪声干扰严重、背景多变且工件型号切换频繁,导致目标分割不准确,影响目标定位与装配的准确度.本文分别从系统软硬件架构和分割算法整合的角度出发,提出了基于最大类间方差与分水岭标记的复杂目标分割与定位算法.首先,搭建网络摄像头、多轴伺服电机和视觉软件系统之间的系统架构,达到可用于现场实际作业要求的目的.然后,根据图像目标与背景的最大类间方差特性,计算最佳分割阈值,达到对工件目标粗分割的目的.最后,构建分水岭分割线特性,提取分割线,优化区域标记准确度,达到对工件精分割,剔除噪声干扰的目的.实验测试结果显示:与当前分割算法相比,本文算法拥有更高的分割准确度与作业稳定性.

Object Location Algorithm Based on Maximum Variance and Watershed Marker

In order to solve the problem that the noise of the workpiece is too much,the background is changeable and the workpiece model is frequently changed,the target segmentation is not accurate,which affects the accuracy of feeding.In this paper,based on the software and hardware architecture of the system and the integration of the segmentation algorithm,the segmentation algorithm of the target is proposed based on the maximum between class variance and watershed mark.First of all,the system structure of network camera,multi axis servo motor and visual software system is set up,which can be used for the purpose of practical operation.Then,according to the maximum variance between the target and the background of the image,the optimal segmentation threshold is calculated to achieve the purpose of the coarse segmentation of the object.Finally,according to the characteristics of the skeleton of the watershed,the segmentation line is extracted,and the accuracy of the region marking is optimized.The experimental results show that compared with the current segmentation algorithm,the system has higher segmentation accuracy and stability.