基于机器视觉的辐射单元中心点提取算法设计

Design of Center Point Extraction Algorithm for Radial Unit on Machine Vision

针对阵面雷达现有装配过程中存在的自动化水平低、检测速度慢等问题,提出一种基于机器视觉引导机械手装配与检测的设计方案,方案核心在于对辐射单元口沿中心的快速、精确提取。首先通过基于形状的模板匹配算法对采集图像进行匹配,实现对辐射单元大致位置的快速定位;然后为实现对辐射单元中心点的精确提取,提出一种基于灰度差与极性变化的方法来重新拟合辐射单元外轮廓线。经Halcon仿真对比验证,该算法的检测效果比单纯进行模板匹配得到的结果有明显的优越性,这为实现阵面雷达的自动化装配与检测提供了有力的基础。

Aiming at the problems of low level of automation and slow detection speed in the existing process of the radar array,a system design of manipulator assembly based on machine vision is proposed,and the core part of the scheme is to extract the center of the radiation unit apace and accurately.Firstly,by template matching algorithm based on shape matches the image after preprocessing to achieve rapid positioning of the radial unit.Then in order to obtain the accurate position of radiation unit,this paper proposes a new method based on gray value and polarity to fit the external profile of radiation unit.The Halcon simulation comparison shows that the proposed algorithm is superior to the template matching,it provides a powerful basis for the automatic assembly and detection of the array radar.In addition,the algorithm can be applied to the fast and accurate extraction of contour lines for any object,including circle,besides the radiating element reconstruction.