基于计算机视觉的蜂窝陶瓷在线检测系统研究

Research and Development of Honeycomb Ceramics’ On-line Automatic Inspection System Based on Computer Vision

摘要:

研发了一种基于计算机视觉的蜂窝陶瓷智能在线检测系统；设计了系统的体系结构,4台高速摄像机在线获取蜂窝陶瓷图像,研究了改进型加权中值滤波处理,论述了基于Fisher评价函数的图像阈值分割算法；最后进行亚像素测量,判断产品尺寸是否在允许范围内,实验验证了系统的有效性和检测算法的可行性,可以很好地代替人工检测。

Abstract:

This paper introduce a design of honeycomb ceramics’ on-line automatic checkout system based on machine vision. The checkout system acquires images of honeycomb ceramics through high-speed video camera. Firstly, the images are processed by using improved weighted median filtering. Then, the images run into threshold segmentation on the basis of Fisher criterion function method. Finally, sub-pixel measurement is applied so as to judge whether the product dimension is within the allowed scope. The experiment proves the effectiveness of the system and feasibility of detecting algorithm. 