基于DSP+FPGA的烟包包装缺陷视觉检测系统研究

Visual Inspection System of Cigarette Package Faults Based on DSP+FPGA

提出了采用机器视觉技术,对卷烟企业生产的烟包进行图像识别及检测的DSP+FPGA硬件平台设计方案,并对烟包包装缺陷检测的软件实现进行研究,提出选取图像强度分量进行灰度处理,选择合适的二值化和滤噪方案对图像进行二值化及滤噪处理,最终在二值化图像中进行包装检测的过程,为烟包高速自动化包装检测提供一种高效可行的实施方案。

A hardware platform design with machine vision to inspect the image of cigarette package was proposed and the software was developed. Firstly, the intensity of pictures should be processed by gray scale, then denoised and binarized with proper binarization. The faults of package tag paper such as mismatch and rake angle are finally identified in the binarized image. This method is efficient in the high speed and automatic package inspection.