为了实现生产线上玻璃质量的自动检测,采用机器视觉的方法对生产线上的玻璃进行缺陷检测.采集图像过程中光照方式对图像质量影响很大,为了避免由光照效果不佳对玻璃质量判断造成的干扰,保证视觉检测系统可识别不同种类的裂痕,研究了机器视觉系统中光源系统相对于生产线上玻璃的位置和光强度,解决上述问题造成的影响.分析光源系统对拍摄玻璃图像中缺陷部分和质量合格部分的相关影响,提出了一种可以检测出不同种类裂痕的光源照射方式,使其能够拍摄满足条件的高质量玻璃图像.通过样本批量测试实验证明:该光源照射方式对玻璃质量在线检测的识别准确率均为90%以上,所采集图像质量满足要求,能够准确检测出不合格产品的缺陷位置、面积大小及缺陷深度.

A method for automatic glass inspection on the production chain based on visual system was adopted.The method of illuminating is crucial to image in acquisition process,in order to avoid the illuminating jamming for judgment to guarantee the accurate identification result of different cracks by machine visual system,we carried out the research on the illuminating source system,including the luminous intensity,the wavelength and the position relative to line scan camera and glass which is on the production chain.We proposed a novel method of illuminating to achieve high quality image to solve problems caused by the influence of illuminating,with the analysis of influence between the defective part and the normal part when illuminating source system operates.It is proved that the method can realize realtime detecting to achieve accurate and reliable results,through a batch of glass samples tests.From the detection of glass image scanned under the condition provided by illuminating source system,the visual system can compute the defect and present the position size and depth,if the glass is not the acceptable product.