基于机器视觉的钢轨表面检测光学模型的研究

在机器视觉的检测中,图像是整个系统中最重要的原始数据,其质量决定了后期图像处理的效果和速度。为探讨高质量图像的采集,提出一种基于线阵CCD相机和线阵光源的钢轨表面缺陷检测的光学理论模型,分析线阵CCD采集系统中振动模糊的原因,推导出图像灰度值与系统振动幅度和缺陷深度的关系,研究光源照射角度和相机拍摄角度对图像灰度和缺陷区域对比度的影响,并通过实验验证模型的合理性。结果表明:缺陷区域图像灰度值随着钢轨表面缺陷深度增大而降低,采用较低的光源照射角度可增大缺陷与背景的对比度,突出缺陷特征便于后期图像处理的缺陷识别。

Study on Optical Model for Rail Surface Detection Based on Machine Vision

In the rail surface defect detection,images are the most important and original data of the entire system, their quality determines the effectiveness and speed of post-image processing. To investigate the gathering of high-quality images,this paper proposes a optical theory model based on the linear CCD camera and the linear light source to detect the defects on rail surface,analyze the causes of the vibration vague in the linear CCD system,deduce the relationship between the amplitude of system,the depth of the defect and the image gray level and to study the influence of the illumination angle and the camera angle on the image gray level and the contrast of defect area. The rationality of the proposed model is proved by experiment. The results show that the image gray level of the defect area decreases with the increasing of the defect depth and a lower illumination angle can highlight the contrast between defects and backgrounds,which facilitates the image identification in late image processing.