基于图像处理的内衬套表面缺陷检测

Detection of surface defects in internal bushing of air conditioning pipe joints based on image processing

为了提高内衬套的检测速度和精度,保证内衬套的使用寿命,提出结合图像处理技术实现内衬套表面缺陷的自动检测.通过采用CMOS相机在近红外背光源暗域照明环境中获取图像并进行处理,实现对内衬套的毛刺及擦痕的自动检测.本检测系统主要通过图像形态学滤波和GrabGut图像分割算法分别实现对内衬套表面毛刺和擦痕的检测,通过轮廓拟合提取检测毛刺和擦痕的图像,从而实现对内衬套的表面缺陷检测.实验表明,所提出的内衬套表面缺陷的自动检测方法具有高效、准确的优点,且该系统运行稳定,因而具有推广价值.

To improve the detection speed and precision of the inner liner of air-conditioning pipe joints, an automatic detection method based on image processing technique was proposed. The images collected by using a CMOS camera in the near infrared backlight dark area illumination environment were processed with digital image processing technique. The image morphological filtering and the GrabGut image segmentation algorithm were used to distinguish the burrs and scratches in the inner liner of air-conditioning pipe joints. Contour extraction was performed to detect the size of the burrs and scratches. Experimental results show that the automatic detection method is effective and accurate; and the system is stable and practically valuable.