基于三维模型和射线探伤二维图像的铸件缺陷检测技术

Casting Defect Detection Technology Based on 3D Model and 2D Radiographic Image

计算机自动铸件缺陷检测被越来越广泛地应用于企业的生产中,但是对于探伤图片中缺陷和铸件结构混杂重叠的图片,计算机识别技术难以从图像中判别出缺陷。为此,提出了基于三维模型和二维射线探伤图像的铸件缺陷检测技术。建立了类X射线照射探伤图像模型,通过模拟射线穿透铸件STL模型的过程,提取出射线穿透模型的信息,输出类平行光源X射线照射的仿真射线探伤图像。通过有缺陷和无缺陷图像灰度值相减结果提取缺陷轮廓,进行边缘检测,最终进行缺陷标定。该方法可从图片中有效识别出缺陷,并将其成功标定位置。

Computer automatic casting defect detection has been widely used in the production of enterprises,how-ever,it is difficult to distinguish defects from the overlapped images mixed with defects and casting structure.Therefore,a casting defect detection technology based on three-dimensional model and two-dimensional radio-graphic image was proposed,and the image model of quasi X-ray irradiation flaw detection was established.By simulating the process of ray penetrating the STL model of castings,the information was extracted,and the sim-ulated ray flaw detection image of quasi parallel light source X-ray irradiation was output.The defect contour was extracted by subtracting gray values of defective and non-defective images for edge detection,and finally the de-fect was calibrated.The method succeeds to identify defects from images and calibrate positions.