基于单幅测量图像的三维缺陷检测技术

为了对缺陷进行更加全面的判断,设计了一种新型三维缺陷检测方法,只需要根据单幅实时测量的工件图像,就可获得0°~180°范围内工件形貌的三维数据,进而对缺陷的平面及深度尺寸进行全面判断。其核心技术是根据单幅测量图像中留下的三维线索—灰度信息,进行亮度分析和转换,利用倾角和偏角计算物体深度信息。在工业现场磁性材料缺陷检测中,该方法在X和Y方向的分辨力达到0.1 mm,Z方向的分辨力达到0.007 mm。实验证明所使用的三维缺陷检测方法,工作方式简单,硬件成本低,处理速度快,精度高,适宜在工业现场应用。

Three-Dimensional Defect Detection Based On Single Measurement Image

In order to acquire overall three-dimensional (3D) defect information, a new 3D defect detection method is designed. It only needs single image captured from real-time measurement line. From the single image, 3D data of the work piece shape ranged from 0° to 180° can be got. Moreover, it can give an overall recognition of plane dimensions and depth information of the defect. The key technique is made use of the 3D clues——“shading value”, which left in the single measurement image, to process luminance analysis and conversion, thus to calculate the depth information by slant and tilt. In the application of defect detection for magnetic work piece, the resolution in X and Y direction can reach 0.1 mm, and the resolution in Z direction can reach 0.007 mm. This experiment proves that the designed 3D defect detection method has the merits of simple operation, low cost, high speed and high precision. This 3D method is fit to use in industrial application.