A novel automatic surface scratch defect detection for fluid-conveying tube of Coriolis mass flow-meter based on 2D-direction filter

The tube outside surface scratch detection is very important to Coriolis mass flow-meter industry, as scratches leads to degradation of the pressure-resistant property of the fluid-conveying tube and malfunction of the flower meter consequently. This study aims to develop an automatic system which can be used to detect surface scratches of the large curvature tubes in order to reduce computational costs. This is achieved by proposed edge detection method which involves two steps: (1) a 2D-direction filter is designed accordingly to enhance the contrast between the background and the scratches then (2) a modified adaptive threshold method is employed to automatically detect the edge of the scratches on a filtered image. To assess the effectiveness of the proposed method, experimental equipment was built and experiments were performed with 100 fluid conveying tubes with surface scratches. Experiment results shows that the proposed method is very efficient and its performance is competitive to the existing state of art methods.