Non-ferrous Metal Defect Recognition Based on Machine Vision

Abstract

It is easy to generate burrs on non ferrous metals in the process of casting. At present, the repair operations of metals, such as grinding and cutting, are carried out manually in the factory. Intelligent and automatic ingot repair methods are needed in which, defect identification is the key point. Based on this, this paper proposes an intelligent defect identification algorithm with the characteristics of high efficiency and high precision. Firstly, the metal ingot image is extracted by edge detection, Hough line detection and parameter calibration. Secondly, HSV color segmentation technology is used to effectively separate the metal ingot from the background, and the mask image reflecting the shape information of the metal ingot is obtained. Then, a new method is applied to screen the preliminarily extracted straight lines to obtain the contour of the metal ingot. Finally, by using the contour information, we can obtain a new mask image, in which the burr position can be accurately located. The results show that the proposed algorithm achieves the success rate of 91.6%.