Method for detecting internal cracks in joints of composite metal materials based on dual-channel feature fusion

Abstract

Many kinds of products are cast from two or more metals to take advantage of the properties of different metals. Due to different thermal properties, an air gap may occur between the interfaces of two metals, which seriously affects the product quality. It is difficult to detect this kind of cracks with conventional metallographic detection methods and the detection efficiency is low. Based on this, a visual detection method based on two-channel feature fusion is proposed to rapidly detect the cracks, using flowable color ink to paint on the end face of the composite metals. If the flowable ink flows into the gap of the metal interfaces, a crack will occur in ink, reflecting the internal defects at the junction of the metals. The effectiveness of the proposed method was verified by testing the motorcycle wheel hubs which are casted with two metals. The wheel hub painted with ink was applied with an alternating load to enlarge the gap. Two images of the ink were captured with a camera before and after the alternating load was applied. Both images were converted from RGB model to HSV model, and both S and V channels of the two images were processed with the dual-channel feature fusion algorithm. A variety of experiments were conducted to determine a set of optimized parameters of the algorithm, and 1000 groups with qualified and defective hub images were tested. The results demonstrate the reliability of the proposed method, which can complete the detection of a product within 10 s, and the crack recognition rate can reach a high value of 97 %.