Quantitative Detection of Wire Rope Based on Three-Dimensional Magnetic Flux Leakage Color Imaging Technology

Abstract

In the field of nondestructive testing (NDT) of wire rope, magnetic flux leakage (MFL) detection is the most widely used method. The traditional single-dimensional MFL detection system has the disadvantages of large volume, heavy weight, low sensitivity and low recognition rate. In order to overcome these problems, this paper designed the three-dimensional MFL acquisition system based on unsaturated magnetic excitation using the tunnel magnetoresistive (TMR) elements and proposed the three-dimensional MFL detection method. We mapped the three-dimensional MFL signals after noise reduction to Red-Green-Blue (RGB) space for color imaging. Then, the localization and segmentation algorithms were used to crop color images of broken wire defects. Finally, the color moment eigenvalues of color images of six types broken wire defects were extracted as the input of the back propagation (BP) neural network to quantitatively identify the broken wires. From the experimental results of the quantitative identification of broken wire defects, it can be seen that it is feasible to use the color images of the three-dimensional MFL signals to quantitatively identify the broken wire defects. Moreover, the three-dimensional MFL color imaging can effectively improve the recognition rate of broken wires.