Abstract
Most traditional wire rope sensors produce the integrated volume of magnetic leakage around the whole circumference of wire rope, which is insensitive to the circumferential distribution of wire rope defects. In this paper the three-dimensional magnetic flux leakage of rope surface is obtained by the aid of Hall sensors array distributed around the wire rope. Then a spatial notch filter is designed to eliminate the strand-waveform signal, and the real defect signal is emphasized. The signal of any defect is transformed to the corresponding gray-scale map. Then an algorithm of two-dimensional image recognition is introduced to extract features from these gray-scale maps and identify defects. The experiment results show the degree and the width of defects, the circumferential distribution of localized flaw such as concentrated or dispersive broken wire can be well distinguished. The discrimination of this method for several typical defects under the condition of laboratory can reach 90%.