Research on Multi-target Recognition Algorithm of Pipeline Magnetic Flux Leakage Signal Based on Improved Cascade RCNN

Magnetic Flux Leakage (MFL)internal detection is the main technology to detect long-distance oil pipelines. Aiming at the low detection accuracy and poor versatility of existing pipeline magnetic flux leakage signal target recognition algorithms, this paper proposes a pipeline magnetic flux leakage signal target recognition algorithm based on improved Cascade RCNN. Firstly, an adaptive image conversion method is proposed to convert the original magnetic flux leakage data into colormap. Secondly, Feature Pyramid Networks (FPN) and Online Hard Example Mining (OHEM) are added to Cascade RCNN to improve target detection accuracy. Finally, the effectiveness of the method is verified through comparative experiments. The results indicate that the method proposed in this paper is effective.