Data Fusion Method for Multi-Sensor Detection of Pipeline Defects

Considering the problem of low fusion accuracy of multi-sensor pipeline defect detection data, a data fusion method of multi-instrument pipeline defect detection is proposed, which combines the improved bird swarm algorithm (IBSA) with the weighted regularized extreme learning machine (WRELM). First, pipeline defect data are collected using electromagnetic ultrasonic guided wave testing equipment, magnetic flux leakage testing equipment, and eddy current testing equipment. The Gaussian kernel function sample weight matrix and the regularization parameter are subsequently introduced into the extreme learning machine, and the WRELM data fusion model is established. The bird swarm algorithm is then optimized by introducing chaotic variables and Gaussian perturbations, which optimizes vigilance behavior and changes the step factor in the flight behavior. The IBSA is used to optimize the connection weight between the input layer and the hidden layer and the bias of the hidden layer of WRELM. Finally, the data fusion platform for multi-instrument pipeline defect detection is utilized for experimental analysis. The experimental results show that the error of the multi-instrument pipeline defect data fusion model using the IBSA to optimize the WRELM is the smallest at just 2. 33%. The fusion accuracy of multi-instrument pipeline defect data is effectively improved.