Target magnetic anomaly signal recognition based on a fusion algorithm

Under the background of ocean complex interference, the target magnetic anomaly detection has always been the focus and difficulty of the research. In particular, it takes a long time to detect the characteristic signals of the target magnetic anomaly in a full period, so it is difficult to analyze and identify the target in real time. A fusion algorithm based on particle swarm optimization (PSO) and least squares support vector regression (LS-SVR) is proposed to predict the target magnetic anomaly signal characteristics based on time series. The fusion algorithm uses the OPS’ virtue that is the property of fast convergence to optimize parameters of LS-SVR algorithm. And root mean square error (RMSE) is applied for loss function to assess prediction model of magnetic anomaly signal based on LS-SVR. The algorithm model is utilized to predict the signal characteristics of target magnetic anomalies with time. Experiments show that the prediction accuracy of the new algorithm outperforms Least Squares(LS), support vector regression (SVR) and least squares support vector regression (LS-SVR). This paper provides an idea for the detection of magnetic targets in the marine environment.