Microseism Detection Method in Coal Mine Based on Spatiotemporal Characteristics and Support Vector Regression Algorithm

In view of the inconsistency guided wave energy in distributed acoustic sensing coal mine maps and difficulty distinguishing vibration levels mines, which leads to poor sensitivity accuracy microseism detection, a detection method based on time–space characteristics support vector regression algorithm is proposed ensure safety operations. The spatiotemporal sliding window was used collect data real-time, continuous attribute discretization entropy discretize data, then were mapped different state spaces build Markov chain; by calculating transition probability matrix cross-state matrix, respectively, temporal spatial microseisms at target node extracted. extracted as input particle-swarm-optimization-improved model, solution results signals output. error penalty factor kernel function parameters improved, particle swarm optimization introduced optimize mines. experimental showed that this can accurately detect real-time mines mining area, effectively control rate missing detections process, stability overall operation. When inertia weight set 0.9 number particles 45, had highest best-detection for