The K-DAS software combines deterministic time-frequency signal processing with neural-network classifiers trained on extensive field libraries of target and background events. By analyzing spectral signatures, duration, movement speed, and spatial localization, the system reliably separates passing vehicles, trains, and grazing cattle from stationary excavator digging, directional drilling, manual trenching, and mechanical work on the pipe wall.