Powered by multi-sensor fusion and a deep learning framework, Siterwell’s algorithm enables highly accurate identification of fire combustion sources. By integrating data from gas sensor arrays, thermal imaging sensors, and multi-spectral labyrinth optical detectors, the system performs spatiotemporal feature analysis to significantly enhance fire detection accuracy in complex environments — especially where interference sources such as steam or dust are present.