Locomotive multi-dimensional fire detection method

The invention discloses a locomotive multi-dimensional fire detection method, which comprises the following steps: acquiring fire characteristic data in a reference application scene, preprocessing the fire characteristic data, and acquiring an average value of the fire characteristic data of initial working states of the reference application scene and an actual application scene; obtaining effective first fire feature data of the actual application scene and a weight coefficient of the first fire feature data of the actual application scene; meanwhile, effective second fire feature data are obtained; and finally, a system fire alarm output judgment parameter W is obtained to provide a basis for locomotive workers to judge whether a fire alarm occurs or not. According to different application scenes of rail transit and reasonable arrangement of the fire detectors in the scenes, multi-source information acquired by various sensors is comprehensively fused, the problems that a traditional detector is high in f