K-nearest neighbor based detection method for predicting underground coal mine water bursting source

The invention discloses a K-nearest neighbor based method for predicting a coal mine water bursting source. On the basis of the K-nearest neighbor, an Nd: YAG solid laser is adopted to emit laser which is focused on a to-be-detected water body; the water body generates a fluorescence spectral signal under the laser excitation; the spectral signal collected by a fluorescent probe is irradiated to a multichannel detector and subjected to signal processing, and signal data reflecting water quality condition of the to-be-detected water body is obtained; the signal data is compared with signal data reflecting water quality condition of a normal water body under the same measurement condition, and then a mode recognition model is established with the K-nearest neighbor in an ARM 7 processor and used for analyzing data of the to-be-detected water body; and water quality change condition of the underground coal mine water source can be accurately obtained through calculation, so that prediction of coal mine water bur