Rainstorm weather detection method, system and terminal

The invention discloses a rainstorm weather detection method, system and terminal, and the method is based on a weighted density adaptive A-DBSCAN algorithm, and comprises the steps: processing a monitoring camera image through employing a Gauss-Laplace LOG operator, and obtaining an LOG value at a corresponding moment; screening out effective LOG values from the LOG values through an adaptive A-DBSCAN algorithm based on weighted density; and detecting the rainstorm weather according to the regression relationship between the effective LOG value and the rainfall data at the corresponding moment. Ambient noise overmuch, an improved A-DBSCAN clustering algorithm is adopted to carry out de-noising processing; the R-Square coefficients before and after denoising are compared to obtain that thedenoising effect of the density-related data based on the weighted density adaptive A-DBSCAN algorithm is obvious, so that the detection accuracy of rainstorm weather in a monitoring camera shootingarea is greatly improved o