Large wingspan unmanned aerial vehicle wing icing detection method and system based on binocular vision

The invention discloses a large-wingspan unmanned aerial vehicle wing icing detection method and system based on binocular vision, and the method comprises the steps: respectively installing two cameras on a supporting rod of a symmetry axis of the front edge of an aircraft, and setting a plurality of code identification points on the surface of the front edge of a wing; performing external parameter calibration and internal parameter calibration of the two cameras by using the spatial three-dimensional coordinates of the coding identification points; wing images are collected through two cameras, feature points are detected, and binocular camera image stereo matching is carried out; and simultaneously solving spatial three-dimensional coordinate values of all the identification points at the front edge of the wing section according to a correlation conversion equation of the homonymy points in a pixel plane and an airframe coordinate system, and subtracting the solved Z coordinate values from the correspondi