Wheat yield estimation method based on unmanned aerial vehicle low-altitude visible light image and target detection network

The invention relates to a wheat yield estimation method based on an unmanned aerial vehicle low-altitude visible light image and a target detection network. The method comprises the following steps: planning a waypoint sampling route of a wheat plot; high-resolution visible light images shot by the unmanned aerial vehicle at low altitude are obtained and preprocessed, and a first training set and a second training set are established; respectively constructing a wheat ear detection counting model and a wheat ear grain detection counting model based on the small target detection neural network; and calculating the pre-estimated yield of the wheat in the measured plot. The device does not need to enter a wheat plot for sampling, wheat ears do not need to be picked, and the sampling range and the sampling efficiency are greatly increased through a fixed-point and fixed-height aerial photography sampling mode of the unmanned aerial vehicle; the wheat ear detection counting model and the wheat ear grain detection