Aerial photography mousehole detection method and device based on improved YOLOv5 algorithm

The invention discloses an aerial mousehole detection method and device based on an improved YOLOv5 algorithm, and the method comprises the steps: obtaining an aerial grassland image, carrying out the image screening and preprocessing, and building an aerial mousehole image data set; mouse hole positions are marked on images in the image data set, and a corresponding label set is constructed; dividing the mouse hole image data set into a training set and a verification set according to a preset proportion; an aerial photography mouse hole recognition model based on YOLOv5 is improved; comprising the following steps: adding an ECA attention module in a backbone network, and compressing a C3 module by using structural re-parameterization during reasoning; training and optimizing the improved YOLOv5 mouse hole detection model by using the marked mouse hole image data set to obtain an optimal detection model; and performing mouse hole detection based on the optimal detection model. According to the method, a gras