3 results about "Downburst" patented technology

This invention discloses a low-altitude wind shear recognition system and method based on convolutional neural networks, belonging to the field of meteorological monitoring technology. The system includes: a low-altitude wind shear dataset construction module, which generates a wind shear image dataset containing crosswind shear, low-altitude jet stream, headwind and tailwind shear, and micro-downbursts; a data preprocessing module, which performs preprocessing on the images; a convolutional neural network module, which incorporates a lightweight convolutional neural network with a progressive channel design, trained using a two-stage transfer learning strategy, and utilizes an improved loss function obtained by weighted summation of cross-entropy loss, flow direction consistency constraint loss, and global vortex intensity constraint loss to obtain a wind shear recognition model; and a wind shear intelligent classification module, used to identify the type of wind shear in low-altitude wind field images in real time. This invention has the advantages of high recognition accuracy, low computational load, good real-time performance, and low hardware cost, and can effectively ensure the flight safety of UAVs in complex low-altitude wind field environments.