Lstm-Based UAV and Bird Classification Method for Multidimensional Track Features

UAV and birds classification is important and difficult for modern radar application. Current attempts still need to be improved in terms of recognition accuracy. This paper adds to characteristic parameters of vertical plane, and constructs a multi-dimensional recognition framework based on LSTM according to the target motion characteristics, the horizontal plane track map and the vertical plane track map. This method is more optimized in terms of recognition accuracy and training time, and is a better solution based on track recognition.