Research on Aerial Target Recognition Method Based on PSO-BP Neural Network

Air target recognition is one of the key links of air defense operation decision support. Aiming at the problems of slow solving speed and poor classification accuracy of traditional methods in solving the problem of aerial target recognition, a neural network classification method combined with particle swarm optimization algorithm is proposed. According to the various attributes of air targets, using the optimization ability of particle swarm optimization (PSO), BP neural network is optimized to obtain the best initial weights and thresholds, and a network model more suitable for air target type recognition task is designed. By comparing the classification accuracy of the model before and after optimization, the effectiveness and correctness of the optimization direction are verified. The model uses the memory, association and fault tolerance functions of neural network to further improve the stability and reliability of air target recognition.