Fusion feature multi-objective classification algorithm based on adaptive scale feature residual network

With the development of science and technology, radar has been widely used in both civil and military fields, and radar target classification and recognition is an important research direction in radar signal processing field. The traditional classification and recognition technology does not utilize all the information of echo, and the recognition rate will decrease obviously when the environment is too complicated and the SNR is low. To solve the above problems, this paper proposes three recognition algorithms. It is a fusion feature multi-objective classification algorithm based on multi-scale feature residual network (MKRN), a fusion feature multi-objective classification algorithm based on adaptive scale feature residual network (AKRN) and a fusion feature multi-objective classification algorithm based on adaptive scale feature multi-level branch residual network (AKSRN). These three algorithms introduce residual structure to avoid network degradation caused by the deepening of network depth, and use convolution kernel of different sizes to enhance the ability of the network to extract multi-scale feature information, which improves the accuracy of network classification and recognition.