Broadband radar target HRRP identification method based on hybrid model fusion

Abstract

The invention relates to a broadband radar target HRRP identification method based on hybrid model fusion, and belongs to the technical field of radar target identification. Comprising the steps of 1) collecting a data set and dividing a test set and a training set; (2) setting the flag bits Ltone and Ctone to be 0; 3) utilizing the training set to respectively train two sub-models of the F-CNN model, namely a CNN model and a LightGBM model; and 4) using the F-CNN model to identify the type of the test set, specifically, performing sub-model processing after training to obtain classification results A and B. Judging whether the flag bits are 1 or not; if not, waiting is carried out; if yes, the result A and the result B are input into the decision tree to be combined, and an F-CNN model recognition result is output; according to the method, detection targets can be effectively classified according to HRRP signals; the recognition accuracy under different signal-to-noise ratios is obviously improved; the complexity of the model is obviously reduced, and the reasoning speed is improved.

Description

technical field [0001] The invention relates to a wideband radar target HRRP recognition method based on hybrid model fusion, and belongs to the technical field of radar target recognition. Background technique [0002] The distance resolution of high-resolution broadband radar is much smaller than the size of the target, and its echoes contain extremely valuable structural information for classification and identification, such as the number of strong scattering centers of the target, the center of gravity of the target scattering, the width of the target, and the distribution entropy of the scattering centers. Broad engineering application prospects. Therefore, the radar target recognition method based on the high-resolution broadband radar to obtain the high-resolution range profile (High Range Resolution Profile, HRRP) of the radar target has gradually become a research hotspot in the field of radar target recognition technology. [0003] In recent years, methods based ...

AI Technical Summary

Problems solved by technology

Compared with the traditional method, the deep learning method based on CNN has improved the accuracy rate considerably, but there are still the following problems: 1) HRRP signals are usually obtained in a non-cooperative environment, due to the non-cooperative nature of the target or the long distance of the target and other problems, there is noise interference, and the existing models have room for improvement in the recognition and classification accuracy of HRRP signals under various SNR conditions; 2) The existing models have the shortcomings of too deep network depth and too many parameters, and the complexity higher

Images