Research of Objection Recognition Based on Radar Cross Section Calculations

Object recognition is one of the main research directions of computer vision. This technique has been widely used in many fields, including application in electromagnetism. Among many objection recognition algorithms, deep learning algorithms occupy the mainstream position by its faster recognition speed and higher recognition accuracy. In the field of electromagnetic measurement, Radar Cross Section (RCS) is an important physical quantity to measure the intensity of the scattered echo. Using Radar Cross Section for objection recognition has great development potential and demand. In the paper, theoretical formulas of Shooting and Bouncing Ray (SBR) method are derived, firstly. And then the RCS of target is calculated by using SBR. Next, the basic theory of neural network model is introduced and an improved LeNet-5 neural network model is constructed by Pytorch. This model is trained by using a data set composed of scattered echo matrices. Finally, this trained neural network model is used to identify similar ship targets and flight targets. The results of objection recognition show that the recognition accuracy of ship target can reach more than 99.5%, and the recognition accuracy of flight target can reach more than 99.7%. It is proved that the improved neural network model can recognize the target successfully by using scattering echo matrix data. The results also show that objection recognition technology has great application potential in the field of electromagnetism. Therefore, the use of artificial intelligence technology in the electromagnetic field will promote the progress of science and the development of society in the future.