Research Of Target Recognition Based On Radar Range Profile

1D Range and 2D Radar Profile can reflect radar targets' subtle structure. And it is very useful for target measurement and recognition. With the purpose of target recognition, it is foundational to do research on 1D Range and 2D Radar Profile. This paper studies on the matching of 2D Radar Profile with its corresponding radar target, and on target recognition based on 2D Radar Profile.Inverse Synthetic Aperture Radar (ISAR) is a kind of high range profile radar. It can be used for radar imaging such as aircraft and space objects and other targets. This technique uses the relative movements between radar and target and the information processing technique to achieve equivalent aperture. And then supplemented by high-expansion of the pulse compression technology, we get 2D Radar Profiles of radar targets.This paper introduces the imaging principle of 1D Range and 2D Radar Profile, and the methods obtaining 1D Range and 2D Radar Profile and rendering these images. We develop a method to match 2D Radar Profile with its corresponding target.PCA (Principal Component Analysis) is a statistical analysis method. It can resolve main influencing factors from multiple things, reveals the essence, and makes them easy. It can project high dimensional data to lower one. Neural network method simulates human brain. Based on connection of nerve cells, distributed storage and handling is implemented, and it also provides self-organizing, adaptive and self-learning abilities. It is particularly applicable to handle the problems such as multivariable, imprecise and vague information processing.This paper applies PCA to lower the dimension of 2D Radar Profile vector. By using Least Distance Method and Neural Networks, we develop the radar target recognition prototype system based on the 2D Radar Profile. At last we test our method on several radar targets, and analyze the experimental results.Through integration of estimating electromagnetic theory and visualization techniques, we develop a high-frequency electromagnetic scattering simulation system. As we use object-oriented methods to design the system, it is proved to be easier to reuse and maintain.