Performance optimization and evaluation system for radar target recognition algorithm

Ying Zhou,Xinggan Zhang,Qiong Wang

DOI: https://doi.org/10.13232/j.cnki.jnju.2017.06.021

2017-01-01

Abstract:Radar target recognition has always been a hotspot in radar and signal processing,wherein the used data and techniques depend heavily on the scenario.For different kinds of targets,different combinations of characteristics and classification methods are needed in order to make target recognition results more accurate.Radar target recognition technology research not only involves complex algorithms,but also is a very experimental system engineering.A system ofperformance optimization and evaluation for radar target recognition algorithms is designed in this paper.The system is designed on Qt and MATLAB platform.The former is used to build the system framework and realize the foreground man-machine interaction.The latter is used to realize the calculation and opti-mization of background data.Andthe target recognition algorithm rule database is established on the MySQL database development support environment.Obj ect-oriented technology and modular design ideasare adopted.At the same time,security,reliability,openness and scalability are taken into account in the design phase.In this system, data importing,calculation and processing are excuted in the background,and multiple functions are intergrated, which including mangement of recognition rules,importing of new recognition methods and other functional modules.The target feature data is imported through the foreground man-machine interaction,and according to the selected computing mode,the specific feature transformation algorithm and recognition algorithm are adopted during the phase of training,testing or identifying.According to the specific requirements of the application,the recognition algorithms are selected manually,as well as the recommended rules are stored on the database automatically.And the system is convenient to add the input feature type and recognition algorithm.The system ofperformance optimization and evaluation for radar target recognition algorithms is convenient to use.It provides an effective research platformto solve the intelligent identification problem of different feature types,and has wide application prospects.