Research On Real-time ISAR Imaging Of Wideband Radar Based On GPU

At present,the bandwidth of radar used in ISAR imaging is gradually increasing,which makes the data volume of echo signal higher and higher.In order to keep the coherent information of the signal,the direct acquisition method is adopted,which cannot reduce the echo data amount as the analog STRETCH processing,and the real-time processing of the echo signal is becoming more and more difficult.In this paper,relying on the development task of a radar wideband data acquisition and processing subsystem,the research on real-time ISAR imaging processing of large bandwidth echo signal is mainly carried out under the target of both processing efficiency and imaging quality.The main contents can be divided into three parts.In the first part,a reasonable imaging processing flow is designed according to the system requirements and the pulse compression method and compensation method used in this paper.The process is divided into two parts: one is to obtain high-quality one-dimensional images,and the other is to obtain high-resolution ISAR images.In order to realize real-time processing of large data volume signals,GPU with high-performance computing capability is selected as the processing platform,and its usage method is introduced.In the second part,the pretreatment process is explained in detail.For the data collected directly,in order to avoid the high calculation amount of the matching filter pulse compression method,the digital STRETCH pulse compression method is selected.A channel compensation method is proposed for digital STRETCH processing.After digital STRETCH and target region extraction to reduce the amount of data,high-speed motion compensation and channel distortion compensation are added to improve the image quality without reducing the processing efficiency.According to the characteristics of GPU programming,the one-dimensional image data processing flow is designed to ensure the real-time signal processing,and the maximum amount of data that can be processed in real-time on the platform is analyzed.In the third part,the high-resolution imaging process is explained in detail.A parallel envelope alignment algorithm based on cross-correlation and minimum mean envelope entropy is proposed,which is applied to GPU platform to improve processing efficiency and obtain ISAR images with higher quality than other parallel algorithms.According to the characteristics of ISAR real-time imaging processing on GPU platform,the optimization technology is put forward,which is divided into the processing methods of reducing data rate and the programming methods of improving data computing efficiency,which achieves the goal of improving the calculation efficiency without lowering the imaging quality.This paper achieves the requirement of both efficiency and quality in ISAR real-time imaging processing.The work done in this paper provides a basis for the research of subsystems,and also provides experience for GPU as a general platform for more radar systems.