Research On Algorithms Of Marine Navigation Radar Detection Based On GPU Parallel Computing

For safe and smooth sailing activities,speed,precision and anti-interference ability of marine navigation radar target detection and localization processing are needed to improved.CFAR(Constant False Alarm Rate)processing technologies widely used in radar target detection makes accuracy of target automatic detection be guaranteed better.However,complex algorithms of automatic detection method make radar systems with strong real-time signal processing capabilities absolutely necessary.For signals’ real-time processing,taking radar signal processing algorithms of automatic detection method and the marine navigation radar target characteristics into consideration,this paper discusses the radar signal processing algorithms and tests the core parts,and then from the project,making use of OpenCL software architecture,the CPU-based and GPU-based marine navigation target detection algorithms are designed and developed under serial and parallel programming modes basing on correct simulations of the algorithms.The analysis of the algorithms’ performance and their results shows that the GPU-based radar signal processing algorithms are more helpful to improve marine navigation radar real-time signal processing capabilities under the premise of high accuracy and robust reliability of the target detection,compared with the CPU-based.