Long-term Coherent Accumulation Algorithm Research And GPU Implementation

The Radon-Fourier transform(RFT)algorithm is an effective means to achieve longterm coherent accumulation for target detection.However,the traditional RFT algorithm is difficult to be applied in practical scenario,because it not only produces a large number of blind velocity side-lobes,but also needs to search the motion parameters of the target,which leads to the high complexity of the algorithm.To address the above-mentioned problems,this thesis carried out the research on blind velocity side-lobe suppression of the RFT algorithm and its GPU(Graphics Processing Unit)parallelization implementation.The main work is summarized as follows:1.The impact of range cell migration caused by moving target on energy accumulation is analyzed,and the application of the RFT algorithm in longterm coherent accumulation is briefly introduced.The feasibility of the RFT algorithm is validated by computer simulations.2.In order to solve the problem of blind speed side-lobe in the RFT algorithm,blind speed side-lobe suppression methods based on PRI(Pulse Repetition Interval)design and CLEAN processing are studied respectively.On this basis,a blind speed side-lobe suppression algorithm based on peak side-lobe ratio is proposed.Compared with the blind speed side-lobe suppression algorithm based on PRI,the proposed algorithm only needs to transmit a set of PRI pulses,which greatly reduces the computational complexity.Compared with the blind speed side-lobe suppression algorithm based on CLEAN processing,the proposed algorithm can suppress the blind speed side-lobes caused by multiple targets in the same distance unit without affecting the detection probability,making its application scenarios more extensive.3.In order to solve the problem that the generalized RFT algorithm needs to search the target motion parameters,which leads to a large amount of computation and is not conducive to real-time detection,a parallel processing algorithm based on GPU is proposed.The thirdorder motion model of the target is established,and the parallel search of the generalized RFT algorithm is completed by using the computer unified device architecture(CUDA),which makes the algorithm speed up by about 2 orders of magnitude.At the same time,the visual design of the generalized RFT algorithm is completed by using Matlab GUI(Graphic User Interface),and the results of the algorithm are displayed through the human-computer interaction interface.