Design Of Small Target Detection Circuit For Array Lidar

Gm-APD(Geiger-Mode of Avalanche Photodiodes)array lidar has single photon detection sensitivity and extremely fast photoelectric reaction speed,which can be applied to distant small target detection.For the remote detection of non-cooperative small target,this thesis designs a corresponding small target detection circuit and a corresponding small target detection algorithm based on the64×64 Gm-APD arrays lidar.With the guarantee of frequency of image acquisition,the small target is captured and tracked in real time,and the system detection rate of the target is improved effectively.This thesis firstly introduces the imaging principle of Gm-APD arrays lidar range and intensity image.A small target detection algorithm is applied based on the tiny feature of the target and the Poisson distribution of the noise.The sparse regularization Poisson denoising algorithm is used to compress a few frames of image data to enhance and reconstruct the intensity information of the target.The position of the target is acquired by the intensity information distribution.As a consequence,the capture of the small target is realized.The correlation information of the distance value between the adjacent frame range image is considered to complete the tracking of the target.Additionally,combined with the application requirements of the project and the analysis of relevant technical indicators,the dual processor mode based on the cooperation of DSP and FPGA is selected as the overall hardware design scheme,and the selection of main chips is completed.Besides,the schematic diagram and PCB design of the small target identification circuit,plate welding and debugging are completed.The principle and routing method that need to pay attention to in PCB design are summarized.The circuit debugging results indicate that it is normal.Finally,the software design of each module applied to FPGA and DSP are introduced,and the test and verification are completed on the circuit.The problems encountered in the course of circuit debugging are analyzed,and the corresponding experience are summarized.The target restoration,signal to noise ratio,calculation complexity and calculation time of the small target detection algorithm applied in this paper and the photon cumulative counting method are analyzed and compared.On the premise of guaranteeing real-time performance,the algorithm applied in this paper can significantly improve the precision rate of target detection.