Research On System Of Detection And Tracking Of Dim And Small Targets In Starry Background

The space debris issue couldn't be ignored in terms of human spaceflight activities.With the door to space opened by human at the first time,a growing number of space debris is been left in out space.There are more than 100 million debris in different size so far.Humans need to deal with space debris together immediately.The existing space debris observation means are foundation on earth and space-based system.Space-based system has obvious advantages over foundation in real-time,capacity and accuracy.This dissertation focus on research of detecting and tracking of space debris by space optical camera,which is of great significance to space flight and national security.The image of the space debris obtained from the space optical camera is called dim and small target in starry background image.This dissertation studies the detection and tracking of dim and small moving targets,and puts forward feasible project in both algorithm implementation and the hardware platform design.While the SNR(signal to noise rate)of dim and small target in starry background is very low,small object,lack of texture,structure and color characteristic,the dissertation improves the detection and tracking algorithm based on Kalman filtering and the detection and tracking algorithm based on improved Particle Filter.When the signal to noise ratio is low(5 ? SNR ? 3.5),the experiments of the detection rate is higher than 95%,the false alarm rate is lower than 10% and the RMS(root mean square)of tracking is prior than 1.5pixel,which has a good prospect for engineering application.In the hardware platform realization,the system takes latest TI eight-core DSP TMS320C6678 as a central processing chip,the maximum operating frequency up to 10 Ghz.The Xilinx's FPGA XC5VLX20 T is used as the master chip.The image data is flowing from USB2.0 interface into the FPGA,through the SRIO FPGA to DSP communication.At last,the processed data will be uploaded to the host computer through UART port.The dissertation gives detail introduction and analysis from the overall structure,power supply design,clock distribution to peripheral interface.The complete design and implementation scheme is given also.Finally,the dissertation introduces the circuit signal integrity analysis and the matter should be pay attention in high frequency PCB design process.At last,the system is given related debug.The dissertation also introduces the USB2.0 firmware and multi-core DSP debugging related operations.Under the support of MCSDK platform,the speed of hardware could reach 25 fps.At last the small target detection and tracking algorithm and hardware design and implementation of the system are summarized and pointting out that the algorithm and hardware platform need be further integrated in order to enhance the overall performance of the system.