The Design Of Infrared Image Recognition And Tracking System Based On FPGA

As the advantages of the passive of infrared radiation and working day and night, infrared imaging has been widely used in the military field. In image processing, target recognition and tracking are always critical but difficult technologies. In the precision-guided weapons, infrared guidance has its unique advantages compared to radar and laser guidance. The paper has designed a system for the target recognition and tracking in infrared imaging, and uses the hardware platform to achieve it. FPGA is included in the system to implement the algorithm for detection and tracking, it's real-time and more stable than the software.At first, the issue discusses the algorithm for detection and tracking in infrared images. In detection, based on the noise characteristics of infrared images, we put forward the corresponding algorithm for noise reduction. Then, to save the resource of the hardware, the threshold segmentation algorithm is proposed to separate the target from the image. In tracking, in order to ensure the accuracy and reliability of tracking, the paper uses the adaptive window to shield the complex background image and selected the centroid-algorithm to find the center of the target by comparison w ith the center-algorithm.According to the system requirements, we have built the hardware platform. The good hardware system can ensure the stability of the system and the quality of image. The design chooses the right codec chips to adjust the parameters of the image and convert the video signal from digital to analog or from analog to digital. For the critical image processing, we choose the Spartan 3A-DSP chip from Xilinx. It can not only meet the needs of resources, its DSP core can also provides a great convenience to achieve these algorithms.Then, the paper has introduced how these algorithms were realized in FPGA in detail. The design started with the overall planning of the functions and features of the system in top-down design method, including how to design the logic circuits, how to improve effectiveness, how to reduce the consumption of the hardware resource and so on. Then, we divide the entire system into a number of smaller modules and build the connection between them. So we can get the detailed performance parameters of the system and optimize them at any time if necessary, such as the exact threshold segmentation point, the appropriate tracking point and so on. In this design, the system is divided into noise reduction, threshold segmentation, noise detection, edge detection, adaptive window, centroid-finding and so on. At the same time, it has offered the design flow chart or the RTL diagram of each module. Finally, against the background or anthropogenic interference sources, the anti-jamming design is added into the system, which can largely improve the stability of the system.After the design finished, it's the verification of each module. Finally, this issue even tests the system above with the actual aircraft video in infrared video.