Automotive Night-vision System Based On Infrared Thermal Imaging

Automotive night-vision system based on infrared thermal imaging is a kind of active safety device of vehicle with infrared thermal imaging technology, assist the driver to see the road ahead at night and visibility, and distinguish the pedestrians in the front of the car, assist the driver to predict the location of the road ahead and pedestrians. prevention of the danger early, reduce the occurrence of traffic accidents.Firstly, this thesis introduces the important role of the vehicle infrared night vision system on the night driving. and the research status of infrared night vision system on the home and abroad. And then do a simple introduction to the discovery of infrared, the infrared thermal imaging principle and the composition of the infrared thermal imager. Next, the basic principle of pedestrian recognition is introduced, including image segmentation, template matching, machine learning, etc.Finally, according to the requirement of on-board infrared night vision system, the overall design of the system and the realization of each function module are given.Vehicle infrared night vision system mainly includes the vehicle infrared thermal imager and the pedestrian recognition module. The design of infrared thermal imaging system includes the components of infrared thermal imaging system, the selection of infrared lens, the selection of infrared detector, the realization of the window heating module,power circuit design,detector driver circuit design,detector temperature control circuit design,sampling circuit design,control and image processing circuit design, video output circuit design, etc.The design of pedestrian recognition system mainly includes system composition, hardware design,and the algorithm of pedestrian recognition.The system uses UL03262 infrared detector that receives 8～14μm infrared ray with 384*288 resolution is made by ULIS. infrared detector receives the analog signal of target infrared radiation after photoelectric conversion output with image information, to collect the signal with high speed ADC, converted to FPGA, the noise, the infrared image blind element replacement, nonuniformity correction, gray-scale transformation, character overlay operation FPGA in conjunction with MCU, output to an external video decoder, eventually transformed into a composite video signal (CVBS) transmitted to the pedestrian detection module,the image segmentation, matching module, pedestrian recognition operation, determining whether there is a pedestrian in the image, and will determine the superposition of information transfer to the car audio system.