Pedestrian Detection And Tracking Technology Of Vehicle Infrared Image

On-vehicle infrared technology, as a new area of application in vehicle infrared technology,has attracted more and more attention. Its function in the driver assistance system can be used to recognize the environment and improve driving safety.This paper mainly presents the on-vehicle infrared technology in the application of driver assistance system, And it is studied from the pedestrian detection, recognition and tracking technology on the on-vehicle infrared image separately.Firstly, preprocessing the infrared image through the method of median filter to filter the noise. Then enhancing the contrast of image through the method of histogram equalization. Laying a solid foundation for the pedestrian detection, recognition and tracking below.Secondly, the threshold method and the symmetrical difference method are using to detect the pedestrian. Through the comparison of the two methods, analysis the shorts of the threshold method and put forward a more effective three frame difference method. The target image which is detecting by the new method is more complete, more details, wide application range, and the contour of the target closer to the target itself.Thirdly, a new method of the combination of the geometry characteristics and movement rate of pedestrian is used to recognize the pedestrian.A lower false alarm rate, and more effective achievement on pedestrian recognition has been found by the new method.Finally, during the period of pedestrian tracking, the HU moment invariant theory is introduced first. Then a contour moment matching algorithm is put forward to track the pedestrian based on the HU moment invariant theory.Through the calculation of translational invariance, rotation invariance, and the scale invariance based on the contour moments realize the tracking of the pedestrian.