IR-UWB 레이더와 차영상 기반의 보행자 검출

Abstract

Security systems using images have a simple monitoring function only. Such systems are difficult to use in low image brightness, because it is not possible to clearly detect the shape, shadow, and outline of an object. Even though the performance of object detection can be improved by using infrared and thermal imaging cameras, commercialization is difficult due to high cost. In the present study, using images together with IR-UWB radar is proposed. Because image processing technologies are sensitive to surrounding environments such as brightness and background texture, detection is made by using IR-UWB radar to supplement it. Distance of a pedestrian is measured on the IR-UWB radar. With the measured distance information, existence of a pedestrian is determined. If a pedestrian is not detected in the image, detection is made by adjusting the threshold with the distance information of the IR-UWB radar. If the pedestrian is not detected even after adjusting the threshold, the existence of a pedestrian can be assumed with the distance information. The improvement of pedestrian detection performance based on image processing and IR-UWB radar has been proved through the real-time experiments detecting a moving target in the low brightness environment.