Modified shape independent method in detecting pedestrians using FIR cameras at night

Major percentage of car accidents occur during nighttime due to the lack of sunlight, making it difficult to see external factors like control devices, traffic flow and road markings between the vehicle and its surroundings. With these parameters, car safety systems; active and passive, are developed to prevent car accidents. Some active systems developed for night vision are for imminent rear end collision, overtaking vehicles, blind spots and pedestrians but pedestrians are more vulnerable in car accidents. This led the researchers to develop the modified shape independent method to accurately detect pedestrians at night time. For night vision, IR cameras provides clear pixels but with regards to pedestrian, FIR camera provides the thermographic patterns that will detect pedestrians from non-pedestrian regions. The study´s main concern in this study is to accurately detect pedestrians on the roadside. High-temperature objects with thermographic patterns assumed to be pedestrians and are detected by processing the images from infrared cameras attached on the vehicle. In the course of this study, the researchers were able to detect pedestrians consistently with an 86.27% accuracy.