Pedestrian detection for automotive night vision using thermal camera

The sudden appearance of pedestrian is a major worry for the driver especially during night time since it may lead to an accident due to poor human eyesight at night. To overcome the problem, a pedestrian detection for automotive night vision is introduced. The analysis is done based on the distance, size and clothing. There are four condition has been taken into consideration that are with average weight of 40 to 50 kg with wearing hijab and without hijab, and to the average weight of 60 to 70 kg wearing hijab and without hijab. The first and second condition have 0% error for emitting the first and second warning while the third condition has 9.3% error during first warning and 0% error for second warning. The fourth condition has 23.5% error for the first warning and 0% error for the second warning.