Improve Visibility of Nighttime Camera Images for Pedestrian Recognition on Curve Road
Ryota Ogura, Takeshi Nagasaki, Hitoshi Matsubar · 2020
The purpose of this study is to improve the visibility of nighttime images for recognition nighttime pedestrians. As a method of converting a dark image at night into a bright image, there is a method of adding a continuous image. However, there is a problem that a moving amount is large in a curve road. In our method, we prepared the nighttime image and past converted image and the corresponding daytime images as reference. We improve brightness of nighttime image. We verified the conversion experiments using camera images. As a result, this result suggests that pedestrian recognition is easier at night time.