The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)
Online ISSN : 2424-3124
2023
Session ID : 1A2-G28
Conference information

Obstacle detection using a monocular camera that can handle bad weather
*Eiki ITAKURAJianming YANG
Author information
CONFERENCE PROCEEDINGS RESTRICTED ACCESS

Details
Abstract

Currently, automated driving has reached level 3, and research and development is progressing toward level 4. Driver and passenger safety must be guaranteed in automated driving, there are various causes of accidents, but the accident rate in bad weather is about five times higher than in fair weather. Focusing on this, AI is used to recognize obstacles in bad weather. Specifically, first, the features of the vehicle are extracted. Next, the original image is fog-processed to create fog image data. Let the AI learn it. In addition, in terms of cost reduction, we aimed at obstacle recognition with a low-cost monocular camera.

Content from these authors
© 2023 The Japan Society of Mechanical Engineers
Previous article Next article
feedback
Top