Seat Belt Fastness Detection Based on Image Analysis from Vehicle In-Cabin Camera

Abstract

Seat belt fastness detection in vehicles is the important factor due to the high protection role in case an accident occurs. Modern vehicles usually have belt fastness detection systems that can be simply tricked. There are also algorithms that can recognise seat belt fastness based on driver visual monitoring. Unfortunately, the existing algorithms are not so efficient and car manufactures do not implement them to vehicles. Most of them based on Hough, Canny, or other edge detection. In this paper, we introduce a new model to classify driver seat belt status using a camera inside the driver cabin. Our model based on YOLO neural network for detection the driver seat belt fastness. Retraining We solve the problem of belt detection in two steps: main part of the belt detection and corner detection. These steps allow us to recognise situation when the seat belt is fasten behind the human body. We use Tiny-YOLO to detect the main part of the belt as the first object as well as belt corner as a second object. We classify belt fastness between three cases: the belt is not fastened, the belt is fasten correctly, and the belt is fastened behind the back.

Keywords

• seat belt detection
• neural networks
• convolution neural network
• yolo
• computer vision
• data processing
• image processing