Affiliation:
1. Auckland University of Technology, New Zealand
Abstract
In this book chapter, the authors propose a low-cost distance estimation approach to develop more accurate predictions from a 3D perspective for vehicle detection and ranging by using inexpensive monocular cameras. This distance estimation model integrates YOLOv7 model with an attention module (CBAM) and transformer, as well as extend the prediction vector as the fundamental architecture to improved high-level semantic understanding and enhanced feature extraction ability. This integration significantly improved detection and ranging performance, offering a more suitable and cost-effective solution for distance estimation.
Reference34 articles.
1. Alexey, B. ChienYao, W., & Mark, L. (2020). YOLOv4: Optimal speed and accuracy of object detection. Image and Video Processing, arXiv:2004.10934.
2. Fully convolutional neural networks for LIDAR–camera fusion for pedestrian detection in autonomous vehicle
3. An Approach for Precise Distance Measuring Using Ultrasonic Sensors.;F.Aliew;Engineering Proceedings,2022
4. AlvaradoS. T.BorjaM. G. B.TorresK. B. (2022). Object Distance Estimation from a Binocular Vision System for Robotic Applications Using Artificial Neural Networks. Control, Mechatronics and Automation. ICCMA.
5. Deep learning–based automated measurements of the scrotal circumference of Norwegian Red bulls from 3D images