An improved YOLOv3‐tiny algorithm for vehicle detection in natural scenes

Abstract

YOLO (You Only Look Once), as a target detection algorithm with good speed and precision, is widely used in the industry. In the process of driving, the vehicle image captured by the driving camera is detected and it extracts the license plate and the front part of the vehicle. Compared with the network structure of YOLOv3-tiny algorithm, the acquisition method of anchor box is improved by combining the Birch algorithm. In order to improve the real-time performance, the original two-scale detection is added to the multi-scale prediction of three-scale detection to ensure its accuracy. Finally, the experimental results show that the improved YOLOv3-tiny network structure proposed in this study can improve the performance of mean-average-precision, intersection over union and speed by 5.99%, 17.52% and 48.4%, respectively, and the algorithm has certain robustness.How to cite this article: Huang, B., et al.: An improved YOLOv3-tiny algorithm for vehicle detection in natural scenes. IET Cyber-Syst. Robot. 1-9 (2021).

Grand challenge of image processing in automatic detection of vehicles running in red lights

“…A view-aware local attention network to reidentify vehicles, assign different weights to different views, and adaptively learn the local features in each view [27]. The use of Tiny-Yolo v3 obtains the vehicle image during driving and extracts the license plate and part of the leading vehicle [28]. Adopting the deep learning-based HyperLPR Chinese license plate recognition framework to recognize a wide range of license plates with high accuracy [29].…”

Real‐time vehicle detection using segmentation‐based detection network and trajectory prediction

“…In ref. [47], a YOLOv3‐based network has been proposed, and the acquisition method of the anchor box has been improved by combining the Birch algorithm. Ref.…”

Enhanced lightweight deep network for efficient livestock detection in grazing areas

“…Furthermore, the arithmetic power and cost limitations of embedded devices applied to pastoral areas require lightweight algorithms, while high-precision deep learning algorithms mostly require high arithmetic power equipment support. 16 Thus, driven by the practical needs, a lightweight, efficient, livestock target detection algorithm based on deep networks and adapting to the pastoral environment is urgently required.…”