Loading [MathJax]/extensions/MathZoom.js
Road Target Detection Based on Lightweight Improved YOLOv5l | IEEE Conference Publication | IEEE Xplore

Road Target Detection Based on Lightweight Improved YOLOv5l


Abstract:

In the current stage of autonomous driving technology, road target detection is a key problem for performing related operations. However, the most of the target detection...Show More

Abstract:

In the current stage of autonomous driving technology, road target detection is a key problem for performing related operations. However, the most of the target detection models are not suitable or difficult to be applied in practical real-time road target detection due to the problems of the large parameter computation and the long discriminative time. To address this series of issues, In this paper, we propose a method based on the lightweight improvement of YOLOv5l used into the target detection of road target. Firstly, the Slim-Neck structure is introduced into the original model's Neck to decrease the volume of operations and increase the accuracy of the original model; secondly, the framework which is used in the backbone network of YOLOv5l is altered by the PP-LCNet network structure to reduce the large volume of operations required by the model in the feature extraction of images and substantially increase the operating speed of the model; Finally, inspired by RFBNet, the SPPF layer of the original model is replaced by the RFB structure in RFBNet to prevent the loss of model accuracy in the case of a large reduction of floating point operations and to increase the original model's detection accuracy. According to the experimental results, the improved YOLOv5l model is compared with the original YOLOv5l model, the size of model is reduced to 43.33% of the original size, the operation parameters of model are reduced to 43.95% of the original size, the floating point operations of model are reduced to 28.23% of the original size, and the accuracy of model is improved by 2.46% to 95.77%. The result of experiments shows that the model which is proposed in this paper based on the lightweight improved YOLOv5l can be more efficiently and accurately applied in the practical scenarios for road target detection.
Date of Conference: 23-25 June 2023
Date Added to IEEE Xplore: 16 November 2023
ISBN Information:

ISSN Information:

Conference Location: Dalian, China

I. NTRODUCTION

According to statistics, in the first eight months of 2020, China's national production of the electric vehicles reached 3.97 million units, corresponding to the sales of electric vehicles also reached 3.86 million units, a year-on-year growth of about 1 times, and China's electric vehicle ownership reached 10 million units, accounting for 22.9% of China's overall vehicle sales, while its export volume also reached 340,000 units [1]. Along with China's new energy strategy, technologies such as autonomous driving, which were previously limited by the lack of electric power supply in the cars themselves, have become possible. Also because of the increase of car ownership in China, how to efficiently detect road traffic targets during driving is of great importance for automatic vehicle driving technology, regular driving of drivers, etc. How to detect road targets more efficiently, quickly and accurately is an important criterion for road target detection.

References

References is not available for this document.