Vision-based road signage recognition for autonomous vehicle in agricultural plantation
The growth of self-driving vehicles demands dependable vision-based traffic sign recognition systems to maintain safety and efficiency in agricultural plantations. This research aims to create an enhanced traffic sign identification system based on the YOLOv3 algorithm, which will solve the limitati...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | en |
| Published: |
ZES Rokman Resources
2024
|
| Online Access: | http://eprints.utem.edu.my/id/eprint/28700/2/0145829122024014161535.pdf http://eprints.utem.edu.my/id/eprint/28700/ https://ijafp.org/wp-content/uploads/2024/12/IJAFP15_066.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The growth of self-driving vehicles demands dependable vision-based traffic sign recognition systems to maintain safety and efficiency in agricultural plantations. This research aims to create an enhanced traffic sign identification system based on the YOLOv3 algorithm, which will solve the limitations of standard human vision-based approaches, notably in low-light circumstances and with occlusions. The system leverages computer vision and machine learning techniques, requiring extensive training on diverse datasets to ensure robustness against environmental variations and regional signage differences. Implemented on platforms like Google Colab, the system was trained and tested using a comprehensive dataset, achieving a mean average
precision (mAP) of 96.96%, precision of 94%, and recall of 95%. Despite its high accuracy and effective real-time processing
capabilities, challenges like handling similar signs and occlusions persist. Future work will concentrate on increasing the dataset, refining the model, enhancing occlusion management approaches, and allowing real-time processing on edge devices like the Jetson Nano and Raspberry Pi, boosting system dependability and developing autonomous driving technologies. |
|---|