ASHFAQ, MUHAMMAD SHARJIL (2025) REAL-TIME TRAFFIC SIGN DETECTION USING YOLO AND DEEP LEARNING FOR AUTONOMOUS VEHICLES. S1 thesis, Universitas Mercu Buana Jakarta.
![[img]](https://repository.mercubuana.ac.id/style/images/fileicons/text.png) |
Text (HAL COVER)
01 COVER.pdf Download (437kB) |
|
Text (BAB I)
02 BAB 1.pdf Restricted to Registered users only Download (147kB) |
|
Text (BAB II)
03 BAB 2.pdf Restricted to Registered users only Download (269kB) |
|
Text (BAB III)
04 BAB 3.pdf Restricted to Registered users only Download (252kB) |
|
Text (BAB IV)
05 BAB 4.pdf Restricted to Registered users only Download (962kB) |
|
Text (BAB V)
06 BABA 5.pdf Restricted to Registered users only Download (102kB) |
|
Text (DAFTAR PUSTAKA)
07 DAFTAR PUSTAKA.pdf Restricted to Registered users only Download (97kB) |
|
Text (LAMPIRAN)
08 LAMPIRAN.pdf Restricted to Registered users only Download (269kB) |
Abstract
Real-time and accurate detection of traffic signs is essential for the safe and efficient operation of autonomous vehicles within intelligent transportation systems. Traffic signs provide critical navigational cues, yet detecting them in realworld conditions remains challenging due to factors such as variable weather, lighting conditions, occlusions, and diverse sign appearances. This study evaluates and compares the performance of two deep learning-based object detection models—You Only Look Once version 8 (YOLOv8) and Single Shot Multibox Detector (SSD300)—for real-time traffic sign recognition. Both models were trained on a diverse traffic sign dataset using data augmentation techniques to improve generalization across various conditions. The models were assessed using precision, recall, and mean Average Precision at a 0.50 IoU threshold (mAP@50), along with confusion matrices to analyze detection accuracy and misclassification trends. Performance comparison results show that YOLOv8 significantly outperformed SSD300, achieving a precision of 0.9517, recall of 0.9002, and mAP@50 of 0.9577, while SSD300 reached a precision of 0.6321, recall of 0.6325, and mAP@50 of 0.7224. YOLOv8 demonstrated faster inference and higher accuracy, making it more suitable for real-time applications, whereas SSD300, despite its lower performance, offers advantages in scenarios with limited computational resources due to its smaller model size. These findings emphasize the trade-offs between speed, accuracy, and model complexity, offering practical insights for selecting appropriate detection models in autonomous driving environments. Keywords: Autonomous Vehicles, Traffic Sign Detection, YOLOv8, SSD300, Realtime Detection, Deep Learning, Precision, Recall, mAP@50, Confusion Matrix.
Actions (login required)
 |
View Item |
