Scholarlink

<p>The Intelligent Transportation System (ITS) enhances transportation safety by addressing traffic challenges, with vehicle detection and classification as key components. Unmanned aerial vehicles (UAVs) have gained prominence in computer vision applications; however, vehicle classification using UAV imagery remains challenging due to factors such as small object sizes, varied orientations, and environmental impacts. This study introduces a robust method tailored to the unique conditions of the Iraqi Kurdistan Region, leveraging a newly developed dataset of 2,919 images categorized into five vehicle classes (motorcycle, personal car, taxi, truck, and bus). A modified YOLOv4 algorithm, enhanced with the Convolutional Block Attention Module (CBAM), was proposed to improve detection accuracy and focus on critical features. Experimental results show that the proposed system achieves a mean average precision (mAP) of 88.25% and an inference speed of 35 frames per second, outperforming Faster Region-Convolutional Neural Network (Faster RCNN), YOLOv3, and baseline YOLOv4 when both speed and accuracy are considered.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

YOLO-Based Approach for Multiple Vehicle Detection and Classification using UAVs in the Kurdistan Region of Iraq

  • 期刊论文

  • 发表日期：

    2025年3月4日

查看全文
  • Nama Ezzaalddin Mustafa,
  • Fattah Alizadeh

摘要

The Intelligent Transportation System (ITS) enhances transportation safety by addressing traffic challenges, with vehicle detection and classification as key components. Unmanned aerial vehicles (UAVs) have gained prominence in computer vision applications; however, vehicle classification using UAV imagery remains challenging due to factors such as small object sizes, varied orientations, and environmental impacts. This study introduces a robust method tailored to the unique conditions of the Iraqi Kurdistan Region, leveraging a newly developed dataset of 2,919 images categorized into five vehicle classes (motorcycle, personal car, taxi, truck, and bus). A modified YOLOv4 algorithm, enhanced with the Convolutional Block Attention Module (CBAM), was proposed to improve detection accuracy and focus on critical features. Experimental results show that the proposed system achieves a mean average precision (mAP) of 88.25% and an inference speed of 35 frames per second, outperforming Faster Region-Convolutional Neural Network (Faster RCNN), YOLOv3, and baseline YOLOv4 when both speed and accuracy are considered.