Scholarlink

<p>Nowadays, airport detection in remote sensing images has attracted considerable attention due to its strategic importance in both civilian and military applications. In particular, uncrewed aerial vehicles must immediately detect safe areas for emergency landings. The previous approaches have suffered from various aspects, including complex backgrounds, scales, and shapes of the airport. Meanwhile, the rapid action and accuracy of the method are confronted with significant concerns. Therefore, this study proposes an effective scheme by extending the YOLOV3 and ShearLet transform. In this way, MobileNetV2 and ResNet18, with fewer layers and parameters retrained on a similar dataset, are parallelly trained as base networks. According to the geometrical characteristics of the airport, the ShearLet filters with different scales and directions are considered in the first convolutional layers of ResNet18 as a visual attention mechanism. Besides, the major extension in YOLOV3 concerns the detection of sub-networks with novel structures that boost object expression ability and training efficiency. In addition, novel augmentation and negative mining strategies are presented to significantly increase the localization phase’s performance. The experimental results on the DIOR dataset reveal that the framework reliably detects different types of airports in varied areas and acquires robust results in complex scenes compared to traditional YOLOV3 and state-of-the-art approaches.</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

Eynet: extended YOLO for airport detection in remote sensing images

  • 期刊论文

  • 发表日期：

    2025年12月27日

查看全文
  • Hengameh Mirhajianmoghadam,
  • Behrouz Bolourian Haghighi

摘要

Nowadays, airport detection in remote sensing images has attracted considerable attention due to its strategic importance in both civilian and military applications. In particular, uncrewed aerial vehicles must immediately detect safe areas for emergency landings. The previous approaches have suffered from various aspects, including complex backgrounds, scales, and shapes of the airport. Meanwhile, the rapid action and accuracy of the method are confronted with significant concerns. Therefore, this study proposes an effective scheme by extending the YOLOV3 and ShearLet transform. In this way, MobileNetV2 and ResNet18, with fewer layers and parameters retrained on a similar dataset, are parallelly trained as base networks. According to the geometrical characteristics of the airport, the ShearLet filters with different scales and directions are considered in the first convolutional layers of ResNet18 as a visual attention mechanism. Besides, the major extension in YOLOV3 concerns the detection of sub-networks with novel structures that boost object expression ability and training efficiency. In addition, novel augmentation and negative mining strategies are presented to significantly increase the localization phase’s performance. The experimental results on the DIOR dataset reveal that the framework reliably detects different types of airports in varied areas and acquires robust results in complex scenes compared to traditional YOLOV3 and state-of-the-art approaches.