Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

LAM-YOLO: Drones-based Small Object Detection on Lighting-Occlusion Attention Mechanism YOLO

Yuchen Zheng, Yuxin Jing, Jufeng Zhao, Guangmang Cui

TL;DR

This work proposes LAM-YOLO, an object detection model specifically designed for drone-based target detection, and introduces a light-occlusion attention mechanism to enhance the visibility of small targets under different lighting conditions and utilizes an improved SIB-IoU as the regression loss function to accelerate model convergence and enhance localization accuracy.

Abstract

Drone-based target detection presents inherent challenges, such as the high density and overlap of targets in drone-based images, as well as the blurriness of targets under varying lighting conditions, which complicates identification. Traditional methods often struggle to recognize numerous densely packed small targets under complex background. To address these challenges, we propose LAM-YOLO, an object detection model specifically designed for drone-based. First, we introduce a light-occlusion attention mechanism to enhance the visibility of small targets under different lighting conditions. Meanwhile, we incroporate incorporate Involution modules to improve interaction among feature layers. Second, we utilize an improved SIB-IoU as the regression loss function to accelerate model convergence and enhance localization accuracy. Finally, we implement a novel detection strategy that introduces two auxiliary detection heads for identifying smaller-scale targets.Our quantitative results demonstrate that LAM-YOLO outperforms methods such as Faster R-CNN, YOLOv9, and YOLOv10 in terms of mAP@0.5 and mAP@0.5:0.95 on the VisDrone2019 public dataset. Compared to the original YOLOv8, the average precision increases by 7.1\%. Additionally, the proposed SIB-IoU loss function shows improved faster convergence speed during training and improved average precision over the traditional loss function.

LAM-YOLO: Drones-based Small Object Detection on Lighting-Occlusion Attention Mechanism YOLO

TL;DR

This work proposes LAM-YOLO, an object detection model specifically designed for drone-based target detection, and introduces a light-occlusion attention mechanism to enhance the visibility of small targets under different lighting conditions and utilizes an improved SIB-IoU as the regression loss function to accelerate model convergence and enhance localization accuracy.

Abstract

Drone-based target detection presents inherent challenges, such as the high density and overlap of targets in drone-based images, as well as the blurriness of targets under varying lighting conditions, which complicates identification. Traditional methods often struggle to recognize numerous densely packed small targets under complex background. To address these challenges, we propose LAM-YOLO, an object detection model specifically designed for drone-based. First, we introduce a light-occlusion attention mechanism to enhance the visibility of small targets under different lighting conditions. Meanwhile, we incroporate incorporate Involution modules to improve interaction among feature layers. Second, we utilize an improved SIB-IoU as the regression loss function to accelerate model convergence and enhance localization accuracy. Finally, we implement a novel detection strategy that introduces two auxiliary detection heads for identifying smaller-scale targets.Our quantitative results demonstrate that LAM-YOLO outperforms methods such as Faster R-CNN, YOLOv9, and YOLOv10 in terms of mAP@0.5 and mAP@0.5:0.95 on the VisDrone2019 public dataset. Compared to the original YOLOv8, the average precision increases by 7.1\%. Additionally, the proposed SIB-IoU loss function shows improved faster convergence speed during training and improved average precision over the traditional loss function.

Paper Structure

This paper contains 27 sections, 18 equations, 11 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Object Detection Methods
  4. Two Stage Detection Methods
  5. One Stage Detection Methods
  6. Transformer-Based Detection Methods
  7. YOLO Series Drone-based Object Detection
  8. Vision Attention Mechanisms
  9. Method
  10. Overall Structure of LAM-YOLO
  11. Improvements In Network Structure
  12. Lighting-occlusion Attention Module
  13. Block of Involution
  14. Soft Inner-section Bounding-box IoU
  15. Experiments and Analysis
  16. ...and 12 more sections

Figures (11)

  • Figure 1: The complex background faced by image target detection, the red and green bounding boxes are zoom in area: (a)Dense target of different types such as pedestrians and cars, (b)Target occlusion by other objects such as trees, (c)Strong lighting conditions of sunlight, and (d) Dim lighting conditions during nights.
  • Figure 2: LAM-YOLO structure. The green part is the backbone, which contains CSPDarkNetwang2020cspnet, the yellow part is FPN lin2017feature, the purple part is PAN PAN, the orange part is Head, the blue detection head is the original YOLOv8, and the green detection head is the auxiliary detection head we introduced.
  • Figure 3: Lighting-occlusion attention module.The gray module represents Residential Vision aware Attention Group (RVAG), the orange module represents Vision aware Attention Block (VAB), the green module represents Overlapping Light aware Attention Block (OLAB)
  • Figure 4: The structure of the Involution block
  • Figure 5: Inner-IoU description
  • ...and 6 more figures