Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

GSO-YOLO: Global Stability Optimization YOLO for Construction Site Detection

Yuming Zhang, Dongzhi Guan, Shouxin Zhang, Junhao Su, Yunzhi Han, Jiabin Liu

TL;DR

GSO-YOLO addresses challenging construction-site object detection by integrating a Global Optimization Module for broad context, a Steady Capture Module for temporal stability, and an AIoU loss that blends CIoU and EIoU for accurate and efficient localization. Built on YOLOv8, the approach reorganizes the architecture to place GOM before SPPF and SCM after SPPF, and applies AIoU throughout, enhancing detection of small and occluded objects under variable lighting and long ranges. Empirical results on SODA, MOCS, and CIS show consistent SOTA gains in mAP50 and mAP50–95, with ablations confirming the additive benefits of GOM, SCM, and AIoU. The work demonstrates strong generalization and practical potential for automated safety monitoring in real-world construction environments.

Abstract

Safety issues at construction sites have long plagued the industry, posing risks to worker safety and causing economic damage due to potential hazards. With the advancement of artificial intelligence, particularly in the field of computer vision, the automation of safety monitoring on construction sites has emerged as a solution to this longstanding issue. Despite achieving impressive performance, advanced object detection methods like YOLOv8 still face challenges in handling the complex conditions found at construction sites. To solve these problems, this study presents the Global Stability Optimization YOLO (GSO-YOLO) model to address challenges in complex construction sites. The model integrates the Global Optimization Module (GOM) and Steady Capture Module (SCM) to enhance global contextual information capture and detection stability. The innovative AIoU loss function, which combines CIoU and EIoU, improves detection accuracy and efficiency. Experiments on datasets like SODA, MOCS, and CIS show that GSO-YOLO outperforms existing methods, achieving SOTA performance.

GSO-YOLO: Global Stability Optimization YOLO for Construction Site Detection

TL;DR

GSO-YOLO addresses challenging construction-site object detection by integrating a Global Optimization Module for broad context, a Steady Capture Module for temporal stability, and an AIoU loss that blends CIoU and EIoU for accurate and efficient localization. Built on YOLOv8, the approach reorganizes the architecture to place GOM before SPPF and SCM after SPPF, and applies AIoU throughout, enhancing detection of small and occluded objects under variable lighting and long ranges. Empirical results on SODA, MOCS, and CIS show consistent SOTA gains in mAP50 and mAP50–95, with ablations confirming the additive benefits of GOM, SCM, and AIoU. The work demonstrates strong generalization and practical potential for automated safety monitoring in real-world construction environments.

Abstract

Safety issues at construction sites have long plagued the industry, posing risks to worker safety and causing economic damage due to potential hazards. With the advancement of artificial intelligence, particularly in the field of computer vision, the automation of safety monitoring on construction sites has emerged as a solution to this longstanding issue. Despite achieving impressive performance, advanced object detection methods like YOLOv8 still face challenges in handling the complex conditions found at construction sites. To solve these problems, this study presents the Global Stability Optimization YOLO (GSO-YOLO) model to address challenges in complex construction sites. The model integrates the Global Optimization Module (GOM) and Steady Capture Module (SCM) to enhance global contextual information capture and detection stability. The innovative AIoU loss function, which combines CIoU and EIoU, improves detection accuracy and efficiency. Experiments on datasets like SODA, MOCS, and CIS show that GSO-YOLO outperforms existing methods, achieving SOTA performance.
Paper Structure (14 sections, 8 equations, 7 figures, 3 tables)

This paper contains 14 sections, 8 equations, 7 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. Method
  4. Global Optimization Module
  5. Steady Capture Module
  6. Augmented Intersection over Union
  7. Overall Architecture Adjustment
  8. Experiment
  9. Datasets
  10. Implement Detail
  11. Comparison with the SOTA Results
  12. Ablation Study
  13. Generalization Study
  14. Conclusion

Figures (7)

  • Figure 1: Comparison between GSO-YOLO and YOLOv8.
  • Figure 2: The GSO-YOLO overall architecture.
  • Figure 3: mAP-Epochs curves.
  • Figure 4: mAP-classes curves. (a)mAP-classes curves on SODA. (b) mAP-classes curves on MOCS. (c) mAP-classes curves on CIS.
  • Figure 5: Original image example.
  • ...and 2 more figures