Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Enhancing Construction Site Safety: A Lightweight Convolutional Network for Effective Helmet Detection

Mujadded Al Rabbani Alif

TL;DR

This work lays a foundational framework for ongoing adjustments and optimization in automated helmet detection technology, with future enhancements expected to address the limitations identified during these initial experiments.

Abstract

In the realm of construction safety, the detection of personal protective equipment, such as helmets, plays a critical role in preventing workplace injuries. This paper details the development and evaluation of convolutional neural networks (CNNs) designed for the accurate classification of helmet presence on construction sites. Initially, a simple CNN model comprising one convolutional block and one fully connected layer was developed, yielding modest results. To enhance its performance, the model was progressively refined, first by extending the architecture to include an additional convolutional block and a fully connected layer. Subsequently, batch normalization and dropout techniques were integrated, aiming to mitigate overfitting and improve the model's generalization capabilities. The performance of these models is methodically analyzed, revealing a peak F1-score of 84\%, precision of 82\%, and recall of 86\% with the most advanced configuration of the first study phase. Despite these improvements, the accuracy remained suboptimal, thus setting the stage for further architectural and operational enhancements. This work lays a foundational framework for ongoing adjustments and optimization in automated helmet detection technology, with future enhancements expected to address the limitations identified during these initial experiments.

Enhancing Construction Site Safety: A Lightweight Convolutional Network for Effective Helmet Detection

TL;DR

This work lays a foundational framework for ongoing adjustments and optimization in automated helmet detection technology, with future enhancements expected to address the limitations identified during these initial experiments.

Abstract

In the realm of construction safety, the detection of personal protective equipment, such as helmets, plays a critical role in preventing workplace injuries. This paper details the development and evaluation of convolutional neural networks (CNNs) designed for the accurate classification of helmet presence on construction sites. Initially, a simple CNN model comprising one convolutional block and one fully connected layer was developed, yielding modest results. To enhance its performance, the model was progressively refined, first by extending the architecture to include an additional convolutional block and a fully connected layer. Subsequently, batch normalization and dropout techniques were integrated, aiming to mitigate overfitting and improve the model's generalization capabilities. The performance of these models is methodically analyzed, revealing a peak F1-score of 84\%, precision of 82\%, and recall of 86\% with the most advanced configuration of the first study phase. Despite these improvements, the accuracy remained suboptimal, thus setting the stage for further architectural and operational enhancements. This work lays a foundational framework for ongoing adjustments and optimization in automated helmet detection technology, with future enhancements expected to address the limitations identified during these initial experiments.
Paper Structure (22 sections, 11 figures, 12 tables)

This paper contains 22 sections, 11 figures, 12 tables.

Table of Contents

  1. Introduction
  2. Literature Review
  3. Methodology
  4. Dataset
  5. Dataset Visualization and Partitioning
  6. Data Augmentation
  7. Crop Augmentation
  8. Rotation Augmentation
  9. Brightness Augmentation
  10. Proposed Architecture
  11. Final Model Architecture
  12. Experimental Results
  13. Hyperparameter Tuning
  14. Initial Architecture Performance with Original Dataset
  15. Initial Architecture Performance with Augmented Dataset
  16. ...and 7 more sections

Figures (11)

  • Figure 1: Comparison of annual fatal injuries across major industry sectors for 2021/22 with averages from the five-year period 2017/18 to 2021/22 RN3
  • Figure 2: Visual Comparison of Construction Site Workers: (a) Equipped with Safety Helmet, (b) Not Wearing Safety Helmet.
  • Figure 3: Illustration of Crop Augmentation Technique Showing 35% Reduction: (a) Original Image, (b) Post-Augmentation
  • Figure 4: Demonstration of 30° Rotation Augmentation on an Image: (a) Original, (b) Rotated Counterclockwise by 30°, (c) Rotated Clockwise by 30°
  • Figure 5: Example of 20° Rotation Augmentation: (a) Original Image, (b) Rotated Counterclockwise 20°, (c) Rotated Clockwise 20°
  • ...and 6 more figures