Construction-Hazard-Detection Computer Vision Project

This project focuses on enhancing safety at construction sites by leveraging AI-driven hazard detection. Utilising the YOLO model for object detection, the system identifies potential hazards such as workers without helmets or safety vests, workers near machinery or vehicles, and workers in restricted areas. The project integrates real-time analysis and alert mechanisms to ensure immediate response to identified hazards.

GitHub Repository:

For more details and to access the source code, visit the Construction Hazard Detection GitHub repository.

Key Features:

• Real-Time Detection: Instant identification of safety violations and potential hazards.
• Multi-Language Support: Notifications and interface available in multiple languages including Traditional Chinese, Simplified Chinese, French, English, Thai, Vietnamese, and Indonesian.
• Integration with Messaging Apps: Real-time notifications and images sent via LINE, Messenger, WeChat, and Telegram.
• Customisable Detection Items: Configurable detection parameters to suit various safety requirements.

Dataset Information:

The dataset used for training includes images from the Construction Site Safety Image Dataset by Roboflow, enriched with additional annotations. The labels include:

• Hardhat
• Mask
• NO-Hardhat
• NO-Mask
• NO-Safety Vest
• Person
• Safety Cone
• Safety Vest
• Machinery
• Vehicle

Models for Detection:

Model	Size (pixels)	mAP (val 50)	mAP (val 50-95)	Params (M)	FLOPs (B)
YOLO11n	640	54.1	31.0	2.6	6.5
YOLO11s	640	70.1	44.8	9.4	21.6
YOLO11m	640	73.3	42.6	20.1	68.0
YOLO11l	640	77.3	54.6	25.3	86.9
YOLO11x	640	82.0	61.7	56.9	194.9

Our comprehensive dataset ensures robust detection capabilities, making construction sites safer and more efficient.

                        @misc{
                            construction-hazard-detection-sp8tx_dataset,
                            title = { Construction-Hazard-Detection Dataset },
                            type = { Open Source Dataset },
                            author = { Computer Vision },
                            howpublished = { \url{ https://universe.roboflow.com/computer-vision-naktq/construction-hazard-detection-sp8tx } },
                            url = { https://universe.roboflow.com/computer-vision-naktq/construction-hazard-detection-sp8tx },
                            journal = { Roboflow Universe },
                            publisher = { Roboflow },
                            year = { 2024 },
                            month = { dec },
                            note = { visited on 2024-12-22 },
                            }