Deep learning for site safety: Real-time detection of personal protective equipment

Nath, Nipun D.; Behzadan, Amir H.; Paal, Stephanie German

doi:10.1016/j.autcon.2020.103085

Cited by 284 publications

(140 citation statements)

References 28 publications

Supporting

Mentioning

132

Contrasting

Unclassified

Order By: Relevance

“…Precision and recall are calculated to evaluate the PPE detection module, which are summarized in Table 2. The PPE detection module achieved higher precision and recall compared to [15], which similarly detected the PPE within the worker's bounding box.…”

Section: Implementation and Case Studymentioning

confidence: 99%

Nested Network for Detecting PPE on Large Construction Sites Based on Frame Segmentation

Mohammad¹,

Zhenhua²,

Amin³

2020

Proceedings of the Creative Construction E-Conference 2020

View full text Add to dashboard Cite

Safety is a main concern for the construction industry because of the high rate of accidents and casualties on construction sites. Personal Protective Equipment (PPE) is a major part of safety regulations to prevent accidents. However, workers may neglect to wear the required PPE while working, which subsequently increases the potential risk for accidents. Currently, safety managers and inspectors on construction sites are responsible for monitoring safety regulations, which is a time-consuming task. To facilitate safety monitoring, a large number of research studies applied computer vision for detecting PPE on construction sites. Nevertheless, detecting workers and PPE is still a challenge in far-field videos. This research proposes an approach for detecting if anyone on the construction site is wearing the required PPE, even when he or she is far from the surveillance cameras. This method uses a frame segmentation technique and a nested network with two Faster R-CNN models to detect safety noncompliances. The first model detects the human bodies on the construction site, and the second one detects if the detected person is wearing a hardhat and a safety vest. The proposed method is applied to videos from a construction site. The experimental results demonstrate the practicality and robustness of the proposed method to detect PPE in far-field videos. Based on three different test videos, the average precision and recall for the worker detection model were 99.67% and 92.92%, respectively. The PPE detection model had the average precision and recall of 91.25% and 94.77%, respectively.

show abstract

Section: Implementation and Case Studymentioning

confidence: 99%

Nested Network for Detecting PPE on Large Construction Sites Based on Frame Segmentation

Mohammad¹,

Zhenhua²,

Amin³

2020

Proceedings of the Creative Construction E-Conference 2020

View full text Add to dashboard Cite

show abstract

“…For less operational constraints, two smartphones have been used as stereo cameras to acquire motion data and extract 3D human skeletons to track people working in construction fields [98]. Real-time machine learning models with CNN frameworks have been proposed to detect whether workers are wearing safety equipment, such as hats and vests, from images/videos [99] and to detect ground objects [100]. CNNs have also been used to detect safety guardrails [101], objects on roof construction sites [102], workers who fail to wear hard hats [103], [104], falls from heights [105], [106], to maintain safe distances among objects for safety to prevent accidents [107] and unsafe behaviours [73].…”

Section: A Related Work In the Construction Fieldmentioning

confidence: 99%

A Deep Learning Approach Towards Railway Safety Risk Assessment

Alawad

Kaewunruen

2020

IEEE Access

View full text Add to dashboard Cite

Railway stations are essential aspects of railway systems, and they play a vital role in public daily life. Various types of AI technology have been utilised in many fields to ensure the safety of people and their assets. In this paper, we propose a novel framework that uses computer vision and pattern recognition to perform risk management in railway systems in which a convolutional neural network (CNN) is applied as a supervised machine learning model to identify risks. However, risk management in railway stations is challenging because stations feature dynamic and complex conditions. Despite extensive efforts by industry associations and researchers to reduce the number of accidents and injuries in this field, such incidents still occur. The proposed model offers a beneficial method for obtaining more accurate motion data, and it detects adverse conditions as soon as possible by capturing fall, slip and trip (FST) events in the stations that represent high-risk outcomes. The framework of the presented method is generalisable to a wide range of locations and to additional types of risks.

show abstract

“…Accordingly, a comparison of the performance of different algorithms performed by Liu et al (2016) reveals that only the YOLO (Redmon et al, 2016;Redmon and Farhadi, 2017) algorithm can perform detection in real time and, hence, fulfills the implementation requirements of this study. In the construction domain, YOLO is used for detecting construction machines (e.g., truck and excavator) (Xiao and Kang, 2019) and personal protective equipment (e.g., hard hat and vest) (Nath et al, 2020).…”

Section: Overview Of Fast Object Detection Algorithmsmentioning

confidence: 99%

“…Next, performance of each model is measured using a commonly used metric in object detection (and information retrieval), mean average precision (mAP), a single numerical value that represents the effectiveness of the entire system (Turpin and Scholer, 2006;Ren et al, 2017). To calculate mAP, first, intersection over union (IoU), i.e., the percentage of overlap between ground-truth boxes and predicted boxes (Nath et al, 2020), is measured using Equation (1) where G and P are the ground-truth and predicted boxes, respectively.…”

Section: Model Testingmentioning

confidence: 99%

“…From planning to execution, the profound use of AI is steadily leading the industry toward automation and autonomous systems. Example includes but are not limited to machine-generated project planning for estimation of project cost (Cheng et al, 2010), or designing the site layout (Yahya and Saka, 2014), automated monitoring of construction activities and safety of workers (Ding et al, 2018;Nath et al, 2018Nath et al, , 2020, and deployment of robots to perform construction tasks (Bock, 2015). While new sensing techniques are used for collecting multi-modal data, digital photography (using cameras, or mobile devices such as smartphones and tablet computers) is by far one of the most common ways of documenting fieldwork.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Deep Convolutional Networks for Construction Object Detection Under Different Visual Conditions

Nath

Behzadan

2020

Front. Built Environ.

Self Cite

View full text Add to dashboard Cite

Sensing and reality capture devices are widely used in construction sites. Among different technologies, vision-based sensors are by far the most common and ubiquitous. A large volume of images and videos is collected from construction projects every day to track work progress, measure productivity, litigate claims, and monitor safety compliance. Manual interpretation of such colossal amounts of data, however, is non-trivial, error-prone, and resource-intensive. This has motivated new research on soft computing methods that utilize high-power data processing, computer vision, and deep learning (DL) in the form of convolutional neural networks (CNNs). A fundamental step toward machine-driven interpretation of construction site scenery is to accurately identify objects of interest for a particular problem. The accuracy requirement, however, may offset the computational speed of the candidate method. While lightweight DL algorithms (e.g., Mask R-CNN) can perform visual recognition with relatively high accuracy, they suffer from low processing efficacy, which hinders their use in real-time decision-making. One of the most promising DL algorithms that balance speed and accuracy is YOLO (you-only-look-once). This paper investigates YOLO-based CNN models in fast detection of construction objects. First, a large-scale image dataset, named Pictor-v2, is created, which contains about 3,500 images and approximately 11,500 instances of common construction site objects (e.g., building, equipment, worker). To assess the agility of object detection, transfer learning is used to train two variations of this model, namely, YOLO-v2 and YOLO-v3, and test them on different data combinations (crowdsourced, web-mined, or both). Results indicate that performance is higher if the model is trained on both crowdsourced and web-mined images. Additionally, YOLO-v3 outperforms YOLO-v2 by focusing on smaller, harder-to-detect objects. The best-performing YOLO-v3 model has a 78.2% mAP when tested on crowdsourced data. Sensitivity analysis of the output shows that the model's strong suit is in detecting larger objects in less crowded and well-lit spaces. The proposed methodology can also be extended to predict the relative distance of the detected objects with reliable accuracy. Findings of this work lay the foundation for further research on technology-assistive systems to augment human capacities in quickly and reliably interpreting visual data in complex environments.

show abstract

Deep learning for site safety: Real-time detection of personal protective equipment

Cited by 284 publications

References 28 publications

Nested Network for Detecting PPE on Large Construction Sites Based on Frame Segmentation

Nested Network for Detecting PPE on Large Construction Sites Based on Frame Segmentation

A Deep Learning Approach Towards Railway Safety Risk Assessment

Deep Convolutional Networks for Construction Object Detection Under Different Visual Conditions

Contact Info

Product

Resources

About