Zuzanna Woźniak

Occupational accidents remain one of the key challenges for modern work systems, both from the perspective of employee health and life, and in terms of economic, social, and organisational losses for society. Despite the significant diagnostic value of near misses, data on such events have not been incorporated as input variables in predictive models. This study aimed to develop, experimentally test, and empirically validate the effectiveness of various mathematical models in predicting occupational accidents. The performance of thirteen machine learning and deep learning algorithms, including Random Forest, logistic regression, SVC, ARIMA, SARIMAX, and a proposed modified convolutional neural network incorporating Bayesian correction and a tailored ReLU activation function, was compared. The best results were achieved with a CNNKT model, yielding an accuracy of 0.939, an F1-score of 0.934, and a G-Mean of 0.942. The findings confirm the superiority of deep convolutional networks over traditional statistical models in occupational safety analysis.

The study investigates repetitive and time-consuming professional activities performed by occupational health and safety (OHS) specialists in the construction sector in Poland and their attitudes toward the use of modern digital tools, including solutions based on artificial intelligence (AI). The research was conducted using a questionnaire survey, with a purposive sample and a snowball method. A total of 102 individuals participated in the study, of whom 94 valid responses were included in the analysis. The data were examined using descriptive statistics and chi-square tests. The results showed that the most repetitive and time-consuming activities include documentation analysis, report preparation, inspections, and communication. Nearly 46% of respondents indicated that selected elements of their work could be automated or supported by digital tools, while 33% reported using AI-based solutions in everyday practice. Statistically significant relationships were identified between respondents’ age and both their level of concern about new technologies and their perception of technological support potential. No significant relationships were found for enterprise ownership or size. The findings indicate substantial potential for the implementation of digital and AI-supported tools in routine OHS activities. Future research should involve larger and more homogeneous samples, incorporate probabilistic sampling, and explore organisational and competence-related factors influencing technology adoption.

Despite preventive measures, the construction industry continues to exhibit high accident rates. In response, visual detection system was developed to support safety management on construction sites and promote sustainable working environments. The solution integrates the YOLOv8 algorithm with asynchronous video processing, incident registration, an open API, and a web-based interface. The system detects the absence of safety helmets (NHD) and worker falls (FD). Its low hardware requirements make it suitable for small and medium-sized construction enterprises, contributing to resource efficiency and digital transformation in line with sustainable development goals. This study advances practice by providing an integrated, low-resource solution that unites multi-hazard detection, event documentation, and system interoperability, addressing a key gap in existing research and implementations. The contribution includes an operational architecture proven to run in real time, addressing a gap between model-centred research and deployable, OHS applications. The system was validated using two independent test datasets, each comprising 100 images: one for NHD and one for FD. For NHD, the system achieved a precision of 0.93, an accuracy of 0.88, and an F1-score of 0.79. For FD, a precision of 1.00, though with a limited recall of 0.45. The results demonstrate the system’s potential for sustainable construction site safety monitoring.

The construction sector records a significant number of occupational accidents (A) and near-misses (NM), making it one of the most dangerous in the economy. In recent years, interest in near-miss events has been growing among researchers and practicing engineers, as they are considered precursors to occupational accidents. Based on a review of the literature on the subject and their own experience, the authors of the article conclude that there is a significant gap in research on near misses in the Polish construction industry. The authors believe that such studies are necessary in the context of accident reduction. The purpose of this article is to analyze the time series of near misses and accidents at work. The data used in the study come from the system of registration of hazardous events implemented in one of the Polish construction companies, recorded in 2015–2022. Due to the specific nature of construction work and the circumstances of the event, 8 categories of hazardous events were specified. For each category, a time series was built to inform about the dynamics of the changes taking place. Box plots were developed for random variables representing the time intervals between consecutive events, informing about the statistical characteristics of a given set of events (SHEi ). This research makes it possible to predict the occurrence of specific events over time and to introduce preventive measures in construction practice.

In most countries worldwide, the construction industry is recognized as one of the sectors with an exceptionally high incidence of accidents. The inherent complexity of construction processes, coupled with the variability of implementation conditions, engenders a multitude of disruptions. These disruptions pose a threat to the life and health of construction employees and can give rise to hazardous events, including accidents and near-misses. The research focuses on identifying direct and indirect causes of hazardous events, intending to determine which factors have the greatest impact on their occurrence. A comprehensive dataset was systematically analyzed, comprising 754 recorded accidents at work (A) and 1717 near-miss events (NM). The research was conducted based on a model of the course of a hazardous event that was developed for the purposes of this study. A categorization system was devised, encompassing eight categories of direct causes and 12 categories of indirect causes. The significance of causes was assessed through the utilization of Pareto-Lorenz analysis and ABC methodology. The most significant factors contributing to occupational accidents were identified: inadequate work organization, improper use of materials by employees, and failure to adhere to instructions. The most significant factors contributing to the occurrence of near misses were identified: improper use of material by an employee, inadequate organization of the workstation, and failure to adhere to instructions.

The construction industry is widely recognised as one of the most hazardous industries due to the complexity of its projects and the rapid implementation of investment. A significant contributing factor to injuries and fatalities in this sector is human falls. The prevalence of accidents at work, particularly those attributed to falls, necessitates the implementation of urgent preventive measures and innovative technological solutions. The objective of the research outlined in this paper is the development of an algorithm capable of detecting human falls in real time. This solution is innovative in its application on construction sites, where its use will facilitate the identification of incidents related to human falls, as well as the provision of rapid assistance to the injured. In the case of such injuries, urgent medical intervention is in most cases crucial to save the health and life of the injured person. The intelligent safety monitoring system is designed to utilise machine learning (ML) algorithms for realtime image analysis with the Python language and the implementation of the proposed algorithm. The YOLOv8 algorithm is specifically responsible for people detection. The proposed detection algorithm is predicated on a two-stage analysis: initial detection of human presence and subsequent verification of whether a human has fallen. In the event of a fall, the developed algorithm automatically generates an alert. The model was tested in construction areas in order to reproduce the natural environment of its target application. A set of images containing both cases of construction worker falls and situations in which no incident occurred was used.