Small-scale construction sites in South Korea account for about 91.5% of all construction workplaces and contribute to 72.3% of the total accidents and fatalities. Safety measures at these sites are often underestimated, and proper safety education is lacking. In particular, the fatality rate is about 4.43 times higher compared to medium-/large-scale construction sites. In this study, a systematic analysis was conducted to examine the causes and trends of industrial accidents in small-scale construction sites to address these issues. This study analyzed industrial accidents in small-scale construction sites using statistical analysis, LDA topic modeling, and network analysis based on data from the Korea Occupational Safety and Health Agency (KOSHA) from 2018 to 2022. The analysis revealed that the most critical cause of accidents in small-scale construction sites is ‘Scaffolding and working platforms’, with accidents primarily involving ‘Fall’. Furthermore, various risk factors and accident trends were identified in apartment construction, new building projects, and mobile scaffolding usage. This study systematically analyzed the causes and trends of industrial accidents at small-scale construction sites, providing important evidence to enhance safety management and preventive measures. The results are expected to play a crucial role in establishing a safety culture at construction sites and ensuring the wellbeing of construction workers.

This study compared the influencing factors affecting safety behavior and perception between main contractor managers (CMs) and subcontractor workers (SWs) at Korean construction sites. The safety level, policies, and management capabilities of the main contractor are closely related to the safety behavior of the SWs since CMs have a key role in ensuring work order authority and safety responsibility during the work process. To compare the difference in safety behavior between CMs and SWs, ten hypotheses were prepared, and surveying was conducted. Using the responses of 1219 CMs and 1191 SWs, a frequency analysis, intensive validity analysis, and confirmatory factor analysis were performed; discriminant validity was verified; and a structural equation model was constructed and compared. The results showed that the factors affecting safety behaviors and perceptions were significantly different between CMs and SWs. The WSB (workers’ safety behavior) for CMs was affected, from most to least, by the MSP (manager’s safety perception; 0.382), incentives (0.166), WSP (workers’ safety perception; 0.143), and the MSB (manager’s safety behavior; 0.134). However, for SWs, the WSB was affected, from most to least, by the MSP (0.440), WSP (0.184), the MSB (0.130), and incentives (0.083). Awarding an incentive as a way to encourage safe behavior is an important influencing factor affecting safety behavior for CMs, while workers’ safety perception (WSP) is an important factor for SWs. The results can contribute to the implementation of effective safety and accident prevention activities at construction sites by comparing the influencing factors for the safety behaviors and perceptions of CMs and SWs, which can increase the effectiveness of the safety climate and reduce the possibility of accidents.

Before starting any construction work, providing workers with awareness about past similar accident cases is effective in preventing mishaps. Based on construction accident reports, this study developed two models to identify past accidents at sites with similar site information. The site information includes 16 parameters, such as type of work, type of accident, the work in which the accident occurred, weather conditions, contract conditions, type of work, etc. The first model, the site information classification model, uses named entity recognition tasks to classify site information, which is extracted from accident reports. The second model, the similar-site accident retrieval model, which finds the most similar accidents that occurred in the past from input site information, uses a semantic textual similarity task to match the classified information with it. A total of 17,707 accident reports from South Korean construction sites were found; these models were trained to use Korean Language Understanding Evaluation–Bidirectional Encoder Representations from Transformers (KLUE-BERT) for processing. The first model achieved an average accuracy of 0.928, and the second model was precisely matched, with a mean cosine similarity score exceeding 0.90. These models could identify and provide workers with similar past accidents, enabling proactive safety measures, such as site-specific hazard identification and worker education, thereby allowing recognition of construction safety risks before starting work. By integrating site information with historical data, the models offer an effective approach to improving construction safety.

With increasing industrial sophistication and complexity, workplaces are increasingly prone to occupational accidents, causing negative impacts on workers and employers, including economic losses and decreased productivity. South Korea occupational safety and health has implemented new policies addressing potential risks to overcome stagnation in industrial accident reduction and predict site accidents from past cases. Cases are human-classified according to rules, including occurrence type or original causal materials. However, human errors, subjective judgments, synonyms, and terms incorrectly used by classifiers reduce original data quality and impede developments or applications of policies, technologies, and methods preventing accidents based on past accidents. This study proposes three artificial intelligence models to objectively classify the occurrence type of accident cases. Models are developed based on a natural language processing model (KoBERT), which considers Korean language characteristics. Each model is tested by sequentially performing sentence preprocessing, keyword replacement, and morphological analysis. The proposed Model 3 exhibits 93.1% accuracy, which was the highest among tested models. Up to three classification categories for occurrence type are allowed to assist objective classification. The accident case-based occurrence type classification model is effective for industrial accident prevention, aiding in strategy development and reducing social costs.

This study proposes a method for observing the inside of a temporary structure covered with a construction safety net using an image processing technique and 3D modeling. Images of the temporary structures were obtained using unmanned aerial vehicles (UAVs). Observing the inside of temporary structures is limited to using UAVs due to construction safety nets that act as obstacles. To solve this problem, the histogram equalization and grayscale stretching techniques were used to visualize the inside of a temporary structure covered by safety nets. After verifying the proposed method with small scale specimens, the proposed method was applied to climbing formwork installed at a construction site. Three-dimensional modeling technique was combined with improved images that were processed using the proposed method. The results indicated that the proposed method could identify internal objects that were not visible owing to safety nets. By comparing the original and improved images obtained through applying the proposed method, four target objects located inside the climbing formwork that were not visible owing to safety nets were sufficiently identified in the improved images. Thus, the proposed method can be effectively applied to identify the internal objects of temporary structures covered with safety nets.