Risk Factor Recognition for Automatic Safety Management in Construction Sites Using Fast Deep Convolutional Neural Networks

Many industrial accidents occur at construction sites. Several countries are instating safety management measures to reduce industrial accidents at construction sites. However, there are few technical measures relevant to this task, and there are safety blind spots related to differences in human resources’ capabilities. We propose a deep convolutional neural network that automatically recognizes possible material and human risk factors in the field regardless of individual management capabilities. The most suitable learning method and model for this study’s task and environment were experimentally identified, and visualization was performed to increase the interpretability of the model’s prediction results. The fine-tuned Safety-MobileNet model showed a high performance of 99.79% (30 ms), demonstrating its high potential to be applied in actual construction sites. In addition, via visualization, the cause of the model’s confusion of classes could be found in a dataset that the model did not predict correctly, and insights for result analysis could be presented. The material and human risk factor recognition model presented in this study can contribute to solving various practical problems, such as the absence of accident prevention systems, the limitations of human resources for safety management, and the difficulties in applying safety management systems to small construction companies.

E-contact tracing of hospital infections – validation of SIRSZ automatic risk assessment

Aim. Contact networks play a crucial role in infectious disease propagation and position in the network mediate risk of acquiring or sending infections. We studied the spread of hospital-associated infections through computer simulations and validated our ‘computer assisted’ risk assessment with ‘human’ risk assessment in a prospective study.Concept. We collected time-varying structure of contacts and covariates reconstructed from Polish Hospitals:1. The organisational structure is mapped by a set of questionnaires, CAD maps integration, functional paths annotation and local vision. It is done mostly by surveys within medical staff through an interactive web application.2. The Cohabitation layer processes data from the registry of patient admissions and discharges from each hospital unit (wards, clinics, etc.) and medical shift register. With simulated infection paths, we were able to compute network centrality measures for patients. We obtained the risk of getting infected, based on the patient’s incoming connections, and the risk of spreading infections resulting from outgoing connections. We compare various standard centrality measures – position of patients and staff in contact networks (‘computer assisted’ risk assessment) of both contacts and paths networks, with a predictor of ‘human’ risk perception (based on 190 patients).Results. We showed that the best predictor of HAI risk is Adjusted Rage Rank on paths (r= 0.42, p < 0.01). However, surprisingly good predictive power in risk assessment was found in the betweenness centrality of the underlying network of contacts (r = 0.30, p < 0.01).Conclusion: We conclude that epidemiology of a given pathogen in a given place and time could be explained only with the contact network only to a large extent. However, further possibility of the collection, processing and storage of the data on individual persons, translated to mathematical modelling could lead in future to satisfactory improvement in risk assessment.

Monitoring and Quantitative Human Risk Assessment of Municipal Solid Waste Landfill Using Integrated Satellite–UAV–Ground Survey Approach

Landfills are the dominant method of municipal solid waste (MSW) disposal in many developing countries, which are extremely susceptible to failure under circumstances of high pore water pressure and insufficient compaction. Catastrophic landfill failures have occurred worldwide, causing large numbers of fatalities. Tianziling landfill, one of the largest engineered sanitary landfills in China, has experienced massive deformation since January 2020, making early identification and monitoring of great significance for the purpose of risk management. The human risk posed by potential landfill failures also needs to be quantitatively evaluated. The interferometric synthetic aperture radar (InSAR) technique, unmanned aerial vehicle (UAV) photogrammetry, and ground measurements were combined to obtain landfill deformation data in this study. The integrated satellite–UAV–ground survey (ISUGS) approach ensures a comprehensive understanding of landfill deformation and evolution. The deformation characteristics obtained using the InSAR technique and UAV photogrammetry were analyzed and compared. A close relationship between the most severe mobility events, precipitation episodes, and was observed. Based on early hazard identification using ISUGS, a quantitative risk assessment (QRA) method and F-N curves were proposed, which can be applied to landfills. The comparison showed that ISUGS allowed a better understanding of the spatial and temporal evolution of the landfill and more accurate QRA results, which could be as references for local governments to take effective precautions.

Adoption of contract farming for managing agricultural risks: a case study in rice production in the Mekong Delta, Vietnam

PurposeFarmers' risk attitudes and risk perceptions play an essential role in shaping risk management strategies to address risks and uncertainties. Contract farming is considered as one of the feasible approaches to tackle farmers' concerns. However, risk perspectives under various categories have not been included in studies on farmers' preferences for contract farming in the literature, especially in Vietnam. This study aims to determine factors affecting farmers' choices of different contract farming practices.Design/methodology/approachThe explanatory factor analysis (EFA) and multinomial logit model (MNL) were applied to explore the impacts of risk perspectives on farmers' preferences for contract farming. Data have been collected from 211 rice farmers in An Giang Province, “the rice bowl” of the Mekong Delta, Vietnam.FindingsThe study found that farm size, cooperatives, extension, market access and trust have significantly impacted on contract participation while a delay payment was a barrier for farmers' motivation to opt for the contract. Farmers' contract choices were also influenced by their risk attitudes and perceptions under different risk dimensions. The financial, policy and human risk-averse behavior predisposed farmers to single out the full contract while the policy and human risk-loving and production, market and finance risk-averse respondents were in favor of the marketing contract. Moreover, the findings indicated that the more farmers concerned about risk of weather and market, the more choices for the full contract, whereas the risk perceptions of weather and policy encouraged farmers to use the limited contract. By contrast, farmers who perceived the impacts of risk of diseases/pests and human were likely to adopt the marketing contract.Research limitations/implicationsThis study just focuses on collecting data from farmers’ perspective. Future studies involving stakeholders such as enterprises and policy makers are strongly recommended so as to design suitable contracts and enforce contract schemes effectively in Vietnam.Originality/valueThe findings also contribute to the literature on different types of contracts and the multidimensional aspect of risk for rice production in Vietnam.