YOLOv4目标检测算法在煤矿工人口罩佩戴监测工作中的应用研究*

为防止煤矿工人吸入过量粉尘而导致职业性尘肺病,基于Keras框架利用YOLOv4 (you only look once)目标检测算法对井下人员佩戴防尘口罩情况进行高精度且快速的检测与识别,并与MTCNN(Multi-task convolutional neural network)和FaceNet构成的人脸识别算法相结合,进行煤矿工人口罩佩戴监测的研究。结果表明:模型对井下人员口罩佩戴有较高的检测精度,识别已佩戴口罩的矿井下作业人员的平均精度达到92.78%,识别未佩戴防尘口罩检测的平均精度为91.63%,与其他主流算法相比算法具有更好的鲁棒性和检测效果。研究结果为预防煤矿工人职业性尘肺病提供1种有效的技术手段。     

Occupational safety and health education under the lifelong learning framework in Serbia

Serbia is aligning with European Union requirements and the occupational safety and health (OSH) administration is one of the most representative sectors of this alignment. Many efforts were made in this field, by introducing new laws and regulations, but it turned out to be insufficient. OSH professionals need to renovate and strengthen their knowledge in accordance with continuous, updated and improved OSH standards and regulation. Lifelong learning (LLL) programmes can contribute to forming professionals who are always up to date. This paper presents an implemented LLL programme, over the duration of two academic years, dedicated to OSH professionals, and investigates whether this programme will be helpful and accepted by professionals. The results from the study show that the given LLL programme had indeed a positive influence on the professional careers of the participants and that the LLL presents the future trend in OSH education.     

How do children learn to cross the street? The process of pedestrian safety training

Objective: Pedestrian injuries are a leading cause of child death and may be reduced by training children to cross streets more safely. Such training is most effective when children receive repeated practice at the complex cognitive–perceptual task of judging moving traffic and selecting safe crossing gaps, but there is limited data on how much practice is required for children to reach adult levels of functioning. Using existing data, we examined how children's pedestrian skills changed over the course of 6 pedestrian safety training sessions, each composed of 45 crossings within a virtual pedestrian environment.

Methods: As part of a randomized controlled trial on pedestrian safety training, 59 children ages 7–8 crossed the street within a semi-immersive virtual pedestrian environment 270 times over a 3-week period (6 sessions of 45 crossings each). Feedback was provided after each crossing, and traffic speed and density were advanced as children's skill improved. Postintervention pedestrian behavior was assessed a week later in the virtual environment and compared to adult behavior with identical traffic patterns.

Results: Over the course of training, children entered traffic gaps more quickly and chose tighter gaps to cross within; their crossing efficiency appeared to increase. By the end of training, some aspects of children's pedestrian behavior was comparable to adult behavior but other aspects were not, indicating that the training was worthwhile but insufficient for most children to achieve adult levels of functioning.

Conclusions: Repeated practice in a simulated pedestrian environment helps children learn aspects of safe and efficient pedestrian behavior. Six twice-weekly training sessions of 45 crossings each were insufficient for children to reach adult pedestrian functioning, however, and future research should continue to study the trajectory and quantity of child pedestrian safety training needed for children to become competent pedestrians.     

An empirical examination of personal learning within the context of teams

Using a sample of 588 employees in 59 work teams, we tested a model that situates personal learning within the context of teams, viewing it as a joint function of teams' leadership climate (i.e., transformational leadership) and task characteristics (i.e., task routineness and task interdependence). Consistent with our hypotheses, we found that the positive relationships between transformational leadership climate and the two dimensions of personal learning (relational job learning and personal skill development) were moderated by the nature of the teams' tasks. Specifically, transformational leadership climate was more strongly associated with personal learning for members of teams working on tasks that were less routine, rather than more routine. However, no significant moderation was found for leadership climate and task interdependence. Our findings underscore the importance of taking into account the contextual conditions within which leadership influence occurs while also demonstrating the potential role that leaders can play in promoting employees' personal learning. Overall, our study bolsters theories that conceptualize adult learning as a transaction between people and their social environments and points to a practical need to match leadership styles with team task characteristics to unleash transformational leadership effects. Copyright © 2015 John Wiley & Sons, Ltd.     

Learning for social-ecological change: a qualitative review of outcomes across empirical literature in natural resource management

Learning is considered as a promising mechanism to cope with rapid environmental change. The implications of learning for natural resource management (NRM) have not been explored in-depth and the evidence on the topic is scattered across multiple sources. We provide a qualitative review of types of learning outcomes and consider their manifestations in NRM across selected empirical literature. We conducted a systematic search of the peer-reviewed literature (N = 1,223) and a qualitative meta-synthesis of included articles, with an explicit focus on learning outcomes and NRM changes (N = 53). Besides social learning, we found several learning concepts used, including policy and transformative learning, and multiple links between learning and NRM reported. We observe that the development of skills, together with a system approach involving multi-level capacities, is decisive for implications of learning for NRM. Future reviews could systematically compare how primary research applies different learning concepts and discusses links between learning and NRM changes.     

湖泊水质富营养化评价的模糊神经网络方法

为了探索人工神经网络用于湖泊营养化评价的可能性,提出了基于多准则学习的模糊神经网络湖泊水质营养化评价模型。该模型应用于我国五大主要湖泊水质营养化的评价结果表明,模糊神经网络用于湖泊水质营养化评价,具有简便、实用、客观性和广泛的通用性。      

RS-PSO-ELM下腐蚀管道失效压力预测

为提高腐蚀管道失效压力的预测精度并简化其计算过程,提出基于粗糙集(RS)和粒子群算法(PSO)融合极限学习机(ELM)的腐蚀管道失效压力预测模型。通过属性约简提取影响失效压力的关键因素,选用PSO优化ELM的输入权值和隐含层偏差,将归一化的核心指标数据代入计算。结果表明:该模型预测结果与实际值基本一致,与单一ELM模型相比,预测结果的均方差(MSE)降至0.255;与其他蚀管道失效压力评价模型相比,该模型预测结果的绝对误差平均值降至0.32。     

Identifying socioeconomic and biophysical factors driving forest loss in protected areas

Protected areas (PAs) are a commonly used strategy to confront forest conversion and biodiversity loss. Although determining drivers of forest loss is central to conservation success, understanding of them is limited by conventional modeling assumptions. We used random forest regression to evaluate potential drivers of deforestation in PAs in Mexico, while accounting for nonlinear relationships and higher order interactions underlying deforestation processes. Socioeconomic drivers (e.g., road density, human population density) and underlying biophysical conditions (e.g., precipitation, distance to water, elevation, slope) were stronger predictors of forest loss than PA characteristics, such as age, type, and management effectiveness. Within PA characteristics, variables reflecting collaborative and equitable management and PA size were the strongest predictors of forest loss, albeit with less explanatory power than socioeconomic and biophysical variables. In contrast to previously used methods, which typically have been based on the assumption of linear relationships, we found that the associations between most predictors and forest loss are nonlinear. Our results can inform decisions on the allocation of PA resources by strengthening management in PAs with the highest risk of deforestation and help preemptively protect key biodiversity areas that may be vulnerable to deforestation in the future.     

MSWNet: A visual deep machine learning method adopting transfer learning based upon ResNet 50 for municipal solid waste sorting

● MSWNet was proposed to classify municipal solid waste. ● Transfer learning could promote the performance of MSWNet. ● Cyclical learning rate was adopted to quickly tune hyperparameters. An intelligent and efficient methodology is needed owning to the continuous increase of global municipal solid waste (MSW). This is because the common methods of manual and semi-mechanical screenings not only consume large amount of manpower and material resources but also accelerate virus community transmission. As the categories of MSW are diverse considering their compositions, chemical reactions, and processing procedures, etc., resulting in low efficiencies in MSW sorting using the traditional methods. Deep machine learning can help MSW sorting becoming into a smarter and more efficient mode. This study for the first time applied MSWNet in MSW sorting, a ResNet-50 with transfer learning. The method of cyclical learning rate was taken to avoid blind finding, and tests were repeated until accidentally encountering a good value. Measures of visualization were also considered to make the MSWNet model more transparent and accountable. Results showed transfer learning enhanced the efficiency of training time (from 741 s to 598.5 s), and improved the accuracy of recognition performance (from 88.50% to 93.50%); MSWNet showed a better performance in MSW classsification in terms of sensitivity (93.50%), precision (93.40%), F1-score (93.40%), accuracy (93.50%) and AUC (92.00%). The findings of this study can be taken as a reference for building the model MSW classification by deep learning, quantifying a suitable learning rate, and changing the data from high dimensions to two dimensions.     

Approaching the upper boundary of driver-response relationships: identifying factors using a novel framework integrating quantile regression with interpretable machine learning

● A novel framework integrating quantile regression with machine learning is proposed. ● It aims to identify factors driving observations to upper boundary of relationship. ● Increasing N:P and TN concentration help fulfill the effect of TP on CHL. ● Wetter and warmer decrease potential and increase eutrophication control difficulty. ● The framework advances applications of quantile regression and machine learning. The identification of factors that may be forcing ecological observations to approach the upper boundary provides insight into potential mechanisms affecting driver-response relationships, and can help inform ecosystem management, but has rarely been explored. In this study, we propose a novel framework integrating quantile regression with interpretable machine learning. In the first stage of the framework, we estimate the upper boundary of a driver-response relationship using quantile regression. Next, we calculate “potentials” of the response variable depending on the driver, which are defined as vertical distances from the estimated upper boundary of the relationship to observations in the driver-response variable scatter plot. Finally, we identify key factors impacting the potential using a machine learning model. We illustrate the necessary steps to implement the framework using the total phosphorus (TP)-Chlorophyll a (CHL) relationship in lakes across the continental US. We found that the nitrogen to phosphorus ratio (N׃P), annual average precipitation, total nitrogen (TN), and summer average air temperature were key factors impacting the potential of CHL depending on TP. We further revealed important implications of our findings for lake eutrophication management. The important role of N׃P and TN on the potential highlights the co-limitation of phosphorus and nitrogen and indicates the need for dual nutrient criteria. Future wetter and/or warmer climate scenarios can decrease the potential which may reduce the efficacy of lake eutrophication management. The novel framework advances the application of quantile regression to identify factors driving observations to approach the upper boundary of driver-response relationships.