Objective expertise evaluation of individuals,as a prerequisite stage for team formation,has been a long-term desideratum in large software development companies.With the rapid advancements in machine learning methods...Objective expertise evaluation of individuals,as a prerequisite stage for team formation,has been a long-term desideratum in large software development companies.With the rapid advancements in machine learning methods,based on reliable existing data stored in project management tools’datasets,automating this evaluation process becomes a natural step forward.In this context,our approach focuses on quantifying software developer expertise by using metadata from the task-tracking systems.For this,we mathematically formalize two categories of expertise:technology-specific expertise,which denotes the skills required for a particular technology,and general expertise,which encapsulates overall knowledge in the software industry.Afterward,we automatically classify the zones of expertise associated with each task a developer has worked on using Bidirectional Encoder Representations from Transformers(BERT)-like transformers to handle the unique characteristics of project tool datasets effectively.Finally,our method evaluates the proficiency of each software specialist across already completed projects from both technology-specific and general perspectives.The method was experimentally validated,yielding promising results.展开更多
Modern manufacturing processes have become more reliant on automation because of the accelerated transition from Industry 3.0 to Industry 4.0.Manual inspection of products on assembly lines remains inefficient,prone t...Modern manufacturing processes have become more reliant on automation because of the accelerated transition from Industry 3.0 to Industry 4.0.Manual inspection of products on assembly lines remains inefficient,prone to errors and lacks consistency,emphasizing the need for a reliable and automated inspection system.Leveraging both object detection and image segmentation approaches,this research proposes a vision-based solution for the detection of various kinds of tools in the toolkit using deep learning(DL)models.Two Intel RealSense D455f depth cameras were arranged in a top down configuration to capture both RGB and depth images of the toolkits.After applying multiple constraints and enhancing them through preprocessing and augmentation,a dataset consisting of 3300 annotated RGB-D photos was generated.Several DL models were selected through a comprehensive assessment of mean Average Precision(mAP),precision-recall equilibrium,inference latency(target≥30 FPS),and computational burden,resulting in a preference for YOLO and Region-based Convolutional Neural Networks(R-CNN)variants over ViT-based models due to the latter’s increased latency and resource requirements.YOLOV5,YOLOV8,YOLOV11,Faster R-CNN,and Mask R-CNN were trained on the annotated dataset and evaluated using key performance metrics(Recall,Accuracy,F1-score,and Precision).YOLOV11 demonstrated balanced excellence with 93.0%precision,89.9%recall,and a 90.6%F1-score in object detection,as well as 96.9%precision,95.3%recall,and a 96.5%F1-score in instance segmentation with an average inference time of 25 ms per frame(≈40 FPS),demonstrating real-time performance.Leveraging these results,a YOLOV11-based windows application was successfully deployed in a real-time assembly line environment,where it accurately processed live video streams to detect and segment tools within toolkits,demonstrating its practical effectiveness in industrial automation.The application is capable of precisely measuring socket dimensions by utilising edge detection techniques on YOLOv11 segmentation masks,in addition to detection and segmentation.This makes it possible to do specification-level quality control right on the assembly line,which improves the ability to examine things in real time.The implementation is a big step forward for intelligent manufacturing in the Industry 4.0 paradigm.It provides a scalable,efficient,and accurate way to do automated inspection and dimensional verification activities.展开更多
The changes in land use in the last 30 years in the territory of agro-forest watershed of Lake Buyo resulted in significant sediment into the lake. Sediments are a preferred means of transportation for certain polluta...The changes in land use in the last 30 years in the territory of agro-forest watershed of Lake Buyo resulted in significant sediment into the lake. Sediments are a preferred means of transportation for certain pollutants, like phosphorus in excess. By mapping the source areas of erosion, the authors can determine the risk areas and help to prioritize interventions in the territory. This mapping is done using the SWAT （soil and water assessment tool） model. Several types of data, including topography, land use, soil and climate data are needed to run the model. In this paper, all different steps are presented, from the designing of HRU （hydrological response units）, basic units to run the SWAT model until the simulations. The establishment of HRU has three main stages： space discretization, land use and soil data integration and HRU distribution： （1） space discretization which consist in extracting the limits and the water network of the watershed from the DEM （digital elevation model） and in subdividing them into sub-basins; （2） land use and soil data integration： it consists in digitizing the physical maps of land use and of soils under Mapinfo 7.5 and in keeping them in ＂shape＂ format; （3） HRU distribution： it leads to subdivide the sub-watersheds in small units that combine a single soil type and one type of land use. It appears from this study to obtain 23 sub-watersheds and 71 HRU. Once the HRU designed, it is necessary to integrate climate data, data on physico-chemical characteristics of soils and agricultural practices, before starting the simulations. This will allow the model to assess the risk of sedimentation and eutrophication of the lake using the MUSLE （modified universal soil loss equation） and phosphorus cycle.展开更多
The Lobo watershed is an agricultural area where the use of fertilizers by farmers is intensive, causing eutrophication problems that deteriorate the quality of drinking water distributed to the population. Since the ...The Lobo watershed is an agricultural area where the use of fertilizers by farmers is intensive, causing eutrophication problems that deteriorate the quality of drinking water distributed to the population. Since the phenomenon of eutrophication is directly linked to runoff, it is essential to model the flow in order to better control the transfer of nutrients responsible for eutrophication. It is within this framework that this study was conducted. The objective of this study is to assess the ability of the semi-distributed SWAT (Soil and Water Assessment Tool) model to simulate runoff in the Lobo watershed. The methodology adopted was based on the use of the QSWAT graphical interface to manipulate and execute the main functions of the SWAT model from QGIS tools. The hydrological modeling was carried out with the QSWAT interface for SWAT 2012. The results showed good performance for the flow calibration (1982-1984) with the evaluation criteria R<sup>2</sup>, NSE and PBIAS respectively of 0.64, 0.64 and 3.1. In the validation period (1984-1987), the model also showed good performance in the streamflow simulation for R<sup>2</sup> and NSE of 0.84 and 0.76 respectively as values. However, for the PBAIS criterion, the result was less good but still remains satisfactory with a value of 19.6. It emerges from this study that the SWAT model is suitable for simulating water transfer and can therefore be used to study the transfer of pollutants in the fight against eutrophication in the Lobo watershed.展开更多
土壤与水评估工具(soil and water assessment tool,SWAT)作为一种广泛用于农业非点源(non-point source,NPS)污染模拟的分布式水文模型,虽在许多地区取得成效,但因参数化和水文响应不足,在干旱地区模拟中仍存在不确定性。为解决这一难...土壤与水评估工具(soil and water assessment tool,SWAT)作为一种广泛用于农业非点源(non-point source,NPS)污染模拟的分布式水文模型,虽在许多地区取得成效,但因参数化和水文响应不足,在干旱地区模拟中仍存在不确定性。为解决这一难题,该研究提出一种“物理约束+数据驱动”的混合建模策略:基于SWAT模型与卷积神经网络-长短期记忆网络(convolutional neural network-the long short term memory network,CNN-LSTM)的耦合方法,并引进改进型粒子群优化算法(improved particle swarm optimization,IPSO)应用于耦合框架中。该方法既同步优化了耦合模型的网络结构与超参数,又将每日输出结果的加权融合权重纳入同一优化向量。通过自适应惯性权重与扰动机制,实现对SWAT模型的误差校正。该研究通过分析单一SWAT模型的局限性,比较了SWAT模型与耦合模型在日尺度模拟精度上的差异,并探讨了IPSO与其他9种元启发式算法在超参数优化中的表现。最终以黄河宁夏段为研究区域,分析耦合模型在模拟总氮(total nitrogen,TN)和总磷(total phosphorus,TP)污染中的性能提升,并对流域NPS污染进行多尺度解析。结果表明,耦合模型在TN和TP模拟中显著优于单一的SWAT模型。TN的决定系数(determination coefficient,R^(2))、纳什效率系数(Nash-Sutcliffe efficiency,NSE)、百分比偏差(percent bias,PBIAS)和中心均方根误差(centered root mean square error,CRMSE)分别提高了14.1%、14.5%、38.6%和32.5%;TP的R^(2)、NSE、PBIAS和CRMSE分别提高了10.7%、12.0%、65.3%和40.7%。基于耦合模型的流域NPS污染时空分异分析显示,丰水期的峰值主要由降水和施肥协同作用导致,枯水期受宁夏冬灌影响。南部子流域的污染主要受降水径流驱动,北部灌区则由农业集约化主导。水系区间NPS污染贡献排名中,引黄灌区贡献31%~37%的TN和TP排放,红柳沟和苦水河水系受集约型农牧业影响,单位面积输出强度较高。研究表明,SWAT-IPSO-CNN-LSTM耦合方法有效降低了SWAT在干旱区站点率定的不确定性,并通过误差修正机制显著提升了氮磷模拟的精度与鲁棒性,为干旱区水环境管理提供了更可靠的技术支持。展开更多
岸边带正广泛应用于世界各地的面源污染治理项目,遥感也逐渐成为面源污染研究的重要手段,但如何将遥感技术与岸边带结合使截污效果更佳仍然是一个挑战。该文以云南省星云湖流域为例,耦合遥感建立土壤水分评估模型(soil and water assess...岸边带正广泛应用于世界各地的面源污染治理项目,遥感也逐渐成为面源污染研究的重要手段,但如何将遥感技术与岸边带结合使截污效果更佳仍然是一个挑战。该文以云南省星云湖流域为例,耦合遥感建立土壤水分评估模型(soil and water assessment tool,SWAT),通过改变土地利用类型的方式建立岸边带进行情景模拟,研究不同宽度和植被类型对污染物消减效果的差异。结果发现,设置岸边带对氮元素的截留效果好于磷元素;当岸边带植被类型不同时,林地的截污效果明显好于草地,并随着岸边带宽度的增加污染物消减率逐渐变大。设置30 m林地加30 m草地的岸边带可减少5.20%的总氮产量和6.03%的总磷产量,且可截留19.83%的有机氮入湖量和21.30%有机磷入湖量,在所有岸边带中截污效果最好。展开更多
The Inland Bays in southern Delaware (USA) are facing eutrophication due to the nutrient loading from its watershed. The source of nutrients in the watershed is predominantly agriculture. The Millsboro Pond, a sub-wat...The Inland Bays in southern Delaware (USA) are facing eutrophication due to the nutrient loading from its watershed. The source of nutrients in the watershed is predominantly agriculture. The Millsboro Pond, a sub-watershed within the Inland Bays basin, was modeled using the Soil and Water Assessment Tool (SWAT) model. It was found that the contribution of ground water from outside the watershed had a signifi-cant impact on the hydrology of the region. Once the model was calibrated and validated, five management scenarios were implemented, one at a time, to measure its effectiveness in reducing the nutrient loading in the watershed. Among the Best Management Practices (BMPs), planting winter cover crops on the agricul-ture land was the most effective method in reducing the nutrient loads. The second most effective method was to provide grassland riparian zones. The BMPs alone were not able to achieve the nutrient load reduc-tion as required by the Total Maximum Daily Loads (TMDLs). Two extra scenarios that involved in replac-ing agriculture land with forest, first with deciduous trees and then with high yielding trees were considered. It is suggested that to achieve the required TMDL for the watershed, some parts of the agricultural land may have to be effectively converted into the managed forest with some high yielding trees such as hybrid poplar trees providing cellulose raw material for bio fuels. The remaining agriculture land should take up the prac-tice of planting winter cover crops and better nutrient management. Riparian zones, either in form of forest or grasslands, should be the final line of defense for reducing nutrient loading in the watershed.展开更多
Modeling tools simulate the functioning of ecosystems and their interactions with human activities,helping decision makers understand how interventions impact ecosystems and evaluate management strategies.This leads t...Modeling tools simulate the functioning of ecosystems and their interactions with human activities,helping decision makers understand how interventions impact ecosystems and evaluate management strategies.This leads to informed decisions that balance human development and environmental protection.Among these models,Soil and Water Assessment Tool(SWAT)stands out for its ability to simulate multiple biophysical processes that can be linked to the provision of ecosystem services(ES).Although SWAT has been successfully applied for the evaluation of ES,the development of complementary approaches that translate the results of SWAT into monetary terms is still in its early stages.To narrow this gap,this review article aims to provide a comprehensive assessment of the literature on the relationship between SWAT model results and economic analysis.Specifically,the review summarizes the research conducted on the use of SWAT model results to estimate economic values,including the different methodologies used and the types of economic values estimated.The review will also discuss the limitations and challenges of these approaches,provide a critical evaluation of the strengths and weaknesses of the research in this area,and provide recommendations to strengthen SWAT application for the economic evaluation of management strategies.展开更多
基金supported by the project“Romanian Hub for Artificial Intelligence-HRIA”,Smart Growth,Digitization and Financial Instruments Program,2021–2027,MySMIS No.334906.
文摘Objective expertise evaluation of individuals,as a prerequisite stage for team formation,has been a long-term desideratum in large software development companies.With the rapid advancements in machine learning methods,based on reliable existing data stored in project management tools’datasets,automating this evaluation process becomes a natural step forward.In this context,our approach focuses on quantifying software developer expertise by using metadata from the task-tracking systems.For this,we mathematically formalize two categories of expertise:technology-specific expertise,which denotes the skills required for a particular technology,and general expertise,which encapsulates overall knowledge in the software industry.Afterward,we automatically classify the zones of expertise associated with each task a developer has worked on using Bidirectional Encoder Representations from Transformers(BERT)-like transformers to handle the unique characteristics of project tool datasets effectively.Finally,our method evaluates the proficiency of each software specialist across already completed projects from both technology-specific and general perspectives.The method was experimentally validated,yielding promising results.
基金National Science and Technology Council,the Republic of China,under grants NSTC 113-2221-E-194-011-MY3 and Research Center on Artificial Intelligence and Sustainability,National Chung Cheng University under the research project grant titled“Generative Digital Twin System Design for Sustainable Smart City Development in Taiwan.
文摘Modern manufacturing processes have become more reliant on automation because of the accelerated transition from Industry 3.0 to Industry 4.0.Manual inspection of products on assembly lines remains inefficient,prone to errors and lacks consistency,emphasizing the need for a reliable and automated inspection system.Leveraging both object detection and image segmentation approaches,this research proposes a vision-based solution for the detection of various kinds of tools in the toolkit using deep learning(DL)models.Two Intel RealSense D455f depth cameras were arranged in a top down configuration to capture both RGB and depth images of the toolkits.After applying multiple constraints and enhancing them through preprocessing and augmentation,a dataset consisting of 3300 annotated RGB-D photos was generated.Several DL models were selected through a comprehensive assessment of mean Average Precision(mAP),precision-recall equilibrium,inference latency(target≥30 FPS),and computational burden,resulting in a preference for YOLO and Region-based Convolutional Neural Networks(R-CNN)variants over ViT-based models due to the latter’s increased latency and resource requirements.YOLOV5,YOLOV8,YOLOV11,Faster R-CNN,and Mask R-CNN were trained on the annotated dataset and evaluated using key performance metrics(Recall,Accuracy,F1-score,and Precision).YOLOV11 demonstrated balanced excellence with 93.0%precision,89.9%recall,and a 90.6%F1-score in object detection,as well as 96.9%precision,95.3%recall,and a 96.5%F1-score in instance segmentation with an average inference time of 25 ms per frame(≈40 FPS),demonstrating real-time performance.Leveraging these results,a YOLOV11-based windows application was successfully deployed in a real-time assembly line environment,where it accurately processed live video streams to detect and segment tools within toolkits,demonstrating its practical effectiveness in industrial automation.The application is capable of precisely measuring socket dimensions by utilising edge detection techniques on YOLOv11 segmentation masks,in addition to detection and segmentation.This makes it possible to do specification-level quality control right on the assembly line,which improves the ability to examine things in real time.The implementation is a big step forward for intelligent manufacturing in the Industry 4.0 paradigm.It provides a scalable,efficient,and accurate way to do automated inspection and dimensional verification activities.
文摘The changes in land use in the last 30 years in the territory of agro-forest watershed of Lake Buyo resulted in significant sediment into the lake. Sediments are a preferred means of transportation for certain pollutants, like phosphorus in excess. By mapping the source areas of erosion, the authors can determine the risk areas and help to prioritize interventions in the territory. This mapping is done using the SWAT （soil and water assessment tool） model. Several types of data, including topography, land use, soil and climate data are needed to run the model. In this paper, all different steps are presented, from the designing of HRU （hydrological response units）, basic units to run the SWAT model until the simulations. The establishment of HRU has three main stages： space discretization, land use and soil data integration and HRU distribution： （1） space discretization which consist in extracting the limits and the water network of the watershed from the DEM （digital elevation model） and in subdividing them into sub-basins; （2） land use and soil data integration： it consists in digitizing the physical maps of land use and of soils under Mapinfo 7.5 and in keeping them in ＂shape＂ format; （3） HRU distribution： it leads to subdivide the sub-watersheds in small units that combine a single soil type and one type of land use. It appears from this study to obtain 23 sub-watersheds and 71 HRU. Once the HRU designed, it is necessary to integrate climate data, data on physico-chemical characteristics of soils and agricultural practices, before starting the simulations. This will allow the model to assess the risk of sedimentation and eutrophication of the lake using the MUSLE （modified universal soil loss equation） and phosphorus cycle.
文摘The Lobo watershed is an agricultural area where the use of fertilizers by farmers is intensive, causing eutrophication problems that deteriorate the quality of drinking water distributed to the population. Since the phenomenon of eutrophication is directly linked to runoff, it is essential to model the flow in order to better control the transfer of nutrients responsible for eutrophication. It is within this framework that this study was conducted. The objective of this study is to assess the ability of the semi-distributed SWAT (Soil and Water Assessment Tool) model to simulate runoff in the Lobo watershed. The methodology adopted was based on the use of the QSWAT graphical interface to manipulate and execute the main functions of the SWAT model from QGIS tools. The hydrological modeling was carried out with the QSWAT interface for SWAT 2012. The results showed good performance for the flow calibration (1982-1984) with the evaluation criteria R<sup>2</sup>, NSE and PBIAS respectively of 0.64, 0.64 and 3.1. In the validation period (1984-1987), the model also showed good performance in the streamflow simulation for R<sup>2</sup> and NSE of 0.84 and 0.76 respectively as values. However, for the PBAIS criterion, the result was less good but still remains satisfactory with a value of 19.6. It emerges from this study that the SWAT model is suitable for simulating water transfer and can therefore be used to study the transfer of pollutants in the fight against eutrophication in the Lobo watershed.
文摘土壤与水评估工具(soil and water assessment tool,SWAT)作为一种广泛用于农业非点源(non-point source,NPS)污染模拟的分布式水文模型,虽在许多地区取得成效,但因参数化和水文响应不足,在干旱地区模拟中仍存在不确定性。为解决这一难题,该研究提出一种“物理约束+数据驱动”的混合建模策略:基于SWAT模型与卷积神经网络-长短期记忆网络(convolutional neural network-the long short term memory network,CNN-LSTM)的耦合方法,并引进改进型粒子群优化算法(improved particle swarm optimization,IPSO)应用于耦合框架中。该方法既同步优化了耦合模型的网络结构与超参数,又将每日输出结果的加权融合权重纳入同一优化向量。通过自适应惯性权重与扰动机制,实现对SWAT模型的误差校正。该研究通过分析单一SWAT模型的局限性,比较了SWAT模型与耦合模型在日尺度模拟精度上的差异,并探讨了IPSO与其他9种元启发式算法在超参数优化中的表现。最终以黄河宁夏段为研究区域,分析耦合模型在模拟总氮(total nitrogen,TN)和总磷(total phosphorus,TP)污染中的性能提升,并对流域NPS污染进行多尺度解析。结果表明,耦合模型在TN和TP模拟中显著优于单一的SWAT模型。TN的决定系数(determination coefficient,R^(2))、纳什效率系数(Nash-Sutcliffe efficiency,NSE)、百分比偏差(percent bias,PBIAS)和中心均方根误差(centered root mean square error,CRMSE)分别提高了14.1%、14.5%、38.6%和32.5%;TP的R^(2)、NSE、PBIAS和CRMSE分别提高了10.7%、12.0%、65.3%和40.7%。基于耦合模型的流域NPS污染时空分异分析显示,丰水期的峰值主要由降水和施肥协同作用导致,枯水期受宁夏冬灌影响。南部子流域的污染主要受降水径流驱动,北部灌区则由农业集约化主导。水系区间NPS污染贡献排名中,引黄灌区贡献31%~37%的TN和TP排放,红柳沟和苦水河水系受集约型农牧业影响,单位面积输出强度较高。研究表明,SWAT-IPSO-CNN-LSTM耦合方法有效降低了SWAT在干旱区站点率定的不确定性,并通过误差修正机制显著提升了氮磷模拟的精度与鲁棒性,为干旱区水环境管理提供了更可靠的技术支持。
文摘岸边带正广泛应用于世界各地的面源污染治理项目,遥感也逐渐成为面源污染研究的重要手段,但如何将遥感技术与岸边带结合使截污效果更佳仍然是一个挑战。该文以云南省星云湖流域为例,耦合遥感建立土壤水分评估模型(soil and water assessment tool,SWAT),通过改变土地利用类型的方式建立岸边带进行情景模拟,研究不同宽度和植被类型对污染物消减效果的差异。结果发现,设置岸边带对氮元素的截留效果好于磷元素;当岸边带植被类型不同时,林地的截污效果明显好于草地,并随着岸边带宽度的增加污染物消减率逐渐变大。设置30 m林地加30 m草地的岸边带可减少5.20%的总氮产量和6.03%的总磷产量,且可截留19.83%的有机氮入湖量和21.30%有机磷入湖量,在所有岸边带中截污效果最好。
文摘The Inland Bays in southern Delaware (USA) are facing eutrophication due to the nutrient loading from its watershed. The source of nutrients in the watershed is predominantly agriculture. The Millsboro Pond, a sub-watershed within the Inland Bays basin, was modeled using the Soil and Water Assessment Tool (SWAT) model. It was found that the contribution of ground water from outside the watershed had a signifi-cant impact on the hydrology of the region. Once the model was calibrated and validated, five management scenarios were implemented, one at a time, to measure its effectiveness in reducing the nutrient loading in the watershed. Among the Best Management Practices (BMPs), planting winter cover crops on the agricul-ture land was the most effective method in reducing the nutrient loads. The second most effective method was to provide grassland riparian zones. The BMPs alone were not able to achieve the nutrient load reduc-tion as required by the Total Maximum Daily Loads (TMDLs). Two extra scenarios that involved in replac-ing agriculture land with forest, first with deciduous trees and then with high yielding trees were considered. It is suggested that to achieve the required TMDL for the watershed, some parts of the agricultural land may have to be effectively converted into the managed forest with some high yielding trees such as hybrid poplar trees providing cellulose raw material for bio fuels. The remaining agriculture land should take up the prac-tice of planting winter cover crops and better nutrient management. Riparian zones, either in form of forest or grasslands, should be the final line of defense for reducing nutrient loading in the watershed.
基金MERLIN project(Mainstreaming Ecological Restoration of freshwater-related ecosystems in a Landscape context:INnovation,upscaling and transformation),funded by the European Union's Horizon 2020 research and innovation programme under grant agreement ID 101036337.
文摘Modeling tools simulate the functioning of ecosystems and their interactions with human activities,helping decision makers understand how interventions impact ecosystems and evaluate management strategies.This leads to informed decisions that balance human development and environmental protection.Among these models,Soil and Water Assessment Tool(SWAT)stands out for its ability to simulate multiple biophysical processes that can be linked to the provision of ecosystem services(ES).Although SWAT has been successfully applied for the evaluation of ES,the development of complementary approaches that translate the results of SWAT into monetary terms is still in its early stages.To narrow this gap,this review article aims to provide a comprehensive assessment of the literature on the relationship between SWAT model results and economic analysis.Specifically,the review summarizes the research conducted on the use of SWAT model results to estimate economic values,including the different methodologies used and the types of economic values estimated.The review will also discuss the limitations and challenges of these approaches,provide a critical evaluation of the strengths and weaknesses of the research in this area,and provide recommendations to strengthen SWAT application for the economic evaluation of management strategies.
