Solar cell defect detection is crucial for quality inspection in photovoltaic power generation modules.In the production process,defect samples occur infrequently and exhibit random shapes and sizes,which makes it cha...Solar cell defect detection is crucial for quality inspection in photovoltaic power generation modules.In the production process,defect samples occur infrequently and exhibit random shapes and sizes,which makes it challenging to collect defective samples.Additionally,the complex surface background of polysilicon cell wafers complicates the accurate identification and localization of defective regions.This paper proposes a novel Lightweight Multiscale Feature Fusion network(LMFF)to address these challenges.The network comprises a feature extraction network,a multi-scale feature fusion module(MFF),and a segmentation network.Specifically,a feature extraction network is proposed to obtain multi-scale feature outputs,and a multi-scale feature fusion module(MFF)is used to fuse multi-scale feature information effectively.In order to capture finer-grained multi-scale information from the fusion features,we propose a multi-scale attention module(MSA)in the segmentation network to enhance the network’s ability for small target detection.Moreover,depthwise separable convolutions are introduced to construct depthwise separable residual blocks(DSR)to reduce the model’s parameter number.Finally,to validate the proposed method’s defect segmentation and localization performance,we constructed three solar cell defect detection datasets:SolarCells,SolarCells-S,and PVEL-S.SolarCells and SolarCells-S are monocrystalline silicon datasets,and PVEL-S is a polycrystalline silicon dataset.Experimental results show that the IOU of our method on these three datasets can reach 68.5%,51.0%,and 92.7%,respectively,and the F1-Score can reach 81.3%,67.5%,and 96.2%,respectively,which surpasses other commonly usedmethods and verifies the effectiveness of our LMFF network.展开更多
Paleobiogeography investigates geographical distributions of fossil organisms and controlling factors that affect their distributions in geological history,to reveal the macro-evolution and coordinated development of ...Paleobiogeography investigates geographical distributions of fossil organisms and controlling factors that affect their distributions in geological history,to reveal the macro-evolution and coordinated development of life and the environment.It is a crucial window for understanding the biosphere and the geographical environment.After two centuries of development,paleobiogeographic studies have led to the accumulation of significant amounts of knowledge and data;however,the voluminous outputs present the characteristics of an“isolated island”with a scattered,limited number of authoritative definitions of terminologies and semantic heterogeneity among them.This makes data queries cumbersome,the rate of data reuse low,and data sharing more difficult.A knowledge graph(KG)has the advantage of expressing concepts and their semantic relations,which is an important tool for achieving data organization and fusion,and data mining;further,it is also a key technology for realizing the unrestricted sharing of paleobiogeographic information.Through our efforts over the past two years,a paleobiogeographic KG was developed based on the established construction procedure of the KG,which contains 273 concepts,172 properties,and 47 rules.Meanwhile,the completion of this KG and the construction of a paleobiogeographic platform for display and analysis are now being carried out.展开更多
Unsupervised methods based on density representation have shown their abilities in anomaly detection,but detection performance still needs to be improved.Specifically,approaches using normalizing flows can accurately ...Unsupervised methods based on density representation have shown their abilities in anomaly detection,but detection performance still needs to be improved.Specifically,approaches using normalizing flows can accurately evaluate sample distributions,mapping normal features to the normal distribution and anomalous features outside it.Consequently,this paper proposes a Normalizing Flow-based Bidirectional Mapping Residual Network(NF-BMR).It utilizes pre-trained Convolutional Neural Networks(CNN)and normalizing flows to construct discriminative source and target domain feature spaces.Additionally,to better learn feature information in both domain spaces,we propose the Bidirectional Mapping Residual Network(BMR),which maps sample features to these two spaces for anomaly detection.The two detection spaces effectively complement each other’s deficiencies and provide a comprehensive feature evaluation from two perspectives,which leads to the improvement of detection performance.Comparative experimental results on the MVTec AD and DAGM datasets against the Bidirectional Pre-trained Feature Mapping Network(B-PFM)and other state-of-the-art methods demonstrate that the proposed approach achieves superior performance.On the MVTec AD dataset,NF-BMR achieves an average AUROC of 98.7%for all 15 categories.Especially,it achieves 100%optimal detection performance in five categories.On the DAGM dataset,the average AUROC across ten categories is 98.7%,which is very close to supervised methods.展开更多
Many observational studies have shown that deformation, like vertical vorticity and divergence, is closely related to the occurrence and distribution of strong precipitation. In this paper, to involve deformation in p...Many observational studies have shown that deformation, like vertical vorticity and divergence, is closely related to the occurrence and distribution of strong precipitation. In this paper, to involve deformation in precipitation diagnosis, a new parameter called potential deformation（PD） is derived and then applied to precipitation detection within a simulated mesoscale convective system（MCS）. It is shown that PD includes both stretching deformation and shearing deformation and shares similar characteristics with deformation insofar as it does not change with the rotating coordinate. Diagnosis of the simulated MCS reveals that PD performs well in tracing the MCS＇ precipitation. In terms of their distributional pattern, the large-value areas of PD are similar to the precipitation in the different development stages of the MCS. A detailed analysis of the physical processes contained within the PD shows that it can reflect the three-dimensional moisture variation,vertical wind shear and wind deformation within the MCS. These structures are usually a comprehensive reflection of the characteristics of the surface cold pool, rear inflow jet, downward cold air flow and upward warm moist flow within the precipitating convective cells. For this reason, the PD shows much stronger anomalies in the precipitating atmosphere than the non-precipitating atmosphere, which implies considerable potential for its application in detecting heavy precipitation within MCSs.展开更多
A knowledge graph(KG)is a knowledge base that integrates and represents data based on a graph-structured data model or topology.Geoscientists have made efforts to construct geosciencerelated KGs to overcome semantic h...A knowledge graph(KG)is a knowledge base that integrates and represents data based on a graph-structured data model or topology.Geoscientists have made efforts to construct geosciencerelated KGs to overcome semantic heterogeneity and facilitate knowledge representation,data integration,and text analysis.However,there is currently no comprehensive paleontology KG or data-driven discovery based on it.In this study,we constructed a two-layer model to represent the ordinal hierarchical structure of the paleontology KG following a top-down construction process.An ontology containing 19365 concepts has been defined up to 2023.On this basis,we derived the synonymy list based on the paleontology KG and designed corresponding online functions in the OneStratigraphy database to showcase the use of the KG in paleontological research.展开更多
Stratigraphic hiatuses of variable time intervals within the Rhuddanian to early Aeronian （Llandovery, Silurian） are identified in the area bordering East Chongqing, West Hubei and Northwest Hunan in central China. ...Stratigraphic hiatuses of variable time intervals within the Rhuddanian to early Aeronian （Llandovery, Silurian） are identified in the area bordering East Chongqing, West Hubei and Northwest Hunan in central China. Their distribution suggested the existence of a local uplift, traditionally named the Yichang Uplift. The diachronous nature of the basal black shale of the Lungmachi Formation crossing different belts of this Uplift signifies the various developing stages during the uplifting process. The present paper defines the temporal and spatial distribution pattern of the Yichang Uplift, which might be one of the important controlling factors for the preservation and distribution of the shale gas in this region, as it has been demonstrated that the shale gas exploration is generally less promising in the areas where more of the basal part of the Lungmachi Formation is missing. Therefore, better understanding of the circumjacent distribution pattern developed throughout the uplifting process may provide the important guidance for the shale gas exploration. The present work is a sister study to the published paper, ＂Stage-progressive distribution pattern of the Lungrnachian black graplolitic shales from Guizhou to Chongqing, Central China＂. These two studies thus provide a complete Ordovician-Silurian black shale distribution pattern in the Middle and Upper Yangtze, a region with the major shale gas fields in China.展开更多
Wide distribution of the black shales and diversification of the graptolite fauna in South China during the Late Ordovician resulted from its unique paleogeographic pattern, which was significantly affected by the pal...Wide distribution of the black shales and diversification of the graptolite fauna in South China during the Late Ordovician resulted from its unique paleogeographic pattern, which was significantly affected by the paleogeographic evolution of the Lower Yangtze region. In the study, 120 Upper Ordovician sections from the Lower Yangtze region were collected, and a unified biostratigraphic framework has been applied to these sections to establish a reliable stratigraphic subdivision and correlation. Under the unified time framework, we delineate the distribution area of each lithostratigraphic unit, outline the boundary between the sea and land, and reconstruct the paleogeographic pattern for each graptolite zone. The result indicates that, with the uplift and expansion of the ‘Jiangnan Oldland' in the beginning of the late Katian, the oldland extended into the Yangtze Sea gradually from south to north, which finally separate the Jiangnan Slope and the Yangtze Platform. Consequently,the longstanding paleogeographic pattern of "platform-slope-basin" in South China was broken. The paleogeographic change led to sedimentary differentiation among the two sides of the ‘Jiangnan Oldland' during the Late Ordovician. This event also led to the closure of the eastern exit of the Upper Yangtze Sea, and formed a semi-closed, limited and stagnant environment for the development of the organic-rich black shales during the Late Ordovician. The major controlling factors of these paleogeographic changes in the Lower Yangtze region were not consistent from the Katian to the Hirnantian. In the late Katian, the sedimentary differentiation between the east and west sides mostly resulted from regional tectonic movement-the Kwangsian Orogeny.However, during the Hirnantian, the whole Yangtze region became shallower, which was mostly influenced by the concentration of the Gondwana ice sheet and the consequent global sea level drop.展开更多
Heavy regional precipitation(HRP)over Beijing,Tianjin,and Hebei Province(the Jing–Jin–Ji region or JJJ)in early October(1–10 October)is a high-impact climate event because of travel and outdoor activities by except...Heavy regional precipitation(HRP)over Beijing,Tianjin,and Hebei Province(the Jing–Jin–Ji region or JJJ)in early October(1–10 October)is a high-impact climate event because of travel and outdoor activities by exceptionally large population during the Chinese National Day Holidays(CNDH).What causes the year-to-year variation of the HRP during early October is investigated through an observational analysis.It is found that the HRP arises from moisture transport by southerly anomalies to the west of an anomalous low-level anticyclone over the subtropical northwestern Pacific(SNWP).Sensitivity numerical experiments reveal that the low-level anticyclonic anomaly is caused by a dipole heating pattern over tropical western and central Pacific associated with a La Niña-like SST anomaly(SSTA)pattern in the Pacific and by a negative heating anomaly over North Europe.The latter connects the SNWP anticyclone through a Rossby wave train.Anomalous ascent associated with a positive heating anomaly over the tropical western Pacific may strengthen the local Hadley Cell,contributing to maintenance of the low-level anomalous anticyclone over SNWP and extending westwards of the western Pacific subtropical high(WPSH).Therefore,both the tropical Pacific and midlatitude heating signals over North Europe may be potential predictors for HRP forecast in the JJJ region in early October.展开更多
Dear Editor,Advances in high-throughput omics technologies,along with methodologies for integrating multi-omics datasets,have substantially enhanced the efficiency of identifying candidate genes in breeding(Gusev et a...Dear Editor,Advances in high-throughput omics technologies,along with methodologies for integrating multi-omics datasets,have substantially enhanced the efficiency of identifying candidate genes in breeding(Gusev et al.,2018;Gupta et al.,2019).However,this process is often complex and laborious.To address this challenge,databases that integrate extensive data and enable convenient and efficient functional genomics studies are being developed(Ma et al.,2021;Yang et al.,2023).展开更多
基金supported in part by the National Natural Science Foundation of China under Grants 62463002,62062021 and 62473033in part by the Guiyang Scientific Plan Project[2023]48–11,in part by QKHZYD[2023]010 Guizhou Province Science and Technology Innovation Base Construction Project“Key Laboratory Construction of Intelligent Mountain Agricultural Equipment”.
文摘Solar cell defect detection is crucial for quality inspection in photovoltaic power generation modules.In the production process,defect samples occur infrequently and exhibit random shapes and sizes,which makes it challenging to collect defective samples.Additionally,the complex surface background of polysilicon cell wafers complicates the accurate identification and localization of defective regions.This paper proposes a novel Lightweight Multiscale Feature Fusion network(LMFF)to address these challenges.The network comprises a feature extraction network,a multi-scale feature fusion module(MFF),and a segmentation network.Specifically,a feature extraction network is proposed to obtain multi-scale feature outputs,and a multi-scale feature fusion module(MFF)is used to fuse multi-scale feature information effectively.In order to capture finer-grained multi-scale information from the fusion features,we propose a multi-scale attention module(MSA)in the segmentation network to enhance the network’s ability for small target detection.Moreover,depthwise separable convolutions are introduced to construct depthwise separable residual blocks(DSR)to reduce the model’s parameter number.Finally,to validate the proposed method’s defect segmentation and localization performance,we constructed three solar cell defect detection datasets:SolarCells,SolarCells-S,and PVEL-S.SolarCells and SolarCells-S are monocrystalline silicon datasets,and PVEL-S is a polycrystalline silicon dataset.Experimental results show that the IOU of our method on these three datasets can reach 68.5%,51.0%,and 92.7%,respectively,and the F1-Score can reach 81.3%,67.5%,and 96.2%,respectively,which surpasses other commonly usedmethods and verifies the effectiveness of our LMFF network.
基金supported by the National Natural Science Foundation of China(Nos.42172174,41802017,42250104)the National Key R&D Program of China(No.2018YFE0204201)the Fundamental Research Funds for the Central Universities(No.0206-14380168)。
文摘Paleobiogeography investigates geographical distributions of fossil organisms and controlling factors that affect their distributions in geological history,to reveal the macro-evolution and coordinated development of life and the environment.It is a crucial window for understanding the biosphere and the geographical environment.After two centuries of development,paleobiogeographic studies have led to the accumulation of significant amounts of knowledge and data;however,the voluminous outputs present the characteristics of an“isolated island”with a scattered,limited number of authoritative definitions of terminologies and semantic heterogeneity among them.This makes data queries cumbersome,the rate of data reuse low,and data sharing more difficult.A knowledge graph(KG)has the advantage of expressing concepts and their semantic relations,which is an important tool for achieving data organization and fusion,and data mining;further,it is also a key technology for realizing the unrestricted sharing of paleobiogeographic information.Through our efforts over the past two years,a paleobiogeographic KG was developed based on the established construction procedure of the KG,which contains 273 concepts,172 properties,and 47 rules.Meanwhile,the completion of this KG and the construction of a paleobiogeographic platform for display and analysis are now being carried out.
基金This work was supported in part by the National Key R&D Program of China 2021YFE0110500in part by the National Natural Science Foundation of China under Grant 62062021in part by the Guiyang Scientific Plan Project[2023]48-11.
文摘Unsupervised methods based on density representation have shown their abilities in anomaly detection,but detection performance still needs to be improved.Specifically,approaches using normalizing flows can accurately evaluate sample distributions,mapping normal features to the normal distribution and anomalous features outside it.Consequently,this paper proposes a Normalizing Flow-based Bidirectional Mapping Residual Network(NF-BMR).It utilizes pre-trained Convolutional Neural Networks(CNN)and normalizing flows to construct discriminative source and target domain feature spaces.Additionally,to better learn feature information in both domain spaces,we propose the Bidirectional Mapping Residual Network(BMR),which maps sample features to these two spaces for anomaly detection.The two detection spaces effectively complement each other’s deficiencies and provide a comprehensive feature evaluation from two perspectives,which leads to the improvement of detection performance.Comparative experimental results on the MVTec AD and DAGM datasets against the Bidirectional Pre-trained Feature Mapping Network(B-PFM)and other state-of-the-art methods demonstrate that the proposed approach achieves superior performance.On the MVTec AD dataset,NF-BMR achieves an average AUROC of 98.7%for all 15 categories.Especially,it achieves 100%optimal detection performance in five categories.On the DAGM dataset,the average AUROC across ten categories is 98.7%,which is very close to supervised methods.
基金supported by the Special Scientific Research Fund of the Meteorological Public Welfare of the Ministry of Sciences and Technology (Grant No. GYHY201406002, GYHY201406001)National Key Technology Support Program (Grant No. 2015BAC03B04)+4 种基金a National Program on Key Basic Research project (Grant No. 2013CB430105)the Major Research Plan of the National Natural Science Foundation of China (Grant No. 91437215)the National Natural Science Foundation of China (Grant Nos. 41505040, 41375052 41405055 and 41575065)the Open Project of the State Key Laboratory of Severe Weather (La SW), the Chinese Academy of Meteorological Sciences (CAMS) (Grant No. 2015LASW-B05)the Beijing Natural Sciences Foundation (Grant No. 8142035)
文摘Many observational studies have shown that deformation, like vertical vorticity and divergence, is closely related to the occurrence and distribution of strong precipitation. In this paper, to involve deformation in precipitation diagnosis, a new parameter called potential deformation（PD） is derived and then applied to precipitation detection within a simulated mesoscale convective system（MCS）. It is shown that PD includes both stretching deformation and shearing deformation and shares similar characteristics with deformation insofar as it does not change with the rotating coordinate. Diagnosis of the simulated MCS reveals that PD performs well in tracing the MCS＇ precipitation. In terms of their distributional pattern, the large-value areas of PD are similar to the precipitation in the different development stages of the MCS. A detailed analysis of the physical processes contained within the PD shows that it can reflect the three-dimensional moisture variation,vertical wind shear and wind deformation within the MCS. These structures are usually a comprehensive reflection of the characteristics of the surface cold pool, rear inflow jet, downward cold air flow and upward warm moist flow within the precipitating convective cells. For this reason, the PD shows much stronger anomalies in the precipitating atmosphere than the non-precipitating atmosphere, which implies considerable potential for its application in detecting heavy precipitation within MCSs.
基金supported by the National Natural Science Foundation of China(Nos.41725007,42250104,41830323,42002015,and 42302001)the Fundamental Research Funds for the Central Universities(Nos.020614380168,JZ2023HGQA0144 and JZ2023HGTA0175)。
文摘A knowledge graph(KG)is a knowledge base that integrates and represents data based on a graph-structured data model or topology.Geoscientists have made efforts to construct geosciencerelated KGs to overcome semantic heterogeneity and facilitate knowledge representation,data integration,and text analysis.However,there is currently no comprehensive paleontology KG or data-driven discovery based on it.In this study,we constructed a two-layer model to represent the ordinal hierarchical structure of the paleontology KG following a top-down construction process.An ontology containing 19365 concepts has been defined up to 2023.On this basis,we derived the synonymy list based on the paleontology KG and designed corresponding online functions in the OneStratigraphy database to showcase the use of the KG in paleontological research.
基金supported by the Strategic Priority Research Program of Chinese Academy of Sciences (Grant No. XDB26000000)the National Natural Science Foundation of China (Grant Nos. U1562213 and 41502025)the National Science and Technology Major Project of China (Grant No. 2017ZX05035002-001)
文摘Stratigraphic hiatuses of variable time intervals within the Rhuddanian to early Aeronian （Llandovery, Silurian） are identified in the area bordering East Chongqing, West Hubei and Northwest Hunan in central China. Their distribution suggested the existence of a local uplift, traditionally named the Yichang Uplift. The diachronous nature of the basal black shale of the Lungmachi Formation crossing different belts of this Uplift signifies the various developing stages during the uplifting process. The present paper defines the temporal and spatial distribution pattern of the Yichang Uplift, which might be one of the important controlling factors for the preservation and distribution of the shale gas in this region, as it has been demonstrated that the shale gas exploration is generally less promising in the areas where more of the basal part of the Lungmachi Formation is missing. Therefore, better understanding of the circumjacent distribution pattern developed throughout the uplifting process may provide the important guidance for the shale gas exploration. The present work is a sister study to the published paper, ＂Stage-progressive distribution pattern of the Lungrnachian black graplolitic shales from Guizhou to Chongqing, Central China＂. These two studies thus provide a complete Ordovician-Silurian black shale distribution pattern in the Middle and Upper Yangtze, a region with the major shale gas fields in China.
基金supported by National Natural Science Foundation of China (Grant Nos. 41502025, U1562213 and 41521061)Chinese Academy of Sciences (Grant No. XDB10010100)+1 种基金the China Geological Survey Project (Grant No. 2016-03019)the "Geobiodiversity Database" and IGCP 653 Project "The onset of the Great Ordovician Biodiversity Event"
文摘Wide distribution of the black shales and diversification of the graptolite fauna in South China during the Late Ordovician resulted from its unique paleogeographic pattern, which was significantly affected by the paleogeographic evolution of the Lower Yangtze region. In the study, 120 Upper Ordovician sections from the Lower Yangtze region were collected, and a unified biostratigraphic framework has been applied to these sections to establish a reliable stratigraphic subdivision and correlation. Under the unified time framework, we delineate the distribution area of each lithostratigraphic unit, outline the boundary between the sea and land, and reconstruct the paleogeographic pattern for each graptolite zone. The result indicates that, with the uplift and expansion of the ‘Jiangnan Oldland' in the beginning of the late Katian, the oldland extended into the Yangtze Sea gradually from south to north, which finally separate the Jiangnan Slope and the Yangtze Platform. Consequently,the longstanding paleogeographic pattern of "platform-slope-basin" in South China was broken. The paleogeographic change led to sedimentary differentiation among the two sides of the ‘Jiangnan Oldland' during the Late Ordovician. This event also led to the closure of the eastern exit of the Upper Yangtze Sea, and formed a semi-closed, limited and stagnant environment for the development of the organic-rich black shales during the Late Ordovician. The major controlling factors of these paleogeographic changes in the Lower Yangtze region were not consistent from the Katian to the Hirnantian. In the late Katian, the sedimentary differentiation between the east and west sides mostly resulted from regional tectonic movement-the Kwangsian Orogeny.However, during the Hirnantian, the whole Yangtze region became shallower, which was mostly influenced by the concentration of the Gondwana ice sheet and the consequent global sea level drop.
基金Supported by the National Natural Science Foundation of China(42088101 and 41875074)China Meteorological Administration Innovation and Development Project(CXFZ2021J030 and CXFZ2021J046)+1 种基金Beijing Meterological Service Science and Technology Project(BMBKJ 201901031)Climate Change Special Fund of China Meteorological Administration(202009).
文摘Heavy regional precipitation(HRP)over Beijing,Tianjin,and Hebei Province(the Jing–Jin–Ji region or JJJ)in early October(1–10 October)is a high-impact climate event because of travel and outdoor activities by exceptionally large population during the Chinese National Day Holidays(CNDH).What causes the year-to-year variation of the HRP during early October is investigated through an observational analysis.It is found that the HRP arises from moisture transport by southerly anomalies to the west of an anomalous low-level anticyclone over the subtropical northwestern Pacific(SNWP).Sensitivity numerical experiments reveal that the low-level anticyclonic anomaly is caused by a dipole heating pattern over tropical western and central Pacific associated with a La Niña-like SST anomaly(SSTA)pattern in the Pacific and by a negative heating anomaly over North Europe.The latter connects the SNWP anticyclone through a Rossby wave train.Anomalous ascent associated with a positive heating anomaly over the tropical western Pacific may strengthen the local Hadley Cell,contributing to maintenance of the low-level anomalous anticyclone over SNWP and extending westwards of the western Pacific subtropical high(WPSH).Therefore,both the tropical Pacific and midlatitude heating signals over North Europe may be potential predictors for HRP forecast in the JJJ region in early October.
基金supported by the National Natural Science Foundation of China(32072573,31872096,32322061,and 32070559)the National Key Research and Development Plan of China(2021YFF1000100,2023YFD1200102-03)+1 种基金the Fundamental Research Funds for the Central University HZAU(2662023XXPY001)the Developing Bioinformatics Platform in Hainan Yazhou Bay Seed Lab(no.JBGS-B21HJ0001)。
文摘Dear Editor,Advances in high-throughput omics technologies,along with methodologies for integrating multi-omics datasets,have substantially enhanced the efficiency of identifying candidate genes in breeding(Gusev et al.,2018;Gupta et al.,2019).However,this process is often complex and laborious.To address this challenge,databases that integrate extensive data and enable convenient and efficient functional genomics studies are being developed(Ma et al.,2021;Yang et al.,2023).
