In the deep learning approach for identifying plant diseases,the high complexity of the network model,the large number of parameters,and great computational effort make it challenging to deploy the model on terminal d...In the deep learning approach for identifying plant diseases,the high complexity of the network model,the large number of parameters,and great computational effort make it challenging to deploy the model on terminal devices with limited computational resources.In this study,a lightweight method for plant diseases identification that is an improved version of the ShuffleNetV2 model is proposed.In the proposed model,the depthwise convolution in the basic module of ShuffleNetV2 is replaced with mixed depthwise convolution to capture crop pest images with different resolutions;the efficient channel attention module is added into the ShuffleNetV2 model network structure to enhance the channel features;and the ReLU activation function is replaced with the ReLU6 activation function to prevent the gen-eration of large gradients.Experiments are conducted on the public dataset PlantVillage.The results show that the proposed model achieves an accuracy of 99.43%,which is an improvement of 0.6 percentage points compared to the ShuffleNetV2 model.Compared to lightweight network models,such as MobileNetV2,MobileNetV3,EfficientNet,and EfficientNetV2,and classical convolutional neural network models,such as ResNet34,ResNet50,and ResNet101,the proposed model has fewer parameters and higher recognition accuracy,which provides guidance for deploying crop pest identification methods on resource-constrained devices,including mobile terminals.展开更多
Deep mixed oils with secondary alterations have been widely discovered in the Tarim Basin,but current methods based on biomarkers and isotopes to de-convolute mixed oil cannot calculate the exact mixing proportion of ...Deep mixed oils with secondary alterations have been widely discovered in the Tarim Basin,but current methods based on biomarkers and isotopes to de-convolute mixed oil cannot calculate the exact mixing proportion of different end-member oils,which has seriously hindered further exploration of deep hydrocarbons in the study area.To solve this problem,we constructed a novel method based on the carbon isotope(δ13C)of the group components to de-convolute mixed liquid hydrocarbons under the material balance principle.The results showed that the mixed oil in the Tazhong Uplift was dominantly contributed at an average proportion of 68% by an oil end-member with heavier d13C that was believed to be generated from the Cambrian-Lower Ordovician source rocks,whereas the mixed oil in the Tabei Uplift was predominantly contributed at an average proportion of 61% by an oil end-member with lighter d13C that was believed to be generated from the Middle-Upper Ordovician source rocks.This indicates that,on the basis of the detailed description of the distribution of effective source rocks,the proposed method will be helpful in realizing differential exploration and further improving the efficiency of deep liquid hydrocarbon exploration in the Tarim Basin.In addition,compared to traditional δ13C methods for whole oil and individual n-alkanes in de-convoluted mixed oil,the proposed method has a wider range of applications,including for mixed oils with variations in color and density,indicating potential for promoting the exploration of deep complex mixed oils in the Tarim Basin and even around the world.展开更多
基金supported by the Guangxi Key R&D Project(Gui Ke AB21076021)the Project of Humanities and social sciences of“cultivation plan for thousands of young and middle-aged backbone teachers in Guangxi Colleges and universities”in 2021:Research on Collaborative integration of logistics service supply chain under high-quality development goals(2021QGRW044).
文摘In the deep learning approach for identifying plant diseases,the high complexity of the network model,the large number of parameters,and great computational effort make it challenging to deploy the model on terminal devices with limited computational resources.In this study,a lightweight method for plant diseases identification that is an improved version of the ShuffleNetV2 model is proposed.In the proposed model,the depthwise convolution in the basic module of ShuffleNetV2 is replaced with mixed depthwise convolution to capture crop pest images with different resolutions;the efficient channel attention module is added into the ShuffleNetV2 model network structure to enhance the channel features;and the ReLU activation function is replaced with the ReLU6 activation function to prevent the gen-eration of large gradients.Experiments are conducted on the public dataset PlantVillage.The results show that the proposed model achieves an accuracy of 99.43%,which is an improvement of 0.6 percentage points compared to the ShuffleNetV2 model.Compared to lightweight network models,such as MobileNetV2,MobileNetV3,EfficientNet,and EfficientNetV2,and classical convolutional neural network models,such as ResNet34,ResNet50,and ResNet101,the proposed model has fewer parameters and higher recognition accuracy,which provides guidance for deploying crop pest identification methods on resource-constrained devices,including mobile terminals.
基金The authors are grateful for the financial supports provided by the National Science and Technology Major Project of the Ministry of Science and Technology of China(2016ZX04004-004)National Natural Science Foundation of China(41672125)。
文摘Deep mixed oils with secondary alterations have been widely discovered in the Tarim Basin,but current methods based on biomarkers and isotopes to de-convolute mixed oil cannot calculate the exact mixing proportion of different end-member oils,which has seriously hindered further exploration of deep hydrocarbons in the study area.To solve this problem,we constructed a novel method based on the carbon isotope(δ13C)of the group components to de-convolute mixed liquid hydrocarbons under the material balance principle.The results showed that the mixed oil in the Tazhong Uplift was dominantly contributed at an average proportion of 68% by an oil end-member with heavier d13C that was believed to be generated from the Cambrian-Lower Ordovician source rocks,whereas the mixed oil in the Tabei Uplift was predominantly contributed at an average proportion of 61% by an oil end-member with lighter d13C that was believed to be generated from the Middle-Upper Ordovician source rocks.This indicates that,on the basis of the detailed description of the distribution of effective source rocks,the proposed method will be helpful in realizing differential exploration and further improving the efficiency of deep liquid hydrocarbon exploration in the Tarim Basin.In addition,compared to traditional δ13C methods for whole oil and individual n-alkanes in de-convoluted mixed oil,the proposed method has a wider range of applications,including for mixed oils with variations in color and density,indicating potential for promoting the exploration of deep complex mixed oils in the Tarim Basin and even around the world.
