Red chilli powder(RCP)is a versatile spice accepted globally in diverse culinary products due to its distinct pungent characteristics and red colour.The higher market demand makes the spice vulnerable to unethical mix...Red chilli powder(RCP)is a versatile spice accepted globally in diverse culinary products due to its distinct pungent characteristics and red colour.The higher market demand makes the spice vulnerable to unethical mixing,so its quality assessment is crucial.The non-destructive application of computer vision for measuring food adulteration has always attracted researchers and industry due to its robustness and feasibility.Following the current era of Food Quality 4.0 and artificial intelligence,this study follows an approach based on 1D-convolutional neural networks(CNN)and 2D-CNN models for detecting RCP adulteration.The performance evaluation metrics are used to analyse the efficiency of these models.The histogram features from the Lab colour space trained on the 1D-CNN model(BS-40 and Epoch 100)show an accuracy of 84.56%.On the other hand,the 2D-CNN model DenseNet-121(AdamW and BS-30)also shows a test accuracy of 84.62%.From the observations of this study,it is concluded that CNN models can be a promising tool for solving the adulteration detection problem in food quality evaluation.Further,internet of things-based systems can be developed to aid the industry and government agencies in monitoring the quality of RCP to harness the unethical practices of food adulteration.展开更多
在CMA-GFS(CMA Global Forecast System)全球四维变分资料同化系统(4DVar)基础上,参照BDA(Bogus Data Assimilation)方法,建立了一个全球模式台风初始化方案。该方案通过4DVar同化窗口吸收诊断处理后的1 h间隔台风中心定位及中心气压信...在CMA-GFS(CMA Global Forecast System)全球四维变分资料同化系统(4DVar)基础上,参照BDA(Bogus Data Assimilation)方法,建立了一个全球模式台风初始化方案。该方案通过4DVar同化窗口吸收诊断处理后的1 h间隔台风中心定位及中心气压信息,利用模式动力物理约束产生台风环流。同时,考虑到模式对台风的分辨能力,中心气压数据误差采用动态调整技术。2016年西北太平洋22个台风的试验表明,新方案不仅可以促进初始场中台风环流的生成,还可以显著减小CMAGFS全球预报系统的台风路径和强度预报平均误差,具有业务应用前景。展开更多
文摘Red chilli powder(RCP)is a versatile spice accepted globally in diverse culinary products due to its distinct pungent characteristics and red colour.The higher market demand makes the spice vulnerable to unethical mixing,so its quality assessment is crucial.The non-destructive application of computer vision for measuring food adulteration has always attracted researchers and industry due to its robustness and feasibility.Following the current era of Food Quality 4.0 and artificial intelligence,this study follows an approach based on 1D-convolutional neural networks(CNN)and 2D-CNN models for detecting RCP adulteration.The performance evaluation metrics are used to analyse the efficiency of these models.The histogram features from the Lab colour space trained on the 1D-CNN model(BS-40 and Epoch 100)show an accuracy of 84.56%.On the other hand,the 2D-CNN model DenseNet-121(AdamW and BS-30)also shows a test accuracy of 84.62%.From the observations of this study,it is concluded that CNN models can be a promising tool for solving the adulteration detection problem in food quality evaluation.Further,internet of things-based systems can be developed to aid the industry and government agencies in monitoring the quality of RCP to harness the unethical practices of food adulteration.
文摘在CMA-GFS(CMA Global Forecast System)全球四维变分资料同化系统(4DVar)基础上,参照BDA(Bogus Data Assimilation)方法,建立了一个全球模式台风初始化方案。该方案通过4DVar同化窗口吸收诊断处理后的1 h间隔台风中心定位及中心气压信息,利用模式动力物理约束产生台风环流。同时,考虑到模式对台风的分辨能力,中心气压数据误差采用动态调整技术。2016年西北太平洋22个台风的试验表明,新方案不仅可以促进初始场中台风环流的生成,还可以显著减小CMAGFS全球预报系统的台风路径和强度预报平均误差,具有业务应用前景。
