TY  - JOUR
TI  - Wheat kernels classification using visible-near infrared camera based ondeep learning
AB  - This paper presents a smart machine learning system for classificationof hyperspectral wheat data based on deep learning methodology. Forthis purpose, the performances of AlexNet and VGG16 models wereinvestigated for the classification of hyperspectral wheat samples. Inthis study, the Support Vector Machine (SVM) and Softmax classifierswere carried out to predict labels of wheat kernels. In order to evaluatethe system performance, a new hyperspectral wheat test dataset wasconstructed using Visible-Near Infrared images (VNIR) including 50wheat species with 220 images per specimen, as 11000 samples in total.With experiments applied on newly created test dataset, overallapproximated accuracy rates of 96.00% and 99.00% determined bylinear SVM classifier, in case of fully connected layer (FC6 and FC7)features for AlexNet and VGG16, respectively. From the Softmaxpredictions, the 92% and 70% of samples were correctly discriminatedbased on trained VGG16 and AlexNet models, respectively. The obtainedsuperior results show that using a deep Convolutional Neural Networks(CNN) architecture is more efficient by the means of accuratediscrimination of wheat species. The proposed deep learning basedcategorization system promises high accuracy results for the qualityanalysis, classification and disease detection in food.
AU  - IŞIK, Şahin
AU  - ÖZKAN, Kemal
AU  - SEKE, Erol
DO  - 10.5505/pajes.2020.80774
PY  - 2021
JO  - Pamukkale Üniversitesi Mühendislik Bilimleri Dergisi
VL  - 27
IS  - 5
SN  - 2147-5881
SP  - 618
EP  - 626
DB  - TRDizin
UR  - http://search/yayin/detay/488104
ER  -