CashNet-15:An Optimized Cashew Nut Grading Using Deep CNN and Data Augmentation

Abstract

Since there is a great demand for the quality of agricultural products in the global market. It is very important to improve the quality and standards of agricultural products to competent in the business world. Furthermore cashew is a significant produce in India as well as it takes the major part in the global export market for cashew nut. But the most of the methods proposed for grading system is wouldn’t reach the better accuracy. Hence to improve the performance, we proposed the optimized cashew nut grading using Deep CNN and Data augmentation. This CashNet-15 work consists of totally 15 layers of CNN. Here we used 8 convolution layer and 4 Max-poolong layer for feature extraction and remaining are 1 fully connected layer, 1activation function and 1dropout layer. To attain the better performance we used data augmentation methods. To optimize the network, hyperparameter like SGD with Beta momentum and Leaky rectified linear unit was used to reduce the loss function and to obtain the non-linear property.