Fruits Freshness Recognition System by Manipulating Versatile Neural Networks

Fruits are the prerequisite authority of nutrition. Consuming fresh fruits on a daily basis is good for health. Fruits are the consequential provenance of income for farmers. Harvesting of fruits takes place after it acquires its matured state. The farmers has to separate the harvested fruit as fresh fruit or rotted or raw fruits before selling them. Predicting the fresh fruits for selling and purchasing for regular use by customers is still an intricate endeavor. The traditional methods of determining the ripeness and rottenness of fruits takes time. It requires a huge man power for detection. Sorting of fruits, peculiarly the ripened fruit first to sold them in market prior to the raw fruits is vital. As the maturated fruit are highly prone to decay than the others. Employing recent techniques in agriculture immensely assists farmers. In some instances, the fresh fruits and rottenness fruits looks similar in texture and color. Disparate fruits upon freshness is an exigent venture. The proposed work makes use of DenseNet architecture so as to concede and grade the freshness of fruits. The initiated work incorporates layers of CNN comprising convolutional layers, maximum pooling layers, flattening and fully interconnected neural nets incorporating numerous layers for detecting and sorting the fresh fruits. The proposed work gains an accuracy of 98 % in recognition of freshness of fruits.