Automatic Fruit Recognition Based On DCNN For Commercial Source Trace System

Automatic fruit recognition-based on machine vision is considered as challenging task due to similarities between various types of fruits and external environmental changes e-g lighting.Fruit and vegetable classification is one of the major applications that can be utilized in supermarket and fruit shops to automatically detect and recognize the kind of fruits and vegetables purchased by customers and to determine its prices.Although automatic fruit recognition is getting more and more important,this technology is still far from being mature.The main work and innovations of this thesis are as follows:(1)A fruit recognition algorithm based on deep convolution neural Network(DCNN)is proposed.Most of the previous techniques have some limitations because they were examined and evaluated under limited dataset,furthermore they have not considered external environmental changes.We evaluated our model in much more complicated dataset which mostly meet all real-world challenges to make our model robust.(2)We established fruit images database spanning 15 different categories which comprise of 44406 images collected in our lab environment during a period of 6 months under different real-world conditions & keeping in view the limitations of existing dataset.(3)We also propose an algorithm for Intra-class recognition of fruits using DCNN(deep convolutional neural network).The results of carrying out these experiments demonstrate that the proposed approach could be automatically recognize the fruit with accuracy of 99%.The Input images were directly input to DCNN for training and recognition without features extraction & the DCNN learn optimal features from images adaptively.The final decision was made based on a fusion of all regions classification using probability mechanism.