Convolutional Neural Network Based Classification Of Walnuts Into Healthy And Diseased
Abstract
Walnut has a substantial nutritional value as one of the most popular dietary nutrients. However, it is hard to judge the quality of the walnuts because of its tough shell. In this study, the combination of photography and machine learning generated a novel way for identifying the quality of walnuts without destruction. First, the image of the walnut samples was obtained using a DSLR camera. The dataset has been collected in “healthy” and “diseased” groups and sorted out. In discriminating against walnuts, the CNN was then utilized. Many models have been trained with various parameters. Finally, the results of the tests have been used to evaluate the models. CNN observed the best performance. The accuracy of our model was found to be 99 per cent and it was able to distinguish accurately between Healthy and Diseases Walnuts. Overall study shows that the image data based on the machine may be used for quick, correct and safe detection of walnuts.