Banana Grading using Deep Learning Techniques

Abstract

The horticulture sector in India accounts for around 33% of agriculture’s Gross Value Added (GVA), making it a substantial asset to the Indian economy. India is the second greatest producer of fruits and vegetables in the world, producing mango, banana, guava, papaya, sapota, pomegranate, and lime. Bananas are a popular fruit because of their inexpensive cost and great nutritional content. It is eaten as a ripe or raw fruit in both fresh and cooked form. Among many commercial varieties of banana, Nendran is very unique as it is the only variety that is consumed in many forms and at different stages of ripening. Major production and marketing in Tamilnadu and Kerala. This research work provides one of the most efficient ways for grading banana fruit using deep learning models like EfficientNetB7 and NasNetLarge, on a banana dataset with an accuracy of 97% and 95%, respectively.