Deep Learning Based Dragon Fruit Classification

Abstract

The dragon fruit is now widely grown all over the world. Its popularity is growing as a result of its nutritious benefits. The predominant colors are red and white. However, because it appears to be the same from the outside, it cannot be identified. The CNN approach is used to detect dragon fruit in this research. For this type of dragon fruit, a detection approach based on shape features is proposed. Xceptional Model used for fruit shape features, which is utilized to detect fruits in potential locations and further locate fruit positions. Images captured under various illuminations were used to test the suggested strategy. The ability to detect fruits covered at various levels is also evaluated. The accuracy values are all greater than 94.89% percent, indicating that the suggested approach can detect a large percentage of covered fruits.