An Intelligent Approach to Identify the Date Palm Varieties Using Leaves and Fruits

EasyChair Preprint 6323

Abstract

Tunisian people lack the knowledge to determine the type of date fruit, many consumers struggle when they go to buy fruit every day, even novice farmers find it difficult to know the type of dates they produce. It is very difficult for the human eye to classify similar things such as date fruit. Thus, it is necessary to develop an accurate solution that can accomplish this task. The goal of this research is to prove the feasibility and test the capability of our solution, we focus on identifying the varieties of dates fruits not only using dates images but also palm leaves images. We implement three CNN models for date fruit and leaf classification which the first classifies fruit and leaf image (binary classification), the second classify fruit varieties, and the third classify the leaves varieties. We have come up with good accuracy 99.97% for the binary model, 99.82% for the fruit model, and 99.73% for the leaf model.

Keyphrases: Classification, Convolution Neuron Network, Date palm fruits, Date palm leaves, Phoenix dactyliferaL

@booklet{EasyChair:6323,
  author    = {Neji Ines and Hamza Hammadi and Ejbali Ridha},
  title     = {An Intelligent Approach to Identify the Date Palm Varieties Using Leaves and Fruits},
  howpublished = {EasyChair Preprint 6323},
  year      = {EasyChair, 2021}}