Effectiveness of Deep Feature Extraction Algorithm in Determining the Maturity of Fruits: A Review

Abstract

Intelligent farming technology helps farmers overcome tough obstacles in the farming process, such as increased sup-plier costs, a lack of labour, customer satisfaction, and more. Artificial Intelligence (AI) is a remarkable technology in smart farming because it deeply understands the issue and can help farmers make decisions. This article's main objective is to identify and examine the concepts and techniques of Convolutional Neural Networks (CNN) technology that could aid in classifying the ripeness stages of fruit in intelligent farming. This paper systematically reviews 18 previous works for classifying the ripeness stages of fruit. This review outlines the most commonly used algorithms, activation functions, optimisation functions, and platforms for algorithm implementation. In addition, found that not all algorithms are suitable for even near-equivalent processes. Therefore, this study suggests the intensity of the CNN algorithms concerning various metrics to find the suitability for the operations/applications. Finally, this paper offers some future research directions in the ripeness classification of fruits.