Comprehensive Systematic Review on Fruit Maturity Detection Technique

Abstract

The objective of this paper is to review the classification of multiple fruits in different environments. The tropical fruits are grown mostly in South East Asia, Thailand and in Malaysia. Most of the fruits are usually recognized by its strong aroma and unique taste. These fruits are identified as the most precious fruit in India for the effective health nourishment benefits. The appearance of the fruits generally identifies the quality and ripeness of the fruits. The Quality of the fruit is also determined by the strength of color, texture, aroma and the maturity level of the fruit. This identification of the maturity levels when done manually leads to many flaws. Thus techniques like computer vision, machine learning are used for easy classification of the fruits in to various stages. A detailed review of the multiple fruit classification methods are done with both Machine Learning and Deep Learning Techniques and a comparison is made on the accuracy obtained by these techniques. The feature extractions are done on the different datasets based on size, shape, color and texture. In image processing techniques there are significant research areas to work on the image but classification of image is one of the most complex area. The classification of multiple fruit images are categorized in two main ways of Supervised and Unsupervised methods in Machine Learning Technique. Both the methods have their properties and functions. The main challenge is to obtain a proper and desired result from a noisy fruit image as compared to that of the normal image. Fruit maturity detection is divided into different stages depending on the type of detection method used. In this study we present an analytical review on various image classification and maturity detection techniques of Multiple Fruits. This review paper describes the description of the non-destructive techniques to estimate the quality of the different fruit and to improve the classification based on maturity level.