Fruit and Vegetable Classification and Freshness Detection using Machine learning

Abstract

Freshness is a key factor in determining a fruit or vegetable’s quality, and it directly influences the physical health and coping provocation of consumers. It ascertains the nutritional value of the specified fruit or vegetable. This paper proposes a well-organized and precise fruit and vegetable classification and freshness detection method. The proposed method employs state-of-the-art deep learning models, specifically convolutional neural networks (CNNs), to analyze images of fruits and vegetables captured through high-resolution cameras. The dataset used for training and evaluation is extensive and diverse, encompassing a wide variety of fruits and vegetables in various conditions. The freshness of a fruit or vegetable can be ascertained by looking at a variety of features, including color, texture, shape, and size. Fresh produce, for instance, is colorful and free of mold or brown spots. Traditional methods for assessing the quality of fruits and vegetables are both time-consuming and error-prone. These methods include inspection and sorting. It is possible to reduce these issues by utilizing automatic detection techniques. In light of this, we proposed an automated fruit-vegetable freshness detection approach that first recognizes whether the image is of a fruit or vegetable, after which it classifies it into one of three freshness categories: rotten, fresh, or mixed. To identify and categorize fruits and vegetables, two deep learning models are employed: You Only Look Once (YOLO) and Visual Geometry Group (VGG-16). The suggested method’s qualitative analysis indicates superior performance on the fruit-vegetable dataset.

Topics & Concepts