An Advanced Method of Identifying Fresh and Rotten Fruits Using Deep Learning

Abstract

Fruit classification is crucial in a variety of industries. This hierarchy aids sellers in various supermarkets in identifying fruit species and, as a result, impacts prices. The classification procedure makes it simple to decide whether the fruit is good or bad. People in the digital age take care of their everyday needs on their own, using a variety of contemporary facilities. If we are unable to export fresh fruits, our economic situation will deteriorate. In this example, a fruit categorization system could be useful in a range of fields, including the creation of smartphone apps for spotting unusual fruit species on the market and autonomous agricultural robots A total of 5658 fruits were used. (2834 fresh and 2824 rotten) in this study, which were divided into 10 classes (i.e 5 fresh fruits of 5 types & 5 rottenHumans become weary after performing the same role several times, so machines do not. As a result, the initiative suggests a plan for lowering costs and decreasing human effort. It is possible to reduce processing costs and time by detecting defects in agricultural fruits. We developed five models for our proposed classification system. The InceptionV3 model had the highest accuracy, at 97.34 percent.