Performance Analysis of Fruit Quality Detection Using Computer Vision and Object Detection

Quality-based image classification is becoming a research topic in a wide range of applications, including the export market, precision agriculture, local vendors, fruit industries, and so on. Fruits are important in agriculture and industry, and many farmers in India rely on fruit farming as a profitable export. In line with the trend towards automation in many fields, machine vision techniques are being used more and more to categorize and assess the quality of fruits. In our study, the fruits are classified using the Yolov7 model, while advanced deep learning model InceptionV3 is used for assessing fruit quality. Further, we have conducted performance analysis of VGG19, ResNet50 and InceptionV3models for ripeness detection of fruits namely unripe, partially ripe, ripe and overripe. It was observed that ResNet50 gives the best result among them. The observed testing accuracy is 95.4%. The use of machine vision techniques in the fruit industry has revolutionized the way fruits are classified and their quality is assessed. It has numerous benefits, including increased accuracy, efficiency, and profitability, making it a crucial tool for the growth and success of the Indian fruit industry. This automation saves time, effort and money.