Maturity Detection of Tomatoes Using Deep Learning

Agriculture 5.0 primarily constitutes the use of artificial intelligence and robotics as a hybrid technology that can automate a major portion of agriculture. Artificial intelligence will provide a cognitive skill to a computer to detect diseases that may occur in various eatables, such as fruits and vegetables, that can lead to a potential loss of crop. Also, the maturity that is the ripening status of these fruits and vegetables can be estimated to decide harvesting time. There are numerous ways to estimate the ripening status based on size, shape, texture, or color. Most of these features can be captured with images or video and decision-making is made possible by applying deep learning and artificial intelligence. After the decision-making stage, the fruit or vegetables can be plucked with a robotic arm. In this paper, we demonstrate the use of deep learning to detect the maturity specific of tomatoes. We create a customized dataset of images and use convolution neural networks along with the popular object detection model, YOLO v3 to detect the maturity of the tomatoes and pinpoint their location. © 2021, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.