@article{article_1562408, title={AI-Powered Subterranean Crop Harvesting Stage Detection and Automation}, journal={Gazi University Journal of Science}, volume={38}, pages={1819–1833}, year={2025}, DOI={10.35378/gujs.1562408}, author={M D, Rakesh and S B, Rudraswamy}, keywords={Agriculture Automation, Deep Learning, CNN, Harvesting, Robotics}, abstract={This work introduces a system that combines deep learning with robotics to automate the detection and harvesting of beetroot crops. The system utilizes a convolutional neural network (CNN) based on the ResNet-50 architecture for image classification and is trained to identify beetroot plants at their ideal harvesting stage. With an accuracy of 99.08% and a precision of 98.39%, the model ensures dependable detection. A robotic platform, equipped with a camera, captures images in the field, which are processed by the ResNet-50 model to assess the readiness of the beetroots. Once a beetroot is confirmed ready for harvest, a robotic arm is triggered to carry out the harvesting operation. This system tackles the difficulty of timely and accurate crop identification, automating a critical aspect of the harvesting process. By leveraging deep learning for detection and robotics for execution, the system aims to minimize manual oversight and improve the effectiveness of beetroot harvesting operations.}, number={4}, publisher={Gazi University}