| dc.description.abstract | Science and technology continue to rapidly develop, including in the field of
computers, which is divided into several branches such as computer vision and
robotics. Computer vision studies, analyzes, and interprets visual information in the
form of videos or images, and YOLO (You Only Look Once) is one of the most
frequently used algorithms due to its fast and accurate detection capabilities.
YOLOv3 is considered a stable and easy-to-use version of YOLO, but it produces
larger and more complex files. Meanwhile, in robotics, robotic arms are commonly
used in industries, but their high cost can be overcome by using 3D-printed robotic
arms. On the other hand, edge computing can be used for better and cheaper
computational processes, which have the potential to produce an automatic object
identification and sorting system for cactus plants at an economic cost. The results of
this research are obtained in the form of accuracy, recall, precision, F1 score, and
mAP. The total accuracy value is 78%, recall is 96%, precision is 80%, F1 score is
97%, and mAP is 83% out of a total of 543 data. It can be concluded that cactus plant
sorting using a combination of YOLOv3 and robotic arms has been successfully
implemented and is quite effective in sorting objects. | en_US |
