Confusion Matrix Using Yolo V3-Tiny on Quadruped Robot Based Raspberry PI 3B +

Abstract

The implementation of Confusion Matrix uses Yolo V3-Tiny to detect objects in the form of person cameras and robots for navigation.It is hoped that the quadrupped robot in this research can be used during natural disasters so that it can recognize people who are still or have died so that they can be identified.However, in the evacuation process there are often obstacles, such as the absence of information on victims and the location of disaster victims and the difficulty of accessing victims of natural disasters.Based on these problems, research was made regarding quadruped robots by applying a confusion matrix using YOLO V3-Tiny to detect victims of natural disasters.The YOLOv3-tiny model or system can detect person objects according to the dataset that has been studied previously and in this research hardware and software design was carried out by implementing the YOLOv3-Tiny model on a Raspberry PI 3B+ which was placed on a quadruped robot.The system (model) that has been created with YOLOv3-Tiny is capable of detecting person objects.The system (model) created in this research has an Accuracy value of 100%, a Precision value of 1, a Recall value of 1, an F1 -score of 1.