The Real-Time Object Detection System on Mobile Soccer Robot using YOLO v3
Authors:
- 1 October 2019
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Abstract
In a Soccer robot, each robot must be able to detect the object such as ball, goal and circle line in the game field through the vision system using a camera. The object detection in the past decade was used color filtering method, then it's methods was develop into neural network based. Neural network detection has also a lot of progress, start from R-CNN, fast R-CNN, and new method is YOLOv3. In this paper we will explain the implementation of YOLO V3 to detect object at Barelang mobile soccer robot. This system is run on SHUTTLE Xl MINI PC has Octa Core Intel Core i7-7700HQ processor, 16 GB of RAM and with 3GB of NVIDIA GeForce GTX 1060 graphics card has cuda core score 1152. The experiments result show the object detection get 28.3 fps. Datasets performance of the proposed method is IOU 71.76%, recall 0.92, precision 0.92 and mAP 87.07%. YOLO v3 capable to detect and distinguish objects in the different lighting condition, with max distance 3 m for ball object and 8 m for goal object.Keywords
This publication has 7 references indexed in Scilit:
- Goal Detection and Opponent Avoidance Algorithm for Wheeled Robot Soccer using Color Filtering and Contour Extraction2018
- Real-time face detection based on YOLO2018
- YOLO9000: Better, Faster, Stronger2017
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal NetworksIEEE Transactions on Pattern Analysis and Machine Intelligence, 2016
- You Only Look Once: Unified, Real-Time Object Detection2016
- Fast R-CNN2015
- The Pascal Visual Object Classes (VOC) ChallengeInternational Journal of Computer Vision, 2009