Project Name:COVID-911
Team:EOEO
Members: 박태양, 김윤환, 전현정, 김승수
Term:08/2020 - 10/2020
Project Video Link : https://youtu.be/5f1vN644pEo
If you have any questions, feel free to ask [email protected]
As the COVID-19 pandemic is getting worse and people are being exposed to the virus, wearing mask in public is considered mandatory, as well as checking people's temperature.
Therefore, we made a self-driving robot that patrols inside the building and detects people with high temperature, or not wearing masks and diagnose the possibility of COVID-19 in advance. If they found someone who's not wearing, they will warn them to wear mask until they wear it properly and then it will say 'thank you'.
This is H/W blcok diagram of our Robot.
Main H/Ws are
- NVIDIA Xavier AGX
- VLP-16 3D LiDAR
- LDS-01 2D LiDAR
- 2 Logitech C920 Cameras
- FLIR ONE Gen3 Thermal Camera
- Raspberry Monitor
-NVIDIA Xavier AGX will be used as a main board with Ubuntu 18.04 Linux environment.
-VLP-16 LiDAR will be used for SLAM.
-Upper Camera will be used for mask detection.
-Lower Camera and 2D LiDAR will be used for collision avoidance and to detect people.
-Theremal Camera will be used to detect people's temperature.
-Monitor will display if people are wearing mask or not and sound an alarm when at least one person is not wearing a mask.
-CUDA, Jetpack4.3, PyTorch, Tensorflow used
#Upper Camera
It detects whether people have wore a mask properly or not.
#Lower Camera and 2D LiDAR It stops when the camera detects people's foot, otherwise if it's just an obstacle or a wall, it executes collision-avoidance self driving.
#SLAM with VLP16 LiDAR
We used BLAM(Berkley Localization and Mapping) for slam which has a loop closure.
#We also used Autoware for Autnomous Driving with 3D LiDAR
Used pcd map file created with BLAM.
Ndt Matching for Localization.
Test Video Link : https://youtu.be/gTM9PTE23JQ
For mask detection, we forked codes from pyimagesearch's face detection and modified it. For foot detection, we also forked codes from jetbot and modified it.