서울시립대학교 지능형로봇 팀 프로젝트
Task : 4구 당구 경기에서 최적의 샷 방향을 계산하여 시각화하는 지능형 로봇 시스템 구현
Project Period : 2024년 11월 ~ 2024년 12월
Tech : Python, OpenCV
Member : 2명
Project Wrap Up Report : click here
Project Slides : click here
Environments : Nvidia Jetson Javier NX, Logitech Web Camera
- 당구대 & 당구공 Detection
- 이미지 좌표계(2D pixel coordinate)에서 당구대의 꼭짓점과 당구공 중심 좌표 계산
- 당구대의 꼭짓점과 당구공의 좌표를 당구대 기준 좌표계(3D world coordinate)로 변환
- 당구대 평면이
Z=0
- 당구대 평면이
- 변환한 좌표를 이용하여 칠 방향 계산(3D world coordinate)
- 칠 방향 위의 한 점을 2D pixel coordinate로 변환 후, 화살표로 시각화
- 1, 2, 4, 5번 반복.
c
를 누르는 경우 3번도 재실행
- Table detection
- Ball detection
- Coordinate transformation (2D image coordinate -> 3D world coordinate)
- Get ball information with 3D coordinate
- Simulation
- Return optimal angle
- https://github.com/Detail-AR/Detail_AR
- Zhengyou Zhang. A Flexible New Technique for Camera Calibration. T-PAMI 2000