https://github.com/francisso/robotcalibration

Find coordinate systems transformation between a RGBD camera and a robot's one.
https://github.com/francisso/robotcalibration

calibration computer-vision coordinate-systems manipulator python3 robotics

Last synced: 30 days ago
JSON representation

Find coordinate systems transformation between a RGBD camera and a robot's one.

• Host: GitHub
• URL: https://github.com/francisso/robotcalibration
• Owner: francisso
• License: mit
• Created: 2019-11-14T10:28:35.000Z (about 5 years ago)
• Default Branch: master
• Last Pushed: 2021-10-12T23:01:09.000Z (over 3 years ago)
• Last Synced: 2024-11-10T09:39:58.256Z (3 months ago)
• Topics: calibration, computer-vision, coordinate-systems, manipulator, python3, robotics
• Language: Jupyter Notebook
• Size: 12.3 MB
• Stars: 2
• Watchers: 3
• Forks: 2
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# RobotCalibration

Find coordinate systems transformation between RGBD camera axes and robot's ones.

![Alt text](readme_files/capture1.png?raw=true "Title")

Compute the transformation based on 4+ positions of any robot with pictures of robot in every position.

To find the positions of a robot's manipulator from a camera, the robot should hold a color ball.

## Usage

Install requirements.txt

1. Put a sequence of calibration images and validation images of a robot in the `images` and 'validation' folders

2. Save the corresponding coordinates of a manipulator in the robot coordinate system to the `positions/positions.json`

3. Set up a RoI in `configs/roi.json` and a matrix of intrinsic camera parameters in  `configs/camera_matrix.json`

4. Run `python3 calibrate.py` or use jupyter notebook `calibrate.ipynb`

5. Result transformation is saved in 'transformation.npy', also its accuracy on the validation set will be printed in calibrate.py

## Dependencies

* python3

* opencv (3.3.1)

* numpy (1.14.2)