Hardware:
- Raspberry Pi 3
- Raspberry Pi Camera module
Software:
- OpenCV 3.1.0
- Python 2.7
- SciPy 0.17
- Ubuntu MATE 15.10 for Raspberry Pi 3
The Calibration Setup
Initial Setup:
- Fix camera for test pictures
- Measure out lab space, put down tape on floor indicating distances from cameras
- Take pictures at certain distances (4-15 feet) of a red paper
- Mask (HSV) the images for target color
- Measures target object pixel area at each distance
- Fit a decaying exponential on the measured points
- Use this new function to map distance given pixel area
- Code for measuring size and area is done and present in our Github Repository.
- Test out mapping on new object with same areas and see if the predicted distance is correct
- Use existing measurements to interpolate points for object with arbitrary area
- Fit a decaying exponential on the interpolated points