SWIMSat-Python

Dependencies

Install the following dependencies:

• numpy
• OpenCV ( Best Version is 2.4, not 3.x)
• matplotlib
• Cython
• scipy
• statsmodels

sudo pip install opencv-python # Installs opencv 3.x
sudo pip install --upgrade matplotlib
sudo pip install cython
sudo pip install scipy statsmodels

Note: If any of the above libraries don't get installed on Windows, use Gohlke's Unofficial Windows Binaries for Python Packages. Note2: If the current version gives errors, use anaconda to install opencv 2.4, by using the menpo channel in conda

Instructions to run

• Install git and run

git clone https://github.com/AadityaPatanjali/SWIMSat-Python.git

• Change directory to SWIMSat-Python

cd SWIMSat-Python

• Connect the PhantomX Pincher and the webcam to the computer.
• Find out the port to which the robotic arm is connected.
  • Eg. /dev/ttyUSB0 for linux systems, COM3 for Windows, /dev/ttyACM0 for some Macs
  • If the port is not known, run

python IntegratedVisionController.py

to find out the port.

• Own the port by using Eg. sudo chown <username> /dev/ttyUSB0 or sudo chmod +x /dev/ttyUSB0
• Run python IntegratedVisionController.py <port> replacing by the appropriate port.
• A question will pop up to move to the Home Position. If the answer is no, you can reset the Home position (Use Ctrl + C to select a home position), automatically saving the new position.
• You will see two windows, Frame and Mask.
• Threshold of objects can be set by running the range_detector by using

python range_detector.py --filter HSV --webcam

or by modifying the 'Image Threshold' file. Open it in any text editor, but don't modify the syntax. Syntax is

[H_upper,S_upper,V_upper,H_lower,S_lower,V_lower]

where H, S and, V stand for Hue, Saturation and, Value respectively.

• To close the program do any of the following,
  • Close Frame, if in a linux machine.
  • Press q while the Frame window is active, if in a windows machine
  • Interrupt the program by pressing Ctrl + C in ther terminal and wait for the program to show the Goodbye message
  • If none of the above work, press Ctrl + Z in the terminal and restart the terminal window

Troubleshooting

If the file does not run correctly,

• Reconnect the power to the arm
• Reconnect the arm to the computer
• Reconnect the camera to the computer
• Run IntegratedVisionController.py and reset the home position and move the arm while doing so.

If you have problems with detecting the objects,

• Reset the thresholds by running

python range_detector.py --filter HSV --webcam

• make sure the required object is white and there's less 'white' noise 😉
• Change the HSV sliders one by one, starting from H upper, H lower, going through S and V alternating between upper and lower values