opencv scripts for pick-and-place
includes research materials where relevant
targeted at mac OSX mavericks python2.7
opencv installed with homebrew.... each dependency was installed manually .. in this case, eigen, qt, and a reinstall of cmake. in some cases, numpy / scipy will need to be installed as well
to do:
- find contours
- test with wires
- mask object
- implement template matching
- find coordinates of masked + matching object
- auto-range for thresholding based on expected values
- integrate the components into a single toolchain
- 3dcam
- supports the kinect through libfreenect, tweaked some fairly straightforward video feed example, powered from a desktop power supply
- cntr
- does single largest blob tracking in video, look for the blue dot using static images --------------------
- cell
- usage: `python cell.py `
- the image should be within ~/projects/cv/images and should be already thresholded
- the bread-and-butter of this project: working on actually identifying the cell on a field
- wire
- usage: `python wire.py `
- this is a clone of cell that works off of wire_thresholding to threshold values for the wire
- wire_thresholding
- usage: `python wire_thresholding.py `
- a direct clone of thresholding as a placeholder for a more dynamic toolchain
- thresholding
- the lynchpin of the rest of the software, performs thresholding operations to reduce noise
- usage: `python thresholding.py `
- runs behind cell, to smooth the transition
- displays the thresholded image and saves it
- corner.py
- foundation functions for edge detection sloppy but reasonably quick
- good_corner
- just a demo from the tutorial... pretty accurate but savagely slow.
- contours
- an attempt to define the longest continuous rect
- square and sq2
- attempts to find black square