Goal: Making a setup for measuring location of dot on 2D surface.

Process 2D setup

Goal: Learning OpenCV and tracking dots

Took a 4 hour course learning the basics of openCV. This lead to a code for tracking dots. All this can be found here: OpenCV and tracking

I also ordered an optical flow sensor which should arrive next Tuesday. And I got a Arduino portenta h7 with a vision shield which could be an alternative to for example a raspberry Pi camera module. This. was at the end of the week, so will look more into the portenta during the next weeks. 

BUSY WITH VOLUNTEER WORK

Goal: Get the PAA5100JE optical flow sensor up and running

After a couple of hours of coding and debugging, I got the sensor to read data using an Arduino micro and codes from the github library found here: Github library - optical flow.

Have yet to test the Portenta, but eager to see if I can use it too, making the camera a viable option if the optical flow proves to be harder than expected.

More about this week here. 

Goal: Research on measuring multilayer movement with optic flow

Secondary goals: 

• Finding something physical to make
• Getting started with the portenta as a second option to optic flow
• Research on sparse and dense optic flow

The portenta proved less cooperative than first thought (after debugging and finding solutions all Monday), and is being put on ice for the time being. 

More on what I did coding and imagewise, can be found here: Multilayer optical flow detection.

To do something physical, I made a CAD model of a leg with movement to mimic actual leg and lasercut it. More about that here.

Goal: Get the chessboard detection working, and trying it with multiple layers. 

Secondary goals:

• Develop setup for movement in transversal direction
• Cast silicone for skin outside on this

Still working on corner detection, this time with Harris corner detection. This can be found here (Harris Corner Detection).

If stuck to the sock, an optical sensor can easily capture how the skin moves relative to the sock!

Goal: Use Ludvik's calibration/ calibrate myself to measure in the curved setup.

Secondary goals:

• Finish setup with skin and grid
• Robustify my code further, or manually go through points to find next point close to last one.

The 3D printed setup for measuring rotations can be seen here.

Update Monday afternoon: The code now works and writes correctly to file. This can be seen in Harris corner detection (here) under "Working code (week 7)".

The next step this week will be differentiating between the red and green grid layer. 

Goal: Finish code for grids, and test on 3D setup.

Most coding updates can be found under "Multilayer grids" here.

Also made a version of the above tracking with video found here, under "Video grid tracking without Harris".

(EASTER) 

Goal: Filtering individual corners in both layers as seen under "Harris new better and best for now" here.

Mostly just filtering and adjusting. But really want to add something more to the project. Something physical. 

Also looking into multiple cameras to detect the 3D-motion with a setup. Researching triangulation and calibration, and write about it here: MultiCamera Movement Detection

Goal: Getting good overview of all my code, and add it to GitHub repository.

Secondary goals:

• Develop alternatives to grid due to the too many corners when using grid.
• Start working on setup for testing.

Developed a steancil to check for alternatives. This one with plusses instead of grid. Lost access to wiki, so lots of the data is on my computer.