simtk.org

Hello,

I was currently conducting 15m sprint data collection using OpenCap. I positioned the checkerboard at the 15m line and utilized two iPhones, each placed frontally with a 45-degree angle from the checkerboard. The gathered displacement data, specifically the step length of the last three steps, was compared to data obtained from OptoJump for validation. However, it appears that the step lengths estimated from OpenCap data are significantly underestimated, with differences exceeding 20 cm. This discrepancy persists even in the case of the last step, which occurs in close proximity to the checkerboard location. I am seeking suggestions on what is the acceptable measurement volume from the location of the checkerboard and how to enhance the accuracy of anterior/posterior displacement measurements.

Thank you.

Hi Jay, can you share a session id so we can see the setup?

Dear Scott,

Here is the session ID. I used the trial named Tb.

758bc67c-7be2-42e4-bbab-2c8feb755897

Thanks!

Errors in translations can be due to inaccurate calibrations. In this case the calibration board has direct sunlight on it, which makes the black-black corners of the board less clear in the image. Two suggestions that I would try in order:
1) Shade the calibration board or face it away from direct sunlight. This will make the calibration more accurate.
2) We often use a larger board for large capture volumes, which increases the calibration accuracy. We use the 800x600 course one from calib.io: https://calib.io/collections/products/p ... 4463429679

I will try to use the lager checkerboard with better light condition. Really appreciate your suggestion.

Make sure any checkerboard you use is very flat, and verify the dimensions of the squares.