adquiring a shoe footprint

[cadcad]

Hello everbody!

I´m a very new user of the NI Vision Assistant. I´m trying to detect the area (m^2) of a footprint captured with a high-speed camera under a glass floor.

How can I detect the contact zone?

I attached a capture of my problem,

thanks in advance!

[muks]

Dont you feel the area of interest is a bit too large fo this? There are lot of unwanted parts focused and thus wasting valuable pixels. Can you change the working distance?

[muks]

I actually thought only the top portion to be the foot print. So ignore my previuos message. Also can the non-uniform illumination be adjusted?

[cadcad]

dear Muks, yes that´s only a test, not a real image. The real system will work only with illumination under the glass, and without the hand moving the shoe.

Thank you! 

[cadcad]

yes, we can adjust the illumination under the glass. The problem is that I don´t have the control of the hardware, I receive the videos.

[muks]

One thing you can try is use pattern matching with the black hole seen. Also include some part of the bright shiny thing. So that you can determine the orientation with pattern amtching. With reference to this match you can draw a roi and extract the print.

[muks]

Try pattern matching.

Or try convex hull and particle analysis

[cadcad]

thank you very much!....I´m going to trying it!

[JesusG]

Hi,

from my point of view, the best solution is to adjust the illumination and binary particle analysis once thresholded.

Jesus

[AnalogKid2DigitalMan]

I am a bit confused- how are you going to detect what parts of the shoe are actually contacting the glass and what parts of the sole are not?

Sounds like an application for some type of reversible pressure sensing film or thin plate of LCD material (see what happens when you poke your finger into your LCD computer monitor screen).

-AK2DM