I am trying to select & learn a specific object with a lot of brown color variations in it, however I realized that this is hard to succeed…
Is it possible to control such an object? If yes, what am I doing wrong? what should be the background in such darker colors? Does a special lighting makes any difference? and finally, is it a better way to avoid the reflections near that darker area?
I’d appreciate any suggestion,
regards + greets to all pixy lovers
Brown will be pretty hard unless you have super consistent lighting. For instance sometimes my brown kitchen cabinets will show up as orange if the direct sunlight is in the room. In low light, dark colors like blues and browns will only register as black. Consistent lighting and setting the sensitivity will be your key.
Hello Vas,
As Scott points out, brown is difficult because it tends to be dark. This page has some hints:
https://docs.pixycam.com/wiki/doku.php?id=wiki:v2:some_tips_on_generating_color_signatures_2
But you might improve things are increasing the Camera brightness so that your brown is brighter and more in the linear region of the image sensor.
Hope this helps!
Edward
thank you both, it’s quite tricky to experiment with the brightness and the sensitivity for the darker objects, so I play with different light sources for now. My major problem is “…the stronger the light source, the bigger the reflections {overexposure?}” so I wonder if an ‘ultraviolet light’ or such, would help… I already tried placing a UV/polarised camera glass filter in front of the pixy lens, but the reflections were just changing place(direction I guess)…
Which version of Pixy are you using? The stock original Pixy has a UV-blocking coating on the lens, and I believe the same is true for Pixy2’s camera assembly…so in short, Pixy does not detect UV light (unless you purchase an IR-LOCK original Pixy, which has a different lens setup).
I’m using Pixy2 with the latest firmware. Not sure if it is de-UV.
However, throwing a UV-light from my flashlight, I see things changing a bit…
e.g.: the brown becomes red-to-brown…
so maybe the point is to find out what color areas are more “vivid” …or more manageable… in terms of camera read performance