Image processing Using my computer

[project maker No.1001628]

so I want use an embedded 3d stereo camera or multiple infrared cameras, and do the processing of the image using my cpu and display it on my screen, how do i start.
>> i want to do something similar to kinect where my motion and gestures can be sensed i wouldnt mind using flourecent paper to make things easier
>> i would also like to use the gpu to make some visualy pleasing graphics

[Anonymous No.1001971]

>>1001628
python have multiple libraries for image processing, but i think only opencv makes use of your gpu

no idea about your cameras resolution but you have to be clever about not to wast too much processing time on every frame, if it wont have to be real time its fine.

blender have a scripting interface in python, that might be a good bet for making pleasing graphics

using markers would be your best bet, just threshold the image and track it, cv would be great for tracking but there are other methods, use some logic to triangulate the 2 image to 3d points and skeletonize the markers, then do whatever

there should be lots of stuff about these on the web

[Anonymous No.1003140]

>>1001628
i think the cheapest DIY mocap involves 4(?) ps3 eye toy cameras hooked up to your PC via USB

[Anonymous No.1003144]

Apart from the obvious:Get 2 cameras and a computer.
You are gonna do a lot of programing, Here is what i would use:
Base:
C++
https://isocpp.org/
Computer Vision:
OpenCV
http://opencv.org/
Display:
OpenGL
https://www.opengl.org/

The last 2 are old aren't really easy to use and aren't as well documented by todays standard. But have a lot of "Helper" Libraries that are newbie friendly. They are also cross platform with just a recompile.

You can use OpenCV to interpret the image you are seeing and OpenGL To display an image based on said interpretation. The only limit is your coding and hardware.If you want to add sound OpenAL