I have a question that if we want to do fast advanced image processing (like stereo image processing , object tracking which high computation,time consuming processing) in real time which software and hardware is preferred? Which will be the best for advanced image processing?
If you mean a library which has many useful functions and it can be used for programming in C/C++ or C# , OPEN CV and EMGU CV is good choices.
If you mean a library which has many useful functions and it can be used for programming in C/C++ or C# , OPEN CV and EMGU CV is good choices.
I would add a Tesla http://www.nvidia.com/object/tesla-servers.html
and you can program in C with CUDA extensions.
I agree with Mehdi. As ana alternative, if you need to develop or to use some algorithms for real-time imaging and you need to debug often or to try different configurations, you can also use Matlab/Simulink.
I would recommend C/C++ rather than Matlab particularly for real-time applications. The choice of hardware would change depending on whether you need an embedded platform or not.
OpenCV is classic decision to create system in one library, but you can search for decision for your current project in other open source like http://www.ensta-paristech.fr/~garrigues/video_extruder.html (good tracking project, but i not use it in practice) . In some cases speed of OpenCV is not enough
I am agree with Serkan Gunal that "The choice of hardware would change depending on whether we need an embedded platform or not." so What type of hardware we should use for embedded platform?
Though there are many alternatives, you can take a look at Texas Instruments DSP kits, Altera FPGAs, or beagleboard.org.
This board http://www.parallella.org/board/ will be available soon. It is an ARM-A9 based system with FPGA and an OpenCL compatible multi-core (16 and 64) co-processor. It runs a Linux OS. Seems promising when it comes to embedded applications.
- Badges
- Science topic