Hi,
I've been extensively working in image processing and computer vision domain since past 2+ yrs. I've been doing programming using OpenCV, C++, QT and using CUDA GPU. CUDA GPU was used for acceleration.
Below is the work done so far in image processing domain:
1. Mosaicing high resolution (1024x770) and low resolution (640x480) images.
2. Hands on with various image registration methods like Phase Correlation, Feature Based (SURF/SIFT), Template Matching, Optical Flow etc.
3. Moving Object Detection
4. Tracking
5. Zitter correction & Image stabilization
6. Image enhancement: De-hazing, De-noising, Light balancing, Intensity equalization,Sharpening, Turbulence correction, De-vignetting etc.
Mosaicing application for low resolution images was developed for live video feed and was running in real-time at 25 FPS.
The solutions developed are for both PC and embedded hardware running on Win and Linux.
The solution for this project, as seems from the given requirement involves following steps:
1. Image enhancement
2. Image registration
3. Image stitching or mosacicing
Let me know if my understanding is wrong.
Query:
1. How much processing time is expected per frame?
2. How much is the image resolution?
3. Is there any scope of GUI?
Look forward to hear back.