Investigates Image data fusion techniques that combine image and track data from multiple sensors to achieve improved accuracies and more specific inferences than could be achieved by the use of a single sensor alone. Our aim is to explore the state-of-the-art image processing algorithms for achieving effective data fusion as in:
1. Introduction
- Super-Resolution is a process that reconstructs high resolution image from one or more low resolution images. This also
includes image denoising, deblurring and up-scaling process.
2. Main algorithm and principle
- General Super-Resolution algorithm structure
1. Compensate motion between low resolution images.
2. Fuses aligned low resolution images to get high resolution image.
3. Apply post-processing for deblurring or other artifacts.
- Sequential Super-Resolution based on Kalman filter approach
1. Recursive image acquisition modeling
2. Consider only translational motion for faster processing
3. Enable to process color video input
- Super-Resolution for moving object
1. Detection/tracking for moving object
2. Image registration using SIFT/Optical flow.
3. Data fusion based on ML estimation.
4. Enhance the reconstruction result by Patch-based Super-Resolution.
3. Experiment Results
- Sequential Super-Resolution
- Super-Resolution for moving object
4. Demo
