Fused Image Ghosting Removal via Spatial Difference Map
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Image stabilization techniques struggle to effectively capture both stationary and moving objects without ghosting artifacts, especially in scenarios with long-exposure times and camera motion, such as in still and video cameras, and astronomical telescopes.
Innovation Solution
The method involves capturing multiple images with both short- and long-exposure times, generating a spatial difference map to identify moving objects, and fusing these images using a weight mask to produce a crisp output image, reducing noise and visual discontinuities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If long-exposure time is used to capture stationary objects clearly, then image quality of stationary objects improves, but moving objects exhibit ghosting artifacts and blur
Solution Approach 1:
The patent segments the image into stationary and moving regions using a spatial difference map. By comparing long-exposure and short-exposure images, the system identifies which pixels correspond to stationary objects and which correspond to moving objects. This segmentation allows different processing strategies to be applied to different regions, resolving the contradiction between capturing clear stationary objects and avoiding ghosting of moving objects.
Solution Approach 2:
The patent applies local quality by using a weight mask to assign different weights to different pixels during image fusion. Pixels corresponding to stationary objects receive higher weights from the long-exposure image, while pixels corresponding to moving objects receive higher weights from the short-exposure image. This local differentiation resolves the contradiction by optimizing image quality for each type of object independently.
2Measurement precision
If short-exposure time is used to capture moving objects clearly, then moving objects are represented crisply, but stationary objects become noisy and lose detail
Solution Approach 1:
The spatial difference map segments the image to identify which regions contain moving objects. This segmentation allows the system to selectively apply short-exposure data only to regions with moving objects, while using long-exposure data for stationary regions. This resolves the contradiction by ensuring each region uses the exposure time most appropriate for its content.
Solution Approach 2:
The weight mask implements local quality by assigning pixel-specific weights based on motion detection. In regions with moving objects, the short-exposure image contributes more to the final output, providing crisp representation. In stationary regions, the long-exposure image contributes more, providing low-noise, detailed representation. This local optimization resolves the contradiction between sharpness and noise.
3Measurement precision
If multiple images are fused to represent both stationary and moving objects, then image quality improves, but processing complexity increases
Solution Approach 1:
The patent uses segmentation to divide the fusion process into distinct steps: generating a spatial difference map, creating a binary difference map through thresholding, and applying a weight mask. This structured segmentation of the processing pipeline makes the complex fusion operation more manageable and efficient, resolving the contradiction between improved image quality and processing complexity.
Data Source
AI summary
Techniques to capture and fuse short- and long-exposure images of a scene from a stabilized image capture device are disclosed. More particularly, the disclosed techniques use not only individual pixel differences between co-captured short- and long-exposure images, but also the spatial structure of occluded regions in the long-exposure images (e.g., areas of the long-exposure image(s) exhibiting blur due to scene object motion). A novel device used to represent this feature of the long-exposure image is a “spatial difference map.” Spatial difference maps may be used to identify pixels in the short- and long-exposure images for fusion and, in one embodiment, may be used to identify pixels from the short-exposure image(s) to filter post-fusion so as to reduce visual discontinuities in the output image.