Motion Invariant Imaging De-Blurring via Parabolic Trajectory
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Motion blur in industrial imaging systems, particularly with lower cost rolling shutter sensors, limits image quality and processing effectiveness, as traditional de-blurring techniques often fail due to un-invertible motion blur and power constraints.
Innovation Solution
Implementing a motion invariant imaging approach by moving optical system elements in a parabolic trajectory during image capture, allowing for post-processing that estimates object speed and reduces de-convolution artifacts, without pre-exposure motion estimation and maintaining high light throughput.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional de-blurring techniques are applied to motion blurred images, then image processing is attempted, but the results are inadequate because motion blur is un-invertible
Solution Approach 1:
Instead of attempting to directly deblur motion-blurred images (which is un-invertible), the patent inverts the approach by intentionally introducing a known blur kernel during capture. This allows the blur to be invertible in post-processing, transforming an unsolvable problem into a solvable one.
Solution Approach 2:
The patent applies preliminary action by pre-introducing a controlled blur kernel during the image capture phase. This preliminary blurring operation enables subsequent de-blurring to be mathematically invertible, whereas direct de-blurring of natural motion blur would be impossible.
2Manufacturing precision
If global shutter sensors are used to eliminate motion blur, then image quality improves, but cost increases significantly
Solution Approach 1:
The patent creates a controlled copy of the blur effect through a known kernel during capture, which then allows perfect reconstruction. This approach copies the essential information needed for de-blurring without requiring expensive global shutter hardware.
Solution Approach 2:
The patent changes the parameter of blur invertibility by introducing a specific blur kernel with known characteristics during capture. This parameter change transforms the un-invertible motion blur into an invertible mathematical operation.
3Ease of manufacture
If rolling shutter sensors are used to reduce cost, then system cost decreases, but motion blur increases and limits processing effectiveness
Solution Approach 1:
The patent converts the harmful motion blur introduced by rolling shutter sensors into a beneficial known blur kernel. By intentionally introducing a controlled blur during capture, the harmful effect becomes a useful piece of information that enables perfect de-blurring.
4Illumination intensity
If high-powered flash is used with global shutter sensors to illuminate target, then image illumination improves, but power consumption exceeds available power in power-constrained applications
Solution Approach 1:
The patent uses a inexpensive rolling shutter sensor instead of expensive global shutter sensors that require high-powered flash. The rolling shutter's sequential reading mechanism allows for longer exposure times with lower instantaneous power requirements.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A method that includes using a point spread function to de-blur an original motion invariant image to create a modified motion invariant image; using an edge detector to find edges in the modified motion invariant image; determining the distances between the edges and corresponding artifacts in the modified motion invariant image; using the distances between the edges and the corresponding artifacts to estimate a velocity of an object in the modified motion invariant image; generating a corrected point spread function corresponding to the estimated velocity of the object; and using the corrected point spread function to de-blur the original motion invariant image and create a resulting image.