Flutter Shutter Camera Invertible Motion Blur Deblurring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Acquiring sharply-focused images of moving subjects is challenging due to motion blur, especially in applications like iris recognition, where traditional shutters limit the tolerance for subject velocity and result in irreversible loss of information, amplification of sensor noise, and hallucination of image content.
Innovation Solution
The method involves pre-computing invertible point spread functions (PSFs) for various motion rates and using a flutter shutter camera to capture images, estimating in-capture motion, and modulating light according to these PSFs to produce invertible motion blur, allowing for effective deblurring and noise reduction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If a traditional shutter is used to capture images, then the image capture process is simple, but motion blur occurs and information is irreversibly lost when subjects move
Solution Approach 1:
The shutter is operated in a periodic fluttering pattern during exposure, opening and closing at controlled intervals. This periodic action encodes motion information into the blur pattern, making it invertible. The shutter cycles through open and closed states systematically, allowing the capture of multiple temporal snapshots that can be computationally reconstructed to recover the sharp image.
Solution Approach 2:
The invention converts the harmful effect of motion blur into a beneficial encoding mechanism. Instead of treating blur as information loss, the systematic fluttering shutter transforms blur into a coded pattern that contains recoverable information. The blur becomes a carrier of temporal-spatial information that can be decoded through computational methods to reconstruct the original sharp image.
2Manufacturing precision
If exposure time is increased to capture clear images in moderate lighting, then image quality improves, but the range of tolerable subject velocities decreases
Solution Approach 1:
The shutter transitions from a static open/closed state to a dynamic fluttering state with controlled temporal variations. This dynamic operation allows the system to adapt to different subject velocities by adjusting the fluttering frequency and pattern. The dynamic shutter creates a time-varying point spread function that encodes motion information, enabling recovery of sharp images even when subjects move at higher velocities during the extended exposure period.
3Measurement precision
If post-processing deblurring is attempted on traditionally captured blurred images, then subjective image quality may improve, but sensor noise is severely amplified and image content may be hallucinated
Solution Approach 1:
The system performs preliminary encoding of motion information during the image capture process itself, rather than attempting to recover information after blur has occurred. The fluttering shutter pre-structures the captured data to contain recoverable information about the sharp image and motion patterns. This preliminary encoding ensures that the information needed for deblurring is preserved in the captured image, allowing noise-free reconstruction without the need for aggressive post-processing that would amplify noise or hallucinate content.
Data Source
AI summary
Methods and systems for capturing an image of a moving subject employ a camera image sensor that captures a blurred image of a moving subject. In-capture positions of the moving object are also determined using a high frame rate camera or other motion sensing device. The PSF for controlling modulation of the light hitting the camera image is successively updated by selecting, from among a plurality of pre-computed invertible PSFs, a pre-computed invertible PSFs for each estimated motion of the moving object. Light hitting the camera image sensor is modulated in capture phase according to one or more of the updated pre-computed invertible PSFs such that the captured blurred image is invertible. The resulting invertible blurred image can be de-blurred using the selected known PSFs to provide a substantially sharp image.