Image Synthesis via Multi-Resolution Capture and Binning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image capture technologies face challenges in reducing motion blur and noise, especially in low-light conditions, due to smaller pixel sizes and increased exposure times, leading to compromised image quality and hardware limitations in processing multiple high-resolution images.
Innovation Solution
The method involves acquiring multiple lower resolution images with longer exposure times and fewer higher resolution images, using binning to increase signal-to-noise ratio, and merging these images to form a synthesized image with reduced motion blur and higher resolution, while minimizing shutter lag through a rolling buffer and synchronized audio data capture.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If exposure time is increased to reduce noise in smaller pixels, then signal-to-noise ratio is improved, but motion blur increases
Solution Approach 1:
The patent segments the image capture process into multiple separate exposures of different resolutions rather than a single high-resolution exposure. Multiple low-resolution images are captured quickly to reduce motion blur, while a single high-resolution image captures fine detail, and these are later merged to achieve both low noise and high resolution.
Solution Approach 2:
The patent applies partial action by capturing multiple images at different resolutions rather than a single full-resolution image. The low-resolution images are used primarily for noise reduction and motion compensation, while the high-resolution image provides the final detail, allowing each image type to be optimized for its specific purpose.
2Measurement precision
If multiple high resolution images are captured to reduce motion blur, then spatial resolution is maintained, but readout time increases causing shutter lag
Solution Approach 1:
The patent segments the image set into a small number of high-resolution images and multiple low-resolution images. The low-resolution images can be read out quickly due to fewer pixels, reducing the overall readout time and shutter lag, while the few high-resolution images maintain spatial resolution in the final merged image.
Solution Approach 2:
The patent uses partial action by capturing only a limited number of high-resolution images (e.g., 1-7 images/sec) rather than continuously capturing high-resolution frames. This reduces the readout burden while still providing sufficient high-resolution data for the final image, balancing quality with speed.
3Speed
If binning is used to increase pixel sensitivity and reduce exposure time, then temporal resolution is improved, but spatial resolution decreases
Solution Approach 1:
The patent segments the resolution requirements by using binning only for low-resolution video/image capture, not for the final high-resolution output. The binned low-resolution images provide temporal resolution and noise reduction, while the un-binned high-resolution images preserve spatial resolution, and the two are merged to achieve both goals simultaneously.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach results in improved image quality with reduced motion blur and increased signal-to-noise ratio, maintaining spatial resolution and providing a video context to the synthesized image, while minimizing the perceived shutter lag.
Implementation Method 1
the available area to capture light during the exposure
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Multiple images of a scene are acquired over a contemporaneous period of time. Most of the multiple images are lower resolution images acquired with a lower resolution than the other of the multiple images. A corrected set of images is formed at least by correcting for motion present between at least some of the lower resolution images. In addition, a synthesized image is formed at least by merging (a) at least a portion of at least one of the images in the corrected set of images, and (b) at least a portion of at least one of the other of the multiple images. The synthesized image is stored in a processor-accessible memory system. The synthesized image exhibits improved image quality including reduced motion blur, a higher signal-to-noise ratio, and higher resolution over conventional techniques.