Weighted Dark Current Filtering for Low-Light Image Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image sensors, particularly CMOS and CCD sensors, suffer from high read noise and motion blur under low light conditions, affecting image quality and other operations in mixed-reality systems, while SPAD sensors face challenges with signal noise due to dark current, leading to degraded image quality.
Innovation Solution
A system for dark current compensation using weighted filtering, which generates a refined dark current image through template matching, non-maximum suppression, and template fitting to account for changes in faulty pixels, and combines it with motion-compensated previous images based on ambient light conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If conventional CMOS or CCD image sensors are used under low light conditions, then the system can capture images, but read noise and motion blur degrade image quality
Solution Approach 1:
The system performs preliminary actions by capturing dark current images during calibration phases and storing reference data about faulty pixels. This preliminary characterization of sensor defects allows the system to compensate for dark current effects during actual operation without requiring additional real-time calibration, thereby improving image quality in low light conditions.
Solution Approach 2:
The system implements feedback mechanisms by continuously monitoring image data for patterns indicative of dark current noise and faulty pixels. The system uses this feedback to dynamically adjust processing parameters and apply corrective filtering, progressively improving image quality as the system adapts to specific sensor defects during operation.
2Reliability
If SPAD sensors are used to reduce read noise, then signal detection improves, but dark current causes signal noise and degraded image quality
Solution Approach 1:
The system extracts and isolates dark current effects from the signal by capturing separate dark current images where only the sensor's inherent noise is present. By extracting these dark current components and storing them as reference data, the system can later subtract or filter them from actual signal images, thereby eliminating dark current noise while preserving SPAD sensors' superior signal detection capability.
Solution Approach 2:
The system introduces intermediary processing steps between the SPAD sensor and the final image output. These intermediary steps include dark current image capture, faulty pixel identification, and weighted filtering that uses reference data to correct sensor defects. This intermediary processing layer mediates between the high signal detection capability of SPADs and the harmful dark current effects, producing clean final images.
3Measurement precision
If the frame rate is reduced to allow sufficient photon detection, then read noise decreases, but motion blur increases
Solution Approach 1:
The system segments the image processing task into multiple components: dark current compensation, motion compensation, and final image synthesis. By handling dark current effects separately through calibration and reference data, the system can operate at higher frame rates without suffering from accumulated noise, as the noise correction is performed independently of the frame rate.
Solution Approach 2:
The system dynamically adjusts processing parameters based on operational conditions. The weighted filtering process adapts to varying scene conditions and sensor states, allowing the system to optimize between frame rate and noise reduction in real-time. This dynamic adjustment enables high frame rates while maintaining image quality through adaptive compensation rather than fixed low-frame-rate capture.
Data Source
AI summary
A system for facilitating dark current compensation by weighted filtering is configurable to (i) receive an input dark current image; (ii) generate a corrected dark current image at least by scaling pixel values of the input dark current image based upon ambient light conditions; and generate a weight map comprising a weight value for each pixel of the corrected dark current image. For each pixel of the corrected dark current image, the weight value of the weight map may be based upon a light level of the pixel of the corrected dark current image. The system is also configurable to generate an output image by utilizing the weight map to filter an input image.