Automatic White Balancing with Chromaticity Measure
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automatic white-balancing (AWB) algorithms for digital cameras face challenges in accurately estimating illumination chromaticity with minimal reliance on camera module characterization (CMC) data and without high-level image processing, especially in devices with limited processing capability or under power constraints.
Innovation Solution
A computer-implemented method that estimates the white point of raw image data by normalizing color channel values, determining chromaticity weights based on disparities and distances in the sensor chromaticity space, and accumulating weighted image data to achieve robust illumination chromaticity measurement, independent of CMC data and high-resolution requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the gray-edge algorithm is used to estimate white point, then white point estimation accuracy is improved, but processing complexity and computational requirements increase significantly
Solution Approach 1:
The patent extracts only the essential chromaticity information from image blocks without requiring full high-resolution image processing or complex edge detection algorithms. By working directly with down-sampled image blocks and their chromaticity coordinates, the method eliminates the need for computationally intensive edge detection while maintaining white point estimation accuracy.
Solution Approach 2:
The patent uses down-sampled image blocks as simplified, low-cost representations of the full image. These reduced-resolution blocks require minimal processing power and memory, making the algorithm suitable for mobile devices with limited resources while still providing accurate white point estimation through chromaticity analysis.
2Measurement precision
If high-resolution image data is used for edge information extraction, then white point estimation accuracy is improved, but memory usage and processing power requirements increase
Solution Approach 1:
The patent divides the image into multiple down-sampled blocks and processes each block independently to determine chromaticity weights. This segmentation allows the algorithm to work with small, manageable data units that require minimal memory, while collectively providing comprehensive white point estimation across the entire image.
Solution Approach 2:
The patent transitions from spatial domain processing (requiring high-resolution edge detection) to chromaticity space processing. By converting image block data into chromaticity coordinates and analyzing color relationships in this transformed space, the method achieves accurate white point estimation without requiring high-resolution spatial information.
3Measurement precision
If camera module characterization data is used for white balancing, then color accuracy is improved, but susceptibility to manufacturing variations and computational cost increase
Solution Approach 1:
The patent enables the camera system to perform self-calibration by analyzing the chromaticity distribution within the captured image itself. Rather than relying on pre-stored camera module characterization data that may contain manufacturing variations, the algorithm derives white point information directly from the image content, making the system adaptive to its specific hardware instance.
Solution Approach 2:
The patent changes the approach from using fixed camera module parameters to dynamically calculating chromaticity weights based on actual image block data. By computing weights from the ratio of color channel values in each block and comparing against expected achromatic values, the method adapts to actual sensor characteristics without requiring pre-characterization data.
Data Source
Figure 1~2
Figure 3A~3B
Figure 4A~4B
AI summary
An apparatus and methods for estimating a chromaticity of illumination from raw image data. In an embodiment, one or more image chromaticity weight is determined based on a distance between the raw image data in a sensor chromaticity space and a nearest point within a locus of sensor illumination chromaticities. In a further embodiment, one or more image chromaticity weight is determined based on a disparity among normalized color channel values. In certain embodiments, image chromaticity estimates are utilized to determine a white point estimate for the raw image data. In embodiments, an electronic device including a camera estimates the chromaticity value of raw image data captured by the camera as part of an AWB pipeline. The electronic device may further determine, for example as part of the AWB pipeline, a white point estimate based, at least in part, on the raw image data chromaticity value estimate(s).