Computational Camera Sensor Fusion for HDR Image Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional camera devices struggle to produce high-resolution, high dynamic range (HDR) images with a high signal-to-noise ratio due to limitations in image processing techniques, particularly in combining monochromatic and color image data effectively.
Innovation Solution
A computational camera system that employs fusion of image sensors using monochromatic and HDR color sensors, along with advanced image processing algorithms such as demosaicing, dynamic shifts, minimum spanning tree calculations, and guided filtering to enhance image quality by establishing precise correspondence and mapping between color and clear image pixels, thereby producing an enhanced, high-resolution HDR color image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional image processing techniques are used to combine monochromatic and color image data, then the processing can be kept simple, but the image quality and signal-to-noise ratio deteriorate
Solution Approach 1:
The patent segments the image processing into distinct stages: demosaicing of color data, establishment of correspondence between color and clear pixels using minimum spanning trees, and fusion of the processed data. This segmentation allows each stage to be optimized independently, achieving high image quality without overwhelming complexity
Solution Approach 2:
The patent performs preliminary demosaicing of the color image data before fusion, and pre-establishes correspondence relationships using minimum spanning trees. These preliminary actions prepare the data in advance, making the final fusion process more efficient and producing higher quality results
2Measurement precision
If advanced image processing algorithms are used to establish precise correspondence between color and clear pixels, then the signal-to-noise ratio improves, but the computational complexity increases
Solution Approach 1:
The patent replaces traditional mechanical or straightforward pixel-by-pixel correspondence methods with a minimum spanning tree algorithm that uses graph theory. This substitution creates more accurate correspondence relationships between color and clear pixels, improving signal-to-noise ratio through smarter computational approaches
Solution Approach 2:
The correspondence establishment using minimum spanning trees is performed as a preliminary step before the actual image fusion. This pre-computation of relationships allows the main fusion process to proceed more efficiently, reducing the overall computational burden
3Manufacturing precision
If multiple image sensors are used to capture both monochromatic and color data, then the HDR image quality improves, but the device complexity increases
Solution Approach 1:
The patent merges monochromatic and color image data through a systematic fusion process. By combining the high dynamic range capabilities of monochromatic sensors with the color information from color sensors, the system achieves superior HDR image quality that leverages the strengths of both sensor types
Solution Approach 2:
The patent segments the multi-sensor data processing into distinct phases: separate demosaicing of color channels, independent processing of clear image data, establishment of correspondence relationships, and final fusion. This segmentation makes the complex multi-sensor processing manageable and systematic
Data Source
AI summary
A camera device includes monochromatic and color image sensors that capture an image as a clear image in monochrome and as a Bayer image. The camera device implements image processing algorithms to produce an enhanced, high-resolution HDR color image. The Bayer image is demosaiced to generate an initial color image, and a disparity map is generated to establish correspondence between pixels of the initial color image and clear image. A mapped color image is generated to map the initial color image onto the clear image. A denoised clear image is applied as a guide image of a guided filter that filters the mapped color image to generate a filtered color image. The filtered color image and the denoised clear image are then fused to produce an enhanced, high-resolution HDR color image, and the disparity map and the mapped color image are updated based on the enhanced, high-resolution HDR color image.