Multi-Camera HDR Imaging via Simultaneous Subsampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image capturing and processing technologies are inefficient in generating high dynamic range (HDR) images with a wide field of view at high framerates, leading to poor visual quality and latency, which is inadequate for immersive extended-reality applications.
Innovation Solution
The implementation of a multi-camera system that employs simultaneous subsampling and HDR imaging techniques, using multiple image sensors with different settings for exposure time, sensitivity, and aperture size, and processing the captured image data using interpolation and demosaicking to generate high-quality HDR images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If full pixel readout is used to achieve wide field of view at high resolution, then image quality is improved, but processing time and computational resources increase significantly
Solution Approach 1:
The patent divides the image sensor into multiple regions with different readout modes. A first region uses full pixel readout for high quality images, while a second region uses subsampled readout for lower resolution images. This segmentation allows the system to process only necessary high-resolution data while reducing overall processing time and computational load.
Solution Approach 2:
Different regions of the image sensor are assigned different quality levels based on their importance. The first region (higher importance) maintains full pixel readout for superior image quality, while the second region (lower importance) uses subsampled readout. This local quality differentiation optimizes the balance between image quality and processing efficiency for different parts of the scene.
2Measurement precision
If full pixel readout is used to achieve wide field of view at high resolution, then image quality is improved, but computational power requirements increase
Solution Approach 1:
The patent segments the image processing workload by applying full pixel readout only to the first region requiring high quality, while using subsampled readout for the second region. This reduces the total number of pixels requiring computationally intensive processing, thereby lowering power consumption while maintaining acceptable image quality for the overall wide field of view.
Solution Approach 2:
The system applies different processing quality levels to different regions based on their importance. By maintaining full quality only where necessary and using reduced quality elsewhere, the system minimizes computational power requirements while preserving image quality in critical areas.
3Productivity
If subsampled readout is used to reduce processing time, then processing speed is improved, but image quality deteriorates
Solution Approach 1:
The patent implements segmentation by dividing the image sensor into a first region processed with full pixel readout (maintaining high quality) and a second region processed with subsampled readout (achieving faster processing). This allows the system to achieve overall processing speed improvements while preserving image quality in the more important first region.
Solution Approach 2:
Different quality levels are applied locally to different regions. The first region maintains high image quality through full pixel readout, while the second region accepts reduced quality in exchange for faster processing. This local differentiation resolves the contradiction by ensuring high quality where needed while achieving speed improvements overall.
4Area of stationary object
If more pixels are arranged on image sensor to achieve wide angle view at high resolution, then field of view is improved, but processing resources required increase
Solution Approach 1:
The patent segments the large image sensor into multiple regions with different processing requirements. By applying full pixel readout only to the first region and subsampled readout to the second region, the system can accommodate a larger total number of pixels for wide field of view while reducing the processing resources required by eliminating redundant processing in less critical areas.
Solution Approach 2:
The system applies different processing quality levels to different regions of the large image sensor. This allows the sensor to maintain a large area for wide field of view while consuming fewer processing resources by processing only the necessary portions at full quality and accepting reduced quality in other portions.
Data Source
AI summary
First image data and second image data are captured by a first image sensor and second image sensor(s), using at least two different settings. The first image data includes first subsampled image data of at least a first part of a first field of view of the first image sensor that has at least a part of overlapping field of view. The first image data and the second image data are processed together, using HDR imaging technique, to generate a first HDR image and a second HDR image. During processing, interpolation and demosaicking is performed on the first subsampled image data, by employing the second image data and using HDR imaging technique. Demosaicking is performed on the second image data, by employing the first image data and using HDR imaging technique.