CMOS Image Sensor Shared Pixel Structure for HDR
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current HDR imaging techniques for image sensors face challenges in achieving high dynamic range with compactness, fast response time, and minimal artifacts, particularly in industrial vision applications, due to high memory requirements and complex pixel structures that compromise resolution and noise ratio.
Innovation Solution
A CMOS sensor design with shared pixel structures and optimized control sequences enables horizontal interpolation and HDR reconstruction, using a global shutter mode with two exposure times applied alternately to columns, and analog summation in shared read nodes, reducing memory needs and improving signal-to-noise ratio.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional HDR techniques use multiple images with different exposure times, then dynamic range is improved, but memory capacity requirements increase significantly
Solution Approach 1:
The patent divides the pixel matrix into multiple groups, where each group contains pixels with different exposure times. This segmentation allows the sensor to capture HDR information across different exposure durations without requiring storage of multiple complete images, thereby reducing memory capacity requirements while maintaining dynamic range.
Solution Approach 2:
The patent introduces a temporal dimension by assigning different exposure times to different pixel groups within a single image capture. This allows HDR reconstruction by combining data from pixels with varying exposure durations, effectively extending dynamic range without requiring multiple sequential images, thus reducing memory needs.
2Measurement precision
If multiple images are captured for HDR reconstruction, then dynamic range is improved, but frame rate decreases
Solution Approach 1:
The patent performs preliminary action by capturing multiple exposure times simultaneously in a single image capture operation. Different pixel groups are exposed for different durations (e.g., short exposure for bright areas, long exposure for dark areas) within the same capture cycle, enabling HDR reconstruction without requiring sequential image acquisition, thereby maintaining high frame rates.
Solution Approach 2:
The patent implements periodic action by systematically alternating between different exposure time configurations across multiple image captures. This periodic variation in exposure parameters allows HDR data collection while maintaining regular capture intervals, ensuring high frame rates are preserved despite the complex HDR processing requirements.
3Productivity
If interlacing and interpolation techniques are applied, then frame rate is improved, but image resolution decreases
Solution Approach 1:
The patent applies local quality by assigning different exposure times to different spatial regions (pixel groups) based on their specific requirements. Bright regions use short exposure times while dark regions use long exposure times, allowing each local area to be optimized for its lighting conditions. This localized approach enables HDR reconstruction without requiring aggressive interpolation that would compromise overall image resolution.
Solution Approach 2:
The patent uses copying by creating multiple copies of the same scene at different exposure times within different pixel groups. Instead of requiring multiple complete images, the sensor captures multiple exposure versions of each local region simultaneously, enabling HDR reconstruction through copying and combining these local copies, thereby maintaining resolution while improving frame rate.
4Measurement precision
If complex pixel structures are used for HDR control, then dynamic range capability is improved, but device complexity increases
Solution Approach 1:
The patent applies universality by designing pixel groups that can function with different exposure times using a standardized pixel structure. The same basic pixel design serves multiple functions by being assigned different exposure durations depending on the region's lighting requirements, eliminating the need for complex specialized structures for each exposure type while maintaining HDR capability.
Solution Approach 2:
The patent implements dynamics by making the exposure time a variable parameter that can be adjusted for different pixel groups within the same capture operation. This dynamic control of exposure parameters allows a single pixel structure to adapt to different lighting conditions and HDR requirements without requiring physical structural changes, thereby maintaining device simplicity while improving dynamic range capability.
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 a compact, high-performance image sensor with improved frame rate and signal-to-noise ratio, capable of capturing high dynamic range images with reduced artifacts and memory usage, suitable for color imaging without compromising field of view.
Implementation Method 1
Each pixel on a respective row and column of the matrix comprises a photodiode
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
The invention proposes an image sensor having a functional architecture and an optimized pixel structure for promoting easy high-dynamic reconstruction without compromising on compactness. It uses a structure of pixels (fig.2) that vertically share, in pairs, at least one read node, and comprises a memory node for temporarily storing charges. It provides a layout of the pixel control rows such that commands to transfer charges (TRA) from the photodiode (PH) to the memory node (MN) apply vertically, to the pixels of one and the same column, and the other signals apply horizontally to the pixels of one and the same row, and said vertical transfer command rows are driven so as to apply two different exposure times, one to the even columns and the other to the odd columns. The pixels and the read circuit are driven so as to perform one read operation per vertical pair of pixels, which supplies a digital value (Vr) that represents the analogue sum of the charges integrated by the two pixels of the pair, over the exposure time associated with the column of the pair. A digital processing block is configured so as to calculate, for each vertical pair of pixels, an interpolated digital value (Vir) corresponding to the other exposure time, and select the value from the one associated with the long exposure time and the one associated with the short exposure time as high-dynamic digital value through comparison with at least one threshold. A periodic pattern of coloured filters makes it possible to create an HDR-mode colour image sensor, by defining a macro-pixel covering 4 blocks of 2 by 2 pixels, with one block per coloured filter colour of the pattern, promoting additional horizontal grouping in each block.