Pixel-Level Conversion Gain Selection for High-Dynamic-Range Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image sensors struggle to capture high-dynamic-range images, often resulting in lost highlight and shadow detail due to limited dynamic range, leading to issues like blown out highlights and overexposed shadows.
Innovation Solution
The implementation of automatic conversion gain selection (ACGS) circuitry on a pixel-by-pixel basis, which adjusts the gain settings based on light intensity, allowing for enhanced dynamic range by combining image data from pixels with different gain settings to capture both bright and dark areas effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a single image sensor with fixed gain setting is used, then the device complexity is low, but the dynamic range is limited resulting in lost highlight and shadow detail
Solution Approach 1:
The image sensor array is divided into multiple segments, where each pixel or group of pixels can independently select between different gain settings (first gain and second gain). This segmentation allows different regions of the image to use different gain values, enabling high dynamic range capture while keeping each individual pixel's circuitry relatively simple.
Solution Approach 2:
The gain setting for each pixel is made dynamic rather than fixed. The circuitry automatically switches between first and second gain settings based on the light intensity detected by each pixel during exposure. This dynamic adaptation allows the system to optimize for both bright highlights and dark shadows within the same image, resolving the contradiction between adaptability and complexity.
2Manufacturing precision
If automatic conversion gain selection circuitry is implemented on a pixel-by-pixel basis, then highlight and shadow detail are retained, but the device complexity increases
Solution Approach 1:
Different parts of the image sensor (individual pixels or pixel groups) are assigned different gain qualities based on their local light conditions. Bright regions use one gain setting while dark regions use another, allowing each local area to have optimized quality without requiring the entire sensor to be complex.
Solution Approach 2:
Each pixel's circuitry automatically determines its own optimal gain setting based on the light intensity it detects during exposure. The pixels self-select between first and second gain values without requiring complex external control, thereby improving image quality while minimizing the added circuitry complexity.
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 enables the capture of high-dynamic-range images by optimizing the gain settings for each pixel, thereby retaining detail in both highlights and shadows, improving the overall quality of the captured image.
Implementation Method 1
Each image sensor pixel may include a photodiode for sensing the intensity of incoming light during an exposure
Data Source
AI summary
An image sensor may include an image pixel array. The image sensor may be provided with automatic conversion gain selection on a pixel-by-pixel basis to produce a high-dynamic-range image. Each image pixel may include a capacitor and a conversion gain transistor coupled in series between a power supply line and a floating diffusion node. The conversion gain transistor may be coupled to a control line through a gating transistor. The gating transistor may have a gate connected to a row select line. The image pixel may have an output line that is coupled to a column amplifier and a comparator. The column amplifier may generate a difference voltage based on reset and image signals. The comparator may compare the difference voltage with a predetermined threshold to determine whether to place the selected pixel in a high or low conversion gain mode.