Pixelated Detector Super-Resolution Bias Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting particulate radiation using semiconductor-based pixelated detectors face limitations in achieving super-resolution beyond half a detector pixel size due to position-dependent bias, leading to artefacts and reduced signal-to-noise ratio near the Nyquist frequency.
Innovation Solution
Adjusting the contribution of each impact position to corresponding image pixels using a mathematical function or look-up table, and spreading the contribution of each particle over multiple pixels to counter position-dependent bias, thereby improving resolution and signal-to-noise ratio.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the extent of the centroid is larger than the detector pixel size, then super-resolution imaging is enabled, but position-dependent bias introduces artefacts and reduces signal-to-noise ratio near the Nyquist frequency
Solution Approach 1:
The patent applies parameter changes by modifying the weighting function parameters used in centroid calculation. Specifically, it adjusts the mathematical model parameters to account for the non-uniform response of detector pixels, thereby correcting position-dependent bias while maintaining super-resolution capability. This resolves the contradiction by changing the computational parameters rather than the physical detector configuration.
2Measurement precision
If the extent of the centroid is larger than the detector pixel size, then super-resolution imaging is enabled, but position-dependent bias introduces artefacts
Solution Approach 1:
The patent converts the harmful position-dependent bias into a beneficial correction factor. By characterizing the non-uniform response pattern and incorporating it into the weighting function, the system transforms the artefact-generating mechanism into a tool for improving measurement accuracy. The position-dependent bias information is repurposed to enhance rather than degrade image quality.
Solution Approach 2:
The patent implements feedback by using the detected position-dependent bias to adjust and refine the centroid estimation algorithm. The system continuously monitors for bias patterns and applies corrective weighting factors based on the estimated particle position, creating a closed-loop correction mechanism that eliminates artefacts while preserving super-resolution capability.
3Ease of operation
If standard centroid estimation is used, then detection is simple, but resolution is limited to half a detector pixel or multiple of whole detector pixel
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing optimal weighting factors in a lookup table before actual particle detection begins. These pre-computed weights account for the detector's non-uniform response characteristics, allowing the system to achieve high resolution simply by applying the pre-prepared weighting factors during centroid calculation, without requiring complex real-time computations.
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 allows for improved resolution and signal-to-noise ratio by accurately distributing the impact position information across multiple pixels, reducing artefacts and enhancing image quality without losing positional information.
Implementation Method 1
each particle of the particulate radiation causing a multitude of electron/hole pairs in the semiconductor material of the detector
Data Source
Figure 1
AI summary
When detecting particulate radiation, such as electrons, with a pixelated detector, a cloud of electron/hole pairs is formed in the detector. Using the signal caused by this cloud of electron/hole pairs a position of the impact is estimated. Inventors found that, when the size of the cloud is comparable to the pixel size, or much smaller, the estimated position shows a strong bias to the center of the pixel and the corners, as well to the middle of the borders. This hinders forming an image with super-resolution. By shifting the position or by attributing the electron to several sub-pixels this bias can be countered, resulting in a more truthful representation. Inventors further found that by spreading the image Moiré-effects and interferences in the image can be countered. As long as the image is a sparse image (almost all pixels representing one or no impact) this spreading is a reversible process. After spreading (effectively a spatial low-pass filtering) a high-pass filtering may be used to crisp the image. It is noted that shifting and/or spreading the information over several image pixels should take place before adding the events per image pixel. When said spreading is done after combining detector images, information is lost. It is noted that shifting cannot take place after combining detector images.