Imaging Fourier Correction for Wafer-Induced Regular Image Defects
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Imaging devices suffer from regular image defects due to manufacturing defects in semiconductor wafers, which affect image quality and are not effectively addressed by existing technologies.
Innovation Solution
An imaging device employing Fourier transform methods to generate corrected images by identifying and correcting regular defects using correction parameters, including angles and amplitudes, through processes involving Fourier transform and inverse Fourier transform operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If Fourier transform method is used to correct regular image defects, then image quality is improved, but device complexity increases
Solution Approach 1:
The patent pre-calculates correction parameters (dark Fourier pattern, correction angle, correction amplitude) during device manufacturing or initialization. These parameters are stored and reused during actual image processing, avoiding repeated complex calculations and reducing real-time processing complexity while maintaining correction effectiveness
Solution Approach 2:
The patent transforms the image correction problem from spatial domain to frequency domain using Fourier transform. By operating on frequency domain parameters (amplitude and angle of dark Fourier pattern), the system simplifies the correction process and enables efficient parameter-based correction of regular defects
2Measurement precision
If correction parameters are calculated from dark images, then correction accuracy is improved, but processing time increases
Solution Approach 1:
The patent calculates correction parameters using dark images captured during manufacturing or initialization phases. These parameters are stored and applied to multiple subsequent images without recalculation, achieving accurate correction while minimizing real-time processing time
Solution Approach 2:
The patent creates a frequency domain representation (dark Fourier pattern) from dark images that serves as a template for correcting regular defects. This copied frequency pattern is reused across multiple images, avoiding repeated full correction calculations
Data Source
AI summary
An imaging device and an image processing method and an image testing method for the same are disclosed. The image processing method includes generating a first transformed image by performing Fourier transform on a raw image; generating a second transformed image obtained by correcting the first transformed image using a correction parameter; and generating a corrected image by performing inverse Fourier transform on the second transformed image.