Abnormal Pixel Detection Using Shifted Reference Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image forming apparatuses face issues with detecting abnormal pixels due to foreign substances on the optical path, leading to streaks and deteriorated image quality, as conventional shading correction methods may not uniformly correct pixel values.
Innovation Solution
An abnormal-pixel detecting device and method that utilize an image sensor and processor to calculate the ratio between pixel values shifted in the main scanning direction, allowing for accurate detection of abnormal pixels by comparing uniform ADF background plate data with reference member data, thereby minimizing erroneous corrections and enhancing detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional shading correction is performed using uniform reference data, then correction simplicity is maintained, but detection accuracy of abnormal pixels deteriorates due to foreign substances on optical path
Solution Approach 1:
The patent divides the reference data into multiple segments with different reference positions in the main scanning direction. Instead of using a single uniform reference, the system creates multiple reference data sets (first reference data, second reference data, etc.) each with shifted reference positions, allowing comparison across different segments to detect abnormal pixels caused by foreign substances
Solution Approach 2:
The patent introduces a new dimension of comparison by shifting reference positions in the main scanning direction. This creates multiple reference frames at different positional offsets, enabling the system to detect abnormalities that would be invisible in a single reference frame, thereby improving detection accuracy without requiring complex additional hardware
2Measurement precision
If reference position is shifted to detect abnormal pixels, then detection accuracy improves, but calculation complexity increases
Solution Approach 1:
The patent performs preliminary calculations by pre-comparing pixel values with reference data at multiple shifted reference positions before final abnormal pixel detection. This preliminary action creates intermediate comparison results that simplify the final detection step, as the system already has pre-processed data showing variations at different reference positions
Solution Approach 2:
The patent uses multiple reference positions (excessive action) to ensure comprehensive coverage of potential abnormal pixel locations. By comparing against more reference positions than the minimum single reference, the system guarantees detection of abnormal pixels regardless of their position, with the trade-off being manageable calculation complexity through systematic comparison
3Reliability
If foreign substance detection is performed on white reference plate, then image quality is improved, but detection reliability deteriorates when foreign substance is on optical path
Solution Approach 1:
The patent introduces an intermediary comparison mechanism where pixel values are compared against multiple reference data sets with different reference positions. This intermediary comparison process acts as a mediator that can identify whether variations are due to foreign substances on the optical path or legitimate image features, thereby maintaining detection reliability even when foreign substances are present
Solution Approach 2:
The patent segments the reference data into multiple portions with different reference positions in the main scanning direction. By dividing the reference into multiple segments rather than using a single unified reference, the system can identify which segment is affected by foreign substances and exclude it from final correction, maintaining overall detection reliability
Data Source
AI summary
An abnormal-pixel detecting device includes an image sensor and an image processor. The image sensor is configured to capture an image of a subject. The image processor is configured to calculate: a ratio between a first plurality of pixel values captured by the image sensor and a second plurality of pixel values whose reference position is shifted relative to the first plurality of pixel values in a main scanning direction to obtain a third plurality of pixel values; and detect an abnormal pixel in the third plurality of pixel values.