Vertical Color Image Sensor Defect Classification and Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting and classifying defects in CMOS color image sensors, such as pixel clusters, chip-surface defects, and window defects, lack effective classification capabilities, leading to high manufacturing costs and yield loss due to the inability to distinguish between different types of defects and correct them efficiently.
Innovation Solution
A method that utilizes all color channel information to classify and correct defects in vertical color pixel image sensors by characterizing each defective pixel, distinguishing between pixel clusters, chip-surface defects, and window defects through multiple characteristics, and implementing a two-pass methodology for defect detection and correction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional defect testing methods are used that only measure size and quantity of defects, then the testing process is simple and fast, but the classification capability is minimal and yield enhancement is limited
Solution Approach 1:
The patent segments the defect analysis process into two distinct passes: a first pass that performs comprehensive multi-characteristic analysis for classification, and a second pass that performs faster verification. This segmentation allows the system to achieve high classification precision while managing complexity by separating detailed analysis from routine verification.
Solution Approach 2:
The patent performs preliminary defect classification in the first pass using multiple characteristics (color channel information, response levels, image type performance) before final verification. This preliminary action enables the system to identify and categorize defects early, allowing for targeted correction strategies and reducing the need for complex repeated analysis.
2Measurement precision
If comprehensive multi-characteristic analysis is performed to accurately classify defects, then classification accuracy improves, but calibration time increases
Solution Approach 1:
The patent implements periodic action through its two-pass methodology: the first pass performs comprehensive multi-characteristic analysis at specific intervals (during initial calibration) to establish accurate defect classification, while the second pass performs faster verification for routine checks. This periodic comprehensive analysis combined with frequent rapid verification achieves high accuracy while managing calibration time efficiently.
Solution Approach 2:
The patent changes analysis parameters between passes: the first pass uses full multi-characteristic analysis (color channels, response levels, image type performance) for comprehensive classification, while the second pass uses simplified verification parameters. This parameter change allows accurate classification when needed while reducing calibration time for routine operations.
3Measurement precision
If all color channel information is utilized for defect classification in vertical color sensors, then classification accuracy improves, but processing complexity increases
Solution Approach 1:
The patent extracts and utilizes all color channel information (red, green, blue channels) specifically for defect classification purposes in vertical color sensors. By taking out and analyzing the color channel characteristics separately (including stuck channels, offset defects, gain defects), the system achieves high classification accuracy while managing processing complexity through targeted extraction of relevant information rather than processing all sensor data uniformly.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method of classifying and correcting defects in vertical color pixel sensors in utilizing defective pixel information, the method includes defining ranges of output levels of normal pixel sensors from exposure to dark and bright flat field light sources. Dark and bright images are captured, and pixel outputs are measured for each image. Pixels in the dark and bright images having outputs in at least one color channel that are outside of the ranges of the output levels of normal pixel sensors are entered into defective pixel maps. The defective pixels are classified into categories and defective pixel sensors that can be corrected are identified. Correction values for identified defective pixel sensors that can be corrected are generated and entered into a calibration memory associated with the vertical color image sensor.