CMOS Image Sensor Black Level Adjustment via Dithering Mask
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional CMOS image sensors face significant color hue shifting and brightness fluctuations due to numerical errors in black level control processing, particularly when high channel gains are applied, as the finite accuracy of digital circuits leads to errors in calculating average pixel values for dark and active rows.
Innovation Solution
A black level adjustment method and device for CMOS image sensors that computes an average pixel value from dark rows as an integer and fraction, generates a dithering mask based on the fraction, applies this mask to active rows, and adds a calculated offset value to the dithered pixels to reduce numerical errors and maintain color consistency across channels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a low-cost divider is used to compute the average value of dark rows, then the device complexity is reduced, but numerical errors increase causing color/brightness flashing
Solution Approach 1:
The patent introduces an intermediary dithering mask that operates between the average value computation and the final pixel adjustment. The mask, generated from the fractional part of the average value, serves as a mediator to distribute quantization errors across multiple pixels, thereby reducing numerical errors without requiring a more complex divider circuit.
2Reliability
If high channel gains are applied in the ISP, then the sensitivity and dynamic range are improved, but numerical errors in color channels are amplified causing significant color hue shifting
Solution Approach 1:
The patent applies preliminary action by computing the dithering mask from the fractional part of the average dark row value before the BLC offset is applied to active rows. This preliminary computation of the mask ensures that when high channel gains are later applied in the ISP, the numerical errors have already been distributed and minimized, preventing color hue shifting while maintaining the benefits of high gain processing.
Data Source
AI summary
A black level adjustment method for a CMOS image sensor is provided. The CMOS image sensor has a pixel array with dark rows and active rows. The method has the following steps of: computing an average value of pixels from the dark rows, wherein the average value is in the form of an integer and a fraction; calculating a black level control (BLC) offset value according to the integer; generating a dithering mask based on the fraction; applying the dithering mask to pixels from the active rows; and adding the calculated BLC offset value to the dithered pixels from the active rows to generate resulting pixels.