Camera Automatic Gain Adjustment via Frequency Domain Illuminance Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing camera technologies face challenges in automatically adjusting gain settings for regions of interest in video content, particularly in scenes with varying illuminance, such as backlit or night scenes, often resulting in unsatisfactory image quality due to the need for manual user intervention and limited accuracy in exposure adjustment.
Innovation Solution
A method for automatic gain adjustment based on illuminance that involves setting preset gain values, performing frequency domain transformation on image data, calculating cumulative high frequency information, and selecting the gain value corresponding to the maximum high frequency information cumulative value for optimal exposure without manual user focus.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If illuminance averaging method is used to calculate gain according to average illuminance of the whole picture, then the operation is simple, but the exposure effect is unsatisfactory in scenes with varying illuminance
Solution Approach 1:
The patent divides the image into multiple regions and performs frequency domain transformation on each region separately. By segmenting the image processing into regional frequency analysis, the system achieves both automated operation and precise exposure control for different illuminance zones without manual intervention.
Solution Approach 2:
The patent transitions from spatial domain processing to frequency domain processing by performing frequency domain transformation on image blocks. This dimensional change allows the system to analyze illuminance characteristics more effectively and automatically determine optimal gain values for different regions, resolving the contradiction between operational simplicity and exposure precision.
2Manufacturing precision
If illuminance weighting method is used with manual region demarcation, then the exposure effect for regions of interest is improved, but the operation becomes complex requiring manual user participation
Solution Approach 1:
The patent implements self-service by automatically performing frequency domain transformation and gain calculation without requiring manual user intervention. The system autonomously identifies regions of interest through frequency analysis and automatically adjusts exposure parameters, eliminating the need for manual region demarcation while maintaining precise exposure control.
Solution Approach 2:
The patent replaces the manual mechanical operation of region demarcation with an automated frequency domain transformation process. By substituting manual user actions with algorithmic frequency analysis, the system achieves precise exposure control for regions of interest while eliminating operational complexity and manual participation requirements.
3Reliability
If illuminance histogram method is used with threshold determination, then overexposure or underexposure is prevented, but accurate exposure for regions of interest cannot be achieved
Solution Approach 1:
The patent segments the image into multiple blocks and performs independent frequency domain transformation on each block. This segmentation allows the system to determine optimal gain values for different regions individually, achieving both reliable exposure control and precise exposure for regions of interest by treating each region separately rather than applying a global threshold.
Solution Approach 2:
The patent transitions from histogram-based threshold determination in the spatial domain to frequency domain transformation. This dimensional change enables the system to capture illuminance characteristics more effectively and achieve both exposure reliability and precision for different regions by analyzing frequency components rather than relying on global histogram thresholds.
Data Source
AI summary
A method for automatic gain adjustment based on illuminance of video content and a camera are provided. The method includes arranging a first gain value for an image sensor according to a list of preset gain values; obtaining frequency domain data using frequency domain transformation on illuminance components of the image; setting a constant percentage value and summing up the percentage of the frequency domain data accumulated from the highest signal amplitude to obtain a first high frequency information cumulative value; repeating above steps according to a preset gain value arrangement order to obtain a number of high frequency information cumulative values; and using the gain value corresponding to a maximum high frequency information cumulative value in gain adjustment for the image sensor.