High Dynamic Range Image Generation Using Brightness Index Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Standard image sensors struggle to capture the full dynamic range of a scene, especially in high-contrast situations, as they can only discriminate a finite number of illumination levels, leading to incomplete image representation when trying to capture both bright and dark areas with a single exposure level.
Innovation Solution
A method is developed to generate a high dynamic range image by determining a brightness index and using specific calculation formulas based on threshold values to combine two images taken at different exposure levels, allowing for more accurate representation of pixel values across varying illumination levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If a standard image sensor uses a single exposure level to capture an image, then the acquisition process is simple and fast, but it cannot capture the full dynamic range of high-contrast scenes
Solution Approach 1:
The image acquisition process is segmented into multiple exposures of the same scene at different exposure levels. Each exposure captures a specific portion of the dynamic range, and these segmented captures are later merged to form a complete high dynamic range image, resolving the contradiction between simple acquisition and comprehensive dynamic range capture.
Solution Approach 2:
The solution transitions from a single-dimension (single exposure level) acquisition approach to a multi-dimensional approach by adding the exposure level dimension. Multiple images are captured at different exposure levels (dimensions), and this additional dimension enables capturing the full dynamic range that cannot be achieved with a single exposure level.
2Loss of information
If multiple images at different exposure levels are merged to create a high dynamic range image, then the dynamic range capture is improved, but visible artifacts and noise appear in the merged image
Solution Approach 1:
The merging process applies local quality by selecting pixel values from different exposure levels based on local brightness characteristics. For each pixel position, the algorithm determines which exposure level provides the optimal signal-to-noise ratio, and selectively uses that value. This local optimization minimizes artifacts and noise while maximizing dynamic range representation.
Solution Approach 2:
The invention changes the parameter selection strategy by introducing a brightness index and multiple thresholds. Instead of uniformly merging images, the algorithm dynamically changes which exposure level's pixel values are used based on the local brightness parameter. This parameter-driven selection optimizes the signal-to-noise ratio across different regions of the image, reducing visible artifacts.
3Productivity
If simple threshold-based merging is used to combine images, then the processing is computationally efficient, but the signal-to-noise ratio curve shows abrupt variations and artifacts
Solution Approach 1:
The merging strategy is made dynamic by introducing a brightness index calculation and multi-threshold comparison system. Instead of using fixed simple thresholds, the algorithm dynamically determines which exposure level to use for each pixel based on local brightness characteristics. This dynamic adaptation smooths the signal-to-noise ratio curve while maintaining processing efficiency.
Solution Approach 2:
The algorithm incorporates feedback by calculating a brightness index for each pixel and using this index to determine the optimal exposure level selection. The brightness index serves as feedback that guides the merging process, ensuring that pixel values are selected from the exposure level that provides the best signal-to-noise ratio, thereby reducing artifacts while maintaining computational efficiency.
Data Source
AI summary
A method of generation, by a digital processing device, of a first high dynamic range digital image from second and third digital images of a same scene, including, for at least one point of the first image: determining a brightness index; comparing this index with at least one of first, second, third, and fourth decreasing thresholds stored in a memory; and determining the value of the point by taking into account the value of the index.