Scaled Two-Band Histogram Equalization for Image Contrast
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional histogram equalization methods fail to preserve brightness and often result in saturated images, lacking effective image enhancement for improved contrast without losing image information.
Innovation Solution
The method involves constructing an input histogram, performing histogram equalization on two bands, and scaling, with the division value at half the pixel count, to generate an enhanced equalization curve applied to each image in a series, allowing for real-time image enhancement while maintaining constant parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If traditional histogram equalization is performed to increase global contrast, then image contrast is improved, but brightness is not preserved and images become saturated
Solution Approach 1:
The histogram is divided into two bands: a first band from the darkest value to a division value (where half the pixels are located), and a second band from the division value to the lightest value. Histogram equalization is performed separately on each band, allowing different contrast enhancement strategies for different tonal ranges while preserving overall brightness distribution.
Solution Approach 2:
Different processing approaches are applied to different regions of the histogram. The first band undergoes histogram equalization to enhance dark region contrast, while the second band is scaled to maintain brightness preservation. This local differentiation resolves the contradiction between contrast enhancement and brightness preservation.
2Illumination intensity
If traditional histogram equalization is applied to enhance image contrast, then global contrast is improved, but the image appears saturated
Solution Approach 1:
By segmenting the histogram into two bands and applying different processing to each, the patent avoids the uniform contrast enhancement that causes saturation. The first band's equalization enhances contrast in dark regions without affecting bright regions, while the second band's scaling prevents over-enhancement and saturation.
Solution Approach 2:
The patent changes the processing parameter applied to different histogram regions: histogram equalization is applied to the first band while scaling is applied to the second band. This parameter differentiation allows contrast enhancement where needed while preventing saturation in other regions.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method of image enhancement includes constructing an input histogram (112) corresponding to an input image (130) received at a focal plane array. The method includes performing histogram equalization on a first band (116) of the input histogram (112) starting from a zero value and ending at a division value (114) representing a pixel bin value where a predetermined fraction of the input histogram (112) by pixel hound is reached to produce a first portion (120) of an equalization curve (123). The method includes performing histogram equalization on a second band (118) of the input histogram (112) starting from the division value (114) and ending at a pixel bin value where all of the input histogram (112) by pixel count is reached to produce a second portion (122) of the equalization curve (124). The method includes applying the equalization curve (124) to the input image (130) to produce a corresponding enhanced image.