Image Fusing Method for Thermal and Visible Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image fusion methods in monitoring systems reduce image definition in relatively bright regions due to high luminance, particularly when combining thermal and visible images.
Innovation Solution
An image fusing method that determines the importance of each pixel based on luminance values, adjusts pixel coefficients, and applies these coefficients to generate a fused image by weighting the thermal and visible images, thereby improving image definition in bright regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If thermal image luminance values are increased to improve visibility in bright regions, then image definition in bright regions is improved, but image definition is reduced in certain portions of the relatively bright region due to high luminance
Solution Approach 1:
The patent applies different luminance adjustment strategies to different regions of the thermal image based on their importance. Important pixels (those corresponding to significant features in the visible image) receive different treatment compared to non-important pixels. Specifically, the luminance of important pixels is adjusted less aggressively or differently than non-important pixels, preventing over-saturation while maintaining definition in critical areas.
Solution Approach 2:
The patent dynamically adjusts the luminance parameter of thermal image pixels based on their importance classification. By changing the luminance parameter selectively - increasing it for some pixels while maintaining or decreasing it for others - the system optimizes image definition in bright regions without causing harmful over-luminance effects in certain portions.
2Illumination intensity
If pixel luminance values in thermal image are uniformly adjusted, then overall visibility is improved, but image definition is reduced in specific portions due to high luminance
Solution Approach 1:
The patent divides the thermal image into different quality zones based on pixel importance. Important pixels (those aligned with significant visible image features) are distinguished from non-important pixels. The luminance adjustment is then applied locally - with different adjustment magnitudes or directions - to maintain overall visibility while preserving definition in specific important portions.
Solution Approach 2:
Instead of uniformly adjusting all pixel luminance values, the patent applies partial adjustment only to non-important pixels or applies excessive adjustment (luminance increase) selectively to certain regions while maintaining or reducing luminance in important regions. This partial action approach improves overall visibility without compromising definition in critical areas.
Data Source
AI summary
Provided is an image fusing apparatus for fusing a thermal image and a visible image of a subject. The image fusing apparatus includes: a determination processor configured to determine importance of each pixel, based on a luminance value thereof, of the thermal image and the visible image; a pixel coefficient setting processor configured to set a pixel coefficient for each pixel of the thermal image based on the importance of each pixel of the thermal image and the visible image; a thermal image processor configured to generate another thermal image by applying the pixel coefficient to the luminance value of each pixel of the thermal image; and an image fusing processor configured to fuse the other thermal image and the visible image to generate a fused image.


