Image Fusion Using Background Reference for Low-Light Color Accuracy

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current image fusion methods fail to ensure high-quality fusion of color and infrared images in low-illumination scenes, resulting in loss of details and poor imaging quality.

Innovation Solution

An image fusion method that includes obtaining color and infrared images shot for the same scene, using a trained model to extract features, and incorporating a background reference image to enhance both foreground and background quality, with adjustable fusion weights for different application scenarios.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If color image and infrared image are fused to improve imaging quality, then texture information and natural color are enhanced, but background blur problem occurs due to flashlight effect of infrared image

Engineering Contradiction:
Improveimage detail lossVSAvoidbackground blur
Core Design Contradiction:
Loss of informationVSObject-affected harmful factors

Solution Approach 1:

A background reference image is introduced as an intermediary component to resolve the background blur issue. The background reference image serves as a mediator that provides accurate background information to counteract the flashlight effect of the infrared image, while allowing the foreground to retain the enhanced texture and color from the fusion of color and infrared images.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The fusion process applies different quality characteristics to different regions of the image. The foreground area benefits from the texture and detail enhancement of infrared fusion, while the background area uses the background reference image to maintain clarity and avoid blur, achieving local optimization of image quality.

Inventive Principle:
Principle #3Local quality

2Manufacturing precision

If infrared fill-in light is used to improve imaging quality in low-illumination scenes, then resolution and contrast are improved, but real color reproduction is lost

Engineering Contradiction:
Improveimaging qualityVSAvoidcolor information
Core Design Contradiction:
Manufacturing precisionVSLoss of information

Solution Approach 1:

The patent merges the advantages of color images (real color reproduction) and infrared images (enhanced resolution and contrast in low-illumination) by fusing them together. The color image preserves accurate color information while the infrared image contributes improved texture and detail, creating a fused image that has both natural color and high resolution.

Inventive Principle:
Principle #5Merging (Combining)

3Reliability

If current fusion methods are used to combine color and infrared images, then some imaging quality improvement is achieved, but a large quantity of details are lost in output image

Engineering Contradiction:
Improvefusion effectVSAvoidimage details
Core Design Contradiction:
ReliabilityVSLoss of information

Solution Approach 1:

The background reference image is prepared in advance through specific processing (such as obtaining from historical data or through dedicated capture) to ensure it contains accurate background information before the fusion process. This preliminary preparation ensures that when fusion is performed, the background details are preserved and not lost, while foreground details benefit from infrared enhancement.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS12573002B2Image fusion method and apparatus and training method and apparatus for image fusion model
Publication Date: 2026.03.10 HUAWEI TECH CO LTD
  • US12573002B2 patent drawing
  • US12573002B2 patent drawing
  • US12573002B2 patent drawing

AI summary

An image fusion method and apparatus and a training method and apparatus for an image fusion model are provided, which relate to the field of artificial intelligence, and specifically, to the field of computer vision. The image fusion method includes: obtaining a to-be-processed color image, an infrared image, and a background reference image, where the infrared image and the to-be-processed color image are shot for a same scene; and inputting the to-be-processed color image, the infrared image, and the background reference image into an image fusion model for feature extraction, and performing image fusion based on extracted features to obtain a fused image. This method can improve image quality of a fused image, and also ensure accurate and natural color of the fused image.