Dual-Camera Image Fusion Using Reference Coordinate Layers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Images captured by dual-camera devices, such as mobile phones, suffer from poor stereoscopic effects and inconsistent definition due to mismatched fields of view between cameras, leading to issues like misaligned and low-quality fused images.
Innovation Solution
An image processing method that adds a reference coordinate image layer to one of the images with different fields of view, processed using a deep learning network model to enhance image quality by retaining details and natural fusion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If images from dual cameras with different fields of view are simply fused, then image information can be complemented, but the fused image has poor stereoscopic effect and poor quality due to field of view mismatch
Solution Approach 1:
The patent segments the image processing into distinct stages: first processing each camera's image separately (first image and second image) to generate intermediate images, then fusing these intermediate images. This segmentation allows each camera's unique field of view characteristics to be optimally processed before combination, resolving the contradiction by maintaining information quantity while improving fusion quality.
Solution Approach 2:
The patent applies preliminary processing to each image before fusion - the first image undergoes first processing and the second image undergoes second processing. This preliminary action prepares each image appropriately for its specific field of view characteristics, ensuring that when they are fused, the stereoscopic effect and overall quality are improved rather than degraded by direct mismatched fusion.
2Ease of manufacture
If images with different fields of view are directly fused, then processing is simple, but the fused image has inconsistent definition between central and surrounding parts
Solution Approach 1:
The patent applies different processing operations to different images based on their specific characteristics. The first image receives first processing while the second image receives second processing, allowing each to be optimized for its field of view. This local quality approach ensures consistent definition across the fused image by treating each input appropriately rather than applying uniform simple fusion.
3Manufacturing precision
If multiple processing steps are applied sequentially to fuse images, then image quality can be improved, but errors accumulate from serial processing
Solution Approach 1:
The patent merges the processing of multiple images into a coordinated multi-input deep learning model. Instead of sequentially processing images one after another (which would accumulate errors), the model processes the first image, second image, and reference image simultaneously in a unified framework. This combining approach maintains high image quality while preventing error accumulation through parallel rather than serial processing.
Data Source
AI summary
This application provides an image processing method and an electronic device. The image processing method includes: acquiring a plurality of frames of original images; adding a reference coordinate image layer to a second field-of-view image; obtaining an image layer set according to a first field-of-view image, the second field-of-view image, and the reference coordinate image layer; processing, by using a deep learning network model, the image layer set to obtain a first enhanced image; and obtaining a second enhanced image according to the first enhanced image. In the method, since the added reference coordinate image layer reflects a mapping relationship between a field of view corresponding to the first field-of-view image and a field of view corresponding to the second field-of-view image, through the addition of the reference coordinate image layer, priori information can be added, so that different adjustments can be made subsequently according to different field-of-view relationships.