Image Stitching via Viewpoint Transformation for Parallax Elimination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image stitching technologies face challenges in aligning images captured by multiple cameras without physical rotation or movement, leading to parallax issues and ghost phenomena due to mismatched optical centers, which are difficult to resolve without using separate optical devices or causing partial distortion.
Innovation Solution
An image stitching method that performs viewpoint transformation using a depth map-based formula to match optical centers and remove parallax between images, allowing seamless stitching without physical camera movement or additional devices, by determining movement parameters based on focal length, baseline, and parallax to align images as if captured at the same location.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If multiple cameras are used to capture images for stitching, then the viewing angle and coverage are improved, but parallax occurs between images due to mismatched optical centers
Solution Approach 1:
The patent replaces the mechanical approach of physically rotating the camera or using separate optical devices with a computational approach. By using viewpoint transformation formulas based on depth maps, the system virtually adjusts the optical centers of multiple cameras to coincide, eliminating parallax through image processing rather than mechanical adjustment.
Solution Approach 2:
The patent changes the parameters of the images by applying viewpoint transformation formulas that adjust pixel coordinates based on depth information. This transforms the images from different viewpoints into a unified perspective, effectively changing the optical center parameters computationally to eliminate parallax effects.
2Manufacturing precision
If a separate optical device such as a mirror is used to match optical centers, then parallax is removed, but the device complexity and bulk are increased
Solution Approach 1:
The patent replaces complex mechanical optical devices with a computational method. Instead of using mirrors or other optical components to physically align optical centers, the system uses viewpoint transformation algorithms that computationally adjust image perspectives, achieving the same effect with simpler hardware.
Solution Approach 2:
The patent creates a virtual copy of the optical center alignment process through mathematical transformation. Rather than physically moving or adjusting the camera positions, the system generates transformed images that appear as if captured from a unified viewpoint, effectively copying the desired optical alignment result through computation.
3Productivity
If images are stitched without viewpoint transformation, then the processing speed is faster, but ghost phenomena and blurry joints occur due to parallax
Solution Approach 1:
The patent performs viewpoint transformation as a preliminary step before stitching images. By pre-adjusting the perspectives of individual images to a unified viewpoint using depth map-based transformation, the system eliminates parallax effects beforehand, ensuring that subsequent stitching operations produce accurate results without ghosting or blurry joints.
Data Source
AI summary
An image stitching method using viewpoint transformation and a system therefor are provided. The method includes obtaining images captured by a plurality of cameras included in the camera system, performing viewpoint transformation for each of the images using a depth map for the images, and stitching the images, the viewpoint transformation of which is performed.