Multi-Camera Image Correction Through Projection Axis Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing manual deployment of cameras for multi-angle and multi-camera shooting in free perspective video results in image jittering due to vertical and horizontal parallax and image scaling, affecting the viewing experience.
Innovation Solution
An image correction method that uses correction parameters to align projection axes of target images from multiple shooting devices on a virtual plane, ensuring the virtual center corresponds to physical optical centers, thereby eliminating image jittering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If manual deployment of cameras is used for multi-angle shooting, then the shooting flexibility is improved, but image jittering occurs due to vertical and horizontal parallax
Solution Approach 1:
The patent applies parameter changes by introducing correction parameters that adjust the projection axes of images from different cameras. These parameters modify the geometric transformation of images to eliminate parallax effects, resolving the contradiction between shooting flexibility and image stability by mathematically correcting the positional deviations caused by manual camera deployment.
Solution Approach 2:
The patent replaces the mechanical precision requirement of camera deployment with a computational correction system. Instead of relying on precise physical alignment of cameras, the system uses image processing and correction parameters to achieve alignment, substituting mechanical precision with algorithmic correction.
2Device complexity
If manual deployment of cameras is used for multi-angle shooting, then the device complexity is reduced, but manufacturing precision is insufficient leading to image jittering
Solution Approach 1:
The patent replaces complex mechanical alignment requirements with a computational correction system. The deployment process remains simple and manual, but the resulting images are corrected using calculated parameters that compensate for the imprecision in manual camera placement, effectively substituting mechanical precision requirements with software-based correction.
Solution Approach 2:
The patent performs preliminary calculation of correction parameters based on the known camera positions and projection geometry. By pre-calculating the necessary correction values before image processing, the system enables simple manual deployment while achieving precise alignment through subsequent application of these pre-computed correction parameters.
3Reliability
If correction parameters are used to align projection axes, then image jittering is eliminated, but processing complexity increases
Solution Approach 1:
The patent replaces complex iterative image alignment processes with a direct computational correction approach. By calculating correction parameters based on geometric relationships between camera positions and projection axes, the system achieves image stabilization through a single transformation step rather than complex multi-step alignment procedures.
Data Source
AI summary
Disclosed in embodiments of the present disclosure are an image correction method and apparatus, an electronic device, and a storage medium. The method may include: for each of a plurality of photographing devices, obtaining a target image photographed by the photographing device; and correcting the target image on the basis of a correction parameter corresponding to the photographing device, wherein when at least two axis points on a rotating axis including a virtual center are projected onto the target image to obtain a projection axis, a projection axis on each corrected target image is consistent, and the virtual center is located on virtual planes corresponding to the plurality of photographing devices and corresponds to physical optical centers of the plurality of photographing devices.