3D Surveillance Video Mesh Mapping Without Camera Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing 3D monitoring systems require cumbersome calibration of intrinsic and extrinsic camera parameters, making deployment inconvenient and inflexible.
Innovation Solution
A rapid deployment method involving manual selection of corresponding point sets in virtual and monitoring videos to establish a mapping, generating a video mesh for superimposition without camera parameter calibration, and allowing for accurate access right control over regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If camera calibration is performed to obtain intrinsic and extrinsic parameters for 3D monitoring system deployment, then the accuracy of video projection in virtual scene is improved, but the deployment complexity and time consumption increase significantly
Solution Approach 1:
The patent uses a video mesh as a simplified copy or representation of the camera's field of view, mapping video pixels directly to virtual scene coordinates without requiring full camera calibration. This copying approach maintains sufficient accuracy for monitoring purposes while dramatically simplifying the deployment process.
Solution Approach 2:
The patent changes the parameter representation from complex camera intrinsic/extrinsic parameters to a simpler video mesh parameter system. By transforming the problem from camera calibration to mesh mapping, the system achieves acceptable projection accuracy with fewer parameters and less computational complexity.
2Loss of time
If manual point selection is used to establish mapping between virtual scene and monitoring video, then the deployment time is reduced, but the mapping precision may be affected
Solution Approach 1:
The patent uses only N key feature points (where N is small, typically 4 or more) to establish the video mesh mapping, rather than requiring comprehensive calibration of all camera parameters. This partial action approach achieves sufficient mapping precision for monitoring applications while significantly reducing deployment time.
Solution Approach 2:
The patent transforms the mapping problem from requiring complete camera parameter sets to using a minimal set of correspondence points between virtual and real spaces. This parameter reduction enables rapid deployment while maintaining adequate precision through the video mesh interpolation method.
3Measurement precision
If one-to-one correspondence between camera device and projection is maintained, then the accuracy of 3D reconstruction is improved, but the system flexibility and adaptability decrease
Solution Approach 1:
The video mesh structure serves multiple functions: it maintains the necessary one-to-one correspondence for accurate projection while also enabling flexible adaptation to different camera positions, angles, and monitoring scenarios. The mesh can be regenerated for different cameras without requiring complete system redesign.
Solution Approach 2:
The patent introduces dynamic adaptability by allowing the video mesh parameters to be adjusted and regenerated based on different monitoring requirements and camera configurations. This dynamic approach maintains reconstruction accuracy while providing flexibility for various deployment scenarios.
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
The present application relates to a video surveillance technology, and discloses a 3D monitoring system and a rapid deployment method for the same. In the present application, two corresponding point sets are obtained by means of an input device respectively in a virtual scene and in a monitoring video, so as to obtain a mapping between the virtual scene and the monitoring video; and a video mesh, to be superimposed and displayed on the virtual scene, is generated based on this mapping. Thus, a 3D monitoring system can be rapidly deployed without calibrations of intrinsic and extrinsic parameters of a camera, and requirements on accuracy in matching between a 3D monitoring scene and an actual monitored site are decreased, greatly reducing the difficulty in system deployment while ensuring the effect of superimposing a monitoring video on a virtual scene.