Scene Space Model Construction via Mobile Point Cloud and Depth Camera
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing indoor 3D reconstruction technologies face challenges in accurately constructing scene space models due to cumulative errors from splicing point clouds collected at different points, leading to serious distortions.
Innovation Solution
A method that integrates mobile point cloud collection equipment and fixed-point depth camera equipment to acquire and process point cloud and depth image information. This involves determining a rotation matrix to align camera coordinates with a global coordinate system, generating panoramic images, and performing surface reconstruction to create a scene space model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If point clouds are spliced at different fixed points using existing indoor 3D reconstruction technology, then a 3D model of the indoor scene can be generated, but cumulative errors occur leading to serious distortions in the scene space model
Solution Approach 1:
The patent introduces a mobile depth camera as an intermediary device that moves continuously through the scene to collect point cloud data. This mobile collection approach acts as a mediator between multiple fixed-point measurements, avoiding the cumulative errors that occur when splicing data from different fixed positions. The mobile camera creates a continuous data stream that can be seamlessly integrated without the discontinuities and errors inherent in fixed-point splicing methods.
Solution Approach 2:
The patent transitions from static fixed-point measurement to dynamic mobile measurement. The depth camera is moved to different positions and orientations dynamically during the scanning process, allowing continuous data collection throughout the scene. This dynamic approach enables the system to capture comprehensive spatial information while maintaining coordinate consistency through continuous motion tracking, thereby eliminating the cumulative errors associated with static splicing operations.
2Productivity
If mobile point cloud collection equipment is used to collect data throughout the scene, then shooting efficiency is improved and equipment requirements are reduced, but coordinate system alignment between different collection points becomes complex
Solution Approach 1:
The patent establishes a universal global coordinate system that serves as a common reference frame for all point cloud data collected at different positions and orientations. This universal coordinate system allows the mobile depth camera to collect data throughout the scene without requiring complex real-time coordinate transformations. All collected point clouds are automatically registered to this global coordinate system, simplifying the alignment process while maintaining high shooting efficiency.
Solution Approach 2:
The patent performs preliminary establishment of the global coordinate system and pre-calculates transformation parameters before actual data collection begins. By setting up the reference coordinate framework in advance and preparing the registration algorithms beforehand, the system eliminates the need for complex real-time coordinate alignment during mobile data collection. This preliminary preparation enables efficient continuous scanning while maintaining accurate spatial relationships.
Data Source
AI summary
According to the present disclosure, a method, device for constructing a scene space model and a storage medium are provided. The method includes: acquiring first point cloud information collected by mobile point cloud collection equipment; acquiring depth image information collected by fixed-point depth camera equipment; determining a rotation matrix of a camera coordinate system of the fixed-point depth camera equipment corresponding to a global coordinate system; generating a first panoramic image based on the depth image information, mapping the first panoramic image onto a three-dimensional unit sphere; rotating the three-dimensional unit sphere based on the rotation matrix to generate a second panoramic image; and generating a scene space model based on the first point cloud information and the second panoramic image.