3D Modeling Movable Objects via Video Interest Points
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for visualizing 3D models are limited to mapping terrain and cannot display simulation results on movable objects or objects not tied to the terrain, nor can they implement additional objects or simulate their interactions.
Innovation Solution
A method that uses a video camera to obtain images, identify interest points, determine their coordinates, and calculate spatial coordinates, allowing for the creation of 3D models of movable objects and their interactions by extracting and controlling additional object models within the 3D environment, enabling real-time image signal formation and display.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If digital elevation maps and satellite imagery are used to visualize 3D models, then terrain mapping capability is improved, but the method cannot be applied to movable objects or objects not tied to terrain
Solution Approach 1:
The patent applies universality by creating a 3D modeling system that works with multiple types of objects (terrain, movable objects, internal spaces) using the same video camera-based methodology. The system extracts geometric features from video frames and constructs 3D models without requiring specialized data sources for each object type, making the solution universally applicable across different scenarios.
Solution Approach 2:
The patent extracts the essential geometric information directly from video camera frames by identifying interest points and tracking their movement. This extraction approach eliminates the dependency on external data sources like digital elevation maps, allowing the system to work with any object that can be captured by a video camera, including movable objects and internal spaces.
2Adaptability or versatility
If traditional mapping methods are used, then terrain visualization is achieved, but additional objects and their interactions cannot be implemented
Solution Approach 1:
The patent applies dynamics by enabling the 3D model to accommodate moving objects and changing scenarios. The system continuously tracks interest points across video frames and updates the 3D model in real-time, allowing objects to move and interact within the modeled space. This dynamic approach contrasts with static terrain mapping methods.
Solution Approach 2:
The patent transitions from 2D video frames to 3D spatial representation by tracking the movement of interest points through multiple frames. This dimensional transformation enables the system to reconstruct three-dimensional models and simulate object interactions in 3D space, going beyond flat terrain mapping.
3Device complexity
If video camera frames are analyzed to create 3D models, then additional data requirements are reduced, but interest point tracking accuracy must be maintained
Solution Approach 1:
The patent applies feedback by continuously monitoring the positions of interest points across multiple video frames and using this information to refine the 3D model construction. The system tracks the movement of interest points and adjusts the spatial coordinates based on accumulated observational data, improving accuracy through iterative refinement.
Solution Approach 2:
The patent applies preliminary action by pre-identifying interest points in video frames before constructing the 3D model. The system selects and marks key features in advance, then uses their tracked positions across multiple frames to determine spatial coordinates. This preliminary identification ensures consistent tracking and improves measurement accuracy.
Data Source
AI summary
The invention relates to the field of information technology. The present invention involves obtaining an image, said image being formed using a video camera, and forming a model, including identifying points of interest on frames of an image and determining the coordinates thereof, shifting the video camera, and receiving data regarding the parameters of the shift; computing forecastable spatial coordinates for sectors in which actual spatial objects may be found, which are displayed on the image using corresponding points of interest; computing the coordinates of an area of a frame for each indicated sector in which a corresponding point of interest is expected to be found, performing a search for the point of interest within the bounds of the indicated area of the frame, and, if the point of interest is found, storing to memory the spatial coordinates of the corresponding sector; correlating images of points of interest to the sectors corresponding thereto on a spatial model, identifying flat and/or curved surfaces on the spatial model, extracting from memory a model of an additional object, and forming signals for controlling same; computing results for the interaction between the model of the additional object and the parameters of the spatial model, and forming a signal for the image of the additional object superimposed on the image formed by the video camera and outputting same to a display.