3D Model Generation from Object and Structure Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The visual volume intersection method for generating three-dimensional models of objects is prone to defects when structures within the image capturing scene obstruct parts of the target object, leading to inaccurate silhouettes and reduced geometric accuracy of the three-dimensional model.
Innovation Solution
A device that acquires area information for both the object and potential obstructing structures, and generates three-dimensional shape data by integrating this information to create a high-quality three-dimensional model, even when parts of the object are hidden by structures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the visual volume intersection method is used to generate three-dimensional models, then the three-dimensional shape can be estimated from multi-viewpoint images, but defects occur in the generated model when structures obstruct parts of the target object
Solution Approach 1:
The patent segments the image capturing scene into multiple regions: object regions (containing the target object), structure regions (containing obstructing structures), and unknown regions. By dividing the scene and processing each region separately with appropriate methods, the system can handle obscured areas differently from visible areas, thereby maintaining model quality even when structures obstruct the target object
Solution Approach 2:
The patent introduces an intermediary classification process that identifies and separates structure regions from object regions before three-dimensional model generation. This intermediary step allows the system to use different processing strategies for different regions, preventing structure-induced defects while preserving geometric accuracy in visible areas
2Reliability
If mask images with lost silhouette portions are discarded to avoid defects, then model quality is maintained, but geometric accuracy is reduced
Solution Approach 1:
The patent applies different quality requirements and processing methods to different regions of the image. In object regions where the target is visible, high geometric accuracy is maintained through precise silhouette extraction. In structure regions where obstruction occurs, the system applies appropriate handling strategies rather than discarding the entire mask image, thus maintaining overall model quality while preserving geometric accuracy where possible
Solution Approach 2:
By segmenting mask images into object regions and structure regions, the system can retain and process object region data for high geometric accuracy while separately handling structure region data to prevent defects, thereby achieving both model quality and geometric accuracy simultaneously
3Productivity
If the silhouette on the mask image is incorrect due to obstruction, then the three-dimensional shape generation process can proceed, but the generated model contains defects
Solution Approach 1:
The patent performs preliminary classification of image regions into object regions, structure regions, and unknown regions before the three-dimensional model generation process. This preliminary action identifies problematic areas in advance, allowing the system to apply appropriate handling strategies during model generation and prevent defects rather than correcting them afterward, thus maintaining both efficiency and quality
Data Source
AI summary
A generation device of a three-dimensional model including: an acquisition unit configured to acquire a first mask image indicating a structure area, which is an object still within each image captured from a plurality of viewpoints, and a second mask image indicating a foreground area, which is a moving object within each image captured from the plurality of viewpoints; a combination unit configured to generate a third mask image that integrates the structure area and the foreground area within the image captured from the plurality of viewpoints by combining the first mask image and the second mask image both acquired; and a generation unit configured to generate a three-dimensional model including the structure and the foreground by a visual volume intersection method using the third mask image.