3D Bird's-Eye Image Layering for Natural Object Placement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing three-dimensional bird's eye view image generation systems struggle to naturally incorporate new objects into the generated image.
Innovation Solution
An image processing device that includes a position information acquiring part to determine the placement of an object in a three-dimensional space, a conversion part to project this position onto a three-dimensional projection surface, and an output part to superimpose the object onto a generated bird's eye view image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple photographed images are projected onto a three-dimensional projection surface to generate a bird's eye view image, then the overall surroundings can be displayed, but new objects cannot be displayed naturally in the image
Solution Approach 1:
The patent introduces a two-layer structure: a background layer containing the projected photographed images and a foreground layer containing the new object to be displayed. By separating the object placement into a distinct foreground layer with different depth information, the system can naturally integrate new objects into the three-dimensional bird's eye view image while maintaining the accuracy of the original projected images.
2Ease of operation
If objects are simply overlaid on the three-dimensional projection surface, then object placement is simple, but the objects do not appear naturally integrated into the scene
Solution Approach 1:
The patent applies different quality characteristics to different layers: the background layer maintains the original image projection properties while the foreground layer is specifically designed to hold new objects with appropriate depth information. This local differentiation allows objects to be easily placed in the foreground layer while achieving natural visual integration through the layered depth structure.
Data Source
AI summary
An image processing device is configured to generate a three-dimensional bird's eye view image by projecting captured images acquired from a plurality of imaging devices onto a three-dimensional projection surface, and includes: an acquiring part configured to acquire position information of a first area, on which an object is to be placed, in a three-dimensional space; a conversion part configured to convert the position information of the first area into position information in accordance with the three-dimensional projection surface; and an output part configured to output the three-dimensional bird's eye view image, on which the object is to be placed on a second area in the three-dimensional space, the second area being specified by the converted position information. By this means, when the three-dimensional bird's eye view image is generated from multiple images, it is possible to display the object naturally in the three-dimensional bird's eye view image.