Edge Detection Using 2D Image and Depth Data Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies fail to accurately and efficiently detect edges and junctions in three-dimensional reality capture data, particularly when combining data from laser scanners with depth-based sensors, leading to inaccurate and time-consuming processing in CAD environments.
Innovation Solution
Combining information from 2D images with range images to improve the extraction of edges and corners in space by using a method that involves eigenvector analysis and least-squares intersection of lines or planes, allowing for the integration of RGB and depth data from sensors like the Kinect™.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If laser scanners are used to capture 3D data, then flat surfaces and planes can be captured well, but edges are not easily determinable and require expensive processing to locate
Solution Approach 1:
The patent combines depth data from laser scanners with color/intensity data from 2D images to detect edges and junctions. By merging these two data types, the system achieves accurate edge detection without requiring expensive neighboring point examination, thus resolving the contradiction between measurement precision and processing time.
Solution Approach 2:
The patent uses 2D images as an intermediary to facilitate edge detection in 3D space. The color/intensity information from 2D images serves as a mediator that helps identify edges that are difficult to detect from sparse laser point data alone, improving edge detection accuracy while reducing processing requirements.
2Loss of information
If depth based sensors are used, then depth information can be obtained, but the resolution is sparse and not as dense as color/intensity based scanners
Solution Approach 1:
The patent merges depth data from depth-based sensors with high-resolution color/intensity data from 2D images. This combination compensates for the sparse resolution of depth sensors by overlaying it with the dense information from 2D images, thereby recovering lost depth information while maintaining measurement precision.
Solution Approach 2:
The patent creates a composite data structure that integrates depth information and color/intensity information. This composite representation combines the strengths of both data types, providing both the depth context from depth sensors and the high-resolution detail from 2D images, thus addressing the resolution sparsity issue.
3Measurement precision
If 2D images are used, then color/intensity data and edge transitions can be visualized, but depth perception is poor especially at edges
Solution Approach 1:
The patent combines 2D images with depth data to create a unified representation that preserves both the excellent edge visualization capabilities of 2D images and the depth information from depth sensors. This merging allows accurate edge detection while recovering the depth information that 2D images lack.
Solution Approach 2:
The patent enhances 2D images by incorporating depth information from another dimension (the third dimension). By projecting or integrating depth data with 2D image data, the system maintains the visual clarity of 2D images while adding depth perception, particularly at edges where depth information is critical.
Data Source
AI summary
A method, system, apparatus, article of manufacture, and computer program product provide the ability to detect junctions. 3D pixel image data is obtained/acquired based on 2D image data and depth data. Within a given window over the 3D pixel image data, for each of the pixels within the window, an equation for a plane passing through the pixel is determined/computed. For all of the determined planes within the given window, an intersection of all of the planes is computed. A spectrum of the intersection/matrix is analyzed. Based on the spectrum, a determination is made if the pixel at the intersection is of 3 or more surfaces, 2 surfaces, or is 1 surface.