3D Cuboid Dimension Refinement Under Noisy Vision Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing 3D vision systems struggle to accurately measure the dimensions of cuboidal objects, such as boxes and polybags, due to noise and coarse resolution, which limits their ability to meet tight accuracy requirements of 2.5 mm or less.
Innovation Solution
A 3D vision system that employs statistical analysis methods, including Points Statistical Analysis (PSA), Boundary Statistical Analysis (BSA), Height from Least Squares Surface Fitting, and Height from Bulginess, to refine the dimensions of cuboidal objects by analyzing 3D image data, and a user interface for setup and operation to ensure accurate measurement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 3D vision systems use standard imaging methods to measure cuboidal objects, then the system structure remains simple, but measurement precision deteriorates due to noise and coarse resolution
Solution Approach 1:
The patent segments the measurement process into distinct stages: initial bounding box detection, identification of candidate face regions, statistical analysis of point distributions, and refinement of dimension estimates. This segmentation allows each stage to focus on specific aspects of the measurement problem, improving overall precision without requiring complete system redesign.
Solution Approach 2:
The patent applies partial action by selectively analyzing only the most relevant portions of the 3D point cloud data - specifically focusing on points near identified face regions rather than processing the entire point cloud. This selective approach improves measurement precision while controlling computational complexity.
2Measurement precision
If 3D vision systems acquire detailed 3D image data to improve measurement accuracy, then measurement precision improves, but loss of information increases due to noise in the data
Solution Approach 1:
The patent extracts useful information from noisy 3D image data by identifying and isolating points that belong to specific face regions of cuboidal objects. By separating signal (points on object faces) from noise (other points), the system improves measurement precision while minimizing information loss.
Solution Approach 2:
The patent changes parameters by applying statistical transformations to the raw 3D point coordinates - computing distributions, variances, and other statistical measures that convert noisy positional data into more reliable dimensional estimates, thereby improving precision while filtering noise.
3Measurement precision
If 3D vision systems process raw 3D point cloud data directly to determine object dimensions, then processing speed remains high, but measurement precision deteriorates due to coarse resolution
Solution Approach 1:
The patent performs preliminary actions by first identifying bounding boxes and candidate face regions before conducting detailed dimensional analysis. This preliminary structuring of the data enables subsequent precise measurements to be performed more efficiently, maintaining productivity while improving precision.
Solution Approach 2:
The patent transitions from analyzing raw 3D point coordinates to examining 2D projections and distributions of points on identified face regions. This dimensional transformation enables more precise measurement calculations while maintaining computational efficiency through optimized algorithms.
Data Source
Figure 1
Figure 1A
Figure 2
AI summary
A system and method for estimating dimensions of an approximately cuboidal object from a 3D image of the object acquired by an image sensor of the vision system processor is provided. An identification module, associated with the vision system processor, automatically identifies a 3D region in the 3D image that contains the cuboidal object. A selection module, associated with the vision system processor, automatically selects 3D image data from the 3D image that corresponds to approximate faces or boundaries of the cuboidal object. An analysis module statistically analyzes, and generates statistics for, the selected 3D image data that correspond to approximate cuboidal object faces or boundaries. A refinement module chooses statistics that correspond to improved cuboidal dimensions from among cuboidal object length, width and height. The improved cuboidal dimensions are provided as dimensions for the object. A user interface displays a plurality of interface screens for setup and runtime operation.