3D Point Cloud Processing Using Bounding Boxes for Noise Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for generating three-dimensional point cloud data are hindered by the presence of noise, which complicates accurate shape recognition and increases data size, and noise removal processes are computationally intensive.
Innovation Solution
A three-dimensional point cloud data processing system that utilizes bounding boxes (BBs) to filter noise by calculating BBs from multiple angles of brightness information, employing AI for point identification, and applying high-accuracy noise removal techniques to reduce data size and computation time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If stratum estimation and standard deviation calculation are used for noise removal, then noise removal accuracy is improved, but calculation time increases significantly
Solution Approach 1:
The patent applies preliminary action by performing noise removal using bounding boxes before conducting stratum estimation and standard deviation calculation. The bounding box method quickly identifies and removes obvious noise points based on geometric constraints, reducing the data volume and computational load for subsequent more accurate but time-consuming stratum-based noise removal processes.
Solution Approach 2:
The patent segments the noise removal process into multiple stages: first using bounding boxes for coarse noise removal, then applying stratum estimation and standard deviation calculation for fine noise removal. This segmentation allows each method to operate on progressively smaller datasets, balancing accuracy requirements with computational efficiency at different processing stages.
2Measurement precision
If design shape data is used for noise removal, then noise removal accuracy is improved, but applicability to objects without design data is reduced
Solution Approach 1:
The patent implements self-service by enabling the system to automatically generate bounding boxes and perform noise removal without requiring pre-existing design shape data. The bounding box method derives geometric constraints directly from the point cloud data itself, allowing the system to serve itself and process any object regardless of whether design data is available.
Solution Approach 2:
The patent achieves universality by developing a noise removal approach that works for all objects without requiring object-specific design data. The bounding box technique and subsequent stratum estimation method are general-purpose approaches that can be applied to any three-dimensional point cloud data, making the system versatile across different applications and object types.
3Loss of information
If primary point data is extracted from image data, then three-dimensional shape information is obtained, but large amounts of noise points are included
Solution Approach 1:
The patent applies the extraction principle by selectively removing noise points from the primary point data while preserving the essential three-dimensional shape information. The bounding box method extracts valid points within geometric constraints, and the stratum estimation further extracts points belonging to the actual object surface, separating useful information from noise.
Solution Approach 2:
The patent implements local quality by applying different processing strategies to different regions of the point cloud data. The bounding box method applies uniform geometric constraints, while the stratum estimation and standard deviation calculation apply localized analysis to identify points belonging to the object surface versus noise, preserving shape information while removing noise.
Data Source
AI summary
A three-dimensional point cloud data processing system includes: one or more cameras that image an object to acquire one or more items of brightness information; and a controller that calculates point cloud data of the object from the one or more items of brightness information. The controller calculates at least one bounding box (BB), which surrounds the object and is formed by combining one or more rectangles, from at least one of the one or more items of brightness information, and calculates the point cloud data based on the at least one BB and the one or more items of brightness information.