Dense Point Cloud Reconstruction With Object Extraction Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating dense point clouds (DPCs) suffer from inclusion of irrelevant objects, leading to lower resolution, missed fine details, and potential distortion, limiting their usefulness.
Innovation Solution
An image processing device and method that separates pixels and datapoints representing the physical object from those not representing it, using object extractors, to generate a modified sparse point cloud (SPC) which is then integrated to create a dense point cloud (DPC) with higher quality and resolution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional 3D reconstruction software is used to generate point clouds of the whole scene, then the complete scene including surrounding environment is captured, but irrelevant objects are included which reduce resolution and miss fine details of the target physical object
Solution Approach 1:
The patent applies object extractors (masks) to extract only the pixels representing the target physical object from the 2D digital images, excluding irrelevant objects. This extraction is then applied to the sparse point cloud to remove irrelevant datapoints, allowing the dense point cloud generation to focus computational resources on the target object, thereby improving resolution and preserving fine details.
2Measurement precision
If bounding boxes are applied to filter out irrelevant objects during feature extraction, then the target area is focused, but distortion of scales occurs and camera pose estimation becomes unreliable
Solution Approach 1:
Instead of applying uniform bounding boxes that mask entire regions, the patent uses object extractors that precisely delineate the boundaries of the target physical object. This allows selective inclusion of only the relevant pixels and datapoints while maintaining the complete scene context for camera pose estimation, thus achieving both focused processing and reliable geometric reconstruction.
3Measurement precision
If photogrammetry software reconstructs many data points on external surfaces to create dense point clouds, then high resolution is achieved, but processing time and computational resources increase significantly
Solution Approach 1:
The patent performs preliminary filtering by applying object extractors to 2D digital images before the dense point cloud reconstruction process. This pre-extraction creates a modified sparse point cloud that already excludes irrelevant objects, so the subsequent dense reconstruction only processes relevant data points, significantly reducing processing time while maintaining high resolution for the target object.
Data Source
AI summary
A dense point cloud (DPC) is generated for a physical object. 2D digital images of the physical object are obtained. Object extractors are defined. A sparse point cloud (SPC) is created from the 2D digital images. A modified SPC is obtained by first applying the object extractors to the 2D digital images to exclude pixels from the 2D digital images and then excluding the datapoints in the SPC that correspond to the pixels that have been excluded from the 2D digital images. The datapoints of the SPC that do not represent the object are excluded from being included in the modified SPC. The DPC is generated by integrating information obtained from the pixels in the 2D digital images with the modified SPC.