3D Point Cloud Alignment Using Visibility Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for aligning lidar images using nearest neighbor criteria are prone to incorrect point associations and have slow execution times due to the consideration of irrelevant points and the need for extensive processing.
Innovation Solution
A method that estimates the visibility of points using a simple calculation, selects visible points for association, and applies spatial transformations to align images, incorporating an odometry step for vehicle position estimation, and optionally weights points based on visibility for refined alignment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If nearest neighbor criteria are used to associate points between images, then alignment can be performed, but incorrect point associations occur and execution time increases
Solution Approach 1:
The patent applies preliminary action by estimating the visibility of points before performing the association step. Points are pre-filtered based on visibility criteria (e.g., based on distance from the sensor or other geometric conditions) to identify only those points that are likely to be visible in the captured image. This preliminary filtering reduces the number of points that need to be processed in subsequent association steps, thereby reducing execution time while improving association accuracy by excluding potentially invisible points from consideration.
2Manufacturing precision
If all points are considered in the alignment process, then comprehensive alignment is achieved, but processing time increases significantly
Solution Approach 1:
The patent applies segmentation by dividing the set of all detected points into two subsets: visible points and invisible points. The visibility estimation step segments the point cloud based on geometric criteria (such as distance thresholds or angular conditions) to separate points that are likely to be visible in the captured image from those that are not. Only the visible points are then used in the association and alignment processes, significantly reducing the computational workload while maintaining alignment accuracy for the relevant features.
3Productivity
If invisible points are included in association, then more points are processed, but incorrect associations increase
Solution Approach 1:
The patent applies the taking out principle by extracting and removing invisible points from the set of points to be processed in the association step. The visibility estimation mechanism identifies points that are unlikely to be visible in the captured image (e.g., points too far from the sensor, points occluded by foreground objects, or points outside the field of view) and removes them from the working set. Only the remaining visible points are subjected to nearest neighbor association, which improves association correctness by eliminating mismatches caused by invisible points while maintaining adequate processing throughput.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach reduces incorrect associations and execution time by focusing on visible points, improving alignment accuracy and efficiency while enabling real-time processing and accurate vehicle localization.
Implementation Method 1
This receiving surface records the time that elapses between the time when the laser pulse is emitted and the time when it is received by the sensor in order to calculate the distance between the sensor and the target.
Implementation Method 2
The reflection of the laser beam from the surface of the target is detected by a receiving surface arranged in the lidar.
Data Source
AI summary
A method for aligning at least a first source image with a second reference image, each image including a set of three-dimensional points. The method being intended to reconstruct a common image by aligning the first source image with the second reference image. The method including at least: an association step of associating, in pairs, at least some of the points of the first source image, forming a first group of interest of points, with the corresponding points of the second reference image, using nearest neighbor criteria, a step of aligning the points associated in pairs by applying a spatial transformation. The method being noteworthy in that it includes a step of estimating the visibility of the points in order to limit point alignment errors.