Flash Lidar Point Cloud De-jittering via 3D Model Fitting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Jitter in point cloud data from flash lidar systems introduces errors that complicate target recognition processes, as it arises from the relative motion of components within the system.
Innovation Solution
The method involves shifting lidar point cloud data in its coordinate system to minimize fit error with pre-defined 3-D models of possible targets, using metrics like Root Mean Square goodness-of-fit, until the best fit is achieved, thereby identifying the target.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If flash lidar systems are used to generate point cloud data, then target recognition capability is enabled, but jitter errors are introduced due to relative motion of components
Solution Approach 1:
The patent applies preliminary action by pre-defining multiple 3-D models of possible targets before the recognition process. These models are prepared in advance with known geometric characteristics, allowing the system to compare actual point cloud data against predetermined references. This pre-preparation enables the de-jittering process to work effectively by having reference frameworks ready for comparison.
Solution Approach 2:
The patent uses copying by creating virtual 3-D models that represent physical targets. These digital models are copies of the expected target geometries, stored in memory for comparison. The system compares the actual point cloud data (which contains jitter) against these idealized copies, allowing it to identify and correct jitter errors by finding the best match between the noisy data and the clean reference models.
2Measurement precision
If point cloud data is shifted to minimize fit error, then de-jittering accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the point cloud data into multiple segments or groups that can be shifted independently. Instead of attempting to process the entire point cloud as a single unit, the system segments the data and applies coordinate shifts to different segments. This segmentation reduces the computational burden of finding the optimal shift for all points simultaneously, while still achieving effective de-jittering through the cumulative effect of segment-level adjustments.
Solution Approach 2:
The patent applies partial action by performing de-jittering operations on selected portions of the point cloud data rather than all points. The system identifies key features or representative segments of the point cloud that, when de-jittered, sufficiently correct the overall data accuracy. This partial approach reduces computational complexity while maintaining adequate de-jittering performance for target recognition purposes.
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 effectively de-jitters point cloud data, enhancing the accuracy of target recognition by aligning the data with the most likely 3-D model, thus improving the precision of identifying targets such as aircraft.
Implementation Method 1
measures distances to a target by illuminating the target with laser light and measuring the reflected light with one or more sensors. Differences in the laser return times and/or wavelengths can then be used to make a digital 3-D representation of the target
Data Source
AI summary
Jitter is removed from point cloud data of a target by fitting the data to 3-D models of possible targets. The point cloud data is de-jittered as a group by shifting the point cloud data in its coordinate system until a minimum fit error is observed between the shifted data and a 3-D model under analysis. Different 3-D models may be evaluated in succession until a 3-D model is identified that has the least fit error. The 3-D model with the least fit error most likely represents the identity of the target.