LIDAR Point Cloud Quality Metric for Aggressor Object Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
LIDAR systems face challenges in generating accurate point clouds due to the presence of 'aggressor objects' such as retroreflective, specular, and active objects, which can cause errors in Time of Flight measurements and result in lower point cloud quality.
Innovation Solution
The development of a point cloud quality metric that quantifies the impact of aggressor objects on LIDAR systems, allowing for calibration and post-processing adjustments to improve point cloud accuracy and quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If LIDAR systems emit light pulses to detect objects and generate point clouds, then object detection capability is provided, but measurement precision deteriorates due to errors from aggressor objects
Solution Approach 1:
The system performs preliminary characterization of aggressor objects by capturing reference point clouds of known objects with precise ground truth measurements. This pre-established knowledge base allows the system to identify and compensate for measurement errors caused by aggressor objects during actual operation, resolving the contradiction between maintaining detection capability and improving measurement accuracy.
Solution Approach 2:
The system implements feedback through iterative optimization where point cloud quality metrics are calculated by comparing detected points against ground truth data. The quality metric provides feedback that guides calibration adjustments and post-processing corrections, continuously improving measurement precision while maintaining the core detection function.
2Loss of information
If LIDAR systems constantly emit and detect light pulses to create point clouds, then object information is obtained, but point cloud quality deteriorates due to errors associated with detected points
Solution Approach 1:
The system extracts and separates erroneous points from the point cloud by calculating quality metrics that identify points affected by aggressor objects. Through point cloud subtraction and comparison with ground truth data, the system removes inaccurate points while preserving valid object information, thus maintaining information completeness while improving accuracy.
Solution Approach 2:
The system changes parameters by introducing quality metric thresholds and calibration factors that adjust point cloud processing. By modifying how points are evaluated and weighted based on their quality scores, the system transforms raw point cloud data into accurate representations without losing essential object information.
3Measurement precision
If a point cloud quality metric is developed to quantify aggressor object impact, then point cloud accuracy is improved, but device complexity increases
Solution Approach 1:
The system uses simplified reference copies of known objects with precise ground truth measurements to characterize aggressor objects. Instead of implementing complex real-time analysis, the system creates and utilizes simplified models or lookup tables from reference measurements, reducing computational complexity while maintaining accurate quality assessment capabilities.
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
The proposed solution enables LIDAR systems to generate higher quality point clouds by accounting for the effects of aggressor objects, thereby enhancing the accuracy of object detection and navigation in autonomous vehicles.
Implementation Method 1
the LIDAR system may be able to determine a distance the object is from the LIDAR system based on the amount of time between the light pulse being emitted and the return light pulse being detected (for example, the 'Time of Flight' (ToF) of the light pulse)
Implementation Method 2
the presence of 'aggressor objects' such as retroreflective, specular, and active objects, which can cause errors in Time of Flight measurements
Data Source
AI summary
A method including capturing valid points on a test target including a retroreflective object, the valid points including a first point cloud for a first region, capturing invalid points outside the edge of the test target, which the invalid points include a first point in a second point cloud within of the acceptable error threshold based on the first region and a second point in the second point cloud outside of the acceptable error threshold based on the first region, recording a plurality of frames of the first point cloud and the second point cloud, evaluating the number of the invalid points and recording the maximum value of the invalid points, generating a first score for the first point and a second score for the second point based on a penalty function and combining the first score and the second score to produce a point cloud quality metric for the invalid points, and calibrating, based on the plurality of the point cloud quality metric, the LIDAR system for the retroreflective object.