Lidar Probability Mapping for Objects Hidden by Particulate Matter
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Lidar systems are prone to false positive detections due to reflections from particulate matter such as fog, smoke, and exhaust, which can lead to inaccurate depth mapping and pose safety risks by misidentifying the presence of solid objects.
Innovation Solution
A multi-stage machine-learning model approach is employed to classify split returns as either associated with particulate matter or true positive surfaces, utilizing edge detection, image classification, and density analysis to differentiate between the two, and suppress false returns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If lidar systems detect reflected light signals to determine distances to surfaces, then distance measurement capability is improved, but false positive detections occur due to reflections from particulate matter such as fog, smoke, and exhaust
Solution Approach 1:
The patent segments the detection process into multiple stages: initial return detection, split return identification, machine learning classification, and probability map generation. This multi-stage segmentation allows the system to distinguish between true surface reflections and particulate matter reflections by analyzing different characteristics at each stage, thereby resolving the contradiction between measurement precision and reliability.
Solution Approach 2:
The patent introduces machine learning models as intermediary components between the raw lidar signals and the final distance measurements. These models act as mediators that classify split returns and generate probability maps, filtering out false positives from particulate matter while preserving true surface detections, thus improving both measurement precision and detection accuracy simultaneously.
2Reliability
If lidar systems suppress returns associated with particulate matter, then false positive detections are reduced, but objects obscured by particulate matter may be missed
Solution Approach 1:
The patent implements dynamic decision-making through probability maps that continuously update the likelihood of object presence behind particulate matter. Rather than static suppression, the system dynamically adjusts detection thresholds based on learned patterns from machine training data, allowing flexible balancing between suppressing false positives and maintaining object detection completeness based on environmental conditions.
Solution Approach 2:
The patent changes the parameter of detection confidence by generating probability maps that quantify the likelihood of object presence at different locations. This parameter transformation allows the system to make informed decisions about which suppressed returns to reinvestigate, optimizing the balance between reducing false positives and maintaining complete object detection.
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 method significantly reduces false positive detections, enhances the accuracy of lidar systems, and improves safety by accurately identifying objects obscured by particulate matter, ensuring precise navigation and control of autonomous vehicles.
Implementation Method 1
When the emitted light is incident on a surface, a portion of the light is reflected and received by the light sensor
Implementation Method 2
the system may measure the propagation time of a light signal as it travels from the laser emitter, to the surface, and back to the light sensor
Implementation Method 3
fine particulate matter may also reflect light
Data Source
AI summary
A machine-learned (ML) model for determining, using at least depth data (e.g., lidar data, radar data), whether an object exists beyond particulate matter, such as dust, steam, smoke, rain, etc. from a sensor. Determining the likelihood may include determining a probability map that varies, distance beyond the particulate matter, reflectance of the object, transmit power, or the like.