LiDAR Pixel Normal Vector Analysis for Particle Discrimination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing LIDAR sensors on vehicles face challenges in accurately distinguishing between suspended particles, such as exhaust gases, and solid objects due to their similar density and elevation profiles, leading to difficulties in precise detection and discrimination.
Innovation Solution
A data processing device and method for LIDAR sensors that identify groups of pixels by determining normal vectors, average distance, intensity, and azimuth angles to differentiate between clouds of particles and solid elements, using thresholds and models to classify pixels as belonging to either category.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If density profile and elevation profile comparison is used to distinguish exhaust gases from solid objects, then classification capability is provided, but measurement precision deteriorates because solid objects can be included in exhaust gases making discrimination difficult
Solution Approach 1:
The patent introduces a new dimension of analysis by computing the distribution of normal vectors associated with each pixel in the point cloud matrix. This normal vector distribution characteristic provides an additional feature space that helps distinguish between exhaust gas particles and solid objects, overcoming the limitation of using only density and elevation profiles.
Solution Approach 2:
The patent changes the parameters used for classification by introducing normal vector distribution as a new distinguishing parameter. Instead of relying solely on density and elevation profiles, the system analyzes the angular distribution of normal vectors to differentiate between particle clouds and solid surfaces, thereby improving measurement precision.
2Measurement precision
If normal vector distribution analysis is introduced to improve discrimination precision, then measurement precision is improved, but device complexity increases due to additional computational steps
Solution Approach 1:
The patent performs preliminary actions by pre-defining thresholds and classification models for normal vector distributions. These pre-computed reference values enable rapid classification during actual operation without requiring complex real-time calculations, thus reducing the operational computational burden while maintaining high precision.
Solution Approach 2:
The patent substitutes complex mechanical/physical analysis with computational geometric analysis. Instead of using complex sensor arrays or multiple physical measurement systems, the invention uses computational geometry to analyze normal vector distributions from standard LIDAR point cloud data, achieving high precision with existing hardware.
Data Source
AI summary
The disclosure relates to a data processing device for a LiDAR sensor mounted on a vehicle. The data processing device is used to determine whether groups of pixels of a matrix of pixels obtained by the LiDAR sensor correspond to clouds of particles, notably based on a distribution of normal vectors associated with these pixels.