LiDAR Point Cloud Noise Removal via Distance Distribution Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for removing noise from three-dimensional point cloud data using LiDAR are ineffective, especially when noise point clouds are continuous and dense, and struggle to accurately distinguish between noise and reflection signals from distant objects.
Innovation Solution
An information processing apparatus that classifies three-dimensional point cloud data into clusters based on distance values and determines which clusters to adopt or reject using a histogram analysis of distance distributions, allowing for accurate noise removal.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If isolated point removal filter or clustering method using reflection intensity is used, then some noise points can be removed, but continuous and dense noise point clouds cannot be completely removed
Solution Approach 1:
The patent changes the parameter used for noise detection from reflection intensity to distance value distribution. By analyzing the statistical distribution of distance values within clusters and comparing them against expected target object distance distributions, the system can identify and remove noise points based on distance anomalies rather than reflection characteristics, effectively handling continuous and dense noise that cannot be distinguished by intensity alone.
2Measurement precision
If clustering method using reflection intensity is used, then noise can be separated, but it is difficult to discriminate noise from distant object reflection signals
Solution Approach 1:
The patent switches from using reflection intensity as the clustering parameter to using distance values. By analyzing the distribution of distance values and comparing them with expected distance distributions of target objects, the system can distinguish noise from distant object signals based on distance characteristics rather than reflection properties, making the method applicable across various scenarios including distant objects with weak reflection signals.
Solution Approach 2:
The patent replaces the intensity-based discrimination mechanism with a distance-based statistical analysis mechanism. Instead of relying on the photoelectric conversion and intensity measurement process, the system uses arithmetic operations on distance values to compute distributions and identify noise, providing a more robust discrimination method that works regardless of reflection intensity variations.
3Productivity
If conventional noise removal methods are used, then processing can be performed, but ranging accuracy decreases due to erroneous ranging from noise signals
Solution Approach 1:
The patent performs preliminary classification of point clouds into clusters based on distance values before conducting noise removal. By pre-organizing the point cloud data into distance-based clusters and analyzing their distribution characteristics, the system identifies noise points in advance through statistical comparison, allowing for accurate noise removal that preserves valid ranging data and maintains overall ranging accuracy.
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
Enables precise removal of noise from three-dimensional point cloud data, even in dense regions, by distinguishing between target and noise point clouds based on distance value distributions, thereby improving ranging accuracy.
Implementation Method 1
a reflected light pulse reflected thereby is photoelectrically converted by an optical receiver inside the device and processed as an electrical signal
Data Source
AI summary
An information processing apparatus of the present disclosure includes: a classifying unit that classifies three-dimensional point cloud data including distance values to a target in a specified region into one or more clusters based on the distance values; and a determining unit that determines the cluster to adopt based on the distribution of the distance values of the three-dimensional point cloud data included by the cluster obtained by the classification.