3D Pole Localization Using Vertical Point Cloud Clusters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing pole-like object detection methods in autonomous vehicles face challenges with sparse point cloud data, leading to inaccurate computation of point features and failure in detecting pertinent structures along roadways, resulting in incorrect exclusion of relevant data points.
Innovation Solution
The proposed solution involves a three-step approach: identifying potential pole-like object clusters, classifying them using machine learning algorithms based on statistical features, and modeling these clusters as cylinders to determine their attributes, thereby improving detection accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional point cloud processing methods are used, then individual data points are processed separately, but sparse point cloud data leads to inaccurate computation of point features and failure to detect pertinent structures
Solution Approach 1:
The patent merges sparse individual point cloud data points into vertical clusters by identifying spatial and temporal relationships between points. This combining approach transforms insufficient individual data into meaningful structural representations of pole-like objects, enabling accurate detection despite sparse input data.
Solution Approach 2:
The patent transitions from processing two-dimensional point cloud data to three-dimensional vertical cluster representations by adding the temporal dimension (across multiple frames). This dimensional expansion allows sparse 2D points to form coherent 3D structures, improving detection accuracy.
2Reliability
If all point cloud data is processed individually, then comprehensive data analysis is achieved, but processing time and computational resources increase significantly
Solution Approach 1:
The patent segments the point cloud data processing into distinct stages: initial cluster identification, vertical cluster formation, and classification. This segmentation allows the system to process only relevant data points for pole-like object detection rather than all points individually, reducing processing time while maintaining reliability.
Solution Approach 2:
The patent performs preliminary clustering and filtering operations before detailed classification. By pre-organizing sparse points into vertical clusters and identifying candidate regions in advance, the system reduces the computational burden of subsequent processing steps, decreasing overall processing time.
3Measurement precision
If detailed classification and modeling of each detected object is performed, then accurate localization is achieved, but storage requirements and computational complexity increase
Solution Approach 1:
The patent applies detailed classification and modeling only to identified vertical clusters that match pole-like object characteristics, rather than all detected objects. This localized detailed processing maintains high localization accuracy for relevant objects while reducing overall system complexity by applying simplified processing elsewhere.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Embodiments include apparatus and methods for automatic detection of pole-like objects for a location at a region of a roadway and automatic localization based on the detected pole-like objects. Pole-like objects are modeled as cylinders and the models are generated based on detected vertical clusters of point cloud data associated to corresponding regions along the region of the roadway. The modeled pole-like objects are stored in a database and associated with the region of the roadway. Sensor data from a user located at the region of the roadway is received. The pole-like object model is accessed and compared to the received sensor data. Based on the comparison, localization of the user located at the region of the roadway is performed.