Environment Element Classification for Obstacle Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current motion planning systems face challenges in accurately distinguishing between vegetation and non-vegetation environment elements, leading to difficulties in obstacle detection and safe navigation, especially in scenarios where vegetation resembles other obstacles in size and outline.
Innovation Solution
A method and system that classify environment elements by identifying edges and non-edges based on height differences, clustering scanned points, and validating vegetation and non-vegetation candidates using reflection features, generating a grid map to differentiate between types of environment elements and obstacles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If motion planning systems use standard obstacle detection methods, then navigation can be performed, but vegetation and non-vegetation environment elements cannot be accurately distinguished
Solution Approach 1:
The patent segments the environment classification problem into multiple processing stages: obtaining scanned points, identifying edges based on height differences, clustering points using sliding windows, classifying vegetation candidates, and validating results. This segmentation allows complex classification to be broken down into manageable steps, improving accuracy without overwhelming system complexity
Solution Approach 2:
The patent introduces height dimension analysis by identifying edges based on height differences between adjacent scanned points. This vertical dimension differentiation enables the system to distinguish vegetation from non-vegetation obstacles that may appear similar in 2D space, thereby improving classification accuracy
2Measurement precision
If the system classifies all scanned points individually, then classification detail is improved, but processing time increases
Solution Approach 1:
The patent merges adjacent scanned points into clusters using sliding window techniques. By processing points in clusters rather than individually, the system maintains classification detail while reducing the total number of processing operations, thereby decreasing processing time
Solution Approach 2:
The patent applies partial action by focusing detailed classification only on edge points and vegetation candidates identified through height difference analysis, rather than processing every scanned point with the same level of detail. This selective approach maintains necessary classification precision while reducing overall processing time
Data Source
AI summary
The presently disclosed subject matter includes, inter alia, a method and a system which enables to classify information of a scanned outdoor or indoor environment to different types of environment elements. The identification of environment element types enables to improve detection of obstacles in an environment. The improved detection of obstacles can assist in motion planning to enable to advance quickly and safely through the environment.