Obstacle Detection Using 3D Point Cloud Height Filtering

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional obstacle detection systems for working machines fail to distinguish between insignificant objects and obstacles that require route changes, leading to unnecessary travel adjustments and increased distance.

Innovation Solution

A method and system using a three-dimensional point cloud filtering technique with a voxel grid to identify and remove irrelevant data points, converting the cloud to a two-dimensional obstacle map, and updating the planned traveling route based on the filtered data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional object detection systems detect all objects around the working machine, then the detection coverage is comprehensive, but the system cannot distinguish between insignificant objects and obstacles, leading to unnecessary route changes and increased travel distance

Engineering Contradiction:

Improveobstacle identification accuracyVSAvoidtraveling efficiency

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The system changes the parameter of obstacle identification by introducing height information through a three-dimensional point cloud generated by a LiDAR sensor. By filtering objects based on their vertical dimension (comparing object height to the working machine's height), the system accurately distinguishes between insignificant objects (below threshold) and obstacles (above threshold), resolving the contradiction between detection comprehensiveness and identification accuracy.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The detection space is segmented into significant and insignificant regions based on height thresholds. The system divides the three-dimensional space around the working machine into zones: objects below a certain height are segmented as insignificant and filtered out, while objects above the threshold are segmented as potential obstacles requiring further route planning consideration. This segmentation enables efficient filtering while maintaining productivity.

Inventive Principle:

Principle #1Segmentation

2Reliability

If the working machine changes traveling route for every detected object, then obstacle avoidance is ensured, but unnecessary route adjustments increase the time and distance traveled

Engineering Contradiction:

Improveobstacle avoidance reliabilityVSAvoidtraveling time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system introduces a height parameter threshold to filter detected objects. By comparing the height of each detected object against the working machine's height, the system reliably identifies which objects truly require avoidance (those taller than the machine) and which can be safely ignored (shorter objects). This parameter-based filtering maintains obstacle avoidance reliability while eliminating unnecessary route changes that would waste time.

Inventive Principle:

Principle #35Parameter changes

3Loss of information

If the system processes all detected data points, then complete object information is available, but processing time and computational load increase

Engineering Contradiction:

Improveobject detection information completenessVSAvoiddata processing time

Core Design Contradiction:

Loss of informationVSLoss of time

Solution Approach 1:

The system extracts only the relevant height parameter from the three-dimensional point cloud data and uses it to filter out insignificant objects. By taking out the critical height information and using it as a filtering criterion, the system maintains complete information about potential obstacles while eliminating processing of insignificant objects, thus reducing computational load and processing time without losing important detection information.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The system performs preliminary filtering of the three-dimensional point cloud data by removing data points that correspond to insignificant objects based on height thresholds before further processing. This preliminary action reduces the volume of data that requires detailed analysis, decreasing computational load and processing time while preserving all information about significant obstacles that require attention.

Inventive Principle:

Principle #10Preliminary action

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

Accurately identifies obstacles, reducing unnecessary travel adjustments and optimizing route planning for working machines.

Implementation Method 1

a sensor that generates a three-dimensional point cloud including a plurality of data points

Methodology Applied

Scientific EffectTime of Flight: Time of Flight

Implementation Method 2

The one or more data points are removed from the plurality of data points based on a position of the one or more data points

Methodology Applied

Scientific EffectLIDAR: LIDAR

Data Source

PatentUS12499619B2Obstacle detection system and method

Publication Date: 2025.12.16 KUBOTA CORP

AI summary

A method includes generating a three-dimensional point cloud including a plurality of data points and filtering the three-dimensional point cloud to remove one or more data points from the plurality of data points. The one or more data points are removed from the plurality of data points based on a position of the one or more data points.