3D Ground Plane Detection From Depth Images With Point Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining a ground plane in sensor data, such as from a time-of-flight sensor, are inefficient and require significant processing power, making them unsuitable for real-time applications in autonomous robots.
Innovation Solution
A method involving preprocessing of sensor data to generate a 3D point cloud, discarding irrelevant points, and iteratively selecting and validating non-collinear points to identify a horizontal plane within specified tolerances, thereby reducing processing load and improving efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If existing methods for determining ground plane are used, then ground plane detection can be performed, but processing time and power consumption are excessive for real-time applications
Solution Approach 1:
The patent segments the ground plane detection process into distinct stages: (1) generating a 3D point cloud from depth image data, (2) filtering points based on spatial location (discarding points above the sensor), (3) iteratively selecting non-collinear points, (4) determining horizontal planes within tolerance thresholds, and (5) comparing plane sizes to identify the largest horizontal plane. This segmentation allows each stage to be optimized independently and enables real-time processing by avoiding unnecessary computations in later stages.
Solution Approach 2:
The patent performs preliminary actions by pre-filtering the point cloud to discard points located above the sensor before initiating the ground plane detection algorithm. This preliminary filtering reduces the number of points that need to be processed in subsequent iterative steps, significantly decreasing processing time and power consumption while maintaining detection accuracy.
2Reliability
If existing ground plane detection methods are used, then ground plane can be identified, but processing power requirements are too high for autonomous robots
Solution Approach 1:
The patent applies partial action by performing only the necessary computational steps required for ground plane detection. Instead of processing all points in the point cloud uniformly, the method selectively processes points based on their spatial location and plane membership, reducing energy consumption while maintaining sufficient accuracy for autonomous navigation.
Solution Approach 2:
The patent utilizes parameter changes by dynamically adjusting tolerance thresholds for horizontal plane determination and iteratively refining plane selections based on the number of points found. This adaptive approach allows the system to achieve reliable ground plane detection with reduced computational resources by stopping iterations once sufficient accuracy is achieved.
3Measurement precision
If iterative selection of non-collinear points is performed, then ground plane accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent implements feedback mechanisms by iteratively selecting non-collinear points, determining horizontal planes within tolerance thresholds, and using the results to guide subsequent iterations. The algorithm compares the number of points found in each plane and uses this feedback to identify the largest horizontal plane, progressively improving precision while managing complexity through structured iteration.
Data Source
AI summary
A system for determining a ground plane in a depth image can include a processor, which can be configured to generate a 3D point cloud using data included in the depth image. The processor can also be configured to receive data related to an orientation of the 3D point cloud. The processor can also be configured to iteratively select at least 3 non-collinear points. The processor can also be configured to iteratively determine whether the at least 3 non-collinear points form a first plane that can be horizontal within a first tolerance and if so: (1) find other points in the first plane, and (2) compare a number of points in the first plane to a number of points in a largest horizontal plane that has already been found.