LIDAR Object Detection Using Z-Domain and Range-Domain Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current machine-vision systems are ineffective in detecting small objects, such as beverage cans or small animals, near the ground plane, which is crucial for autonomous vehicles to navigate safely.
Innovation Solution
A system utilizing LIDAR sensors to emit light beams and process three-dimensional data points, identifying inlier and outlier points in both the z-domain and range-domain using a joint RANSAC algorithm, allowing for the detection of objects based on outliers that do not lie in the hypothesized ground plane.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current machine-vision systems are used for object detection, then the system complexity remains manageable, but the detection precision for small objects near the ground plane deteriorates
Solution Approach 1:
The patent segments the object detection process into multiple specialized modules: a ground-plane detection module that identifies the ground plane using a first error condition, and an object detection module that identifies objects using a second error condition. This segmentation allows each module to specialize in specific detection tasks, improving overall detection precision for small objects while managing system complexity through modular architecture.
Solution Approach 2:
The patent introduces a vertical dimension (z-axis) to the traditional 2D image plane analysis. By detecting the vertical distance of objects from the ground plane and using this z-coordinate information alongside x and y coordinates, the system creates a 3D detection space that significantly improves the ability to detect small objects near the ground that would otherwise be indistinguishable from the ground plane in 2D.
2Reliability
If traditional 2D image analysis is used, then the ease of operation is maintained, but the reliability of detecting small objects deteriorates
Solution Approach 1:
The processing is segmented into distinct stages: ground-plane detection using inlier points with a first error condition, followed by object detection using outlier points with a second error condition. This segmentation improves detection reliability by systematically separating ground from objects, while the automated modular processing maintains ease of operation.
Solution Approach 2:
The patent replaces traditional mechanical/image-processing approaches with a computational geometry approach based on RANSAC (Random Sample Consensus) algorithms. This substitution uses mathematical error conditions and statistical methods to automatically distinguish ground plane from objects, improving reliability while the automated computational process maintains operational simplicity.
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 the reliable detection of small objects as close as 1 or 2 cm in height above the ground, improving the ability of autonomous vehicles to recognize and respond to obstacles that might otherwise be missed.
Implementation Method 1
one or more LIDAR sensors disposed within the vehicle and configured to emit a plurality of light beams
Data Source
AI summary
Systems and methods described herein relate to detecting objects. One embodiment receives a plurality of three-dimensional (3D) data points from a plurality of light beams emitted by one or more sensors; identifies, among the plurality of 3D data points, a first set of inlier points that satisfy a first predetermined error condition with respect to a plane hypothesis and a first set of outlier points that fail to satisfy the first predetermined error condition; identifying, among the first set of inlier points, a second set of outlier points, the second set of outlier points failing to satisfy a second predetermined error condition in a range domain with respect to a plurality of line hypotheses corresponding, respectively, to the plurality of light beams; and detecting an object based, at least in part, on at least one of the first set of outlier points and the second set of outlier points.