Obstacle Height Detection Using Reference Obstacles and 3D Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Unmanned ground vehicles face challenges in accurately evaluating the height of obstacles due to the foreshortening effect in images, making it difficult to determine the actual size and height of obstacles from image frames.
Innovation Solution
A method and device that acquire observation data of reference obstacles to calculate evaluation values for the height of detected obstacles using formulas that account for the observed height, reference obstacle height, and distance from the optical axis, allowing for the determination of a second evaluation value through filtering and denoising, and fitting functions to model road geometry for 3D dimension detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If obstacle height is evaluated directly from image frame sizes, then the evaluation process is simple, but the measurement precision is poor due to foreshortening effect
Solution Approach 1:
The patent introduces reference obstacles with known fixed heights as intermediaries between the camera and the target obstacle. By comparing the observed height of the target obstacle with the observed heights of multiple reference obstacles at different distances, the system can calculate the target obstacle's actual height while compensating for foreshortening effects. The reference obstacles serve as mediators that provide known height references at various depths in the scene.
Solution Approach 2:
The patent transitions from two-dimensional image plane measurements to three-dimensional height estimation by incorporating depth information. Multiple reference obstacles at different distances from the camera create a depth reference framework. The system uses the relationship between observed height and distance across multiple reference objects to establish a depth-height mapping that enables accurate 3D height reconstruction of the target obstacle.
2Measurement precision
If multiple reference obstacles are used to improve height detection accuracy, then the measurement precision improves, but the device complexity increases
Solution Approach 1:
The patent makes the reference obstacles multi-functional by using them for both depth estimation and height calibration simultaneously. The same set of reference obstacles with known heights serves dual purposes: establishing the depth scale in the scene and providing the height reference for calculating target obstacle dimensions. This eliminates the need for separate calibration objects and reduces overall system complexity.
Solution Approach 2:
The system performs self-calibration using the reference obstacles without requiring external calibration equipment or complex setup procedures. The known heights of the reference obstacles are stored in advance, and the system automatically uses these stored values along with observed heights to calculate scaling factors and detect target obstacle heights. The reference obstacles effectively calibrate the system themselves through their known geometric properties.
Data Source
AI summary
A method, device, apparatus and a computer-readable storage medium for detecting a height of an obstacle are provided. The method can include: acquiring observation data of a plurality of reference obstacles from a frame; according to the observation data of each of the reference obstacles, fitting a function F: Z=F(ymax), wherein the observation data of a reference obstacle comprises a longitudinal coordinate of a bottom of the reference obstacle in the frame, and a distance between the reference obstacle and a camera capturing the frame; and determining a distance between the obstacle to be detected and the camera, according to the longitudinal coordinate of the bottom of the obstacle in the frame and the function F; and determining an evaluation value of the height of the obstacle to be detected according to the distance between the obstacle to be detected and the camera.