Monocular UAV Target Distance Estimation Without Ground Assumptions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for estimating distance to a target using a monocular camera on an unmanned aerial vehicle are inaccurate due to reliance on unstable height data and assumptions about level ground, and cannot estimate distance when the vehicle is hovering.
Innovation Solution
A method and device that utilize image processing to calculate the distance to a target by acquiring a current frame of image, determining the target's location and attitude information, and applying geometric calculations to estimate the distance based on angles and focal lengths of a monocular camera.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If height data is used to calculate distance, then distance estimation is possible, but the accuracy deteriorates due to unstable height data and ground level assumptions
Solution Approach 1:
The patent introduces an intermediary geometric model (smallest circumscribed rectangle, angle calculations) to bridge the gap between monocular image data and three-dimensional distance estimation, avoiding direct reliance on unstable height data and ground level assumptions
Solution Approach 2:
The patent replaces the mechanical/physical measurement system (height sensors, ground level detection) with an optical-geometric calculation system based on monocular image analysis and angular relationships
2Measurement precision
If triangulation method is used, then distance estimation accuracy improves, but the system cannot estimate distance when the unmanned aerial vehicle is hovering without displacement
Solution Approach 1:
The patent performs preliminary geometric model setup and parameter calibration during the hovering state, allowing distance estimation to be performed without requiring subsequent displacement or movement of the unmanned aerial vehicle
Solution Approach 2:
The patent transforms the static triangulation method (requiring displacement) into a dynamic geometric calculation method that can operate in both hovering and moving states by using real-time angular measurements and image processing
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
The present invention relates to a method and device for estimating a distance to a target, and a drone. The method is applicable to a drone including a photographing device. The method includes: acquiring a current frame of image of the target captured by the photographing device; acquiring location information of the target according to the current frame of image, where the location information includes a height of the target and two-dimensional pixel coordinates of the target in the image; acquiring attitude information of the photographing device, where the attitude information includes a pitch angle of the photographing device; and acquiring a distance between the target and the drone according to the location information of the target and the attitude information of the photographing device. The present invention can accurately calculate the distance between the target and the drone, is not based on a strong assumption, requires a small amount of calculation, can estimate the spatial location information of the target in real time, and has high accuracy.