Drone Landing Position Determination Using Image and Sentence Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in accurately landing autonomous drones at user-designated positions, particularly in narrow areas like a house entrance, due to GPS errors and the impracticality of user-created markers.
Innovation Solution
An information processing method that analyzes a captured image using object identification and sentence analysis to determine a target arrival position based on user input, including latitude/longitude data and a descriptive sentence, enabling precise landing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If GPS position information is used for landing, then the drone can autonomously navigate to the destination, but the landing precision is insufficient due to several meters error in latitude and longitude
Solution Approach 1:
The landing system is segmented into multiple components: GPS for coarse positioning, camera for visual capture, and image processing for fine positioning. This segmentation allows each component to operate at its optimal precision level, with the camera capturing detailed ground features that GPS cannot resolve.
Solution Approach 2:
The camera acts as an intermediary between GPS and the landing target. It captures visual information of the ground area around the GPS-coarse-position, and image processing extracts precise location features (like doors, windows, or specific landmarks) that serve as the final landing reference, bridging the gap between low-precision GPS and high-precision landing requirements.
2Measurement precision
If a marker is recorded at the landing target position for the drone to capture, then the landing precision can be improved, but the user burden becomes heavy and it becomes practically difficult
Solution Approach 1:
The system uses naturally existing ground features (doors, windows, walls, vegetation, roads) as landing markers instead of requiring users to install artificial markers. The image processing algorithm automatically identifies and selects these features as landing targets, making the system self-sufficient and eliminating the need for user intervention in marker placement.
Solution Approach 2:
The image processing system is designed to recognize multiple types of ground features (buildings, doors, windows, natural landmarks, roads) as potential landing targets. This universality allows the drone to adapt to various environments without requiring environment-specific marker installation, making the system applicable to diverse locations worldwide.
3Measurement precision
If the drone lands at a charging system position with special equipment, then the landing precision can be ensured, but the system cannot be applied to positions without special equipment such as user's yard or in front of entrance
Solution Approach 1:
The patent replaces mechanical/specialized landing guidance equipment (transmitters, charging system infrastructure) with an optical/image-based system. The camera captures visual information and image processing algorithms identify landing targets based on visual features alone, eliminating the need for specialized mechanical infrastructure at the landing site.
Solution Approach 2:
The system changes the parameter basis for landing from electromagnetic signals requiring special transmitters to visual parameters (color, shape, texture, patterns) that can be detected by standard cameras. This parameter change enables the system to operate in any environment with visual features, dramatically increasing adaptability to different locations.
Data Source
AI summary
A configuration is described that is capable of determining a target landing position of a drone according to a sentence indicating a landing position generated by a user. Aspects including identifying an object type of an object contained in a captured image of a camera mounted on a mobile body; sentence analysis processing analyzes a sentence indicating a designated arrival position; word-corresponding subject selection processing collates results of the sentence analysis processing with the object identification processing and selecting a subject corresponding to a word included in the sentence indicating the designated arrival position from the captured image; and target arrival position determination processing extracts an area corresponding to the designated arrival position indicated by the sentence indicating the designated arrival position from the captured image on the basis of a result of word-corresponding subject selection processing, and area determination is made as target arrival position of mobile body.