Virtual Drone Camera Image Matching for Precise Landing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous drones face challenges in accurately landing at specific positions, especially in narrow areas like user's gardens or house entrances, due to limitations in GPS position measurement accuracy and the need for special equipment setups.
Innovation Solution
An information processing system and method that generates a 'virtual drone camera image' based on a captured image from a user terminal, allowing the drone to align with the planned landing position using pixel positional relationships, enabling accurate landing without requiring special equipment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If GPS position control is used for autonomous drone landing, then the drone can fly autonomously without visual monitoring, but the landing position accuracy is limited to about 1m error which is insufficient for narrow areas
Solution Approach 1:
The patent introduces an image recognition system as an intermediary between GPS positioning and the drone's navigation. The system captures images of the landing area, identifies feature points, and calculates precise position corrections based on the relationship between GPS coordinates and actual image features. This intermediary system bridges the gap between coarse GPS positioning and fine landing accuracy requirements.
Solution Approach 2:
The patent creates a digital copy of the landing area by capturing images and extracting feature points. This digital model is then used to calculate position corrections and guide the drone to the precise landing location. The feature point coordinates serve as a copied reference that allows the drone to compensate for GPS errors and achieve accurate landing in narrow areas.
2Measurement precision
If special equipment such as transmitters and optical sensors are added to achieve accurate landing, then landing position control accuracy improves, but the device complexity and cost increase significantly
Solution Approach 1:
The patent makes the smartphone camera serve multiple functions: it acts as both the imaging device for capturing the landing area and the processing unit for image recognition and position calculation. This universal use of existing equipment eliminates the need for specialized landing guidance hardware while maintaining high positioning accuracy.
Solution Approach 2:
The system uses the drone's own onboard camera to capture images of the landing area and perform self-positioning. The drone independently processes its own imaging data to calculate position corrections and guide its landing, without requiring external transmitters, optical sensors, or other specialized equipment. This self-service approach simplifies the overall system configuration.
3Measurement precision
If external transmitters and optical sensors are deployed at the landing position, then accurate landing can be achieved, but the ease of operation and deployment difficulty increase for user locations without special equipment
Solution Approach 1:
The system performs preliminary image capture and feature point extraction at the desired landing location before the drone arrives. The smartphone takes pictures of the target area, identifies feature points, and stores their coordinates. When the drone approaches, it uses these pre-prepared reference data to achieve accurate landing without requiring any physical equipment to be installed at the location.
Solution Approach 2:
Instead of requiring physical transmitters or optical markers to be installed at the landing site, the system creates a digital copy of the location using smartphone camera images. This digital reference model can be easily deployed at any location without physical infrastructure, making the system highly accessible for user locations such as gardens or house entrances.
Data Source
AI summary
A user terminal generates a virtual drone camera image, as an estimated captured image where it is assumed that a virtual drone camera mounted on a drone has captured an image of a planned landing position on the basis of a captured image obtained by capturing the planned landing position of the drone with the user terminal, and transmits the generated virtual drone camera image to the drone. The drone collates the virtual drone camera image with the image captured by the drone camera and lands at the planned landing position in the image captured by the drone camera. The user terminal generates a corresponding pixel positional relationship formula indicating a correspondence relationship between a pixel position on the captured image of the user terminal and a pixel position on the captured image of the virtual drone camera, and generates the virtual drone camera image using the generated relationship formula.