Predicting Virtual Road Sign Locations Using Deep Neural Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Augmented reality navigation systems face challenges in accurately placing virtual road signs in relation to the physical environment, especially at complex intersections, leading to inconsistencies between the location of augmenting information and displayed scene images.
Innovation Solution
A computer-implemented method using a deep neural network trained with aerial and satellite images and geocentric positions of key point markers to predict virtual road sign locations, allowing for precise superimposition of virtual road signs onto environmental data, ensuring accurate placement and visibility.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If augmented reality navigation systems overlay digital road database data with scene images, then navigation information can be displayed, but accurate placement of virtual road signs becomes difficult at complicated intersections
Solution Approach 1:
The system uses aerial and satellite images as reference copies of the physical environment to train the deep neural network. These reference images contain labeled key point marker positions that serve as ground truth for teaching the network how to accurately locate virtual road signs in complex intersections, thereby improving placement accuracy without increasing real-time processing complexity
Solution Approach 2:
The patent replaces traditional mechanical/map-based navigation systems with a deep neural network-based image processing system. The neural network automatically learns to identify and locate key point markers in aerial/satellite images, substituting complex manual mapping processes with automated AI-based recognition and prediction
2Measurement precision
If high-definition maps are used for precise navigation, then navigation accuracy improves, but memory space requirements increase
Solution Approach 1:
The system extracts only the essential information needed for navigation from large aerial and satellite images - specifically the locations of key point markers such as intersections and landmarks. By taking out only these critical positional data points rather than storing complete high-definition maps, the system achieves navigation accuracy while significantly reducing memory space requirements
Solution Approach 2:
Instead of using expensive, large-scale high-definition maps that consume significant memory, the system uses a database of predicted key point marker locations derived from processed aerial/satellite images. These condensed location data serve as sufficient navigation references without requiring the extensive storage space of full HD maps
Data Source
AI summary
Provided are a computer-implemented method and apparatus for predicting virtual road sign locations of virtual road signs that may be superimposed onto environmental data of a vehicle. The method includes collecting, as a first training data subset, one or more aerial and/or satellite images of a pre-determined region; obtaining, as a second training data subset, geocentric positions of key point markers in the pre-determined region; supplying the first and second training data subsets to a deep neural network as training dataset; training the deep neural network on the training dataset to predict key point marker locations in a region of interest, the key point marker locations corresponding to virtual road sign locations; defining a region of interest as input dataset; and processing the input dataset by the trained deep neural network to predict key point marker locations within the defined region of interest.