Topology Map Generation Using OCR and Polygon Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional mapping technologies require significant time and financial resources to create precise maps, limiting their ability to provide quick positioning or route-finding services, especially in rapidly changing environments.
Innovation Solution
A topology map generation apparatus and method that uses optical character recognition (OCR) to detect polygons and characters from guide maps, automatically creating vertices and edges, and storing location information without collecting sensor data at the service site, enabling offline map creation and expansion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional mapping technologies use LIDAR or camera data collection to create precise maps, then mapping precision is improved, but time and financial resources are significantly increased
Solution Approach 1:
The patent uses OCR (optical character recognition) to copy and extract text information from guide maps instead of using LIDAR or camera data collection. This allows precise extraction of location names and pathways from existing map images, avoiding the need for expensive and time-consuming sensor data collection while maintaining mapping precision for position estimation and route finding services
Solution Approach 2:
The patent replaces the mechanical/data collection system (LIDAR, camera) with an information extraction system (OCR, image processing). Instead of physically collecting sensor data at service sites, the system processes guide map images to extract topological information, significantly reducing time and financial resources while achieving the same mapping objectives
2Measurement precision
If data collection is performed at the service site to create maps, then map accuracy is improved, but costs are driven up
Solution Approach 1:
The system copies topological information from existing guide map images rather than collecting new data at service sites. By extracting text and spatial relationships from published guide maps, the system achieves accurate mapping without incurring the high costs associated with on-site sensor data collection, making services like position estimation and route finding economically feasible
Solution Approach 2:
The patent uses inexpensive guide map images as the data source instead of expensive sensor collection equipment. These guide maps serve as disposable, low-cost inputs that provide sufficient topological information for creating accurate service maps, eliminating the need for expensive LIDAR or camera systems while maintaining map accuracy
3Reliability
If services are provided after map generation is completed, then service reliability is improved, but response time is delayed
Solution Approach 1:
The system performs preliminary processing of guide map images to extract topological information before services are needed. By pre-processing the map data into a usable format, the system enables rapid response to position estimation and route finding requests without delaying service provision for map generation, thus improving both reliability and speed
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables efficient and cost-effective creation and expansion of maps for position and route finding services, allowing for convenient services in dynamic environments without additional costs, as multiple maps can be combined and modified easily.
Implementation Method 1
recognizing a character and a character position from the guide map image by using an optical character recognition (OCR)
Implementation Method 2
fill the first polygon with a black color and fill a portion excluding the polygon with a white color, in the guide map image, in order to binarize the guide map image into black and white
Implementation Method 3
generate the skeleton image including a passage, which is an object classified as the white color excluding the black color through a skeletonization algorithm
Implementation Method 4
extract the vertex and the edge from the skeleton image by using a Voronoi diagram
Data Source
AI summary
A topology map generation apparatus may include an image processor configured to detect a polygon from a guide map image, and generate a vertex and an edge based on the detected polygon, and a character processor configured to recognize a character and a character position from the guide map image by using an optical character recognition (OCR), and map the character and the polygon based on the character position, where the image processor stores location information matched to the character in a vertex closest to the character.