Image Boundary Detection for Splitting Merged Property Polygons
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing deep learning based image segmentation techniques struggle to accurately separate and split adjacent properties in high-resolution aerial images, leading to polygon structures that cover multiple properties, which affects applications like urban planning and property evaluation.
Innovation Solution
A machine learning model, specifically a Variational Autoencoder (VAE), is used to revise polygon structures extracted from image segmentation models by determining if and how to split them based on coordinate locations and graphical indications, using a R* search algorithm for selection and a UNET-VGG16 neural network for initial segmentation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If deep learning based image segmentation techniques are used to extract polygon structures from aerial images, then automation and processing speed are improved, but measurement precision and manufacturing precision deteriorate due to inability to accurately separate adjacent properties
Solution Approach 1:
The patent applies segmentation by dividing the polygon structure into multiple portions based on detected boundaries. The boundary detector identifies separation lines between adjacent properties, and the polygon generator creates distinct polygon portions for each property, resolving the issue of merged polygons while maintaining automated processing.
Solution Approach 2:
The patent introduces an intermediary boundary detection module between the initial polygon extraction and final polygon generation. This intermediary step analyzes aerial images to detect boundaries between adjacent properties, allowing the system to refine initial polygon structures and achieve accurate property separation while maintaining automation.
2Extent of automation
If deep learning based image segmentation techniques are used to extract polygon structures from aerial images, then automation is improved, but manufacturing precision deteriorates due to polygon structures covering multiple properties
Solution Approach 1:
The system automatically segments polygon structures into distinct portions corresponding to individual properties. The boundary detector and polygon generator work together to automatically divide merged polygon structures, maintaining high automation while improving the precision of extracted polygon structures.
Solution Approach 2:
An automated boundary detection intermediary is introduced that processes aerial images to identify property boundaries. This intermediary module enables the system to automatically refine initial polygon extractions, ensuring each polygon structure accurately represents a single property without manual intervention.
3Ease of operation
If polygon structures are used to represent properties in aerial images, then ease of operation and data representation are improved, but measurement precision deteriorates when properties are not accurately separated
Solution Approach 1:
The patent segments polygon structures into distinct portions for each property, ensuring that each polygon accurately represents a single property. This segmentation maintains the ease of polygon-based data representation while improving the precision of property separation and boundary definition.
Data Source
AI summary
A computer-implemented system and method for detecting a boundary in an image are provided. The system includes at least one processor and memory in communication with said at least one processor, wherein the memory stores instructions, when executed at said at least one processor, cause said system to: receive or access a first image comprising a first polygon structure; generate, using a data model representing a neural network, a second image based on the first image by splitting the first polygon structure in the first image, wherein the second image comprises a first portion and a second portion partitioned by a line across the first polygon structure; and generate, based on the second image, a geo-image comprising corresponding spatial-reference information for one or more pixels in the geo-image, the geo-image comprising one of the first portion and the second portion in the second image.