Automatic Detection of Dead-End Roads in High-Resolution Satellite Image

2025-01-7165

02/21/2025

Event
2024 International Conference on Smart Transportation Interdisciplinary Studies
Authors Abstract
Content
There are dead-end roads in the road network, and many of them have the function of indicating specific target clues, which is of great significance in the fields of military, urban construction, and disaster relief and rescue. However, many of the important cut-offs are in mountainous or wilderness areas, and surveying them is difficult and costly. The research objective of this project is to extract the breakpoints in the road network using high-resolution Google satellite imagery, so as to provide clues and indications for the subsequent relevant work. Firstly, the image is corrected and pre-processed to highlight road edge information.Then the phase grouping method is improved by setting a double-angle threshold, filtering the edge operator to reduce the calculation error of the gradient angle, and the road network is extracted by the improved phase grouping method. And finally screens out the dead-end road points with the eight-neighbourhood method, and marks them on the experimental image, so as to realize the identification and positioning of the dead-end roads.The algorithm has a certain degree of universality, after replacing the experimental area and image, it can still locate the dead-end road breakpoints accurately, and the universal accuracy rate reaches more than 80%, and after adjusting the threshold value for a specific area, the extraction accuracy rate can reach more than 90%. This project realizes the technology of identifying and locating the dead-end road based on the phase grouping method, which can intelligently and automatically identify and locate the dead-end roads on the remote sensing image, and the improvement of the phase grouping method significantly improves the extraction precision of the road network and the extraction precision of the broken point, which is of practical value.
Meta TagsDetails
Pages
11
Citation
Liu, R., Haoping, Q., Jingjie, K., Yanyan, W. et al., "Automatic Detection of Dead-End Roads in High-Resolution Satellite Image," SAE Technical Paper 2025-01-7165, 2025, .
Additional Details
Publisher
Published
Feb 21
Product Code
2025-01-7165
Content Type
Technical Paper
Language
English