We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
基于无人机遥感和SAM 的山区房屋提取研究.
- Authors
张卢奔
- Abstract
In order to solve the application difficulties of house information extraction in mountainous areas efficiently and correctly, this paper used drone remote sensing as a geographical environment monitoring method, and adopted emerging artificial intelligence and information technology to construct house extraction model based on the Segment Anything artificial intelligence large model. An efficient and accurate method for extracting information on houses in mountainous areas was developed, and experiments and result analysis on house extraction in typical research areas was conducted. The final extraction method was applied to unmanned aerial vehicle remote sensing images with a CPA of 0. 995 6, house IoU of 0. 861 1, and non-house IoU of 0. 995 4. The experimental results show that, the mountain house extraction method based on drone remote sensing and Segment Anything large model has good analysis and adaptability to complex geographical environments, and can effectively mine the relevant feature information of houses. The extracted houses have the characteristics of complete morphology and accurate edges, meeting the requirements for obtaining information on mountain houses. The method demonstrates the potential application of remote sensing technology and artificial intelligence in mountain area governance, and provides new information technology support for railway route selection and prevention and control of geological disaster.
- Publication
Railway Investigation & Surveying, 2024, Vol 50, Issue 3, p21
- ISSN
1672-7479
- Publication type
Article
- DOI
10.19630/j.cnki.tdkc.202310180002