计算机与数字工程2019，Vol.47Issue(12)：3182-3186,5.DOI:10. 3969/j. issn. 1672-9722. 2019. 12. 046
基于深度学习的遥感图像新增建筑物语义分割
New Semantic Segmentation of Remote Sensing Image Based on Deep Learning
陈一鸣 1彭艳兵 2高剑飞3
作者信息
- 1. 武汉邮电科学研究院 武汉 430074
- 2. 烽火通信科技股份有限公司南京研发部 南京 210019)
- 3. 烽火通信科技股份有限公司南京研发部 南京 210019
- 折叠
摘要
Abstract
One of the most important tasks in land supervision is to supervise the construction,demolition,renovation and ex?pansion of buildings on the ground. For large cities and their suburbs,it is difficult to rely on staff to patrol the city every day,rely?ing on high-resolution satellite images and deep learning algorithms to innovate existing workflows. This paper designs a new net?work structure based on the deep learning method U-net neural network principle,and adopts the optimization method of stochastic gradient descent and Momentum combination to train the deep learning model to achieve semantic level image segmentation. Using an improved U-net based approach,experiments show that more and more complex new buildings can be identified more accurately.关键词
深度学习/图像语义分割/卫星遥感图像/U-net/神经网络Key words
deep learning/semantic-level image segmentation/satellite remote sensing image/U-net/neural networks分类
信息技术与安全科学引用本文复制引用
陈一鸣,彭艳兵,高剑飞..基于深度学习的遥感图像新增建筑物语义分割[J].计算机与数字工程,2019,47(12):3182-3186,5.
