| 基于深度学习的居民区识别在电力线规划中的应用 |
| |
| 引用本文： | 马旭,李宝昕,乔新辉,赵晶辉,邰建豪,李怡瑾,李默煊. 基于深度学习的居民区识别在电力线规划中的应用[J]. 电网与清洁能源, 2019, 35(8): 25-33 |
| |
| 作者姓名： | 马旭 李宝昕 乔新辉 赵晶辉 邰建豪 李怡瑾 李默煊 |
| |
| 作者单位： | 1. 北京洛斯达数字遥感技术有限公司,2. 国网陕西省电力公司,1. 北京洛斯达数字遥感技术有限公司,1. 北京洛斯达数字遥感技术有限公司,3. 河南财经政法大学 资源与环境学院，4. 河南财经政法大学 城乡数据挖掘院士工作站,5.国网陕西省电力公司检修公司,5.国网陕西省电力公司检修公司 |
| |
| 基金项目： | 河南省高等学校重点科研项目计划项目（19A420003）；河南省重点研发与推广专项（科技攻关）项目（192102210138） |
| |
| 摘 要： | 针对高压输电线路规划设计中居民区数据难以精准获取的问题,采用人工智能算法,提出了基于深度学习的遥感影像居民区自动识别方法。在Matconvnet框架下,利用迁移学习技术在小样本情况下开发了一套针对多源遥感影像的居民区检测系统。以陕西省宝鸡市为例,对算法进行验证,结果表明,该算法居民区识别精度优于93%,能很好地满足高压线规划设计中对居民区数据的需求。
|
| 关 键 词： | 高压线路设计；深度学习；遥感影像；居民区识别；全卷积网络 |
Application of Residential Area Recognition Based on Deep Learning in Power Line Planning |
| |
| MA Xu,LI Baoxin,QIAO Xinhui,ZHAO Jinghui,TAI Jianhao,LI Yijin and LI Moxuan. Application of Residential Area Recognition Based on Deep Learning in Power Line Planning[J]. Power system and clean energy, 2019, 35(8): 25-33 |
| |
| Authors: | MA Xu LI Baoxin QIAO Xinhui ZHAO Jinghui TAI Jianhao LI Yijin LI Moxuan |
| |
| Abstract: | Aiming at the problem that residential area data are difficult to obtain accurately in the planning and design of high voltage transmission lines, an automatic residential area recognition method based on deep learning is proposed by using artificial intelligence algorithm. In the framework of Matconvnet, a residential area detection system with multi-resolution remote sensing images is developed by using the transfer learning technology in the case of small samples. Taking Baoji City of Shaanxi Province as an example, the system is verified. The results show that the accuracy of residential area extraction is better than 93%, which can better meet the needs of residential area data in high-voltage transmission line planning and design. |
| |
| Keywords: | |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《电网与清洁能源》浏览原始摘要信息 |
|
点击此处可从《电网与清洁能源》下载免费的PDF全文 |
|
