摘要
面对观测任务复杂化、观测目标多样化的趋势,亟需发展观测水平更高、成像效能更优的空间光学遥感技术,而深度学习作为人工智能领域的研究热点,其在事件决策、图像处理等方面能力优越,已经被用于各类应用场景。文章从空间光学遥感技术目前存在的主要问题入手,说明深度学习在天基光学遥感中应用的必要性,提出深度学习的应用设想,特别结合深度学习的能力优势,重点介绍深度学习在天基智能遥感中的创新应用及相关技术的研究进展,最后对天基智能遥感技术未来的发展趋势进行了展望。
In the face of complex observation tasks and diversified observation targets,it is urgent to develop the space optical remote sensing technology with higher observation level and better imaging efficiency.Deep learning is a hot research content in the field of artificial intelligence.It has advantages in event decision-making,image processing and so on.It has been used in various application scenarios and promoted the technical development of in related technical fields.This paper starts with the main problems of the space optical remote sensing technology,explains the necessity of deep learning in space-based optical remote sensing,puts forward the application ideas of deep learning,especially introduces the innovative application of deep learning in space-based intelligent remote sensing and the research progress of related technologies,finally makes the summary and prospect.
作者
李维
刘勋
张维畅
阮宁娟
LI Wei;LIU Xun;ZHANG Weichang;RUAN Ningjuan(Beijing Institute of Space Mechanics&Electricity,Beijing 100094,China)
出处
《航天返回与遥感》
CSCD
2020年第6期56-65,共10页
Spacecraft Recovery & Remote Sensing
关键词
智能技术
深度学习
虚拟样本
自动生成
全光学
航天遥感
intellectual technology
deep learning
virtual sample
automatic generation
all optical
space remote sensing