摘要
传统的目标检测识别方法难以适应海量高分辨率遥感影像数据,需要寻求一种能够自动从海量影像数据中学习最有效特征的方法,充分复挖掘数据之间的关联。本文针对海量高分辨率遥感影像数据下典型目标的检测识别,提出一种分层的深度学习模型,通过设定特定意义的分层方法建立目标语义表征及上下文约束表征,以实现高精度目标检测。通过对高分遥感影像目标检测的试验,证明了该方法的有效性。
The traditional object detection and recognition methods do not work well for massive high-resolution remote sensing image data. Thus,we are expected to find an efficient way to automatically learn the presentations from the massive image data,and mine the relationships among the data. This paper proposes a hierarchical deep learning model for object detection and recognition in massive high-resolution remote sensing data. To achieve good performance for object detection,this hierarchical model considers the semantic representation and context information of each layer. The experimental results demonstrated the good performance of the proposed method for object detection in high-resolution remote sensing image data.
出处
《测绘通报》
CSCD
北大核心
2014年第S1期108-111,共4页
Bulletin of Surveying and Mapping
基金
国家自然科学基金(61105014
61170093)
关键词
目标检测
深度学习
高分辨率遥感影像
上下文
object detection
deep learning
high-resolution remote sensing image
context