Computer Engineering and Applications ›› 2021, Vol. 57 ›› Issue (18): 65-74.DOI: 10.3778/j.issn.1002-8331.2105-0034

Previous Articles     Next Articles

Review of Remote Sensing Image Target Detection Research Combining Knowledge and CNN

GAO Yuge, YANG Haitao, WANG Jinyu, LI Gaoyuan, ZHANG Changgong, FENG Bodi   

  1. 1.Graduate School, Space Engineering University, Beijing 101416, China
    2.School of Space Information, Space Engineering University, Beijing 101416, China
  • Online:2021-09-15 Published:2021-09-13

联合知识与CNN的遥感影像目标检测研究综述

高宇歌,杨海涛,王晋宇,李高源,张长弓,冯博迪   

  1. 1.航天工程大学 研究生院,北京 101416
    2.航天工程大学 航天信息学院,北京 101416

Abstract:

Target detection is an important part of the intelligent interpretation of remote sensing images and a key link in converting images into information. The knowledge-based method is the traditional method of remote sensing image target detection, and the deep learning method based on convolutional neural network is the mainstream method that has gradually emerged in recent years and is rapidly applied on a large scale. This paper introduces the methods based on geometric knowledge, context knowledge, auxiliary knowledge, and comprehensive knowledge, as well as one-stage and two-stage convolutional neural network methods, focusing on the new method of joint knowledge and convolutional neural network. In addition, three specific application forms such as improving remote sensing image data sets, adjusting algorithm network framework, and achieving target context reasoning are introduced in detail. The remote sensing image target detection method based on the joint knowledge and convolutional neural network method is prospected.

Key words: target detection, remote sensing image, knowledge, convolutional neural network

摘要:

目标检测是遥感影像智能解译的重要内容,是将影像转换为信息的关键环节。基于知识的方法是遥感影像目标检测的传统经典方法,而基于卷积神经网络的深度学习方法则是近年来逐步兴起并迅速大范围应用的主流方法。介绍了基于几何知识、上下文知识、辅助知识、综合知识的方法,以及一阶段、两阶段的卷积神经网络方法,重点论述了联合知识与卷积神经网络的新方法,并对改进遥感影像数据集、调整算法网络框架、实现目标上下文推理等三种具体应用形式进行了详细介绍。对联合知识与卷积神经网络方法的遥感影像目标检测方法进行了展望。

关键词: 目标检测, 遥感影像, 知识, 卷积神经网络