跳到主要导航 跳到搜索 跳到主要内容

The detection of typical targets under the background of land war

  • Beijing Institute of Technology

科研成果: 书/报告/会议事项章节会议稿件同行评审

摘要

This paper investigates whether advanced neural network techniques can be applied to the detection and identification of typical targets in the context of land warfare. We collected 13 typical targets and built a detection data set. Based on the Faster R-CNN framework, we improve the detection accuracy by two ways. First, we design a neural network model with strong local modeling capabilities. Second, we combine middle layers and the last layer of feature maps as the detection features to enhance the detection ability and improve the detection accuracy.

源语言英语
主期刊名Proceedings of the 29th Chinese Control and Decision Conference, CCDC 2017
出版商Institute of Electrical and Electronics Engineers Inc.
4313-4318
页数6
ISBN(电子版)9781509046560
DOI
出版状态已出版 - 12 7月 2017
活动29th Chinese Control and Decision Conference, CCDC 2017 - Chongqing, 中国
期限: 28 5月 201730 5月 2017

出版系列

姓名Proceedings of the 29th Chinese Control and Decision Conference, CCDC 2017

会议

会议29th Chinese Control and Decision Conference, CCDC 2017
国家/地区中国
Chongqing
时期28/05/1730/05/17

指纹

探究 'The detection of typical targets under the background of land war' 的科研主题。它们共同构成独一无二的指纹。

引用此