| 基于深度学习的目标检测定位系统设计与实现 |
| |
| 引用本文： | 王俊强,李建胜,程相博,杨戬峰.基于深度学习的目标检测定位系统设计与实现[J].测绘与空间地理信息,2020(1):133-136,144. |
| |
| 作者姓名： | 王俊强 李建胜 程相博 杨戬峰 |
| |
| 作者单位： | 78123部队;信息工程大学 |
| |
| 摘 要： | 针对传统遥感影像目标检测方法效率不高,并且无有效手段对检测信息进行管理利用的问题,提出了在B/S构架下基于深度学习的目标检测及定位方法。通过集成深度学习框架、WebGIS以及数据库,实现了集遥感影像目标检测、展示及管理于一体的目标检测定位系统,满足多用户基于前端浏览器的并发目标检测需求。利用网格划分策略,实现了基于前端的大区域范围的目标快速检测。基于某机场飞机目标及某城市区域运动场目标检测结果表明:本文设计的目标检测定位系统能够在前端实现目标快速检测定位,具有较高检测精度,并可有效管理检测信息,为深度学习循环再利用提供数据支撑。
|
| 关 键 词： | 深度学习 B/S架构 目标检测 卷积神经网络 网格 数据库 |
Design and Implement of Object Detection and Location System Based on Deep Learning |
| |
| WANG Junqiang,LI Jiansheng,CHENG Xiangbo,YANG Jianfeng.Design and Implement of Object Detection and Location System Based on Deep Learning[J].Geomatics & Spatial Information Technology,2020(1):133-136,144. |
| |
| Authors: | WANG Junqiang LI Jiansheng CHENG Xiangbo YANG Jianfeng |
| |
| Institution: | (78123 Troops,Chengdu 610000,China;Information Engineering University,Zhengzhou 450000,China) |
| |
| Abstract: | According to the fact that the efficiency of the method of traditional remote sensing image detection is not high,and the management and utilization of the object information is not effective,this paper presented the method of object detection and location based on deep learning under the framework of Browser/Server(B/S).By integrating the deep learning framework,WebGIS and database,the object detection and location system which integrates remote sensing object detection,display and management was realized to meet the requirement of concurrent object detection based on the multi-user browser.By using the strategy of mesh generation,the rapid detection of object in large area based on the front-end is realized.The results of plane detection in one airport and playground detection in one city indicated that the object detection and location system designed in this paper could realize fast detection and positioning in the browser,which had high precision,and can manage the detection information effectively and provide data support for the reuse of the deep learning cycle. |
| |
| Keywords: | deep learning Browser/Server object detection convolution neural network grid database |
| 本文献已被 CNKI 维普 等数据库收录！ |
|
