摘要
山岳型景区游客信息实时感知是智慧景区建设的核心内容之一,本文充分利用山岳型景区分布式摄像头近场感知网的优势,提出了联合分布式摄像头的山岳型景区游客目标动态精准被动感知方法体系。首先,通过注意力模块(CBAM)和自适应空间特征融合(ASFF)技术改进YOLOX网络,完成复杂场景视频流的游客目标初步识别;进而,引入目标动态跟踪算法改善山岳景区游客动态目标遮挡问题,进一步提升模型检测精度和稳定性;在此基础上,通过检测目标的像素坐标空间到地理坐标空间的转化定位算法推导出游客的时间位置信息。结果表明,该方法能够很好地协助完成山岳型景区复杂场景下游客的近场被动感知与空间定位,游客动态目标被动感知总体识别精度达到90%以上,空间定位误差RMSE小于1.1094,为山岳型景区无/弱卫星导航信号情况下区域游客目标的被动实时动态精准感知与安全管理提供了技术方案和信息支持。
Mountainous scenic areas rely on distributed camera networks for real-time visitor perception,a crucial aspect of intelligent scenic area development.This paper exploits these networks to propose a method for dynamic and precise passive perception of visitors in mountainous scenic areas through distributed cameras.By enhancing the YOLOX network with a convolutional block attention module(CBAM)and adaptive spatial feature fusion(ASFF)techniques,and incorporating a dynamic target tracking algorithm,this method achieves accurate detection and tracking of visitors'movements.Subsequently,a localization algorithm is utilized to derive real-time positional information of visitors.The results demonstrate the method's effectiveness in near-field passive perception and spatial positioning within mountainous scenic areas,achieving a detection precision over 90%,spatial positioning accuracy within 1 m,and a root mean square error(RMSE)of less than 1.1094.This provides a technical solution and information support for the passive real-time dynamic precise perception and safety management of visitors in areas with weak or no satellite navigation signals.
作者
施坤涛
朱长明
张新
杨帆
张琨
高宏进
SHI Kuntao;ZHU Changming;ZHANG Xin;YANG Fan;ZHANG Kun;GAO Hongjin(School of Geography,Geomatics and Planning,Jiangsu Normal University,Xuzhou 221116,China;State Key Laboratory of Remote Sensing Science,Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100101,China;Qilu Institute of Aerospace Information,Jinan 250101,China)
出处
《测绘通报》
北大核心
2025年第1期72-77,共6页
Bulletin of Surveying and Mapping
基金
国家重点研发计划(2021YFB1407004,2023YFE0103800)
江苏高校优势学科建设工程资助
江苏师范大学研究生科研与实践创新计划校级项目(2019XKT041)。
关键词
目标识别
近场感知
时空分布
游客信息
山岳景区
object detection
near-field perception
spatio-temporal distribution
dynamic management
mountainous scenic areas
