摘 要:针对人员密集场所人流量统计准确度较低、实时性较差的问题,设计开发一种在一定区域内可以根据视频来统计人流量的系统。提出采用基于深度学习框架 Tensorflow 的物体识别算法(SSD 算法)进行人流量分析,以达到在人流量密集的公共场所对人流量进行监测的目的,为管理人员提供更加准确的、直观的人流量数据信息,方便管理人及时进行调控与管理,以做出更为合理的决策。
关键词:深度学习;OpenCV;SSD
DOI:10.19850/j.cnki.2096-4706.2022.011.003
基金项目:广西研究生教育创新计划项目(YCSW2022125)
中国分类号:TP391.4 文献标识码:A 文章编号:2096-4706(2022)11-0011-04
Design of Pedestrian Flow Calculation System Based on Tensorflow Framework
HUANG Yingjie
(Guangxi Normal University, Guilin 541006, China)
Abstract: Aiming at the problem of lower accuracy and poorer real-time performance of pedestrian flow statistics in densely populated places, a system that can count the pedestrian flow according to the video in a certain area is designed and developed. This paper proposes to analyze the pedestrian flow by using Single Shot MultiBox Detector algorithm (SSD algorithm) based on deep learning framework Tensorflow, so as to achieve the purpose of monitoring the pedestrian flow in densely populated public places. And it provides managers with more accurate and intuitive pedestrian flow data information, convenient for managers to timely control and management, in order to make more reasonable decisions.
Keywords: deep learning; OpenCV; SSD
参考文献:
[1] 冉涌 . 基于百度 AI 的人流量监测系统 [J]. 黄冈职业技术学院学报,2020,22(6):143-146.
[2] 郭敏钢,宫鹤 . 基于 Tensorflow 对卷积神经网络的优化研究 [J]. 计算机工程与应用,2020,56(1):158-164.
[3] 邢萌.基于TensorFlow手写数字识别模型设计与实现 [J].电子技术与软件工程,2019(2):56.
[4] 毛世昕,李捍东 . 改进 SSD 算法在交通场景中的检测研究[J]. 微处理机,2022,43(1):26-29+33.
[5] 孙超,温蜜,景俐娜 . 改进 SSD 算法在交通标志检测中的应用 [J/OL]. 计算机工程与应用,2021:1-10 [2022-04-09].https://www.cnki.net/KCMS/detail/detail.aspx?dbcode=CA PJ&dbname=CAPJLAST&filename=JSGG20210926005&v=MTc2 NzNORE1wbzFEWk9zS1l3azd2QkFTNmpoNFRBemxxMkEwZk xUN1I3cWRaZVpvRml6bFZydklJMVk9THo3TWFiRzRI.
[6] 任焕海 . 基于 B/S 架构的在线考试系统设计与实现 [J]. 现代信息科技,2021,5(22):13-16.
作者简介:黄盈洁(1999—),女,汉族,湖北孝感人,硕士在读,研究方向:人工智能教育应用。