Real-Time Object Detection and Tracking Based on Embedded Edge Devices for Local Dynamic Map Generation

被引:1
|
作者
Choi, Kyoungtaek [1 ]
Moon, Jongwon [2 ]
Jung, Ho Gi [3 ]
Suhr, Jae Kyu [4 ]
机构
[1] Daegu Catholic Univ, Dept Robot Engn, 13-13 Hayang Ro,Hayang Eup, Gyongsan 38430, South Korea
[2] LIG Nex1, IIR Seeker R&D Ctr, Mabuk Ro 207, Yongin 16911, South Korea
[3] Korea Natl Univ Transportat, Dept Elect Engn, 50 Daehak Ro, Chungju 27469, South Korea
[4] Sejong Univ, Dept Intelligent Mechatron Engn, 209 Neungdong Ro, Seoul 05006, South Korea
关键词
edge computing; network slimming embedded devices; object detector; object tracker; Qualcomm SoC; camera for ITS;
D O I
10.3390/electronics13050811
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
This paper proposes a camera system designed for local dynamic map (LDM) generation, capable of simultaneously performing object detection, tracking, and 3D position estimation. This paper focuses on improving existing approaches to better suit our application, rather than proposing novel methods. We modified the detection head of YOLOv4 to enhance the detection performance for small objects and to predict fiducial points for 3D position estimation. The modified detector, compared to YOLOv4, shows an improvement of approximately 5% mAP on the Visdrone2019 dataset and around 3% mAP on our database. We also proposed a tracker based on DeepSORT. Unlike DeepSORT, which applies a feature extraction network for each detected object, the proposed tracker applies a feature extraction network once for the entire image. To increase the resolution of feature maps, the tracker integrates the feature aggregation network (FAN) structure into the DeepSORT network. The difference in multiple objects tracking accuracy (MOTA) between the proposed tracker and DeepSORT is minimal at 0.3%. However, the proposed tracker has a consistent computational load, regardless of the number of detected objects, because it extracts a feature map once for the entire image. This characteristic makes it suitable for embedded edge devices. The proposed methods have been implemented on a system on chip (SoC), Qualcomm QCS605, using network pruning and quantization. This enables the entire process to be executed at 10 Hz on this edge device.
引用
收藏
页数:16
相关论文
共 50 条
  • [1] BED: A Real-Time Object Detection System for Edge Devices
    Wang, Guanchu
    Bhat, Zaid Pervaiz
    Jiang, Zhimeng
    Chen, Yi-Wei
    Zha, Daochen
    Reyes, Alfredo Costilla
    Niktash, Afshin
    Ulkar, Gorkem
    Okman, Erman
    Cai, Xuanting
    Hu, Xia
    [J]. PROCEEDINGS OF THE 31ST ACM INTERNATIONAL CONFERENCE ON INFORMATION AND KNOWLEDGE MANAGEMENT, CIKM 2022, 2022, : 4994 - 4998
  • [2] An Edge-based Real-Time Object Detection
    Ahmadinia, Ali
    Shah, Jaabaal
    [J]. 2022 21ST IEEE INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND APPLICATIONS, ICMLA, 2022, : 465 - 470
  • [3] Real-time embedded object detection and tracking system in Zynq SoC
    Qingbo Ji
    Chong Dai
    Changbo Hou
    Xun Li
    [J]. EURASIP Journal on Image and Video Processing, 2021
  • [4] Real-time embedded object detection and tracking system in Zynq SoC
    Ji, Qingbo
    Dai, Chong
    Hou, Changbo
    Li, Xun
    [J]. EURASIP JOURNAL ON IMAGE AND VIDEO PROCESSING, 2021, 2021 (01)
  • [5] A Hardware/Software Co-Design Approach for Real-Time Object Detection and Tracking on Embedded Devices
    Bui, Casey
    Patel, Nirali
    Patel, Dishant
    Rogers, Samuel
    Sawant, Adarsh
    Manwatkar, Rishiraj
    Tabkhi, Hamed
    [J]. IEEE SOUTHEASTCON 2018, 2018,
  • [6] Real-Time Object Detection and Tracking
    Naeem, Hammad
    Ahmad, Jawad
    Tayyab, Muhammad
    [J]. 2013 16TH INTERNATIONAL MULTI TOPIC CONFERENCE (INMIC), 2013, : 148 - 153
  • [7] Real-time Surveillance based Crime Detection for Edge Devices
    Venkatesh, Sai Vishwanath
    Anand, Adithya Prem
    Sahar, Gokul S.
    Ramakrishnan, Akshay
    Vijayaraghavan, Vineeth
    [J]. VISAPP: PROCEEDINGS OF THE 15TH INTERNATIONAL JOINT CONFERENCE ON COMPUTER VISION, IMAGING AND COMPUTER GRAPHICS THEORY AND APPLICATIONS, VOL 4: VISAPP, 2020, : 801 - 809
  • [8] Real-time detection and tracking of moving object
    Tao, Jianguo
    Yu, Changhong
    [J]. 2008 INTERNATIONAL SYMPOSIUM ON INTELLIGENT INFORMATION TECHNOLOGY APPLICATION, VOL II, PROCEEDINGS, 2008, : 860 - 863
  • [9] Real-time attention-based embedded LSTM for dynamic sign language recognition on edge devices
    Sharma, Vaidehi
    Sharma, Abhishek
    Saini, Sandeep
    [J]. JOURNAL OF REAL-TIME IMAGE PROCESSING, 2024, 21 (02)
  • [10] Real-time attention-based embedded LSTM for dynamic sign language recognition on edge devices
    Vaidehi Sharma
    Abhishek Sharma
    Sandeep Saini
    [J]. Journal of Real-Time Image Processing, 2024, 21