The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Publications Copernicus
Download
Citation
Articles | Volume XLIII-B2-2021
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLIII-B2-2021, 793–800, 2021
https://doi.org/10.5194/isprs-archives-XLIII-B2-2021-793-2021
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLIII-B2-2021, 793–800, 2021
https://doi.org/10.5194/isprs-archives-XLIII-B2-2021-793-2021

  28 Jun 2021

28 Jun 2021

A NOVEL DEEP LEARNING BASED METHOD FOR DETECTION AND COUNTING OF VEHICLES IN URBAN TRAFFIC SURVEILLANCE SYSTEMS

J. J. Majin1, Y. M. Valencia1, M. E. Stivanello3, M. R. Stemmer1, and J. D. Salazar2 J. J. Majin et al.
  • 1Automation and Systems Department, Universidade Federal de Santa Catarina (UFSC), Florianopólis, SC, Brazil
  • 2Mechanical Engineering Department, Labmetro, Universidade Federal de Santa Catarina (UFSC), Florianópolis, SC, Brazil
  • 3Academic Department of Metal-Mechanics, Instituto Federal Santa Catarina (IFSC), Florianópolis, SC, Brazil

Keywords: Deep learning, DeepSORT, Object detection, Traffic monitoring, Vehicle counting, YOLOv4

Abstract. In intelligent transportation systems (ITS), it is essential to obtain reliable statistics of the vehicular flow in order to create urban traffic management strategies. These systems have benefited from the increase in computational resources and the improvement of image processing methods, especially in object detection based on deep learning. This paper proposes a method for vehicle counting composed of three stages: object detection, tracking and trajectory processing. In order to select the detection model with the best trade-off between accuracy and speed, the following one-stage detection models were compared: SSD512, CenterNet, Efficiedet-D0 and YOLO family models (v2, v3 and v4). Experimental results conducted on the benchmark dataset show that the best rates among the detection models were obtained using YOLOv4 with mAP = 87% and a processing speed of 18 FPS. On the other hand, the accuracy obtained in the proposed counting method was 94% with a real-time processing rate lower than 1.9.