Skip to main content

ORIGINAL RESEARCH article

Front. Neurorobot.
Volume 18 - 2024 | doi: 10.3389/fnbot.2024.1448538
This article is part of the Research Topic Multi-source and Multi-domain Data Fusion and Enhancement: Methods, Evaluation, and Applications View all articles

Target Detection and Classification via EfficientDet and CNN over Unmanned Aerial Vehicles

Provisionally accepted
Muhammad O. Yusuf Muhammad O. Yusuf 1Muhammad Hanzla Muhammad Hanzla 1Naif Al Mudawi Naif Al Mudawi 2Touseef Sadiq Touseef Sadiq 3*Bayan Alabdullah Bayan Alabdullah 4Hameedur Rahman Hameedur Rahman 1*Asaad Algarni Asaad Algarni 5
  • 1 Air University, Islamabad, Pakistan
  • 2 Najran University, Najran, Saudi Arabia
  • 3 University of Agder, Kristiansand, Vest-Agder, Norway
  • 4 Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia
  • 5 Northern Border University, Arar, Northern Borders, Saudi Arabia

The final, formatted version of the article will be published soon.

    Advanced traffic monitoring systems face major vehicle detection and classification challenges. Conventional methods need significant computational resources and struggle to adapt to various data collection methods. This research introduces an innovative technique for classifying and recognizing vehicles in aerial image sequences. The proposed model consists of five phases. The first stages of image enhancement include noise reduction and Contrast Limited Adaptive Histogram Equalization (CLAHE). Contour-Based segmentation and Fuzzy C-means segmentation (FCM) are used to distinguish foreground items, followed by EfficientDet for vehicle detection and identification in each image. Accelerated KAZE (AKAZE), Oriented FAST, Rotated BRIEF (ORB) and Scale Invariant Feature Transform (SIFT) are used for feature extraction. A convolutional neural network (CNN) and ResNet Residual Network are used to classify objects into several classes. This method outperforms the previous approaches, according to experiments conducted on several datasets such as Vehicle Aerial Imagery from a Drone (VAID), Unmanned Aerial Vehicle Intruder Dataset (UAVID), and others. The proposed model operates with 96.6% accuracy on UAVID and 97% accuracy on VAID. Our approach outperforms existing methods in the industry in terms of vehicle detection and classification in aerial images.

    Keywords: deep learning, unmanned aerial vehicles, remote sensing, Dynamic environments, path planning, multi-objects recognition Deep learning, multi-objects recognition

    Received: 13 Jun 2024; Accepted: 15 Aug 2024.

    Copyright: © 2024 Yusuf, Hanzla, Al Mudawi, Sadiq, Alabdullah, Rahman and Algarni. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

    * Correspondence:
    Touseef Sadiq, University of Agder, Kristiansand, 4604, Vest-Agder, Norway
    Hameedur Rahman, Air University, Islamabad, Pakistan

    Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.