Object detection in adverse weather condition for autonomous vehicles

被引:0
|
作者
Emmanuel Owusu Appiah
Solomon Mensah
机构
[1] University of Ghana,Department of Computer Science
来源
Multimedia Tools and Applications | 2024年 / 83卷
关键词
Deep Learning; Object Detection; Autonomous vehicles; Adverse weather condition;
D O I
暂无
中图分类号
学科分类号
摘要
As self-driving or autonomous vehicles proliferate in our society, there is a need for their computing vision systems to be able to identify objects accurately, no matter the weather condition. One major concern in computer vision is improving an autonomous car’s capacity to discern between the components of its environment under challenging conditions. For instance, inclement weather like fog and rain can corrupt images which eventually affect how well autonomous vehicles navigate and localise themselves. To provide an efficient and effective approach for autonomous vehicles to accurately detect objects during adverse weather conditions. The study employed the combination of two deep learning approaches, namely YOLOv7 and ESRGAN. The use of ESRGAN is to first learn from a set of training data and adjust for the unfavourable weather conditions in the images before the YOLOv7 detector performs detection of objects. The use of the ESRGAN allowed for the adaptive enhancement of each image for improved detection performance by the YOLOv7. In both good and bad weather, the employed hybrid approach (YOLOv7 + ESRGAN) works well with about 80% accuracy in detecting all objects during adverse weather conditions. We would recommend further study on the methodology utilised in this paper to tackle the trolley-dilemma problem during inclement weather.
引用
收藏
页码:28235 / 28261
页数:26
相关论文
共 50 条
  • [21] Assessing the Impact of Adverse Weather on Performance and Safety of Connected and Autonomous Vehicles
    Abuzwidah, Muamer
    Elawady, Ahmed
    Wang, Ling
    Zeiada, Waleed
    CIVIL ENGINEERING JOURNAL-TEHRAN, 2024, 10 (09): : 3070 - 3089
  • [22] ARODNet: adaptive rain image enhancement object detection network for autonomous driving in adverse weather conditions
    Qiu, Yongsheng
    Lu, Yuanyao
    Wang, Yuantao
    Jiang, Haiyang
    OPTICAL ENGINEERING, 2023, 62 (11) : 118101
  • [23] Roadside Unit-based Unknown Object Detection in Adverse Weather Conditions for Smart Internet of Vehicles
    Chen, Yu-Chia
    Jhong, Sin-Ye
    Hsia, Chih-Hsien
    ACM TRANSACTIONS ON MANAGEMENT INFORMATION SYSTEMS, 2022, 13 (04)
  • [24] Study of Filtering the Weather Adverse Effects to Object Detection
    Shtekhin, S.
    Karachev, D.
    Stadnik, A.
    PHYSICS OF PARTICLES AND NUCLEI, 2024, 55 (03) : 329 - 333
  • [25] Rethinking LiDAR Object Detection in adverse weather conditions
    Vattem, Teja
    Sebastian, George
    Lukic, Luka
    2022 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION (ICRA 2022), 2022, : 5093 - 5099
  • [26] Uncertainty Evaluation of Object Detection Algorithms for Autonomous Vehicles
    Peng, Liang
    Wang, Hong
    Li, Jun
    AUTOMOTIVE INNOVATION, 2021, 4 (03) : 241 - 252
  • [27] Uncertainty Evaluation of Object Detection Algorithms for Autonomous Vehicles
    Liang Peng
    Hong Wang
    Jun Li
    Automotive Innovation, 2021, 4 : 241 - 252
  • [28] Multimodal Fusion Object Detection System for Autonomous Vehicles
    Person, Michael
    Jensen, Mathew
    Smith, Anthony O.
    Gutierrez, Hector
    JOURNAL OF DYNAMIC SYSTEMS MEASUREMENT AND CONTROL-TRANSACTIONS OF THE ASME, 2019, 141 (07):
  • [29] An Adaptive Sequencing Approach for Object Detection in Autonomous Vehicles
    Wolfe, Christopher
    Mohammadi, Khatereh
    Ferdowsi, Hasan
    INTERNATIONAL JOURNAL OF INTELLIGENT TRANSPORTATION SYSTEMS RESEARCH, 2024, 22 (03) : 629 - 647
  • [30] Deep Learning-Based Object Detection and Classification for Autonomous Vehicles in Different Weather Scenarios of Quebec, Canada
    Sharma, Teena
    Chehri, Abdellah
    Fofana, Issouf
    Jadhav, Shubham
    Khare, Siddhartha
    Debaque, Benoit
    Duclos-Hindie, Nicolas
    Arya, Deeksha
    IEEE ACCESS, 2024, 12 : 13648 - 13662