Vehicle detection technology is a key component of an intelligent transportation system, but most of the current vehicle detection technologies are based on road monitoring cameras. Compared with these fixed cameras, Unmanned Aerial Vehicles (UAVs) seem to have a lot of advantages such as more flexible, broader vision, higher speed, which make the vehicle detection more challenging. In this paper, a new dataset built on UAV traffic videos and a neural network which could fuse multi-layer features are proposed. Different from some networks with only a single layer, the proposed network merges the features from multiple layers firstly. Then a convolution layer is used to reduce the feature dimensions and a deconvolution layer is employed to do upsampling and enhance the response information. Finally, multiple fully connected layers are used to finish the detection. Furthermore, the proposed method combines the detecting and tracking for optimization and high detection speed. Experiments on the self-built UAV traffic video dataset demonstrate that the proposed method gets better results and higher speed.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Vehicle Detection in UAV Traffic Video Based on Convolution Neural Network


    Contributors:
    Li, Shulin (author) / Zhang, Weigang (author) / Li, Guorong (author) / Su, Li (author) / Huang, Qingming (author)


    Publication date :

    2018-04-01


    Size :

    1249230 byte




    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Vehicle Counting based on Convolution Neural Network

    Anil, Jenna Maria / Mathews, Liz / Renji, Rajeswari et al. | IEEE | 2023



    Optimized Graph Convolution Recurrent Neural Network for Traffic Prediction

    Guo, Kan / Hu, Yongli / Qian, Zhen et al. | IEEE | 2021


    Wildfire detection CubeSat based on convolution neural network

    Azami, Muhammad Hasif Bin / Orger, Necmi Cihan / Schulz, Victor Hugo et al. | British Library Conference Proceedings | 2021