A Synthetic Integrative Algorithm of Fast Concurrent Detection-Recognition with Effective Tracking of Aerial Objects based on Deep Learning of CNN

In this paper, a synthetic integrative algorithm of fast concurrent detection-recognition with effective tracking of aerial objects is presented, in which the deep learning of CNN is employed to carry out high-performance detection, recognition and tracking of aerial objects. In the synthetic integrative algorithm, a concurrent detection-recognition mechanism is designed and implemented based on DarkNet-53 with FPN structure and RFB module. Furthermore, the object movement is predicted by well-trained Siamese RPN with image transformation so as to achieve a high-performance tracking for aerial objects. In view of the shortcoming of missing target in the tracking stage and some false positive results in the stage of concurrent detection-recognition a discriminative classifier is designed to deal with and eliminate the false positive ones. Finally, a synthetic optimization strategy is proposed towards the integration of concurrent detection-recognition with tracking by filtering out disturbances with the discriminative classifier and confirming effective results by the IOU matching so as to make the whole algorithm works well and achieve a superior performance. Based on the simulation demonstration with a software of X-Plane 10, two datasets, Aerial Object Detection (AOD) and Aerial Object Tracking (AOT), are generated, and a series of experiments are conducted on the AOD and AOT to verify and validate the performance advantages of our proposed algorithms.

[1]  Kaiming He,et al.  Feature Pyramid Networks for Object Detection , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[2]  Adam Van Etten,et al.  You Only Look Twice: Rapid Multi-Scale Object Detection In Satellite Imagery , 2018, ArXiv.

[3]  Wei Wu,et al.  High Performance Visual Tracking with Siamese Region Proposal Network , 2018, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[4]  Ali Farhadi,et al.  You Only Look Once: Unified, Real-Time Object Detection , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[5]  Charles A. Bouman,et al.  Deep Learning for Moving Object Detection and Tracking from a Single Camera in Unmanned Aerial Vehicles (UAVs) , 2018, IMAWM.

[6]  Wei Liu,et al.  SSD: Single Shot MultiBox Detector , 2015, ECCV.

[7]  Jitendra Malik,et al.  Region-Based Convolutional Networks for Accurate Object Detection and Segmentation , 2016, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[8]  Kaiming He,et al.  Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks , 2015, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[9]  Zhenyu He,et al.  The Visual Object Tracking VOT2016 Challenge Results , 2016, ECCV Workshops.

[10]  Li Fei-Fei,et al.  ImageNet: A large-scale hierarchical image database , 2009, CVPR.

[11]  Dumitru Erhan,et al.  Going deeper with convolutions , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[12]  Matthieu Guillaumin,et al.  Non-maximum Suppression for Object Detection by Passing Messages Between Windows , 2014, ACCV.