Improvement of moving object detection accuracy on aerial imagery using sensor geometry

In this paper, we propose an efficient approach to improve moving object detection accuracy which is adaptive to changes in object size according to the altitude of the aircraft. The proposed algorithm can effectively detect moving objects with various pixel-scale from 8x8 to 100x100 in full-HD motion imagery. It consists of two stages which are detection and fusion. At the first stage, two algorithms are performed simultaneously: One is one-stage object detection network for detecting large objects, and the other is optical flow method for detecting small moving objects. In the second stage, results from the first stage are fused with a Ground Sample Distance (GSD) of imagery. We have conducted experiments using aerial imagery taken at a height between 130 meters and 400 meters. We evaluated the detection performance of our method in terms of precision, recall and normalized multiple object detection accuracy (N-MODA). Through experiments, we proved that the superiority of the proposed method.

[1]  Ali Farhadi,et al.  YOLO9000: Better, Faster, Stronger , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[2]  Joachim Höhle,et al.  Photogrammetric Measurements in Oblique Aerial Images , 2008 .

[3]  Thomas Brox,et al.  FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networks , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[4]  Linfu Xie,et al.  An asymmetric re-weighting method for the precision combined bundle adjustment of aerial oblique images , 2016 .

[5]  Mark A. Tarbell,et al.  Tier-scalable reconnaissance: the future in autonomous C4ISR systems has arrived: progress towards an outdoor testbed , 2017, Defense + Security.

[6]  Martin Jägersand,et al.  MODNet: Moving Object Detection Network with Motion and Appearance for Autonomous Driving , 2017, ArXiv.

[7]  Thomas Brox,et al.  FlowNet: Learning Optical Flow with Convolutional Networks , 2015, 2015 IEEE International Conference on Computer Vision (ICCV).

[8]  Ross B. Girshick,et al.  Fast R-CNN , 2015, 1504.08083.

[9]  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.

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

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

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

[13]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[14]  Andreas Geiger,et al.  Object scene flow for autonomous vehicles , 2015, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).