Deep-Patch Orientation Network for Aircraft Detection in Aerial Images

The aerial target detection and recognition are very challenging due to large appearance, lighting and orientation variations. We propose a Deep-patch Orientation Network (DON) method, which is general and can learn the encoded orientation information based on any off the-shelf deep detection framework, e.g., Faster-RCNN and YOLO, and result into higher performance in airplane target detection and classification tasks. Most existing methods neglected the orientation information, which in DON is obtained based on the structure information contained in the patch training samples. In testing process, we introduce an orientation based method to exploit patches for whole target localization. Also, we analyzed how to improve agnostic-target detection framework by tailoring the reference boxes. Experimental results on two datasets show that, our proposed DON method improves the recall at high precision rates for the deep detection framework and provide orientation information for detected targets.

[1]  Philip H. S. Torr,et al.  BING: Binarized normed gradients for objectness estimation at 300fps , 2014, Computational Visual Media.

[2]  Rongrong Ji,et al.  Bounding Multiple Gaussians Uncertainty with Application to Object Tracking , 2016, International Journal of Computer Vision.

[3]  Qixiang Ye,et al.  Orientation robust object detection in aerial images using deep convolutional neural network , 2015, 2015 IEEE International Conference on Image Processing (ICIP).

[4]  Yu Li,et al.  Automatic Target Detection in High-Resolution Remote Sensing Images Using Spatial Sparse Coding Bag-of-Words Model , 2012, IEEE Geoscience and Remote Sensing Letters.

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

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

[7]  Alessio Del Bue,et al.  Adaptive Local Movement Modelling for Object Tracking , 2015, 2015 IEEE Winter Conference on Applications of Computer Vision.

[8]  Xian Sun,et al.  Aircraft Recognition in High-Resolution Satellite Images Using Coarse-to-Fine Shape Prior , 2013, IEEE Geoscience and Remote Sensing Letters.

[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]  Liang Lin,et al.  Is Faster R-CNN Doing Well for Pedestrian Detection? , 2016, ECCV.

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

[12]  Chen Chen,et al.  Output Constraint Transfer for Kernelized Correlation Filter in Tracking , 2016, IEEE Transactions on Systems, Man, and Cybernetics: Systems.

[13]  Hui Wu,et al.  Fast aircraft detection in satellite images based on convolutional neural networks , 2015, 2015 IEEE International Conference on Image Processing (ICIP).

[14]  Wei Li,et al.  Robust airplane detection in satellite images , 2011, 2011 18th IEEE International Conference on Image Processing.

[15]  Luc Van Gool,et al.  The Pascal Visual Object Classes (VOC) Challenge , 2010, International Journal of Computer Vision.

[16]  Frédéric Jurie,et al.  Discriminative Autoencoders for Small Targets Detection , 2014, 2014 22nd International Conference on Pattern Recognition.

[17]  Anil Kumar Gupta,et al.  Effect of Different Distance Measures on the Performance of K-Means Algorithm: An Experimental Study in Matlab , 2014, ArXiv.

[18]  Trevor Darrell,et al.  Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation , 2013, 2014 IEEE Conference on Computer Vision and Pattern Recognition.

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