Marine Object Recognition Based on Deep Learning

In the research of unmanned surface vessel (USV), accurately perceiving the environment around the USV and recognizing the obstacles in real time are the major difficulties. The existing methods based on lidar or unmanned air vehicle have got good performance, but time and money costs are not what we can afford. After analyzing the difficulties existed in the obstacle avoidance test for USV, we propose a new method called marine object detection based on Single Shot MultiBox Detector (SSD). It solves these difficulties well, and the time and money costs are acceptable to us. After modifying and optimizing the SSD model, its average precision is 93.5% and its time cost is 45ms per image (1280*760), which means that it has much better performance than any existing method. The experimental results show that the method can detect object in real time and have great precision, which ensures the safety of USV during the navigation.

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

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

[3]  Yoshua Bengio,et al.  Gradient-based learning applied to document recognition , 1998, Proc. IEEE.

[4]  N. Kolev,et al.  An application of automatic target recognition in marine navigation , 1995, Proceedings International Radar Conference.

[5]  Jinwhan Kim,et al.  Persistent automatic tracking of multiple surface vessels by fusing radar and lidar , 2017, OCEANS 2017 - Aberdeen.

[6]  Jinbiao Chen,et al.  Target recognition for marine search and rescue radar , 2010, 2010 Sixth International Conference on Natural Computation.

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

[8]  Asif Sabanoviç,et al.  Variable Structure Systems With Sliding Modes in Motion Control—A Survey , 2011, IEEE Transactions on Industrial Informatics.

[9]  Gaemus Collins,et al.  Enabling technologies for autonomous offshore inspections by heterogeneous unmanned teams , 2017, OCEANS 2017 - Aberdeen.

[10]  ImageNet Classification with Deep Convolutional Neural , 2013 .

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

[12]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[13]  Jian Sun,et al.  Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition , 2014, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[14]  Sergey Ioffe,et al.  Rethinking the Inception Architecture for Computer Vision , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

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

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

[17]  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.

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

[19]  Michael Blaich,et al.  Mission integrated collision avoidance for USVs using laser range finder , 2015, OCEANS 2015 - Genova.

[20]  Jianhua Wang,et al.  Unmanned surface vessel for monitoring and recovering of spilled oil on water , 2016, OCEANS 2016 - Shanghai.

[21]  Yongming Li,et al.  NN-based adaptive dynamic surface control for a class of nonlinear systems with input saturation , 2012, 2012 7th IEEE Conference on Industrial Electronics and Applications (ICIEA).

[22]  Andrew Zisserman,et al.  Very Deep Convolutional Networks for Large-Scale Image Recognition , 2014, ICLR.

[23]  George W. Irwin,et al.  A review on improving the autonomy of unmanned surface vehicles through intelligent collision avoidance manoeuvres , 2012, Annu. Rev. Control..