DeepUNet: A Deep Fully Convolutional Network for Pixel-Level Sea-Land Segmentation

Semantic segmentation is a fundamental research in optical remote sensing image processing. Because of the complex maritime environment, the sea-land segmentation is a challenging task. Although the neural network has achieved excellent performance in semantic segmentation in the last years, there were a few of works using CNN for sea-land segmentation and the results could be further improved. This paper proposes a novel deep convolution neural network named DeepUNet. Like the U-Net, its structure has a contracting path and an expansive path to get high-resolution optical output. But differently, the DeepUNet uses DownBlocks instead of convolution layers in the contracting path and uses UpBlock in the expansive path. The two novel blocks bring two new connections that are U-connection and Plus connection. They are promoted to get more precise segmentation results. To verify the network architecture, we construct a new challenging sea-land dataset and compare the DeepUNet on it with the U-Net, SegNet, and SeNet. Experimental results show that DeepUNet can improve 1–2% accuracy performance compared with other architectures, especially in high-resolution optical remote sensing imagery.

[1]  Tao Zhang,et al.  Extraction of Coastline in Aquaculture Coast from Multispectral Remote Sensing Images: Object-Based Region Growing Integrating Edge Detection , 2013, Remote. Sens..

[2]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[3]  H. Liu,et al.  Automated extraction of coastline from satellite imagery by integrating Canny edge detection and locally adaptive thresholding methods , 2004 .

[4]  Gholamreza Akbarizadeh,et al.  A new computer vision algorithm for classification of POLSAR images , 2015, 2015 7th Conference on Information and Knowledge Technology (IKT).

[5]  Hui Zhou,et al.  A Novel Hierarchical Method of Ship Detection from Spaceborne Optical Image Based on Shape and Texture Features , 2010, IEEE Transactions on Geoscience and Remote Sensing.

[6]  Roberto Cipolla,et al.  SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation , 2015, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[7]  Gholamreza Akbarizadeh,et al.  Semi-supervised Hyperspectral image classification using spatial-spectral features and superpixel-based sparse codes , 2017, 2017 Iranian Conference on Electrical Engineering (ICEE).

[8]  Gholamreza Akbarizadeh,et al.  A level set based method for coastline detection of SAR images , 2017, 2017 3rd International Conference on Pattern Recognition and Image Analysis (IPRIA).

[9]  N. Otsu.  A threshold selection method from gray level histograms , 1979 .

[10]  Iasonas Kokkinos,et al.  DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs , 2016, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[11]  Gholamreza Akbarizadeh,et al.  Efficient Combination of Texture and Color Features in a New Spectral Clustering Method for PolSAR Image Segmentation , 2017 .

[12]  Fevzi Karsli,et al.  Automatic detection of shoreline change on coastal Ramsar wetlands of Turkey , 2011 .

[13]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[14]  Lihua Yue,et al.  A Novel Sea-Land Segmentation Algorithm Based on Local Binary Patterns for Ship Detection , 2014 .

[15]  G. S. Dwarakish,et al.  Automatic Shoreline Detection and Change Detection Analysis of Netravati-GurpurRivermouth Using Histogram Equalization and Adaptive Thresholding Techniques☆ , 2015 .

[16]  Gholamreza Akbarizadeh,et al.  Coastline extraction from SAR images using spatial fuzzy clustering and the active contour method , 2017 .

[17]  한보형,et al.  Learning Deconvolution Network for Semantic Segmentation , 2015 .

[18]  Vladlen Koltun,et al.  Multi-Scale Context Aggregation by Dilated Convolutions , 2015, ICLR.

[19]  Gholamreza Akbarizadeh,et al.  A New Statistical-Based Kurtosis Wavelet Energy Feature for Texture Recognition of SAR Images , 2012, IEEE Transactions on Geoscience and Remote Sensing.

[20]  Iasonas Kokkinos,et al.  Semantic Image Segmentation with Deep Convolutional Nets and Fully Connected CRFs , 2014, ICLR.

[21]  U. R. Aktas,et al.  Edge-aware segmentation in satellite imagery: A case study of shoreline detection , 2012, 7th IAPR Workshop on Pattern Recognition in Remote Sensing (PRRS).

[22]  Gaofeng Meng,et al.  SeNet: Structured Edge Network for Sea–Land Segmentation , 2017, IEEE Geoscience and Remote Sensing Letters.

[23]  Zhenwei Shi,et al.  Maritime Semantic Labeling of Optical Remote Sensing Images with Multi-Scale Fully Convolutional Network , 2017, Remote. Sens..

[24]  Gaofeng Meng,et al.  Efficient sea-land segmentation using seeds learning and edge directed graph cut , 2016, Neurocomputing.

[25]  Aaas News,et al.  Book Reviews , 1893, Buffalo Medical and Surgical Journal.

[26]  Gholamreza Akbarizadeh,et al.  A Two-Phase Algorithm Based on Kurtosis Curvelet Energy and Unsupervised Spectral Regression for Segmentation of SAR Images , 2016, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

[27]  Gholamreza Akbarizadeh,et al.  A new ensemble clustering method for PolSAR image segmentation , 2015, 2015 7th Conference on Information and Knowledge Technology (IKT).

[28]  Zheng Zhang,et al.  MXNet: A Flexible and Efficient Machine Learning Library for Heterogeneous Distributed Systems , 2015, ArXiv.

[29]  Gholamreza Akbarizadeh,et al.  A New Curvelet-Based Texture Classification Approach for Land Cover Recognition of SAR Satellite Images , 2014 .

[30]  P ? ? ? ? ? ? ? % ? ? ? ? , 1991 .

[31]  Kaichang Di,et al.  AUTOMATIC SHORELINE EXTRACTION FROM HIGH-RESOLUTION IKONOS SATELLITE IMAGERY , 2003 .

[32]  Thomas Brox,et al.  U-Net: Convolutional Networks for Biomedical Image Segmentation , 2015, MICCAI.

[33]  Weihai Li,et al.  A sea-land segmentation scheme based on statistical model of sea , 2011, 2011 4th International Congress on Image and Signal Processing.

[34]  Gholamreza Akbarizadeh,et al.  Segmentation of SAR Satellite Images Using Cellular Learning Automata and Adaptive Chains , 2013 .

[35]  Gholamreza Akbarizadeh,et al.  Unsupervised feature learning based on sparse coding and spectral clustering for segmentation of synthetic aperture radar images , 2015, IET Comput. Vis..

[36]  G. G. Stokes.  "J." , 1890, The New Yale Book of Quotations.

[37]  Gholamreza Akbarizadeh,et al.  Segmentation parameter estimation algorithm Based on curvelet transform coefficients energy for feature extraction and texture description of SAR images , 2015, 2015 7th Conference on Information and Knowledge Technology (IKT).