Logo 2016 Dragon 3 Final Results Symposium

Conference Agenda

Overview and details of the sessions and sub-session of this conference. Please select a date or session to show only sub-sessions at that day or location. Please select a single sub-session for detailed view (with abstracts and downloads if available).

 
Session Overview
Session
B2: ID.10580 Marine Safety & Security
Time:
Tuesday, 05/Jul/2016:
9:00am - 10:00am

Session Chair: Werner Rudolf Alpers
Session Chair: DanLing Tang
Workshop: Oceans & Coastal Zones
Location: Affilated Building 3-202#, School of Remote Sensing and Information Engineering, Wuhan University

Show help for 'Increase or decrease the abstract text size'
Presentations
Oral presentation

Progress of Ocean and Coastal Monitoring by the X-band Spaceborne SAR of TerraSAR-X and TanDEM-X

XiaoMing Li

Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, China, People's Republic of;

A half-century ago, it was recorded that ocean swells can propagate up to halfway around the globe. However, from a global perspective, how ocean swells propagate in the global oceans has yet to be depicted. To date, synthetic aperture radar (SAR) is the only available remote sensing instrument to measure the two-dimensional information of ocean surface waves. Here, a ten-year (2002-2012) global wave dataset of the spaceborne Advanced SAR (ASAR) onboard the European Space Agency’s satellite ENVISAT and the global wind dataset of the WindSat were used to (1) depict the propagation routes of ocean swells in the global oceans, (2) discover four distinguished crossing swell “pools”, and (3) interpret how these pools are formed. Together, these findings yield a new insight into ocean swells propagation and the consequent occurrence of crossing swells on a global basin scale from space, which will further deepen our cognition on nature of ocean.

Li-Progress of Ocean and Coastal Monitoring by the X-band Spaceborne SAR of TerraSAR-X and TanDEM-X_Cn_version.pdf

Oral presentation

Statistical Analysis and Oil Spill Recognition of Dark Patches Extracted from SAR Imagery

Kan Zeng, Xingtao Ding, Youjun Ma, MingXia He

Ocean University of China, China, People's Republic of;

Spaceborne SAR is widely used in sea surface oil spill monitoring because of its high sensitivity about oil spills and its advantage of working all-weather and all-day. Oil spills appear as remarkable dark patches on SAR images. But dark areas may also result from other oceanic phenomena, which are usually called look-alikes. The crucial technique of SAR oil spill monitoring is to correctly discriminate between oil spills and look-alikes.

The number of look-alikes in the dark patch sample set extracted from SAR images by a segmentation algorithm is much larger than that of oil spills because there are many oceanic phenomena other than oil spill making dark areas on SAR images. It results in an imbalance classification problem, i.e., the classifier trends to high detection rate for the look alike instead of oil spills. Following techniques are adopted to solve the imbalance problem and improve the detection rate of oil spills.

The 77 original features extracted from 16451 dark patches obtained from SAR images by a segmentation algorithm are statistically analyzed. The histograms of oil spills and look-alikes for each feature and the pairwise scatter plots of all the features are plotted. Some non-linear coupled feature pairs are found by looking through those plots. Some new features are constructed to decouple the non-linearity couplings. Then principle component analysis(PCA) is used to remove all linear couplings. Finally, The components whose distributions are significant different between oil spills and look-alikes are selected out from the principle components containing 99% of total variances. The selected components are used to generated a new sample set to train the classifier.

A sample weighted pre-classifier is inserted before the main classifier to reduce the imbalance of the sample set. The pre-classifier obtains its extremely high detection rate of oil spills at the cost of high false alarm, therefore the look-alikes identified by the classifier are considered to be reliable so that can be removed out from the sample set. The remainder nearly balanced sample set is then be sent to the main classifier.

For main classifier, adaboost decision tree(ADT) , multi-layer perceptron(MLP) and Adaboost MLP( AMLP), are compared on the classify performance and computation performance.

The effects of the feature selection based on statistic analysis, the pre-classifier and the main classifier are evaluated, respectively. The sample set are automatically generated from 63 ERS/SAR, 143 Envisat/ASAR and 128 Cosmo Sky-Med/SAR images by an adaptive threshold segment algorithm, and their class properties are manually decided by remote sensing experts.

Zeng-Statistical Analysis and Oil Spill Recognition of Dark Patches Extracted_Cn_version.pdf

Poster

Detection and Characters Analysis of Sea Surface Temperature Fronts in The Northwest Pacific Ocean

Meng Bao1, Jin Wang1,,2

1First Institute Of Oceanography, SOA; 2College of Physics, Qingdao University;

Sea Surface Temperature (SST) front is a typical mesoscale phenomenon in the ocean. Temperature fronts detection based on remote sensing data plays an important role in marine fishery resources utilization and global water cycle research. Owing to the influence of Kuroshio and Oyashio, the SST field of the Northwest Pacific Ocean is complex and variable and the SST fronts are well developed. In this research, the accuracy of SST fusion data product of 2013 is validated by Argo buoy and the standard deviation is found to be 0.6K. Additionally, performance of two front detection methods (temperature gradient and Jensen–Shannon divergence method) is compared and assessed by a simulated temperature field. Results show that the temperature gradient method is superior to Jensen–Shannon divergence method when the random error of SST product is 0.6K. Consequently, SST fronts in the study area are detected by temperature gradient method and some characters such as evolution process, seasonal variation, fronts intensity and length is studied. The results show that the front intensity is maximum in January and reaches its minimum in August. In the area affected by the Kuroshio, SST front is steady during the whole year.

Bao-Detection and Characters Analysis of Sea Surface Temperature Fronts_Cn_version.pdf

Poster

Analysis of the SAR Polarimetic Parameters over oil-covered sea surface and clean sea surface

Chen Wang, Chaofang Zhao, Kan Zeng, Mingxia He

Ocean University of China, Ocean Remote Sensing Institute;

With the development of the polarimetric synthetic aperture radar (PolSAR), various polarimetic parameters are exploited to distinguish oil-covered sea surface from clean sea surface. However, different parameters lead to different results with different PolSAR imageries and no comprehensive comparison of those parameters have been processed, which hamper the investigating on scattering mechanisms of oil slicks detection over sea surface. In this work, 14 polarimetric parameters are analyzed based on the Huynen Decomposition. Then, characteristics of polarimetic parameters over oil-covered sea surface and clean sea surface are analyzed with Radarsat-2, ALOS-PALSAR, SIR-C and UAVSAR imageries. Results show that degree of polarization (DoP), pedestal height (PH) and entropy (H) are efficient in oil slicks detection with those four kinds of PolSAR imageries. In addition, we find that the size of average window during calculating polarimetric parameters in C-band PolSAR imagery is expected to be larger than that of L-band PolSAR imagery.

Wang-Analysis of the SAR Polarimetic Parameters over oil-covered sea surface and clean sea surface_Cn_version.pdf

Poster

Monitoring Shallow Water Depth in Coastal Waters Around Islands and Reefs in the South China Sea using Sentinel-2 and landsat-8 Data

Lianbo Hu, Ming-Xia HE

Ocean University of China, China, People's Republic of;

A new shallow water depth inversion method was developed for coastal multi-spectral satellite data in the South China Sea (SCS). In this study, this method applied to Sentinl-2/MSI and Landsat-8/OLI satellite data and retrieved shallow depth and bottom reflectance in the coastal waters around islands and reefs in the SCS. The results were validated with in situ measurements and compared among different coastal multi-spectral satellite data. Six islands and reefs in the SCS were monitored using time series Landsat data.

Hu-Monitoring Shallow Water Depth in Coastal Waters Around Islands and Reefs in the South China Sea using_Cn_version.pdf


 
Contact and Legal Notice · Contact Address:
Conference: 2016 Dragon 3 Final Results Symposium
Conference Software - ConfTool Pro 2.6.94
© 2001 - 2016 by H. Weinreich, Hamburg, Germany