Independent of daylight and weather conditions, synthetic aperture radar (SAR) imagery is widely applied to detect ships in marine surveillance. The shapes of ships are multi-scale in SAR imagery due to multi-resolution imaging modes and their various shapes. Conventional ship detection methods are highly dependent on the statistical models of sea clutter or the extracted features, and their robustness need to be strengthened. Being an automatic learning representation, the RetinaNet object detector, one kind of deep learning model, is proposed to crack this obstacle. Firstly, feature pyramid networks (FPN) are used to extract multi-scale features for both ship classification and location. Then, focal loss is used to address the class imbalance and to increase the importance of the hard examples during training. There are 86 scenes of Chinese Gaofen-3 Imagery at four resolutions, i.e., 3 m, 5 m, 8 m, and 10 m, used to evaluate our approach. Two Gaofen-3 images and one Constellation of Small Satellite for Mediterranean basin Observation (Cosmo-SkyMed) image are used to evaluate the robustness. The experimental results reveal that (1) RetinaNet not only can efficiently detect multi-scale ships but also has a high detection accuracy; (2) compared with other object detectors, RetinaNet achieves more than a 96% mean average precision (mAP). These results demonstrate the effectiveness of our proposed method.

Dynamics of surface water is of great significance to understand the impacts of global changes and human activities on water resources. Remote sensing provides many advantages in monitoring surface water; however, in large scale, the efficiency of traditional remote sensing methods is extremely low because these methods consume a high amount of manpower, storage, and computing resources. In this paper, we propose a new method for quickly determining what the annual maximal and minimal surface water extent is. The maximal and minimal water extent in the year of 1990, 2000, 2010 and 2017 in the Middle Yangtze River Basin in China were calculated on the Google Earth Engine platform. This approach takes full advantage of the data and computing advantages of the Google Earth Engine’s cloud platform, processed 2343 scenes of Landsat images. Firstly, based on the estimated value of cloud cover for each pixel, the high cloud covered pixels were removed to eliminate the cloud interference and improve the calculation efficiency. Secondly, the annual greenest and wettest images were mosaiced based on vegetation index and surface water index, then the minimum and maximum surface water extents were obtained by the Random Forest Classification. Results showed that (1) the yearly minimal surface water extents were 14,751.23 km2, 14,403.48 km2, 13,601.48 km2, and 15,697.42 km2, in the year of 1990, 2000, 2010, and 2017, respectively. (2) The yearly maximal surface water extents were 18,174.76 km2, 20,671.83 km2, 19,097.73 km2, and 18,235.95 km2, in the year of 1990, 2000, 2010, and 2017, respectively. (3) The accuracies of surface water classification ranged from 86% to 93%. Additionally, the causes of these changes were analyzed. The accuracy evaluation and comparison with other research results show that this method is reliable, novel, and fast in terms of calculating the maximal and minimal surface water extent. In addition, the proposed method can easily be implemented in other regions worldwide.

Permafrost is widely distributed in the Tibetan Plateau. Seasonal freeze–thaw cycles of permafrost result in upward and downward surface displacement. Multitemporal interferometric synthetic aperture radar (MT-InSAR) observations provide an effective method for monitoring permafrost displacement under difficult terrain and climatic conditions. In this study, a seasonal sinusoidal model-based new small baselines subset (NSBAS) chain was adopted to obtain a deformation time series. An experimental study was carried out using 33 scenes of Sentinel-1 data (S-1) from 28 November 2017 to 29 December 2018 with frequent revisit (12 days) observations. The spatial and temporal characteristics of the surface displacements variation combined with different types of surface land cover, elevation and surface temperature factors were analyzed. The results revealed that the seasonal changes observed in the time series of ground movements, induced by freeze–thaw cycles were observed on flat surfaces of sedimentary basins and mountainous areas with gentle slopes. The estimated seasonal oscillations ranged from 2 mm to 30 mm, which were smaller in Alpine deserts than in Alpine meadows. In particular, there were significant systematic differences in seasonal surface deformation between areas near mountains and sedimentary basins. It was also found that the time series of deformation was consistent with the variation of surface temperature. Based on soil moisture active/passive (SMAP) L4 surface and root zone soil moisture data, the deformation analysis influenced by soil moisture factors was also carried out. The comprehensive analysis of deformation results and auxiliary data (elevation, soil moisture and surface temperature et al.) provides important insights for the monitoring of the seasonal freeze-thaw cycles in the Tibetan Plateau.

No abstract

No abstract

With the capability to automatically learn discriminative features, deep learning has experienced great success in natural images but has rarely been explored for ship classification in high-resolution SAR images due to the training bottleneck caused by the small datasets. In this paper, convolutional neural networks (CNNs) are applied to ship classification by using SAR images with the small datasets. First, ship chips are constructed from high-resolution SAR images and split into training and validation datasets. Second, a ship classification model is constructed based on very deep convolutional networks (VGG). Then, VGG is pretrained via ImageNet, and fine tuning is utilized to train our model. Six scenes of COSMO-SkyMed images are used to evaluate our proposed model with regard to the classification accuracy. The experimental results reveal that (1) our proposed ship classification model trained by fine tuning achieves more than 95% average classification accuracy, even with 5-cross validation; (2) compared with other models, the ship classification model based on VGG16 achieves at least 2% higher accuracies for classification. These experimental results reveal the effectiveness of our proposed method.

No abstract

No abstract