This paper studies the use of the Fully Convolutional Networks (FCN) model in the extraction of water bodies from Very High spatial Resolution (VHR) optical images in the case of limited training samples. Two different seasonal GaoFen-2 images with a spatial resolution of 0.8 m in the south of the Beijing metropolitan area were used to extensively validate the FCN model. Four key factors including input features, training data, transfer learning, and data augmentation related to the performance of the FCN model were empirically analyzed by using 36 combinations of various parameter settings. Our findings indicate that the FCN-based method can work as a robust and cost-effective tool in the extraction of water bodies from VHR images. The FCN-based method trained on a small amount of labeled L1A data can also significantly outperform the Normalized Difference Water Index (NDWI) based method, the Support Vector Machine (SVM) based method, and the Sparsity Model (SM) based method, even when radiometric normalization and spatial contexts are introduced to preprocess the input data for the latter three methods. The advantages of the FCN-based method are mainly due to its capability to exploit spatial contexts in the image, especially in urban areas with mixed water and shadows. Though the settings of four key factors significantly affect the performance of the FCN based method, choosing a qualified setting for the FCN model is not difficult. Our lessons learned from the successful use of the FCN model for the extraction of water from VHR images can be extended to extract other land covers.

Infrared small-target detection under heterogeneous background, as a challenging task, plays an important role in many applications. In practice, there are not only bright targets but also dim targets, e.g., rescue aircraft and vehicles in the forest fire scene. Considering that most existing infrared small-target detection methods are merely aimed at bright targets, a novel method using multiple morphological profiles (MMP) is proposed, which can detect various types of targets whose brightness varies greatly. In the designed morphological feature extraction, different attributes, i.e., area attribute and height attribute, are applied to extract spatial size and contrast information of small-target in the max-tree and min-tree, respectively. Furthermore, discontinuous pruning values are further utilized for different attributes, and a designed fusion strategy of different pruning values results in more robust detection performance. Experimental results validated on two synthetic data and six real data sets demonstrate that the proposed MMP can not only detect a variety of brightness of targets and different types of targets and kinds of spatial sizes of targets but also further improve the contrast between targets and background, and the background clutter is significantly suppressed.

Existing infrared (IR) small target detection methods are divided into local priors-based and nonlocal priors-based ones. However, due to heterogeneous structures in IR images, using either local or nonlocal information is always suboptimal, which causes detection performance to be unstable and unrobust. To solve the issue, a comprehensive method, exploiting both local and nonlocal priors, is proposed. The proposed method framework includes dual-window local contrast method (DW-LCM) and multiscale-window IR patch-image (MW-IPI). In the first stage, DW-LCM designs dual-window to compensate for the shortcomings of local priors-based methods, which easily mistake some isolated weak-signal targets. In the second stage, MW-IPI utilizes several small windows with various sizes, which can not only decrease the redundant information generated by sliding windows, but also extract more discriminative information to prevent some pixels in the strong-border edge from being falsely detected. Then, multiplication pooling operation is employed to enhance the target separation and suppress the background clutter simultaneously. Experimental results using five real IR datasets with various scenes reveal the effectiveness and robustness of the proposed method.

High-rise buildings (HRBs) as a modern and visually distinctive land use play an important role in urbanization. Large-scale monitoring of HRBs is valuable in urban planning and environmental protection and so on. Due to the complex 3D structure and seasonal dynamic image features of HRBs, it is still challenging to monitor large-scale HRBs in a routine way. This paper extends our previous work on the use of the Fully Convolutional Networks (FCN) model to extract HRBs from Sentinel-2 data by studying the influence of seasonal and spatial factors on the performance of the FCN model. 16 Sentinel-2 subset images covering four diverse regions in four seasons were selected for training and validation. Our results indicate the performance of the FCN-based method at the extraction of HRBs from Sentinel-2 data fluctuates among seasons and regions. The seasonal change of accuracy is larger than that of the regional change. If an optimal season can be chosen to get a yearly best result, F1 score of detected HRBs can reach above 0.75 for all regions with most errors located on the boundary of HRBs. FCN model can be trained on seasonally and regionally combined samples to achieve similar or even better overall accuracy than that of the model trained on an optimal combination of season and region. Uncertainties exist on the boundary of detected results and may be relieved by revising the definition of HRBs in a more rigorous way. On the whole, the FCN based method can be largely effective at the extraction of HRBs from Sentinel-2 data in regions with a large diversity in culture, latitude, and landscape. Our results support the possibility to build a powerful FCN model on a larger size of training samples for operational monitoring HRBs at the regional level or even on a country scale.

Wetlands have been degrading and reducing under the influences of human activity and climate change. Landsat long-term observations can help us better track the changing evidences of wetland habitats that would be valuable for guiding the restoration and conservation of wetland. In this study, we demonstrated the results of tracking the changing evidence of wetland habitats using Landsat observations from 1984 to 2017 through the case study of Baiyangdian wetland in China. We extracted the open water and classified the wetland habitats using collected 190 scenes from Landsat observations. As a result, we found that the yearly variations of wetland present phasic changes in three phases: 1988–1998, 1999–2011 and 2013–2017. The landscape of wetland habitats presented during 1989–1999 mostly show us the natural spatial pattern with less human disturbance traces compared to that during 2013–2017. The water environment, moreover, changed for the better after the 2010s, which indicated the encouraging effects of the environmental restoration project implemented from the year 2010. The current landscapes of wetland habitats, however, present lots of linear belts that are blocking the water cycles and ecological channels of aquatic plants and animals in the wetland. The areas in the northwestern wing and around the northeastern edge of the wetland are changing to be drier due to cropping activities and are at risk of wetland loss. These historical changing evidences could be a guideline for planning and designing restoration for the wetland.

Cladophora qinghaiensis, an endemic species of Cladophora in saltwater lakes, was scientifically named in 2021 (hereafter referred to as Cladophora). Cladophora exists in different morphologies, including attached submerged Cladophora (AC), grown floating Cladophora (GFC), and death floating Cladophora (DFC). Previous satellite remote sensing has mainly focused on identifying floating algae. In this study, Qinghai Lake served as a case study, and a classification decision tree model (CDTM) was proposed. The model employed the chlorophyll spectral index (CSI) and the normalized difference vegetation index (NDVI) to differentiate AC, Floating Cladophora (FC), and water. Additionally, the floating Cladophora index (FCI) was introduced to further distinguish GFC and DFC within FC. The method was applicable to Sentinel-2 images from 2016–2023. Visual interpretation methods were used for Landsat series images from the summer months (July to September) to obtain the AC and FC. The results demonstrate that over the past 30 years, the areas inhabited by AC and FC have increased gradually. The three morphologies of Cladophora also exhibited seasonal variations, with growth observed annually in May–June, reaching peaks in August–September, and gradually declining in October. In addition, by combining factors such as water surface area and climatic factors, we analyzed the driving forces influencing the changes in Cladophora. In this research, AC and FC showed significant correlations with the water surface area, with correlation coefficients (r) of approximately 0.9 and 0.7, respectively. These new findings provide valuable insights regarding the spatiotemporal changes and underlying causes for different morphologies of Cladophora in global saline lakes.