Search

Salient object detection has made substantial progress due to the exploitation of multi-level convolutional features. The key point is how to combine these convolutional features effectively and efficiently. Due to the step by step down-sampling operations in almost all CNNs, multi-level features usually have different scales. Methods based on fully convolutional networks directly apply bilinear up-sampling to low-resolution deep features and then combine them with high-resolution shallow features by addition or concatenation, which neglects the compatibility of features, resulting in misalignment problems. In this paper, to solve the problem, we propose an alignment integration network (ALNet), which aligns adjacent level features progressively to generate powerful combinations. To capture long-range dependencies for high-level integrated features as well as maintain high computational efficiency, a strip attention module (SAM) is introduced into the alignment integration procedures. Benefiting from SAM, multi-level semantics can be selectively propagated to predict precise salient objects. Furthermore, although integrating multi-level convolutional features can alleviate the blur boundary problem to a certain extent, it is still unsatisfactory for the restoration of a real object boundary. Therefore, we design a simple but effective boundary enhancement module (BEM) to guide the network focus on boundaries and other error-prone parts. Based on BEM, an attention weighted loss is proposed to boost the network to generate sharper object boundaries. Experimental results on five benchmark datasets demonstrate that the proposed method can achieve state-of-the-art performance on salient object detection. Moreover, we extend the experiments on the remote sensing datasets, and the results further prove the universality and scalability of ALNet. [ABSTRACT FROM AUTHOR]

Particle size distribution is an important characteristic of reclaimed soil in arid and semi-arid mining areas in western China, which is important in the ecological environment protection and control of the Yellow River Basin. Large-scale coal resource mining disturbances have caused serious damage to the fragile ecological environment. The timely and accurate dynamic monitoring of mining area topsoil information has practical significance for ecological restoration and management evaluation. Investigating Wuhai City in the Inner Mongolia Autonomous Region of China, this study uses Landsat8 OLI multispectral images and measured soil sample particle size data to analyze soil spectral characteristics and establish a particle size content prediction model to retrieve the particle size distribution in the study area. The experimental results and analysis demonstrate that: (1) the 6SV (Second Simulation of the Satellite Signal in the Solar Spectrum Vector version) atmospheric correction model is more accurate than the FLAASH (Fast Line-of-sight Atmospheric Analysis of Hypercubes) model in arid and semi-arid areas with undulating terrain; (2) 0–40 cm is the optimum soil thickness for modeling and predicting particle size content in this study; and (3) the multi-band prediction model is more precise than the single-band prediction model. The multi-band model's sequence of advantages and disadvantages is SVM (Support Vector Machine) > MLR (Multiple Linear Regression) > PLSR (Partial Least Squares Regression). Among them, the 6SV-SVM model has the highest precision, and the prediction precision R2 of the 3 particle sizes' contents is above 0.95, which can effectively predict the soil particle-size distribution and provide effective data to support topsoil quality change monitoring in the mine land reclamation area. [ABSTRACT FROM AUTHOR]

The ocean behaves as a typical multiscale fractal structure, whose dynamic and thermal variabilities extend over a wide range of spatial scales, r, spanning from 10−3 to 107 m. Studying the statistical characteristics of multiscale fractal structures is crucial to understanding the interactions and energy cascade processes between different spatial scales. Remote sensing data are one of the best choices for revealing these statistical characteristics. This work analyzes the multiscale (1–1000 km) fractal structures of sea surface temperature (SST) from the Level-2+ (L2P) satellite orbit Visible Infrared Imaging Radiometer Suite (VIIRS) products over the Kuroshio Extension (KE) region (145°E–160°W, 20°N–50°N), using a conventional method (second-order structure function, D(r)) and a newly developed statistical method (spatial variance, V(r)). The results show that both the power-law distribution slopes of D(r) and V(r) are close to 2/3, which is equivalent to the −5/3 wavenumber spectrum. V(r) is found to be more robust when depicting the fractal structure and variance density, V'(r), compared with D(r). V'(r) is slightly larger at the mesoscale (50–150 km) than at the large scale (higher than 150 km) and is much smaller than that at the submesoscale (smaller than 50 km). Additionally, V'(r) has an indiscernible diurnal variation but remarkable seasonal and latitudinal variations. For the seasonal variability, the maximum V'(r) appears in winter at the large scale and mesoscale, and gradually shifts towards spring at the submesoscale, which implies that a forward energy cascade process may occur during this period. The maximum of the latitude-dependent V'(r) appears around 40°N for all the scales. It is consistent with the latitude of the strongest background SST gradient, indicating that the background SST front is the main source of the strong SST multiscale spatial variabilities over the KE region. This work benefits the application of other high-resolution remote sensing data in this research field, including the forthcoming Surface Water Ocean Topography (SWOT) satellite product. [ABSTRACT FROM AUTHOR]

Multiple upwelling systems develop in the Indian Ocean during the summer monsoon. The Sri Lanka dome (SLD), which occurs in the open ocean off the east coast of Sri Lanka from June to September, is distinct from those near the coast. The SLD is characterized by uplifted thermocline and increased chlorophyll concentration. Mechanisms of the upwelling and its biogeochemical response are not well understood. Here, we explored the dynamics of the chlorophyll and sea-to-air CO2 flux in the SLD using ocean color and altimetry remote sensing data, together with other reanalysis products. We found that the occurrence of high chlorophyll concentration and sea-to-air CO2 flux happens along the pathway of the southwest monsoon current (SMC). The annual cycle of chlorophyll in the SLD has a one-month lag relative to that in the southern coast of Sri Lanka. The positive wind stress curl that forms in the SLD during the summer does not fully explain the seasonal chlorophyll maximum. Transport of the SMC, eddy activity, and associated frontal processes also play an important role in regulating the variability in chlorophyll. In the SLD, upwelled subsurface water has excess dissolved inorganic carbon (DIC) relative to the conventional Redfield ratio between DIC and nutrients; thus, upwelling and sub-mesoscale processes determine this region to be a net carbon source to the atmosphere. [ABSTRACT FROM AUTHOR]

Sixteen years of satellite observational data in the Northwestern Pacific Ocean are used to describe the variability in the sea surface temperature (SST) gradient and its impact on chlorophyll-a concentrations (Chl-a). Spatially, a meridional dependence is identified in which the SST gradient increases to the north in association with elevated Chl-a. Temporally, the seasonal variability shows a large SST gradient and high Chl-a in winter and spring, while the SST gradient and Chl-a are much lower in summer. The seasonal variability in Chl-a leads the variability in the SST gradient by one month. A significant correlation between the SST gradient and Chl-a in the anomalous field is obtained only in the western section of the Kuroshio extension (KE) and the highest correlation is identified without any lags. An index for the section is defined as the proportion of the number of times that the SST gradient magnitude is anomalously large in each year, and the index is highly related to the stability of the KE and has a prominent influence on Chl-a in the region. An anomalously large positive (negative) SST gradient magnitude occurs when the KE is unstable (stable) and the corresponding Chl-a is high (low). [ABSTRACT FROM AUTHOR]