Saliency detection in synthetic aperture radar (SAR) image is a difficult problem. This paper proposed a multitask saliency detection (MSD) model for the saliency detection task of SAR image. Firstly, we extract four features of SAR image as the input of the MSD model, which include the intensity, orientation, uniqueness and global contrast. Then, the saliency map is generated by the multitask sparsity pursuit (MTSP) which integrates the multiple features collaboratively. Subjective and objective evaluation of the MSD model verifies its effectiveness. Based on the saliency maps of the source images, an image fusion method is proposed for the SAR and color optical image fusion. The experimental results of real data show the proposed image fusion method is superior to the presenting methods in terms of several universal quality evaluation indexes, as well as in the visual quality. The salient areas in the SAR image can be highlighted and the spatial and spectral details of color optical image can also be preserved in the fusion result.

The present study investigates the failure mechanisms of industrial fiber materials, using a custom designed fiber cutting performance test bench. The fracture morphologies of single PA6 fibers are examined by scanning electron microscopy. The analysis reveals that fiber cutting can be distinguished according to four distinct stages of fiber failure represented by shearing, cutting, brittle fracture, and tensile failure, which are the result of different mechanisms active during the processes of crack initiation, extension and fracture. The results of fractographic analysis are further verified by an analysis of the blade assembly speed with respect to time over the entire fracture failure process based on high-speed camera data. The results of fractographic analysis and blade assembly speed are fully consistent.

The aim of our study is to utilise longitudinal and representative national data to explore the extent that the association between the fast food environment and overweight in adolescents is confounded by neighbourhood deprivation. Longitudinal data from the Millennium Cohort Study for England were obtained for waves 5 (ages 11/12; 2011/12; n=13,469) and 6 (ages 14/15; 2014/15; n=11,884). Our outcome variable was overweight/obesity defined using age and sex-specific International Obesity Task Force cut points. Individuals were linked, based on their residential location, to data on the density of fast food outlets and neighbourhood deprivation. Structural Equation Models were used to model associations at both ages and explicitly test for confounding. While we found some evidence for an association between the number of fast food outlets and overweight, any associations disappeared following accounting for the confounding nature of neighbourhood deprivation. Neighbourhood deprivation was consistently associated to overweight, with adolescents who resided in deprived areas more likely to be overweight. Results were largely consistent depending on different methodological decisions. Our findings suggest that policy efforts should prioritise focusing on tackling the social determinants of excess body mass which will be more effective than interventions aimed at the built environment.

Deep learning has recently attracted much attention due to its excellent performance in processing audio, image, and video data. However, few studies are devoted to the field of automatic modulation classification (AMC). It is one of the most well-known research topics in communication signal recognition, which remains challenging for traditional methods due to the complex disturbance from other sources. This paper proposes a heterogeneous deep model fusion (HDMF) method to solve the problem in a unified framework. The contributions include: 1) The convolutional neural network (CNN) and long short-term memory (LSTM) are combined by two different ways without prior knowledge involved; 2) A large database, including eleven types of single-carrier modulation signals with various noises as well as a fading channel, is collected with various signal-to-noise ratios (SNRs) based on a real geographical environment; and 3) Experimental results demonstrate that HDMF is super capable of copping with the AMC problem, and achieves much better performance when compared with the independent network. The source code and the database will be publically available.

As the lifespan of porous asphalt pavement extends, it invariably experiences performance degradation and an increase in tire/pavement noise compared with the original surface when it is newly built. To understand the medium-term performance of porous asphalt pavement during service and reveal the influence of different pavement performance indicators on tire/pavement noise generation, a seven-year continuous observation and data analysis study was conducted. Porous asphalt with different designed material parameters was taken into account, including thin-layer porous asphalt overlay with a maximum nominal aggregate size of 10mm (PUC-10), porous asphalt with a maximum nominal aggregate size of 13mm (PAC-13), and a two-layer porous asphalt incorporating both PUC-10 and PAC-13 (PUC-10+PAC-13). Using automated pavement technology testing equipment, key performance indexes, including damage rate (DR), rut depth (RD), international roughness index (IRI), and sideway force coefficient (SFC), were measured and calculated. The on-board sound intensity (OBSI) method was used to test noise levels in three types of porous asphalt pavement structures and on a dense graded easy compact asphalt concrete (ECA) for comparison. The progressions of the medium-term performances were analyzed according to the measured results. The correlations between tire/pavement noise levels and testing speed, testing lane, pavement structure, and road condition indicators were investigated by regression. The research results revealed that the DR and SFC of the three types of porous asphalt pavement displayed more significant attenuation in the service period, while the IRI and RD showed small changes. Among them, the SFC demonstrated significant attenuation over the first 4 years, decreasing between 13.8% and 17.0%. From the 3-7 years of operation, the transverse crack disease in porous asphalt pavement progressed rapidly. The three porous asphalt pavement structures tested exhibited commendable noise reduction capabilities. Compared to ECA, the OBSI noise levels of these structures were lower by 2.09 dB, 1.53 dB, and 2.88 dB, respectively. Compared to other pavement condition indicators, RD had a more significant impact on the OBSI value of the PUC-10 porous asphalt pavement. For the PUC-10+PAC-13 porous asphalt pavement, there was a significant linear relationship between the OBSI value and SFC, mainly due to polishing of the coarse aggregate surface. It leads to a decrease of the micro-texture, which lies in increasing tire/pavement noise at high-frequency ranges and decreasing SFC. This study will have guiding significance for the research of the attenuated law and durability improvement of porous asphalt tire/pavement noise.

Precipitation type is a key parameter used for better retrieval of precipitation characteristics as well as to understand the cloud-convection-precipitation coupling processes. Ice crystals and water droplets inherently exhibit different characteristics in different precipitation regimes (e.g., convection, stratiform), which reflect on satellite remote sensing measurements that help us distinguish them. The Global Precipitation Measurement (GPM) Core Observatory’s Microwave Imager (GMI) and Dual-Frequency Precipitation Radar (DPR) together provide ample information on global precipitation characteristics. As an active sensor, DPR provides an accurate precipitation type assignment, while passive sensors like GMI are traditionally only used for empirical understanding of precipitation regimes. Using collocated precipitation type flags from DPR as the “truth”, this paper employs machine learning (ML) models to train and test the predictability and accuracy of using passive GMI-only observations together with ancillary information from reanalysis and GMI surface emissivity retrieval products. Out of six ML models, four simple ones (Support Vector Machine, Neural Network, Random Forest, and Gradient Boosting) and the 1-D convolutional neural network (CNN) model are identified to produce 90% - 94% prediction accuracy globally for 5 types of precipitation (convective, stratiform, mixture, no precipitation, and other precipitation), which is much more robust than previous similar effort. One novelty of this work is to introduce data augmentation (subsampling and bootstrapping) to handle extremely unbalanced samples in each category. Careful evaluation of Impact matrices demonstrate that polarization difference (PD) and surface emissivity at high-frequency channels dominate the decision process, which are consistent with the physical understanding of polarized microwave radiative transfer over different surface types, as well as in snow and liquid clouds with different microphysical properties. Furthermore, the view-angle dependency artifact that DPR precipitation flag bears with does not propagate into the conical-viewing GMI retrievals. This work provides a new and promising way for future physics-based ML retrieval algorithm development.