ALBERT

Description: Recently, deep convolutional neural networks (DCNN) have obtained promising results in single image super-resolution (SISR) of remote sensing images. Due to the high complexity of remote sensing image distribution, most of the existing methods are not good enough for remote sensing image super-resolution. Enhancing the representation ability of the network is one of the critical factors to improve remote sensing image super-resolution performance. To address this problem, we propose a new SISR algorithm called a Deep Residual Squeeze and Excitation Network (DRSEN). Specifically, we propose a residual squeeze and excitation block (RSEB) as a building block in DRSEN. The RSEB fuses the input and its internal features of current block, and models the interdependencies and relationships between channels to enhance the representation power. At the same time, we improve the up-sampling module and the global residual pathway in the network to reduce the parameters of the network. Experiments on two public remote sensing datasets (UC Merced and NWPU-RESISC45) show that our DRSEN achieves better accuracy and visual improvements against most state-of-the-art methods. The DRSEN is beneficial for the progress in the remote sensing images super-resolution field.

Description: Object detection on very-high-resolution (VHR) remote sensing imagery has attracted a lot of attention in the field of image automatic interpretation. Region-based convolutional neural networks (CNNs) have been vastly promoted in this domain, which first generate candidate regions and then accurately classify and locate the objects existing in these regions. However, the overlarge images, the complex image backgrounds and the uneven size and quantity distribution of training samples make the detection tasks more challenging, especially for small and dense objects. To solve these problems, an effective region-based VHR remote sensing imagery object detection framework named Double Multi-scale Feature Pyramid Network (DM-FPN) was proposed in this paper, which utilizes inherent multi-scale pyramidal features and combines the strong-semantic, low-resolution features and the weak-semantic, high-resolution features simultaneously. DM-FPN consists of a multi-scale region proposal network and a multi-scale object detection network, these two modules share convolutional layers and can be trained end-to-end. We proposed several multi-scale training strategies to increase the diversity of training data and overcome the size restrictions of the input images. We also proposed multi-scale inference and adaptive categorical non-maximum suppression (ACNMS) strategies to promote detection performance, especially for small and dense objects. Extensive experiments and comprehensive evaluations on large-scale DOTA dataset demonstrate the effectiveness of the proposed framework, which achieves mean average precision (mAP) value of 0.7927 on validation dataset and the best mAP value of 0.793 on testing dataset.

Description: A three-dimensional finite element model of rail grinding was established to explore the effects of grinding passes and grinding direction on the material removal behaviour of grinding rails during the grinding process. The results indicate that as the number of grinding passes increases, a decrease in the grinding force reduces both the amount of removed rail material and the surface roughness. There is a decrease in the grinding ratio caused by the increase in the wear on the grinding wheel and the decreased removal of the rail material. When the grinding direction changes, the wear of the grinding wheel decreases, which is contrary to the increasing trend of the amount of removed rail material, the grinding ratio, the surface roughness and the grinding force.

Description: The authors wish to make the following corrections to this paper [...]

Description: The Tiaoxi River is the main source of water for Lake Taihu and can result in plumes in the lake after heavy precipitation events. These plumes have played a crucial role in the water quality changes within the lake. High temporal resolution GOCI (Geostationary Ocean Color Imager) data were used to study the spatial distribution of the total suspended matter concentration in Lake Taihu after heavy precipitation events in the Tiaoxi River Basin via an empirical model. The plumes were analyzed after two heavy precipitation events in 2011 and 2013 using 16 GOCI images, which indicated that the Tiaoxi River had a great influence on the spatial distributions of total suspended matter and algal blooms. It was concluded that the main factors affecting the plumes in the Tiaoxi River were precipitation intensity, runoff, and total suspended matter concentration. Human activity, such as sand excavation also played a crucial role in sediment discharge. The results of this study demonstrate that the visualization of GOCI data makes it possible to use remote sensing technology to continuously monitor an inland water environment on an hourly scale, which is of great significance for studying the diffusion and evolution of river plumes.

Description: In multi-sensor fusion (MSF), the integration of multi-sensor observation data with different observation errors to achieve more accurate positioning of the target has always been a research focus. In this study, a modified ensemble Kalman filter (EnKF) is presented to substitute the traditional Kalman filter (KF) in the multiple hypotheses tracking (MHT) to deal with the high nonlinearity that always shows up in multiple target tracking (MTT) problems. In addition, the multi-source observation data fusion is also realized by using the modified EnKF, which enables the low-precision observation data to be corrected by high-precision observation data, and the accuracy of the corrected data can be calibrated by the statistical information provided by the EnKF. Numerical studies are given to demonstrate the effectiveness of our proposed method and the results show that the MHT-EnKF method can achieve remarkable enhancement in dealing with nonlinear movement variation and positioning accuracy for MTT problems in MSF scenario.

Description: Effective determination of trends in sulfur dioxide emissions facilitates national efforts to draft an appropriate policy that aims to lower sulfur dioxide emissions, which is essential for reducing atmospheric pollution. However, to reflect the current situation, a favorable emission reduction policy should be based on updated information. Various forecasting methods have been developed, but their applications are often limited by insufficient data. Grey system theory is one potential approach for analyzing small data sets. In this study, an improved modeling procedure based on the grey system theory and the mega-trend-diffusion technique is proposed to forecast sulfur dioxide emissions in China. Compared with the results obtained by the support vector regression and the radial basis function network, the experimental results indicate that the proposed procedure can effectively handle forecasting problems involving small data sets. In addition, the forecast predicts a steady decline in China’s sulfur dioxide emissions. These findings can be used by the Chinese government to determine whether its current policy to reduce sulfur dioxide emissions is appropriate.

Description: Intense interest has been given to the fabrication of micro-through-holes with smaller tapering and higher aspect ratios in engineering ceramics due to their wide range of applications in MEMS and aerospace. A semi-water-immersed laser micro-trepanning (SWILT) scheme is proposed and investigated in this paper with alumina ceramics as the target material, and its performance is assessed and compared with the direct laser trepanning method. Relevant processing parameters influencing the trepanning process are explored through an orthogonally designed experiment, and their effects on hole profiles are adequately discussed to yield optimized parameters. It is revealed that SWILT is capable of producing micro-through-holes with minimized hole tapering and much straighter sidewalls compared with the direct trepanning results, whereas the ablated surface quality is relatively rougher. The micro-through-hole formation mechanisms are also amply analyzed, where the transition hole development may be purely attributed to the laser-material interaction in the direct laser trepanning condition, while the SWILT case features an enhanced material-removal rate, especially at the lower part of the through-hole. The latter is due to the strengthened mechanical effects coming from the water-confined plasma zone and the following cavitation bubble collapse, which may efficiently expel the molten material from sidewalls and result in significantly reduced hole tapering.

Description: The fabrication of micro-holes in silicon substrates that have a proper taper, higher depth-to-diameter ratio, and better surface quality has been attracting intense interest for a long time due to its importance in the semiconductor and MEMS (Micro-Electro-Mechanical System) industry. In this paper, an experimental investigation of the machining performance of the direct and chemical-assisted picosecond laser trepanning of single crystalline silicon is conducted, with a view to assess the two machining methods. The relevant parameters affecting the trepanning process are considered, employing the orthogonal experimental design scheme. It is found that the direct laser trepanning results are associated with evident thermal defects, while the chemical-assisted method is capable of machining micro-holes with negligible thermal damage. Range analysis is then carried out, and the effects of the processing parameters on the hole characteristics are amply discussed to obtain the recommended parameters. Finally, the material removal mechanisms that are involved in the two machining methods are adequately analyzed. For the chemical-assisted trepanning case, the enhanced material removal rate may be attributed to the serious mechanical effects caused by the liquid-confined plasma and cavitation bubbles, and the chemical etching effect provided by NaOH solution.

Description: Rhabdomyosarcoma is subclassified by the presence or absence of a recurrent chromosome translocation that fuses the FOXO1 and PAX3 or PAX7 genes. The fusion protein (FOXO1-PAX3/7) retains both binding domains and becomes a novel and potent transcriptional regulator in rhabdomyosarcoma subtypes. Many studies have characterized and integrated genomic, transcriptomic, and epigenomic differences among rhabdomyosarcoma subtypes that contain the FOXO1-PAX3/7 gene fusion and those that do not; however, few investigations have investigated how gene co-expression networks are altered by FOXO1-PAX3/7. Although transcriptional data offer insight into one level of functional regulation, gene co-expression networks have the potential to identify biological interactions and pathways that underpin oncogenesis and tumorigenicity. Thus, we examined gene co-expression networks for rhabdomyosarcoma that were FOXO1-PAX3 positive, FOXO1-PAX7 positive, or fusion negative. Gene co-expression networks were mined using local maximum Quasi-Clique Merger (lmQCM) and analyzed for co-expression differences among rhabdomyosarcoma subtypes. This analysis observed 41 co-expression modules that were shared between fusion negative and positive samples, of which 17/41 showed significant up- or down-regulation in respect to fusion status. Fusion positive and negative rhabdomyosarcoma showed differing modularity of co-expression networks with fusion negative (n = 109) having significantly more individual modules than fusion positive (n = 53). Subsequent analysis of gene co-expression networks for PAX3 and PAX7 type fusions observed 17/53 were differentially expressed between the two subtypes. Gene list enrichment analysis found that gene ontology terms were poorly matched with biological processes and molecular function for most co-expression modules identified in this study; however, co-expressed modules were frequently localized to cytobands on chromosomes 8 and 11. Overall, we observed substantial restructuring of co-expression networks relative to fusion status and fusion type in rhabdomyosarcoma and identified previously overlooked genes and pathways that may be targeted in this pernicious disease.