基于轻量化快速卷积与双向加权特征融合网络的船舶裂纹检测

[目的]针对人工目视与超声波方法的船舶裂纹检测存在效率低下、成本高昂和危险性高的特点,提出一种基于深度学习的船舶裂纹检测方法。[方法]首先,在原模型YOLOv5s的主干网络中使用轻量化卷积结构(GSConv)替代标准卷积并融入注意力机制,在降低主干网络参数量与计算量的同时,提升主干网络对裂纹特征的提取能力;然后,在网络的颈部使用基于PConv构建的C3_Faster替代原C3模块,提升模型的图像处理速度,增强模型快速性;最后,设计一种简化的双向加权特征融合网络(BiFFN)以改进原模型YOLOv5s中的特征聚合网络,提升裂纹的语义信息与位置信息的融合效果,以及模型对裂纹的识别准确度与定位精度。[结果]通过对船舶裂纹原始数据与增强数据的学习,所提改进模型实现了94.11%的检测精度和93.50%的召回率,模型的计算量降低了17.93%,参数量降低了15.81%。[结论]研究表明,基于轻量化快速卷积与双向加权特征融合网络(MLF-YOLO)的船舶裂纹检测方法,实现了模型轻量化与较高的检测精度和召回率,结果可为开发自主无人机船舶检测提供参考。

快速预测混沌控制效果的一个量化指标

本文对Pyragas的时滞自反馈控制方法作了深入详细的研究，并从瞬态Lyapunov特性指数（ILCE）和短期平均Lyapunov特性指数（SLCE）的角度进一步揭示了该方法的控制机理，提出了一个利用瞬态Lyapunov特性指数实时反映混沌控制效果的量化指标P。

基于动态跟踪步长的直线快速矢量化算法

工程图纸扫描识别一直是模式识别中的难点问题。为了克服细化算法逐像素处理的局部特性,已提出了一些非细化算法。这些算法在不同程度上突破了细化算法在像素层次做局部操作的局限,力图抓住图像的宏观特征进行整体识别,但在处理小线素、曲线图像方面有较大局限性。文章采用动态扫描的概念,不对图像编码,利用随机稀疏扫描抽取跟踪头,利用跟踪头信息指导后续跟踪。跟踪时采用动态调节步长,得到了更精确的、完备的低级矢量化结果,为后续矢量处理提供了良好的基础。文章提出的算法可适用于各种不同工程图纸的处理,地理信息数据的提取。

同跨距、不同起升重量门式起重机金属结构快速轻量化设计

为实现跨距相同、起升重量不同的系列门式起重机金属结构快速轻量化设计,在对门式起重机进行全参数化建模和求解分析的基础上,对门式起重机主梁、上下端梁及支腿的截面尺寸和厚度等设计变量做灵敏度分析,得出主梁的截面尺寸及厚度对整机金属结构的重量影响最大,主梁和上端梁的截面尺寸对起重机悬臂刚度影响最大。以门式起重机重要截面尺寸结合灵敏度分析结果选取关键设计变量,同时结合零阶优化算法对整机金属结构进行系列化设计,在满足设计要求下重量比优化前减轻了11.1%。

水利地形信息快速矢量化方法浅谈

水利信息调查中,经常需要对相应区域地形起伏情况有所掌握。测绘部门一般只提供纸质版区域地形图,购买电子版矢量化图纸需要较多资金。本文结合作者工作经验,利用google earth和autocad civil3D两款软件以河北大黑汀水库上游左岸松岭村为例,介绍获取相应区域地形矢量图以及之后与其它软件间数据传递的方法。

基于相轨迹的机组阻尼系数最小二乘量化方法

水电机组的负阻尼特性是诱发水电主导电力系统超低频振荡的内在原因,准确高效地量化机组的阻尼系数对超低频振荡主导机组的定位与控制至关重要。针对现有非线性系统机组阻尼量化方法难以兼顾操作简便性与计算准确性的问题,依据阻尼转矩理论分析单机和多机系统的阻尼特性,提出了一种新的阻尼系数快速量化方法,一方面将阻尼的量化结果与理论值对比验证方法的计算精度,另一方面利用多机系统超低频振荡实验分析与实际电站的测量数据计算验证方法的有效性。结果表明方法准确、简单,可为强非线性系统的阻尼研究与超低频振荡时机组阻尼的在线评估提供有效的技术手段。

起重机快速轻量化设计系统研究及应用

参数化设计和结构优化设计是复杂结构快速轻量化设计系统的核心,也是促进起重机轻量化设计技术在企业推广应用的关键。但是,现有起重机快速设计系统大多以实现参数化设计为主,难以实现结构的自动优化设计。而已公开的优化方法实现过程复杂,且大多采用连续变量优化,结果难以直接运用于实际生产。进行系列化起重机的参数化建模,建立多位置、多状态的载荷模型,集成含零阶优化、参数圆整、局部再优化的二次优化设计方法,开发了具有参数化设计、轻量化优化等功能的起重机快速轻量化设计系统,并完成了箱型门架的快速轻量化设计系统与软件。应用结果表明,该系统显著缩短了设计周期,减少了设计人员的大量繁琐计算工作量。仅对部分截面尺寸进行优化,虽然门架应力应变在容许范围内有所增加,但整机质量下降10.5%,节约了后期成本。

Effects of Interdiffusion on Luminescence of InAs/GaAs Quantum Dots Covered by InGaAs Overgrowth Layer

WT8.BZ]The effects of postgrowth rapid thermal annealing have been studied on the optical properties of 3-nm-height InAs/GaAs quantum dots covered by 3-nm-thick In xGa 1-x As (x=0,0 1 and 0 2) overgrowth layer.At a higher annealing temperature (T≥750℃),the photoluminescence peak of InGaAs layer has been observed at the lower-energy side of InAs quantum-dot peak.In addition,a similar blueshift in photoluminescence (PL) emission energy is observed for all samples when the annealing temperature increases from 650 to 850℃.However,the trend of photoluminescence linewidth towards narrowing is totally different for InAs quantum dots with different In mole fraction in InGaAs overgrowth layer.The results suggest that the intermixing in the lateral direction plays an important role in obtaining a better understanding of the modification of optical properties induced by the rapid thermal annealing.

Comparison on Soil Carbon Stocks Between Urban and Suburban Topsoil in Beijing, China

The urban population and urbanized land in China have both increased markedly since the 1980 s. Urban and suburban developments have grown at unprecedented rates with unknown consequences for ecosystem functions. In particular, the effect of rapid urbanization on the storage of soil carbon has not been studied extensively. In this study, we compared the soil carbon stocks of different land use types in Beijing Municipality. We collected 490 top-soil samples(top 20 cm) from urban and suburban sites within the Sixth Ring Road of Beijing, which cover approximately 2400 km2, and the densities of soil organic carbon(SOC), soil inorganic carbon(SIC), and total carbon(TC) were analyzed to determine the spatial distribution of urban and suburban soil carbon characteristics across seven land use types. The results revealed significant differences in soil carbon densities among land use types. Additionally, urban soil had significantly higher SOC and SIC densities than suburban soil did, and suburban shelterbelts and productive plantations had lower SIC densities than the other land use types. The comparison of coefficients of variance(CVs) showed that carbon content of urban topsoil had a lower variability than that of suburban topsoil. Further findings revealed that soil carbon storage increased with built-up age. Urban soil built up for more than 20 years had higher densities of SOC, SIC and TC than both urban soil with less than 10 years and suburban soil. Correlation analyses indicated the existence of a significantly negative correlation between the SOC, SIC, and TC densities of urban soil and the distance to the urban core, and the distance variable alone explained 23.3% of the variation of SIC density and 13.8% of the variation of TC density. These results indicate that SOC and SIC accumulate in the urban topsoil under green space as a result of the conversion of agricultural land to urban land due to the urbanization in Beijing.

Characterization of Air Pollution in Urban Areas of Yangtze River Delta,China

The hallmark of development in the Yangtze River Delta(YRD) of East China has been sprawling urbanization. However, air pollution is a significant problem in these urban areas. In this paper, we investigated and analyzed the air pollution index(API) in four cities(Shanghai, Nanjing, Hangzhou and Ningbo) in the YRD from 2001 to 2012. We attempted to empirically examine the relationship between meteorological factors and air quality in the urban areas of the YRD. According to the monitoring data, the API in Shanghai, Nanjing, Hangzhou slightly declined and that in Ningbo increased over the study period. We analyzed the inter-annual, seasonal, and monthly variations of API, from which we found that the air quality had different temporal changes in the four cities. It was indicated that air quality was poor in winter and spring and best in summer. Furthermore, different weather conditions affected air quality level. The wind direction was considered as an important and influential factor to air pollution, which has an impact on the accumulating or cleaning processes of pollutants. The air quality was influenced by the different wind directions that varied with seasons and cities.

低速风洞试验模型轻量化快速制造技术

介绍了某飞机模型后机身和滑轨整流罩零部件制造过程,并与机加零件进行经济性对比分析。风洞试验模型桶段类后机身部件采用轻量化树脂泥模具及VARI工艺制造方法,重量减轻58%,周期缩短16%,成本增加20%。滑轨整流罩等异形零件采用选区熔化增材制造方式,与机加滑轨整流罩相比减重86%,研制周期和成本略有增加;与复材滑轨整流罩相比减重48%,制造周期缩短42%,成本降低28%。

Development of a real-time PCR assay(SYBR Green I) for rapid identification and quantification of scyphomedusae Aurelia sp.1 planulae

The complicated life cycle ofAurelia spp., comprising benthic asexually-reproducing polyps and sexually-reproducing medusae, makes it hard for researchers to identify and track them, especially for early stage individuals, such as planulae. To solve this problem, we developed a real-time PCR assay （SYBR Green I） to identify planulae in both cultured and natural seawater samples. Species-specific primers targeting Aurelia sp.1 mitochondrial 16S rDNA （mr 16S rDNA） regions were designed. Using a calibration curve constructed with plasmids containing the Aurelia sp. 1 mt 16S rDNA fragment and a standard curve for planulae, the absolute number of mt 16S rDNA copies per planula was determined and from that the total number ofplanulae per sample was estimated. For the field samples, a 100-fold dilution of the sample DNA combined with a final concentration of 0.2 μg/μL BSA in the PCR reaction mixture was used to remove real- time PCR inhibitors. Samples collected in Jiaozhou Bay from July to September 2012 were subsequently analyzed using this assay. Peak Aurelia sp.1 planula abundance occurred in July 2012 at stations near Hongdao Island and Qingdao offshore; abundances were very low in August and September. The real-time PCR assay （SYBR Green I） developed here negates the need for traditional microscopic identification, which is laborious and time-consuming, and can detect and quantify jellyfish planulae in field plankton samples rapidly and specifically.

Simulation of Volume and Heat Transport along 26.5°N in the Atlantic

The observed meridional overtuming circula- tion （MOC） and meridional heat transport （MHT） estimated from the Rapid Climate Change/Meridional Circu- lation and Heat Flux Array （RAPID/MOCHA） at 26.5°N are used to evaluate the volume and heat transport in the eddy-resolving model LASG/IAP Climate system Ocean Model （LICOM）. The authors find that the Florida Cur- rent transport and upper mid-ocean transport of the model are underestimated against the observations. The simulated variability of MOC and MHT show a high correlation with the observations, exceeding 0.6. Both the simulated and observed MOC and MHT show a significant seasonal variability. According to the power spectrum analysis, LICOM can represent the mesoscale eddy characteristic of the MOC similar to the observation. The model shows a high correlation of 0.58 for the internal upper mid-ocean transport （MO） and a density difference between the western and eastern boundaries, as noted in previous studies.

Scanning electron microscopy coupled with energydispersive X-ray spectrometry for quick detection of sulfuroxidizing bacteria in environmental water samples

Detection of sulfur-oxidizing bacteria has largely been dependent on targeted gene sequencing technology or traditional cell cultivation, which usually takes from days to months to carry out. This clearly does not meet the requirements of analysis for time-sensitive samples and/or complicated environmental samples. Since energy-dispersive X-ray spectrometry(EDS) can be used to simultaneously detect multiple elements in a sample, including sulfur, with minimal sample treatment, this technology was applied to detect sulfur-oxidizing bacteria using their high sulfur content within the cell. This article describes the application of scanning electron microscopy imaging coupled with EDS mapping for quick detection of sulfur oxidizers in contaminated environmental water samples, with minimal sample handling. Scanning electron microscopy imaging revealed the existence of dense granules within the bacterial cells, while EDS identified large amounts of sulfur within them. EDS mapping localized the sulfur to these granules. Subsequent 16S rRNA gene sequencing showed that the bacteria detected in our samples belonged to the genus Chromatium, which are sulfur oxidizers. Thus, EDS mapping made it possible to identify sulfur oxidizers in environmental samples based on localized sulfur within their cells, within a short time(within 24 h of sampling). This technique has wide ranging applications for detection of sulfur bacteria in environmental water samples.

Robust Digital Audio Watermarking Scheme Using Blind Source Separation with Global Optimal Property

The paper proposes a robust digital audio watermarking scheme using blind source separation(BSS) based on the global optimization of independency metric(IM),which is formulated as a generalized eigenvalue(GE) problem.Compared with traditional information-theoretical approaches used in digital audio watermarking,such as fast independent component analysis(FastICA),the proposed scheme has lower complexity without timeconsuming iteration steps used in FastICA.To make full use of the multiresolution characteristic of discrete wavelet transform(DWT) and the energy compression characteristic of discrete cosine transform(DCT),the watermark is embedded in the middle DWT-DCT coefficients and the independent component analysis(ICA) approach based on IM is used in the detecting scheme.Simulation results based on Stirmark for Audio v02 show that the proposed scheme has strong robustness as well as the imperceptibility and security.

Near-infrared spectroscopy and chemometric modelling for rapid diagnosis of kidney disease

Rapid diagnosis is important for efficient treatment in clinical medicine.This study aimed at development of a method for rapid and reliable diagnosis using near-infrared(NIR)spectra of human serum samples with the help of chemometric modelling.The NIR spectra of sera from 48 healthy individuals and 16 patients with suspected kidney disease were analyzed.Discrete wavelet transform(DWT)and variable selection were adopted to extract the useful information from the spectra.Principal component analysis(PCA),linear discriminant analysis(LDA)and partial least squares discriminant analysis(PLSDA)were used for discrimination of the samples.Classification of the two-class sera was obtained using LDA and PLSDA with the help of DWT and variable selection.DWT-LDA produced 93.8%and 83.3%of the recognition rates for the validation samples of the two classes,and 100%recognition rates were obtained using DWT-PLSDA.The results demonstrated that the tiny differences between the spectra of the sera were effectively explored using DWT and variable selection,and the differences can be used for discrimination of the sera from healthy and possible patients.NIR spectroscopy and chemometrics may be a potential technique for fast diagnosis of kidney disease.