期刊文献+
共找到267篇文章
< 1 2 14 >
每页显示 20 50 100
Light Vector Mesons Production in Ultrarelativistic Heavy Ion Collisions from the Color Glass Condensate
1
作者 余功明 杨海涛 李云德 《Communications in Theoretical Physics》 SCIE CAS CSCD 2015年第12期683-686,共4页
We calculate inclusive light vector mesons(ρ,ω,and φ) production in relativistic heavy ion collisions based on the idea of gluon saturation in the color glass condensate(CGC) framework.At high energies,the value of... We calculate inclusive light vector mesons(ρ,ω,and φ) production in relativistic heavy ion collisions based on the idea of gluon saturation in the color glass condensate(CGC) framework.At high energies,the value of saturation momentum becomes larger than the Quantum Chromodynamics(QCD) confinement scale ∧_(QCD),which implies thatα_s(Q_s) <<1.Using the relativistic kinetic theory,we derive the production rate for the light vector mesons produced by gluon-gluon fusion in the k_T-factorization approach.The numerical results indicate that the light vector mesons from the color glass condensate become prominent in p-p,p-Pb,and Pb-Pb collisions at Large Hadron Collider(LHC) energies. 展开更多
关键词 light vector MESON COLOR GLASS CONDENSATE relativi
原文传递
基于最小二乘向量寻优法的三线结构光标定
2
作者 陈琳 江山 +2 位作者 朱训 潘海鸿 申毅莉 《应用激光》 北大核心 2026年第3期71-79,共9页
为解决三线结构光传感器标定过程中,光平面拟合精度低的问题,提出一种基于最小二乘向量寻优的光平面标定方法。首先,使用Steger算法和随机抽样一致性(random sample consensus,RANSAC)算法提取并拟合光条中心线;然后,提取光条与标定板... 为解决三线结构光传感器标定过程中,光平面拟合精度低的问题,提出一种基于最小二乘向量寻优的光平面标定方法。首先,使用Steger算法和随机抽样一致性(random sample consensus,RANSAC)算法提取并拟合光条中心线;然后,提取光条与标定板的所有列角点拟合直线的交点作为标定点,并通过PnP(perspective-n-point)法计算这些标定点在相机坐标系下的表示;接着,将这些标定点转换为空间向量,增加标定数据的样本量;最后,使用改进的最小二乘法求解空间向量组的最优解,进而求解出光平面参数。使用该方法对两种不同尺寸的垫片进行重复测量,平均测量误差达到0.021 mm,均方根误差(RMSE)达到0.025 mm。此外,与传统最小二乘法和RANSAC算法相比,最小二乘向量寻优法的RMSE降低了约20%~40%,验证了其优越性。 展开更多
关键词 最小二乘向量寻优 三线结构光 光平面标定 平面拟合
原文传递
人LIGHT基因的克隆及其重组腺病毒载体的构建 被引量:8
3
作者 吴东 沈锋 +2 位作者 娄永华 焦炳华 吴孟超 《细胞与分子免疫学杂志》 CAS CSCD 北大核心 2002年第2期121-124,共4页
目的  克隆人LIGHT基因,构建其重组腺病毒载体,以研究LIGHT过表达与腺病毒感染对细胞生长的影响。方法用RT-PCR法从人外周血单个核细胞(PBMC)中克隆人的LIGHT全长基因。将LIGHT cDNA克隆到穿... 目的  克隆人LIGHT基因,构建其重组腺病毒载体,以研究LIGHT过表达与腺病毒感染对细胞生长的影响。方法用RT-PCR法从人外周血单个核细胞(PBMC)中克隆人的LIGHT全长基因。将LIGHT cDNA克隆到穿棱载体pAdTrack-CMV-LIGHT重组质粒中,经酶切线性化后,将重组质粒pAdTrack-CMV-LIGHT和骨架质粒pAdEasy-1,以电穿孔法共转染大肠杆菌BJ5183,获得重组腺病毒质粒。最后,将线性化的重组腺病毒质粒转染293细胞包装获得重组腺病毒,用荧光显微镜、PCR及Western blot分析LIGHT基因的表达。 结果 用RT-PCR法从人的PBMC中,扩增出723 bp的cDNA,测序证实为人LIGHT基因。荧光显微镜、PCR及Western blot证实,LIGHT重组腺病毒可感染293细胞,并在293细胞内进行有效的复制。结论成功地克隆了人LIGHT基因,并构建了其重组腺病毒载体,为进一步研究LIGHT基因的功能提供了条件。 展开更多
关键词 腺病毒 表达载体 light基因 基因克隆 肿瘤 基因治疗
原文传递
人LIGHT基因重组慢病毒的构建以及在脐血间质干细胞中的表达
4
作者 马贵亮 宣世英 +1 位作者 朱新红 毛伟征 《临床普外科电子杂志》 2013年第1期16-21,共6页
目的构建带有人LIGHT基因的慢病毒载体,观察其在人脐血间质干细胞中的表达。方法通过逆转录聚合酶链反应从PCD DNA-LIGHI、质粒中获得人LIGHT基因,利用infusion技术重组构建慢病毒载体质粒pGC-FU-LIGHT,在脂质体lipofectamine 2000介导... 目的构建带有人LIGHT基因的慢病毒载体,观察其在人脐血间质干细胞中的表达。方法通过逆转录聚合酶链反应从PCD DNA-LIGHI、质粒中获得人LIGHT基因,利用infusion技术重组构建慢病毒载体质粒pGC-FU-LIGHT,在脂质体lipofectamine 2000介导下与结构质粒pHelper1.0及包膜蛋白质粒pHelper2.0共转染293T细胞包装生产慢病毒。将人脐血间质干细胞分为实验组(pGC-FU-LIGHT)、空载体对照组(pGC-FU-EGFP)及空白组(脐血间质干细胞),分别用重组慢病毒、空载慢病毒、PBS感染后,采用RT-PCR以及Elisa检测IIGHT表达情况。结果所获LIGHT基因经测序后与Gene Bank报道序列完全一致;重组慢病毒载体质粒pGC-FU-LIGHT经鉴定正确;三质粒共转染293T细胞成功,收集、浓缩病毒后测定其滴度为2×107TUJ/L,感染UJCBMSCs后RT-PCR、Elisa检测各组细胞均有LIGHT的表达,其中试验组pGC-FU-LIGHT组更大最表达LIGHT,与其余2组(pGC-FU-EGFP、UJCBMSCs)比较差异具有统计学意义。结论成功构建带有LIGHT基因的慢病毒载体并实现在脐血间质干细胞中的表达,为间充质干细胞移植治疗胃癌的应用奠定了基础。 展开更多
关键词 慢病毒载体 light基因 人脐血间质干细胞
在线阅读 下载PDF
基于LightGBM的车载激光雷达点云分类 被引量:2
5
作者 赵佩佩 张卫星 《应用激光》 CSCD 北大核心 2023年第10期148-155,共8页
车载激光扫描系统获取的海量三维点云自动分类对目标识别和重建有着重要意义,传统点云分类需要人工干预,而现有的自动分类算法大多存在分类准确率低和运算成本高等问题。对此,提出了一种基于LightGBM模型的车载点云自动分类方法,该方法... 车载激光扫描系统获取的海量三维点云自动分类对目标识别和重建有着重要意义,传统点云分类需要人工干预,而现有的自动分类算法大多存在分类准确率低和运算成本高等问题。对此,提出了一种基于LightGBM模型的车载点云自动分类方法,该方法首先计算点云的表面变化三维特征、密度特征、高程特征以及快速点特征直方图,并计算点云法向量与其邻域点的法向量夹角及与水平面的夹角作为约束特征,将所有结果合并得到48维特征向量,然后构建LightGBM模型对点云特征向量训练完成分类预测。试验表明该算法能够准确高效地完成车载雷达点云的自动分类,比对照组算法总精度平均提高8.1%,Kappa系数平均提高18.9%,计算时间平均减少73.7%。 展开更多
关键词 车载激光雷达 点云分类 特征向量 轻量级梯度提升机 法向量
原文传递
基于偏振光谱成像技术的物体表面粗糙度检测方法应用探讨
6
作者 董三主 《天津科技》 2026年第2期56-59,共4页
为避免对物体表面进行粗糙度检测时易因操作失误导致检测结果异常的问题,提出基于偏振光谱成像技术的物体表面粗糙度检测方法。利用电矢量对Mach-Zehnder干涉仪光波的偏振态进行深入分析,以获取物体表面的偏振光谱图像;接着利用穆勒矩... 为避免对物体表面进行粗糙度检测时易因操作失误导致检测结果异常的问题,提出基于偏振光谱成像技术的物体表面粗糙度检测方法。利用电矢量对Mach-Zehnder干涉仪光波的偏振态进行深入分析,以获取物体表面的偏振光谱图像;接着利用穆勒矩阵定量分析光与表面相互作用后的去极化效应,建立偏振态与粗糙度之间的精确映射关系。测量结果的波动范围稳定在0.03 nm以内,显著优于其他两种主流视觉检测方法,验证了该方法在纳米级粗糙度检测方面的高精度与高稳定性,为超精密制造领域的表面质量控制提供了可靠的技术支持。 展开更多
关键词 偏振光谱成像 表面粗糙度 电矢量 光波偏振态 穆勒矩阵
在线阅读 下载PDF
低光场景下考虑卷积生成对抗网络的人脸识别方法研究
7
作者 徐学钰 《计算机应用文摘》 2026年第6期52-55,共4页
针对现有低光环境下人脸识别方法因图像模糊、细节丢失而导致识别准确率显著下降的问题,提出一种基于卷积生成对抗网络(Convolutional Generative Adversarial Network,CGAN)的低光人脸识别方法。该方法首先利用CGAN重建清晰人脸图像,... 针对现有低光环境下人脸识别方法因图像模糊、细节丢失而导致识别准确率显著下降的问题,提出一种基于卷积生成对抗网络(Convolutional Generative Adversarial Network,CGAN)的低光人脸识别方法。该方法首先利用CGAN重建清晰人脸图像,以缓解低光引起的图像模糊;其次通过鲁棒编码技术对重建图像进行修复,提升图像质量与抗干扰能力;进一步提取光照不变特征并编码为紧凑向量,增强特征对光照变化的适应性;最后通过计算特征相似度实现身份识别。实验结果表明,和3种现有方法相比,该方法在低光场景下具有更高的识别精度与鲁棒性,验证了其在复杂光照条件下的有效性和优越性。 展开更多
关键词 低光场景 CGAN 人脸识别 鲁棒编码 紧凑向量
在线阅读 下载PDF
基于SVM-LightGBM算法的上市公司财务数据异常识别模型 被引量:4
8
作者 邵永运 张立莹 《沈阳师范大学学报(自然科学版)》 CAS 2023年第5期464-468,共5页
上市公司是否依法准确披露财务数据对资本市场的稳定发展有重要影响。建立财务数据异常识别模型,对规范财务数据报表和避免财务数据造假具有十分重要的意义。运用机器学习相关技术,将上市公司财务报表数据按照行业予以划分并使用特征工... 上市公司是否依法准确披露财务数据对资本市场的稳定发展有重要影响。建立财务数据异常识别模型,对规范财务数据报表和避免财务数据造假具有十分重要的意义。运用机器学习相关技术,将上市公司财务报表数据按照行业予以划分并使用特征工程完成各行业财务异常指标的选取,然后使用支持向量机算法和轻量级梯度提升算法,建立双层财务数据异常识别的混合模型,对2667家上市公司财务数据进行了实证研究。结果表明,在财务异常识别方面,与其他模型相比,该模型的准确率等指标均有较大提高。 展开更多
关键词 财务造假 机器学习 支持向量机算法 轻量级梯度提升算法
在线阅读 下载PDF
The Third Polarization of Light
9
作者 Richard A. Hutchin 《Optics and Photonics Journal》 2015年第2期29-32,共4页
We are all taught that there are only two polarizations of light because Maxwell’s equations only support two polarizations. This is mathematically true for the electromagnetic fields, but we have learned since the d... We are all taught that there are only two polarizations of light because Maxwell’s equations only support two polarizations. This is mathematically true for the electromagnetic fields, but we have learned since the days of Maxwell that the “real” electromagnetic field is not the electromagnetic field tensor Fμv (composed of Electric and Magnetic field terms) but rather the electromagnetic vector potential Aμ. When considered carefully, this requires a third polarization of light with very unusual properties. This third polarization of light does not generate electric or magnetic fields but should be detectable by its impact on supercurrents or quantum interference. It is also unavoidable since it automatically appears under Lorentz transformations to different moving frames. 展开更多
关键词 LOCAL THEORY of light THIRD POLARIZATION vector Potential
暂未订购
Impact of Different Light Characteristics on the Growth and Lipid Content of Diatom Phaeodactylum tricornutum Transconjugant Strains
10
作者 Nikunj Sharma Elisa Ines Fantino +4 位作者 Fatima Awwad Natacha Mérindol Arun Augustine Fatma Meddeb Isabel Desgagné-Penix 《American Journal of Plant Sciences》 CAS 2023年第1期41-63,共23页
Light regulates important metabolic processes in microalgal cells, which can further impact the metabolism and the accumulation of biomolecules such as lipids, carbohydrates, and proteins. Different characteristics of... Light regulates important metabolic processes in microalgal cells, which can further impact the metabolism and the accumulation of biomolecules such as lipids, carbohydrates, and proteins. Different characteristics of light have been studied on various strains of the model diatom Phaeodactylum tricornutum, but not on transconjugant cells and information on wild-type strains is still limited. Therefore, we studied the impact of different light characteristics such as spectral quality, light intensity and light shift on the growth, and the composition in lipids and fatty acids of P. tricornutum cells to provide a comprehensive context for future applications. Initially, we tested the impact of spectral quality and light intensity on P. tricornutum transformed with an episomal vector (Ptev), harboring the resistance gene Sh ble. Results indicated that Ptev cells accumulated more biomass and overall lipids in spectral quality Red 1 (R1: 34% > 600 nm > 66%) more effectively as compared to Red 2 (R2: 8% > 600 nm > 92%). It was also detected that cell granularity was higher in R1 as compared to R2. Furthermore, by testing two light intensities 65 μmol&middot;m<sup>-2</sup>&middot;s<sup>-1</sup> and 145 μmol&middot;m<sup>-2</sup>&middot;s<sup>-1</sup> light, it was observed that 145 μmol&middot;m<sup>-2</sup>&middot;s<sup>-1</sup> led to an increase in growth trend, total biomass and lipid content. Combining spectral qualities and light intensities, we show that the lipid accumulation raised by 2.8-fold. Studying the light intensity and spectral quality allowed us to optimize the light conditions to R1 spectral quality and light intensity 145 μmol&middot;m<sup>-2</sup>&middot;s<sup>-1</sup>. These initial results showed that red light R1 at 145 μmol&middot;m<sup>-2</sup>&middot;s<sup>-1</sup> was the best condition for biomass and total lipids accumulation in Ptev cells. Next, we further combined these two-light optimizations with a third light characteristics, i.e. light shift, where the cultures were shifted during the early stationary phase to a different light environment. We studied Red light shift (Rs) to investigate how light condition variations impacted P. tricornutum transconjugants Ptev and with an episomal vector containing the reporter gene YFP (PtYFP). We observed that Rs induced growth and fatty acid eicosapentaenoic acid (EPA) in Ptev as compared to PtYFP. Altogether, the study shows that red light shift of R1 at 145 μmol&middot;m<sup>-2</sup>&middot;s<sup>-1</sup> promoted biomass and total lipids accumulation in Ptev and PtYFP cells. The study provides a comprehensive approach to using different light characteristics with the aim to optimize growth and lipids, as well as to fatty acid production. 展开更多
关键词 Phaeodactylum tricornutum light Condition Episomal vector Dia-toms Fatty Acids Biomass
暂未订购
Time-varying classical local entanglement of ultrafast vector light pulses
11
作者 ZHANGYU ZHOU SHUOSHUO ZHANG +5 位作者 JIELEI NI YIXUAN CHEN YUQUAN ZHANG YEJUN HE CHANGJUN MIN XIAOCONG YUAN 《Photonics Research》 2026年第2期I0047-I0057,共11页
As structured optical fields with spatially inhomogeneous polarization distributions,vector beams have attracted widespread attention for their great applications across multiple fields.The coupling between spatial mo... As structured optical fields with spatially inhomogeneous polarization distributions,vector beams have attracted widespread attention for their great applications across multiple fields.The coupling between spatial modes and polarization states in vector beams results in nonseparability analogous to quantum entangled states,referred to as classical entanglement.This concept has garnered considerable scholarly interest,due to its capacity to facilitate quantum-inspired approaches for the characterization of vector optical fields,leading to noteworthy applications such as metrological sensing,communication encryption,and quantum computing.However,most investigations concerning classical entanglement have focused on vector optical fields within the spatial domain,while the classical entanglement characteristics of ultrafast vector beam pulses in the temporal domain remain to be revealed.Here,we investigate the classical local entanglement of ultrafast vector beam pulses in the temporal domain.Through theoretical interpretation and experimental verification,time-varying nonseparability has been demonstrated throughout the pulse duration.The dynamic evolution of spatially vectorial polarization states on femtosecond timescales was measured.The temporal evolution of the classical entanglement and the associated density matrix were analyzed using quantum state tomography.This research provides,to our knowledge,novel perspectives for investigating quantum-classical optical analogies in the temporal domain,advancing deeper fundamental understanding of entanglement. 展开更多
关键词 classical entanglementthis quantum state tomography temporal domain ve structured optical fields ultrafast vector light pulses classical entanglement vector beams
原文传递
Attenuation characteristics of a light attenuator combined by polarizers with different extinction ratios
12
作者 黄翀 邓鹏 +1 位作者 赵爽 陈海清 《Chinese Physics B》 SCIE EI CAS CSCD 2011年第8期246-255,共10页
This paper deals with a systematical analysis and an algorithm of attenuation characteristics of a light attenuator combined by n pieces of polarizers (n-LACP) whose extinction ratios are different from each other. ... This paper deals with a systematical analysis and an algorithm of attenuation characteristics of a light attenuator combined by n pieces of polarizers (n-LACP) whose extinction ratios are different from each other. The attenuation ratio expression of a two-LACP is deduced. We find that the monotonic attenuation interval depends on the first polarizer and that the attenuation range depends on the second one. For the three-LACP, a method for obtaining a monotonic attenuation interval is proposed. Moreover, the attenuation ratio expression is demonstrated. Analysis and experiment show that when the initial status of the three-LACP is at the maximum output, if the second or third polarizer rotates alone, the minimum attenuation ratios can reach K2^-1 and K3^-1, respectively, and if the first polarizer rotates, a minimum attenuation ratio of K2^-1K3^-1 can be obtained (K1, K2 and K3 represent the extinction ratios of the three polarizers in turn). Furthermore, the attenuation ratio expression of n-LACP and the relevant attenuation characteristics are proposed. The minimum attenuation ratio of an n-LACP is (K2K3 ... Kn)-1. 展开更多
关键词 light attenuator POLARIZER extinction ratio Jones vector
原文传递
Human Habitats and Malaria Vector Control in Benin: Current Situation and Implications for Effective Control
13
作者 Barikissou G. Damien Armel Djènontin +1 位作者 Badirou Aguemon Franck Remoué 《Advances in Infectious Diseases》 2023年第4期706-721,共16页
Long-lasting insecticide-treated nets (LLINs) and indoor residual spraying (IRS) are the main indoor malaria vector tools control tools. The study examined housing characteristics and investigated the relationship bet... Long-lasting insecticide-treated nets (LLINs) and indoor residual spraying (IRS) are the main indoor malaria vector tools control tools. The study examined housing characteristics and investigated the relationship between the total volume of household possessions, the volume of the sleeping room, and the hanging of LLINs. A total of 831 bedrooms were randomly selected in Benin in 2015. The findings showed that mud walls were predominant in rural areas (more than 75%), while metal roofs were common (77.3% - 97.9%). Battery-powered lighting was prevalent in rural areas in Northern (97%), while open-flame oil lamps were commonly used in rural areas in Southern (86%). The availability of correct bedding was low, ranging from 1% to 10% in all households. 20% of the bedrooms had at least 50% of their volume occupied by household possessions in urban areas. In rural areas, bedrooms without LLINs had a lower mean rate ratio of the volume occupied by possessions per the total volume of the room compared to bedrooms with at least one LLIN installed (p < 0.0001). The characteristics of human habitats are not favourable to the correct use of vector control intervention indoors. It is therefore important to improve people’s living conditions as the next step for malaria elimination. 展开更多
关键词 HOUSING lightING BEDDING Possessions Malaria vector Control LLINs IRS
暂未订购
The Light during Gravitational Super-Compressibility
14
作者 Kholmurad Khasanov 《Journal of Modern Physics》 2013年第4期468-473,共6页
The interaction of electromagnetic and gravitational fields and gravitational super-compressibility were investigated experimentally. Dynamic emitter provides conditions for the generation of eigenfunctions with eigen... The interaction of electromagnetic and gravitational fields and gravitational super-compressibility were investigated experimentally. Dynamic emitter provides conditions for the generation of eigenfunctions with eigenvalues for the various fields, including: acoustic, gravitational and electromagnetic. We observe the gravitational waves in gas flowing from the dynamic emitter and their interaction with electromagnetic waves. The gravitational field energy was decreasing when electromagnetic field was emitted through the excitation of condensed medium. The direction of maximum change of the emitted energy of excited medium was strongly opposed to gravity vector at that point. The frequency of radiation against the gravity vector in given point of space exceeded radiation frequency of same source in opposite direction. 展开更多
关键词 GRAVITATIONAL Super-Compressibility Interaction of Electric and GRAVITATIONAL FIELDS light Frequency INCREASING against GRAVITY vector
暂未订购
Solving large-scale multiclass learning problems via an efficient support vector classifier 被引量:1
15
作者 Zheng Shuibo Tang Houjun +1 位作者 Han Zhengzhi Zhang Haoran 《Journal of Systems Engineering and Electronics》 SCIE EI CSCD 2006年第4期910-915,共6页
Support vector machines (SVMs) are initially designed for binary classification. How to effectively extend them for multiclass classification is still an ongoing research topic. A multiclass classifier is constructe... Support vector machines (SVMs) are initially designed for binary classification. How to effectively extend them for multiclass classification is still an ongoing research topic. A multiclass classifier is constructed by combining SVM^light algorithm with directed acyclic graph SVM (DAGSVM) method, named DAGSVM^light A new method is proposed to select the working set which is identical to the working set selected by SVM^light approach. Experimental results indicate DAGSVM^light is competitive with DAGSMO. It is more suitable for practice use. It may be an especially useful tool for large-scale multiclass classification problems and lead to more widespread use of SVMs in the engineering community due to its good performance. 展开更多
关键词 support vector machines (SVMs) multiclass classification decomposition method SVM^light sequential minimal optimization (SMO).
在线阅读 下载PDF
水辅助矢量激光表面加工特性的实验研究 被引量:1
16
作者 许佳康 王吉明 +1 位作者 欧盛美 王建磊 《激光技术》 北大核心 2025年第4期501-506,共6页
为了研究水辅助下调控矢量光场在硅片表面的加工特性,搭建了调控纳秒矢量激光水辅助表面加工系统。采用Nd:YAG脉冲调Q激光器,对厚度为3 mm的单晶硅进行激光打孔实验,在静水和水射流辅助下,对矢量激光的脉冲次数、光束的偏振态、拓扑荷... 为了研究水辅助下调控矢量光场在硅片表面的加工特性,搭建了调控纳秒矢量激光水辅助表面加工系统。采用Nd:YAG脉冲调Q激光器,对厚度为3 mm的单晶硅进行激光打孔实验,在静水和水射流辅助下,对矢量激光的脉冲次数、光束的偏振态、拓扑荷数以及水流流量对单晶硅表面加工的影响进行了系统研究。结果表明,在水辅助的情况下,热影响区和重铸层的影响基本消失,矢量光束相较于线偏振光,在高脉冲次数下会获得更大孔径;高拓扑荷数的广义柱矢量偏振光与径向和角向偏振光相比,打孔更为光滑;在水射流辅助的情况下,当水流流量超过0.8 L/min时,激光加工的孔的圆对称性会有所破坏,广义柱矢量偏振光相较于径向与角向偏振光能够更好保持孔的圆对称性。将水辅助与矢量激光相结合,可以完成更多精密加工的工作,在工业加工中具有实际应用的价值。 展开更多
关键词 激光技术 矢量光场 表面加工 水射流
在线阅读 下载PDF
基于ε-SVR在线结构光V型坡口焊缝特征点提取应用 被引量:1
17
作者 李勇齐 薛瑞雷 +2 位作者 刘宏胜 夏磊 聂龙 《组合机床与自动化加工技术》 北大核心 2025年第9期138-143,共6页
在线结构光的管道V型坡口的焊接应用中,管道V型坡口反光是焊缝检测的重要影响因素,而焊缝检测是影响整个焊接性能的重要因素之一。因此,提出一种超参数支持向量回归(epsilon-support vector regression,ε-SVR)的方法来提取结构光数据,... 在线结构光的管道V型坡口的焊接应用中,管道V型坡口反光是焊缝检测的重要影响因素,而焊缝检测是影响整个焊接性能的重要因素之一。因此,提出一种超参数支持向量回归(epsilon-support vector regression,ε-SVR)的方法来提取结构光数据,通过序列最小优化(sequential minimal optimization,SMO)算法对ε-SVR进行优化,并对预测结果进行特征点提取。实验结果表明ε-SVR算法针对母材表面反光条件下的特征点提取拥有较好的效果,均方根误差小于0.2 mm,单帧图像处理时间小于20 ms以内,同时适用于U型、J型管道焊缝坡口,满足实际焊接工作要求。 展开更多
关键词 V型坡口 线结构光 超参数支持向量回归 序列最小优化 抑制反光
在线阅读 下载PDF
基于单光束矢量光场调控的紧聚焦均匀三维光泡的产生与探测
18
作者 郑泽灿 王思聪 +7 位作者 冯嘉浩 周志凯 陈树康 宋世超 邓子岚 秦飞 曹耀宇 李向平 《光电工程》 北大核心 2025年第5期2-9,共8页
三维光泡,即紧聚焦三维中空暗场光斑分布,在光学操控以及激光加工等领域中具有重要的应用价值。在已报道的工作中,三维光泡一般采用多光束干涉叠加的方式来产生,光路复杂,不利于系统集成应用,且能量利用率低。利用单光束矢量光场调控技... 三维光泡,即紧聚焦三维中空暗场光斑分布,在光学操控以及激光加工等领域中具有重要的应用价值。在已报道的工作中,三维光泡一般采用多光束干涉叠加的方式来产生,光路复杂,不利于系统集成应用,且能量利用率低。利用单光束矢量光场调控技术,产生高强度均匀性紧聚焦三维光泡,并利用探测光偏振转换技术实现对三维光泡的探测。通过调节旋向偏振入射光与0/π二元相位调制的径向偏振入射光的能量比,在实验上实现边缘强度与中心暗斑强度比值大于10:1、边缘强度均匀度接近90%的三维光泡,为双光束超分辨激光加工与光存储、粒子操控等领域提供实用的技术途径。 展开更多
关键词 矢量光场调控 三维光泡 纵向偏振分量探测
在线阅读 下载PDF
Tip-enhanced Raman scattering of glucose molecules 被引量:3
19
作者 Zhonglin Xie Chao Meng +3 位作者 Donghua Yue Lei Xu Ting Mei Wending Zhang 《Opto-Electronic Science》 2025年第5期2-9,共8页
Glucose molecules are of great significance being one of the most important molecules in metabolic chain.However,due to the small Raman scattering cross-section and weak/non-adsorption on bare metals,accurately obtain... Glucose molecules are of great significance being one of the most important molecules in metabolic chain.However,due to the small Raman scattering cross-section and weak/non-adsorption on bare metals,accurately obtaining their"fingerprint information"remains a huge obstacle.Herein,we developed a tip-enhanced Raman scattering(TERS)technique to address this challenge.Adopting an optical fiber radial vector mode internally illuminates the plasmonic fiber tip to effectively suppress the background noise while generating a strong electric-field enhanced tip hotspot.Furthermore,the tip hotspot approaching the glucose molecules was manipulated via the shear-force feedback to provide more freedom for selecting substrates.Consequently,our TERS technique achieves the visualization of all Raman modes of glucose molecules within spectral window of 400-3200 cm^(-1),which is not achievable through the far-field/surface-enhanced Raman,or the existing TERS techniques.Our TERS technique offers a powerful tool for accurately identifying Raman scattering of molecules,paving the way for biomolecular analysis. 展开更多
关键词 tip-enhanced Raman scattering scanning near-field optical microscope fiber vector light field tip nanofocusing light source
在线阅读 下载PDF
4种机器学习算法构建的临床预测模型在预测结直肠癌患者术前营养不良中的价值 被引量:1
20
作者 刘瑶 刘娟 葛玉红 《护士进修杂志》 2025年第9期939-945,967,共8页
目的采用4种机器学习算法分别构建结直肠癌患者术前营养不良的临床风险预测模型,探讨其预测价值。方法回顾性收集2023年1月—2024年5月在新疆医科大学附属肿瘤医院胃肠外科就诊的412例结直肠癌患者的术前资料;按7∶3的比例随机分为训练... 目的采用4种机器学习算法分别构建结直肠癌患者术前营养不良的临床风险预测模型,探讨其预测价值。方法回顾性收集2023年1月—2024年5月在新疆医科大学附属肿瘤医院胃肠外科就诊的412例结直肠癌患者的术前资料;按7∶3的比例随机分为训练集(n=288)和验证集(n=124),采用单因素分析及二元logistic回归分析筛选出术前营养不良的预测因子;基于逻辑回归(LR)、支持向量机(SVM)、轻量级梯度提升(LightGBM)、多层感知机(MLP)4种机器学习算法分别构建结直肠癌患者术前营养不良风险预测模型,绘制ROC曲线评价4种算法模型预测效能,通过Delong检验比较4种模型的AUC差异。选择最优算法模型,采用校准曲线和临床决策曲线(DCA曲线)进行验证。结果(1)结直肠癌患者术前营养不良发生率为33.7%,年龄、Braden评分是其独立危险因素;(2)训练集中LightGBM算法模型预测结直肠癌患者术前发生营养不良的AUC高于LR、SVM、MLP算法模型(0.941 VS 0.874、0.830、0.831);(3)ROC曲线结果提示,LightGBM算法模型验证集中预测结直肠癌患者术前发生营养不良的AUC为0.926(95%CI:0.882~0.969);校准曲线显示,LightGBM算法模型预测结直肠癌患者术前发生营养不良的曲线与实际发生营养不良一致性良好;DCA曲线结果显示,LightGBM算法模型在阈值概率区间为0.16~0.79可以提供显著临床净收益。结论基于LightGBM算法构建的临床预测模型在预测结直肠癌患者术前发生营养不良中有较高价值,可以为临床人员实施营养管理提供参考。 展开更多
关键词 结直肠肿瘤 营养不良 机器学习 预测模型 逻辑回归 支持向量机 轻量级梯度提升 多层感知机
暂未订购
上一页 1 2 14 下一页 到第
使用帮助 返回顶部