Institutional Repository of Xi'an Institute of Optics and Precision Mechanics, CAS
Not a member yet
30565 research outputs found
Sort by
A Feature Aggregation Convolutional Neural Network for Remote Sensing Scene Classification
Remote sensing scene classification (RSSC) refers to inferring semantic labels based on the content of the remote sensing scenes. Recently, most works take the pretrained convolutional neural network (CNN) as the feature extractor to build a scene representation for RSSC. The activations in different layers of CNN (named intermediate features) contain different spatial and semantic information. Recent works demonstrate that aggregating intermediate features into a scene representation can significantly improve the classification accuracy for RSSC. However, the intermediate features are aggregated by some unsupervised feature encoding methods (e.g., Bag-of-Visual-Words). Little attention has been paid to explore the information of semantic labels for the feature aggregation. In this paper, in order to explore the semantic label information, an end-to-end feature aggregation CNN (FACNN) is proposed to learn a scene representation for RSSC. In FACNN, a supervised convolutional features' encoding module and a progressive aggregation strategy are proposed to leverage the semantic label information to aggregate the intermediate features. The FACNN integrates the feature learning, feature aggregation, and classifier into a unified end-to-end framework for joint training. In FACNN, the scene representation is learned by considering the information of semantic labels, which can result in better performance for RSSC. Extensive experiments on AID, UC-Merged, and WHU-RS19 databases demonstrate that FACNN performs better than several state-of-the-art methods.</p
Structural Design of Carbon Fiber-reinforced Plastics Barrel for Space Remote Sensing Camera
设计并研制了碳纤维复合材料主次镜连接筒.根据光学设计中的公差分配,结合复合材料层合板理论,分析了复合材料铺层设计对面内刚度和轴向及周向热膨胀系数的影响,确定了最终的铺层方式.利用有限元软件分析了重力作用下及温度变化时主次镜间的位置变化及连接筒的模态分布.最后,完成了碳纤维复合材料薄壁连接筒的成型和精加工,检测了主次镜系统光学性能,并对装配完成后的相机进行了鉴定级力学试验.分析结果表明:在1g重力载荷及2℃温升耦合作用下,次镜偏心小于0.002mm,次镜倾斜小于2″,主次镜连接筒及次镜系统组合体基频达到265Hz,满足光学设计要求和结构稳定性要求.光学性能检测结果及力学试验结果表明,主次镜系统波像差满足装配要求,主次镜连接筒能够承受鉴定级力学试验考核,连接主次镜的两端响应放大仅1.7倍,体现了良好的阻尼性能.本文中研制的主次镜连接筒重量仅为6.4kg,实现了高轻量化和高刚度,满足空间相机对主次镜位置精度和稳定性的要求. </p
Tunable Fiber Laser Using an ErYb Codoped Fiber Based Modal Interferometer Filter
A tunable erbium-doped fiber laser (EDFL) is proposed and demonstrated experimentally. It is based on a tunable fiber filter using the pump induced refractive index change characteristics of the ErYb codoped phosphosilicate fiber (EYDF) based modal interferometer (MI). The laser achieves a tunability of 2.39nm and 1.60nm, with the wavelength linearly tuned from 1534.05nm to 1536.44nm and from 1554.36nm to 1555.96nm by changing the pump power injected into the MI. Its side mode suppression ratios are 30dB and 33dB, respectively. The stability of the tunable EDFL is getting better with the continuous improvement of the power of pump 1, and the dual-wavelength power fluctuations are less than 0.22dB and 0.12dB within 20 min at 250mW of pump 1, respectively.</p
SISTEM OPTOELECTRONIC PENTRU DETERMINAREA TEMPERATURII METALULUI DIN PLASMA TERMICĂ A CORDOANELOR DE SUDURĂ
The invention relates to an optoelectronic system for determining the temperature of the metal in thermal plasma of the welding seams. The claimed system consists of a laser diode () and a laser telemeter () which emit a punctiform radiation in visible red spectral range, the laser diode () being used for marking the start point of the welding seam and the telemeter () being used for targeting and tracking the area of the electric arc on the distance d of the welded material as well as an optoelectronic structure () which represents a unit for the conversion of the electric arc radiation into units of temperature by separating therefrom, with an interference optical filter () in near infrared range, a monochromatic radiation subsequently converted into a proportional electric current with a photoelectric detector (), the said current being converted into temperature units by means of an electronically memorized calibration curve
Laser mode hop detection in a heat-assisted magnetic recording device using a slider sensor and laser modulation
An apparatus comprises a laser diode configured to generate modulated light during a write operation in response to receiving modulated current having a mean amplitude that varies or is constant. A slider is configured for heat-assisted magnetic recording and to receive the modulated light. A writer heater of the slider is configured to receive power during the write operation having a magnitude that varies or is constant. A sensor is situated on or within the slider. The sensor is configured to produce a sensor signal representative of output optical power of the laser diode. Measuring circuitry is coupled to the sensor and configured to measure a change in the sensor signal indicative of a laser mode hop during the write operation
Laser swept source with controlled mode locking for OCT medical imaging
An optical coherence analysis system uses a laser swept source that is constrained to operate in a mode locked condition. This is accomplished by synchronously changing the laser cavity's gain and/or phase based on the round trip travel time of light in the cavity. Many high-speed wavelength swept laser sources emit pulses synchronized with the round trip time of the cavity as part of a nonlinear optical frequency red shifting process. Stable pulsation is associated with smooth tuning and low relative intensity noise. Addition of mode-locking methods to this class of lasers can control and stabilize these lasers to a low clock jitter and RIN state, and in specific cases allow long-to-short wavelength tuning in addition to the usual short-to-long (red shifting). The laser may comprise a SOA (410), a tunable Fabry-Perot-Filter (412) as one reflector and an Output coupler (405) in an optical fiber (406) to adjust the cavity length
基于分波前双闪耀平面反射光栅的光谱成像系统及方法
本发明涉及一种光谱成像技术,具体涉及一种基于分波前双闪耀平面反射光栅的光谱成像系统及方法。本发明的目的是解决现有超宽光谱探测完成超宽光谱成像时仪器体积和重量过大的问题,提供一种基于分波前双闪耀平面反射光栅的光谱成像系统及方法。该系统包括沿光线入射方向依次同轴排布的前置望远镜、狭缝、镜像成像镜头、分波前双闪耀平面反射光栅,以及位于狭缝平面上的第一、第二光学接收装置;狭缝位于前置望远镜像面和镜像成像镜头物面的重合处;分波前双闪耀平面反射光栅的入射面为周期性三角形线槽刻划面。该方法中分波前双闪耀平面反射光栅将入射光束波前W反射并分成正级次闪耀波前W1和负级次闪耀波前W2两部分并在光学接收装置上成像
一种基于配置文件的卫星地面系统通信协议实现方法
本发明提出一种基于配置文件的卫星地面系统通信协议实现方法,充分考虑了通信协议将来的变动的可能和将通信协议与卫星地面系统完全解耦的需求,有利于通信协议未来的改动和扩展。该方法采用JSON格式对通信协议进行编码并存储为作为地面管控设备通信协议的配置文件;地面管控软件在C++架构下定义JSON格式文件各字段的解析方法,建立与通信协议具体内容的联系;地面管控软件通过读取、解析所述JSON格式文件,自动将通信协议转换成计算机可识别的二进制码流进行编码发送与接收解析,完成数据传输或界面显示;当通信协议发生改变时,通过更改JSON格式文件对应字段的设定信息,实现通信协议的更新
基于卷积神经网络的图像描述方法、计算机可读存储介质、电子设备
本发明提供一种基于卷积神经网络的图像描述方法、计算机可读存储介质、电子设备,解决现有递归神经网络方法不能并行处理序列信号以及运算耗时的问题。该方法包括以下步骤:1)将卷积神经网络预训练好;2)提取图像的全局特征和局部特征,将图像的全局特征和局部特征投影到多模态映射空间中;3)对步骤2)中多模态映射空间下的图像表达进行卷积编码;4)单词特征表达;5)对步骤4)的描述语句进行卷积编码,6)计算注意力,得到输入图像对应生成单词的概率;7)构建输入输出之间的目标损失函数,利用损失函数进行神经网络训练,得到神经网络的位置参数;8)将测试图像输入训练的神经网络系统,得到测试图片对应的描述性自然语句