Optimization of the operational parameters in a fast axial flow CW CO2 laser using artificial neural networks and genetic algorithms

This paper presents an artificial intelligence approach for optimization of the operational parameters such as gas pressure ratio and discharge current in a fast-axial-flow CW CO₂ laser by coupling artificial neural networks and genetic algorithm. First, a series of experiments were used as the learning data for artificial neural networks. The best-trained network was connected to genetic algorithm as a fitness function to find the optimum parameters. After the optimization, the calculated laser power increases by 33% and the measured value increases by 21% in an experiment as compared to a non-optimized case.     

Planar segmentation of indoor terrestrial laser scanning point clouds via distance function from a point to a plane

Terrestrial laser scanners are frequently used in most of measurement application, particularly in documentation and restoration studies of indoor historical structures, and in acquiring facade reliefs. When compared to a photogrammetric method, terrestrial laser scanners have the ability to give three dimensional point cloud data directly in a fast and detailed way. High data density of point cloud data is a challenging factor in texture-map operations during documentation and restoration of historical artifacts with more indoor spaces. When coordinate information for terrestrial laser scanner point cloud data is documented, it is seen that there is no regular order and classification for the data. The aim of this study is to suggest the mathematical filtering algorithm for segmentation work towards separation of planar surfaces which have different depths and parallel to each other and which can be frequently encountered in the indoor spaces from the data of terrestrial laser scanner. Filtering function for segmentation used, is based on the distance of a point to the plane. This algorithm has been chosen for the advantage of the rapid and easy results for extracting 3D coordinate data in texture mapping process. The MatLAB interface has been developed for using this method and analyzing the results for application which is detected how many different surfaces exist according to the statistical deviation amount. In the application, test data with 21932 points was segmented by separating it into 16 points in total with four different planes and four corner points per plane. Surfaces with four different depths were obtained as the result of the research. Each of them included four points. These segmented surfaces consisting of four points will facilitate integrated data production by integrating vectorial terrestrial laser scanner data into raster camera data, without the need to conventional measurements that accelerate particularly documentation and modeling in the fields of historical indoor areas.     

高效选取线性独立的原子特征构建精确机器学习势函数

机器学习势由于具有与第一性原理计算相当的准确性,且低得多的计算成本,在原子模拟中极具前景. 然而原子机器学习势的可靠性、速度和可迁移性在很大程度上取决于原子构型的表示. 适当地选取用作机器学习程序输入的描述符是一个成功的机器学习表示的关键. 本文发展了一种简单有效的方法,可以基于训练数据固有的相关性,从大量待选的描述符中自动选取一组最佳的线性独立原子特征. 通过对几个具有较少冗余线性独立嵌入密度描述符的基准分子构建嵌入原子神经网络势的应用,证明了这种新方法的有效性和准确性. 该算法可以大大简化原子特征的初始选取,并极大地提高原子机器学习势的性能.     

基于RBF网络的断路器电流自适应保护研究

针对传统断路器电流保护方法存在受系统运行方式影响、整定困难、智能化低等问题,本文提出了基于RBF的断路器电流自适应保护算法,并给出了算法的模型。该算法融合了RBF神经网络的故障检测和电流自适应保护。首先通过RBF网络检测负载线路的电流故障,然后用电流自适应算法进行保护。在对神经网络进行训练时,利用PSO算法对RBF神经网络的参数进行优化以此来提高网络的泛化能力和学习能力;然后采用优化后的PSO-RBF神经网络对电流故障进行诊断。实验表明,该算法较大地提高了断路器智能化管理水平。     

Adaptive variable step-size neural controller for nonlinear feedback active noise control systems

Adaptive filter techniques and the filtered-x least mean square (FxLMS) algorithm have been used in Active Noise Control (ANC) systems. However, their effectiveness may degrade due to the nonlinearities and modeling errors in the system. In this paper, a new feedback ANC system with an adaptive neural controller and variable step-size learning parameters (VSSP) is proposed to improve the performance. A nonlinear adaptive controller with the FxLMS algorithm is first designed to replace the traditional adaptive FIR filter; then, a variable step-size learning method is developed for online updating the controller parameters. The proposed control is implemented without any offline learning phase, while faster convergence and better noise elimination can be achieved. The main contribution is that we show how to analyze the stability of the proposed closed-loop ANC systems, and prove the convergence of the presented adaptations. Moreover, the computational complexities of different methods are compared. Comparative simulation results demonstrate the validity of the proposed methods for attenuating different noise sources transferred via nonlinear paths, and show the improved performance over classical methods.     

立体成型中扫描误差的分析

对立体成型技术中广泛使用的振镜扫描系统进行了理论分析,指出了振镜的二维扫描偏转角和平面坐标之间存在着非线性映射关系.提出了一种实用的校正算法，该算法用理论数据与激光束在相应点的实际数据进行比对，制成校正表，按插值算法校正实际的枕形畸变及其它误差.该算法适用于使用振镜实现二维扫描的系统，能对应用这种原理工作的激光成型设备进行现场校正.     

非线性荧光光谱的神经网络分析及其应用

大功率超快脉冲激光和气体相互作用可产生非线性荧光光谱,不同的气体分子具有不同的非线性荧光光谱。因而这种光谱可以作为物质的指纹模式加以识别分类,进而获知气体的成分。由于不同气体分子的光谱在同一波段上有很大的交叉重叠,用传统的光谱分析方法分析存在困难,采用神经网络方法分析上述非线性荧光光谱,利用经过预处理的荧光光谱数据作为模式样本,其中一部分样本作为学习样本对级联神经网络进行训练,用训练好的网络对所有样本进行实时识别,学习样本和测试样本的的正确识别率均可达100％,结果表明此方法可实时判断混合气体的组分。     

Dynamic infrared scene projection: a review

Since the early 1990s, there has been major progress in the developing field of dynamic infrared scene projection, driven principally by the need for hardware-in-the-loop simulation of the oncoming generation of imaging infrared missile seekers and more recently by the needs for realistic simulation of the new generation of thermal imagers and forward-looking infrared systems. In this paper the current status of the dynamic infrared projection field is reviewed, commencing with an outline of its history. The requirements for dynamic infrared scene projection are examined, allowing a set of validity criteria to be developed. Each class of infrared projector that has been investigated—emissive, transmissive, reflective, laser scanner and phosphor—together with the specific technology initiatives within the class is described and examined against the validity criteria. In this way the leading dynamic infrared scene projection technologies are identified.     

A deep learning–based method for improving reliability of multicenter diffusion kurtosis imaging with varied acquisition protocols

Multicenter magnetic resonance imaging is gaining more popularity in large-sample projects. Since both varying hardware and software across different centers cause unavoidable data heterogeneity across centers, its impact on reliability in study outcomes has also drawn much attention recently. One fundamental issue arises in how to derive model parameters reliably from image data of varying quality. This issue is even more challenging for advanced diffusion methods such as diffusion kurtosis imaging (DKI). Recently, deep learning–based methods have been demonstrated with their potential for robust and efficient computation of diffusion-derived measures. Inspired by these approaches, the current study specifically designed a framework based on a three-dimensional hierarchical convolutional neural network, to jointly reconstruct and harmonize DKI measures from multicenter acquisition to reformulate these to a state-of-the-art hardware using data from traveling subjects. The results from the harmonized data acquired with different protocols show that: 1) the inter-scanner variation of DKI measures within white matter was reduced by 51.5% in mean kurtosis, 65.9% in axial kurtosis, 53.7% in radial kurtosis, and 61.5% in kurtosis fractional anisotropy, respectively; 2) data reliability of each single scanner was enhanced and brought to the level of the reference scanner; and 3) the harmonization network was able to reconstruct reliable DKI values from high data variability. Overall the results demonstrate the feasibility of the proposed deep learning–based method for DKI harmonization and help to simplify the protocol setup procedure for multicenter scanners with different hardware and software configurations.     

Bidimensional measurement of an underwater sediment surface using a 3D-Scanner

The application of laser technologies to experimental fluvial hydraulics is a relatively new field with a promising future. In the area of sediment transport, these technologies will be used here to describe the evolution of an underwater sediment bed profile. This paper presents the application of a three-dimensional laser scanner (3D-Scanner), which measures elevation through the intersection of the laser beam with the sediment surface, correcting the effect of its refraction as it enters the water. Results were compared with those obtained with a particle image velocimetry (PIV) in recording the longitudinal profile of the bed. In this way, the utility of this high-precision, continuous and non-intrusive technique has been demonstrated in fluvial hydraulic studies with sediment transport.     

Adaptive pulse compression by two-photon absorption in semiconductors

We investigate the adaptive optimization of broadband laser pulses, using a closed-loop learning algorithm in which the merit function is derived from two-photon absorption in semiconductors. Photoluminescence experiments with CdS thin films and photocurrent measurements of a GaAsP photodiode have been performed. The experimental data demonstrate that reliable and accurate pulse compression to the bandwidth limit can be achieved, unperturbed by nontrivial phase effects. Therefore two-photon absorption proves to be an easy-to-implement alternative to second-harmonic generation for the compression of broadband laser pulses.     

模糊条码图像的正则化复原算法

由条码扫描仪获得条码图像的过程可以用理想条码信号与扫描仪光学系统点扩散函数的卷积模型来描述。反卷积是消除由光学系统点扩散带来的模糊现象的最好办法。为克服反卷积的病态问题,研究了反卷积的正则化方法;针对条码信号的特点,构建了适合于条码信号复原的惩罚项,提出了条码信号的正则化复原算法及其适合于计算机运算的迭代算法。通过实验研究了算法在不同情况下的抗干扰能力。实验结果表明,正则化条码信号复原算法在消除系统点扩散函数的影响的同时能够很好地抑制噪声。     

基于三角测量的三维激光扫描仪设计

基于三角测量原理,设计了一套三维激光扫描仪.分析了光信号在不同表面上的反射性质,设计出较为准确的亚像素光点中心定位算法.实验表明即使在红外光较强的环境下,也具备较好的抗干扰能力,具有较高性价比.     

基于深度学习的无波前探测自适应光学系统研究进展

近年来自适应光学（AO）系统向着小型化和低成本化趋势发展,无波前探测自适应光学（WFSless AO）系统由于结构简单、应用范围广,成为目前相关领域的研究热点。硬件环境确定后,系统控制算法决定了WFSless AO系统的校正效果和系统收敛速度。新兴的深度学习及人工神经网络为WFSless AO系统控制算法注入了新的活力,进一步推动了WFSless AO系统的理论发展与应用发展。在回顾前期WFSless AO系统控制算法的基础上,全面介绍了近年来卷积神经网络（CNN）、长短期记忆神经网络（LSTM）、深度强化学习在WFSless AO系统控制中的应用,并对WFSless AO系统中各种深度学习模型的特点进行了总结。概述了WFSless AO技术在天文观测、显微成像、眼底成像、激光通信等领域的应用。     

Wavelength-Tunable Laser-Diode Interferometer

Laser diodes (LDs) have been applied to a phase-measuring interferometer through the wavelength tunability of LDs by controlling their currents. Laser-diode interferometers based on a heterodyne technique are reviewed. A two-wavelength laser-diode interferometer is demonstrated with current control of dual LDs in opposite directions. A synthetic wavelength makes it possible to extend the range of interferometric measurements. The wavelength is controlled by the laser injection current and is stepwise or rampwise changed to introduce a time-varying phase difference between the two beams of an interferometer with unbalanced optical path lengths. The optical output is demodulated with a phase-extraction algorithm. Systematic phase errors caused by the LD-power variation and by the difference between the beat frequency and ramp frequency are analyzed. A feedback interferometer with electronics is used to eliminate the phase error by locking the interferometer on a preset phase. Typical experimental results are shown.     

Dual-wedge scanning confocal reflectance microscope

A confocal reflectance microscope has been developed that incorporates a dual-wedge scanner to reduce the size of the device relative to current raster scanning instruments. The scanner is implemented with two prisms that are rotated about the optical axis. Spiral and rosette scans are performed by rotating the prisms in the same or opposite directions, respectively. Experimental measurements show an on-axis lateral resolution of 1.6 microm and optical sectioning of 4.7 microm, which compares with a diffraction-limited resolution of 0.8 and 1.9 microm, respectively.     

Mössbauer spectra processing using neural networks

A Mössbauer spectrum can be understood as a sum of Lorentz functions. The aim of this work is to develop an automated algorithm for Mössbauer spectra processing. The required output of this algorithm is the number of Lorentz functions contained in the spectrum and the values of parameters of these functions, i.e. values of their amplitudes, positions and halfwidths. For this purpose a modified brute-force method was used. There are tested all possible combinations of parameters of the Lorentz functions whether they properly fit measured spectrum or not. A set of parameters having the best fit with the spectrum is the output of this method. Testing all possible combinations takes a huge computational power so an optimization process was developed to speed up the computational process. This is achieved by decreasing the number of tested combinations. The optimization process is based on a mathematical analysis and neural network processing. An evolutional algorithms were used for proper setting up of the neural weights.     

激光诱导击穿光谱结合神经网络测定土壤中的Cr和Ba

提出了一种基于人工神经网络的激光诱导击穿光谱技术实现元素成分高准确度定量分析的方法.采用基于动量和自适应学习速率梯度下降算法的反向传播神经网络,结合激光诱导击穿光谱技术的方法测定土壤中Cr和Ba元素的含量,得到了Cr和Ba的含量以及多次重复预测的相对标准偏差,并与采用传统的内标法得到的检测结果相比较.研究结果表明:基于动量和自适应学习速率梯度下降算法的反向传播神经网络分析方法,与激光诱导击穿光谱技术相结合能更好地实现对土壤样品中Cr和Ba元素的定量检测.相对内标法,神经网络分析方法与激光诱导击穿光谱技术相结合可以很明显地提高检测准确度和精密度,对采用激光诱导击穿光谱技术定量检测土壤重金属污染具有很好的应用价值.     

High speed ghost imaging based on a heuristic algorithm and deep learning

We report an overlapping sampling scheme to accelerate computational ghost imaging for imaging moving targets,based on reordering a set of Hadamard modulation matrices by means of a heuristic algorithm. The new condensed overlapped matrices are then designed to shorten and optimize encoding of the overlapped patterns, which are shown to be much superior to the random matrices. In addition, we apply deep learning to image the target, and use the signal acquired by the bucket detector and corresponding real image to train the neural network. Detailed comparisons show that our new method can improve the imaging speed by as much as an order of magnitude, and improve the image quality as well.     

Exponential speedup of quantum newton optimization algorithm for general polynomials

Quantum optimization algorithms can outperform their classical counterpart and are key in modern technology. The second-order optimization algorithm(the Newton algorithm) is a critical optimization method, speeding up the convergence by employing the second-order derivative of loss functions in addition to their first derivative. Here, we propose a new quantum second-order optimization algorithm for general polynomials with a computational complexity of O(poly(log d)). We use this algorithm to solve the nonlinear equation and learning parameter problems in factorization machines. Numerical simulations show that our new algorithm is faster than its classical counterpart and the first-order quantum gradient descent algorithm. While existing quantum Newton optimization algorithms apply only to homogeneous polynomials, our new algorithm can be used in the case of general polynomials, which are more widely present in real applications.