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。     

Designing a genetic watermarking approach for 3D scenes

A new watermarking algorithm based on genetic algorithm (GA) in the transform domain is proposed. Unlike the existing computer-generated integral imaging based watermarking methods, the proposed method utilizes GA searching to the optimized transform domain to serve as a trade-off for watermark embedding. In this paper, 3D scene to be captured by using a virtual pinhole array and be computationally recorded as an elemental image array (EIA), watermarking with GA optimization and computer-generated holography is implemented. In the proposed GA optimization process, we utilize the fitness function to improve the visual quality of watermarked images and the robustness. Simulation results show that the proposed algorithm yields a holographic watermark that is imperceptibility to human eyes and robust to standard watermarking attacks. A comparison of the proposed watermarking method to the existing similar watermarking methods demonstrated that the proposed method generally outperforms completing methods in terms of imperceptibility and robustness.     

Improving the performance of DCT-based fragile watermarking using intelligent optimization algorithms

The performance of a fragile watermarking method based on discrete cosine transform (DCT) has been improved in this paper by using intelligent optimization algorithms (IOA), namely genetic algorithm, differential evolution algorithm, clonal selection algorithm and particle swarm optimization algorithm. In DCT based fragile watermarking techniques, watermark embedding can usually be achieved by modifying the least significant bits of the transformation coefficients. After the embedding process is completed, transforming the modified coefficients from the frequency domain to the spatial domain produces some rounding errors due to the conversion of real numbers to integers. The rounding errors caused by this transformation process were corrected by the use of intelligent optimization algorithms mentioned above. This paper gives experimental results which show the feasibility of using these optimization algorithms for the fragile watermarking and demonstrate the accuracy of these methods. The performance comparison of the algorithms was also realized.     

A Spatial Domain Quantum Watermarking Scheme

This paper presents a spatial domain quantum watermarking scheme. For a quantum watermarking scheme, a feasible quantum circuit is a key to achieve it. This paper gives a feasible quantum circuit for the presented scheme. In order to give the quantum circuit, a new quantum multi-control rotation gate, which can be achieved with quantum basic gates, is designed. With this quantum circuit, our scheme can arbitrarily control the embedding position of watermark images on carrier images with the aid of auxiliary qubits. Besides reversely acting the given quantum circuit, the paper gives another watermark extracting algorithm based on quantum measurements. Moreover, this paper also gives a new quantum image scrambling method and its quantum circuit. Differ from other quantum watermarking schemes, all given quantum circuits can be implemented with basic quantum gates. Moreover, the scheme is a spatial domain watermarking scheme, and is not based on any transform algorithm on quantum images. Meanwhile, it can make sure the watermark be secure even though the watermark has been found. With the given quantum circuit, this paper implements simulation experiments for the presented scheme. The experimental result shows that the scheme does well in the visual quality and the embedding capacity.     

Compressive optical image watermarking using joint Fresnel transform correlator architecture

A new optical image watermarking technique based on compressive sensing using joint Fresnel transform correlator architecture has been presented. A secret scene or image is first embedded into a host image to perform optical image watermarking by use of joint Fresnel transform correlator architecture. Then, the watermarked image is compressed to much smaller signal data using single-pixel compressive holographic imaging in optical domain. At the received terminal, the watermarked image is reconstructed well via compressive sensing theory and a specified holographic reconstruction algorithm. The preliminary numerical simulations show that it is effective and suitable for optical image security transmission in the coming absolutely optical network for the reason of the completely optical implementation and largely decreased holograms data volume.     

Time series prediction using wavelet process neural network

In the real world, the inputs of many complicated systems are time-varying functions or processes. In order to predict the outputs of these systems with high speed and accuracy, this paper proposes a time series prediction model based on the wavelet process neural network, and develops the corresponding learning algorithm based on the expansion of the orthogonal basis functions. The effectiveness of the proposed time series prediction model and its learning algorithm is proved by the Macke-Glass time series prediction, and the comparative prediction results indicate that the proposed time series prediction model based on the wavelet process neural network seems to perform well and appears suitable for using as a good tool to predict the highly complex nonlinear time series.     

Neural Network Used for the Fusion of Predictions Obtained by the K-Nearest Neighbors Algorithm Based on Independent Data Sources

The article concerns the problem of classification based on independent data sets—local decision tables. The aim of the paper is to propose a classification model for dispersed data using a modified k-nearest neighbors algorithm and a neural network. A neural network, more specifically a multilayer perceptron, is used to combine the prediction results obtained based on local tables. Prediction results are stored in the measurement level and generated using a modified k-nearest neighbors algorithm. The task of neural networks is to combine these results and provide a common prediction. In the article various structures of neural networks (different number of neurons in the hidden layer) are studied and the results are compared with the results generated by other fusion methods, such as the majority voting, the Borda count method, the sum rule, the method that is based on decision templates and the method that is based on theory of evidence. Based on the obtained results, it was found that the neural network always generates unambiguous decisions, which is a great advantage as most of the other fusion methods generate ties. Moreover, if only unambiguous results were considered, the use of a neural network gives much better results than other fusion methods. If we allow ambiguity, some fusion methods are slightly better, but it is the result of this fact that it is possible to generate few decisions for the test object.     

Intelligent reversible watermarking technique in medical images using GA and PSO

The advances in recording, editing, and broadcasting multimedia contents in digital form motivate to protect these digital contents from illegal use, such as duplication, manipulation, and redistribution. However, watermarking algorithms are designed to satisfy requirements of applications, as different applications have different concerns. We intend to design a watermarking algorithm for applications which require high embedding capacity and imperceptibility, to maintain the integrity of the host signal as well as embedded information. Reversible watermarking is a promising technique which satisfies our requirements. In this paper, we concentrate on improving the watermark capacity and reducing the perceptual degradation of an image. We investigated the Luo's [1] additive interpolation-error expansion algorithm and enhanced it by incorporating with two intelligent techniques: genetic algorithm (GA), and particle swarm optimization (PSO). Genetic algorithm is applied to exploit the correlation of image pixel values to obtain better estimation of neighboring pixel values, which results in optimal balance between information storage capacity and imperceptibility. Particle swarm optimization (intelligent technique) is also applied for the same purpose. Experimental results show that PSO and GA nearly give the same results, but GA outperforms the PSO. Experimental results also reveal that the proposed strategy outperforms the state of art works in terms of perceptual quality and watermarking payload.     

基于L-M算法的反向传播网络的湿度传感器输出误差补偿研究

针对湿度传感器的输出非线性问题,提出了基于L-M算法建立BP神经网络进行补偿校正,实现电阻湿度传感器的输入与输出非线性补偿,并与共轭梯度算法、拟牛顿算法所建立的神经网路模型进行对比,重点比较了模型误差性能、收敛速度。结果表明：基于L-M算法建立的神经网络模型,在收敛速度、误差性能等方面具有更高效的表现,更适合湿度传感器的非线性特性的补偿校正。     

基于BP神经网络的排爆机械臂逆运动学分析

机械臂逆运动学是已知末端执行器的位姿求解机械臂各关节变量,主要用于机械臂末端执行器的精确定位和轨迹规划,如何高效的求解机械臂运动学逆解是机械臂轨迹控制的难点。针对传统的机械臂逆运动学求解方法复杂且存在多解等问题,提出一种基于BP神经网络的机械臂逆运动学求解方法。以四自由度机械臂为研究对象,对其运动学原理进行分析,建立BP神经网络模型并对神经网络算法进行改进,最后使用MATLAB进行仿真验证。仿真结果表明：使用BP神经网络模型求解机械臂逆运动学问题设计过程简单,求解精度较高,一定程度上避免了传统方法的不足,是一种可行的机械臂逆运动学求解方法。     

基于神经免疫生长可免域网络的红外光谱图像分割算法

管道运输对远距离输送石油天然气有着较大优势,而与之伴随的管道安全问题使得管道安全检测至关重要。为确保任何时间下管道状况的有效检测,红外成像技术由于其根据对象的热辐射信息反映目标特征的特殊性,能够忽视可见光的影响检测管道状态,因而在管道检测领域有重要意义。但由于户外环境的多样性,交错的管道和复杂环境使得采集的红外管道图像具有目标特征分布不均匀,目标遮挡和背景类目标干扰等问题。这些问题增加了提取管道目标的难度,不利于管道的分割和检测。生物免疫系统在抗原检测、提取和消除上表现出识别、学习、记忆、耐受和协调配合等目前复杂系统优化策略所缺乏的优异特性,借鉴生物神经系统调控免疫系统的机理,设计一种基于神经免疫网络的复杂背景下红外管道目标的检测与提取算法。根据生物神经网络在免疫系统中的调控机制,利用基础管道形状特征模型构建用于红外管道目标定位的神经网络,并将最优神经免疫可免域和区域种子生长结合,解决管道遮挡影响提取目标完整性的问题。选择三种典型的红外管道图像,将传统目标检测算法与基于神经免疫网络的算法进行了效果对比分析。结果表明,传统算法的平均真阳性率为40.56%,Jaccard相似性指数为27.18%,绝对误差率为11.75%,而基于神经免疫网络算法的真阳性率为98.05%,Jaccard相似性指数为94.44%,绝对误差率为1.18%。对比可知,神经免疫网络算法的真阳性率比传统方法高57.49%,绝对误差率则低10.57%,验证了复杂背景下,本文算法相比传统方法能够更加准确地提取完整的红外管道目标,这对管道安全检测效率的提高有着重要意义。     

基于小波变换和奇异值分解的图像水印算法研究

利用小波变换频谱特性和图像奇异值分解特征,提出了一种结合人眼对比感知特性的图像水印算法。并通过结合人眼视觉特性,将置乱的水印以一定的强度嵌入到图像的奇异值矩阵中,采用其逆过程提取水印,通过仿真进行了验证。对其实施了压缩、剪切、高斯噪声和中值滤波攻击测试,与前人提出的8种水印算法的抗攻击性能进行对比分析。结果表明,在质量因子为20%的较强压缩攻击下,提取水印的NC值仍能达到0.8359,含水印图的PSNR和SSIM达到25.0908dB和0.8451,且比8种水印算法具有更好的鲁棒性。综合表明,提出的算法有效地解决了水印嵌入过程中鲁棒性、视觉透明性与水印嵌入量之间的平衡问题。     

基于SOM和RBF网络的软测量方法研究

针对氢粉碎过程中钕铁硼粉碎状态不可知,为有效预测合金的反应状态,提出了一种基于自组织特征映射(SOM)神经网络和径向基函数(RBF)神经网络结合构建的网络模型。在该模型中,SOM神经网络作为聚类网络,采用无教师学习算法对输入样本进行自组织分类,并将分类中心及其对应的权值向量传递给RBF神经网络,作为径向基函数的中心;RBF神经网络作为基础网络,采用高斯函数作为径向基函数实现从输入到隐含层的非线性映射,输出层则采用有教师学习算法训练网络的权值,从而实现输入层到输出层的线性映射。并以钕铁硼氢粉碎过程合金中氢含量为检测对象,运用上述方法在MATLAB平台上建立了合金中氢含量预测模型,并完成了仿真验证。     

神经网络模式识别系统互连权重二值化研究

在增量算法的基础上，利用截断方法和蒙塔卡罗算法，对以四类飞行目标旋转投影作为学习样本的级联神经网络互连权重进行了二值优化处理，并用非学习样本进行了容错性检验，计算机木匠虱到了满意的结果。     

Pulse coding off-chip learning algorithm for memristive artificial neural network

Memristive neural network has attracted tremendous attention since the memristor array can perform parallel multiply-accumulate calculation (MAC) operations and memory-computation operations as compared with digital CMOS hardware systems. However, owing to the variability of the memristor, the implementation of high-precision neural network in memristive computation units is still difficult. Existing learning algorithms for memristive artificial neural network (ANN) is unable to achieve the performance comparable to high-precision by using CMOS-based system. Here, we propose an algorithm based on off-chip learning for memristive ANN in low precision. Training the ANN in the high-precision in digital CPUs and then quantifying the weight of the network to low precision, the quantified weights are mapped to the memristor arrays based on VTEAM model through using the pulse coding weight-mapping rule. In this work, we execute the inference of trained 5-layers convolution neural network on the memristor arrays and achieve an accuracy close to the inference in the case of high precision (64-bit). Compared with other algorithms-based off-chip learning, the algorithm proposed in the present study can easily implement the mapping process and less influence of the device variability. Our result provides an effective approach to implementing the ANN on the memristive hardware platform.     

Numeral eddy current sensor modelling based on genetic neural network

This paper presents a method used to the numeral eddy current sensor modelling based on the genetic neural network to settle its nonlinear problem. The principle and algorithms of genetic neural network are introduced. In this method, the nonlinear model parameters of the numeral eddy current sensor are optimized by genetic neural network （GNN） according to measurement data. So the method remains both the global searching ability of genetic algorithm and the good local searching ability of neural network. The nonlinear model has the advantages of strong robustness, on-line modelling and high precision. The maximum nonlinearity error can be reduced to 0.037% by using GNN. However, the maximum nonlinearity error is 0.075% using the least square method.     

An enhanced non-uniformity correction algorithm for IRFPA based on neural network

Influenced by detector materials’ non-uniformity, growth and etching techniques, etc., every detector’s responsivity of infrared focal plane arrays (IRFPA) is different, which results in non-uniformity of IRFPA. And non-uniformity of IRFPA generates fixed pattern noises (FPN) that are superposed on infrared image. And it may degrade the infrared image quality, which greatly limits the application of IRFPA. Non-uniformity correction (NUC) is an important technique for IRFPA. The traditional non-uniformity correction algorithm based on neural network and its modified algorithms are analyzed in this paper. And a new improved non-uniformity correction algorithm based on neural network is proposed in this paper. In this algorithm, the desired image is estimated by using three successive images in an infrared sequence. And blurring effect caused by motion is avoided by applying implicit motion detection and edge detection. So the estimation image is closer to real image than the estimation image estimated by other algorithms, which results in fast convergence speed of correction parameters. A comparison is made to these algorithms in this paper. And experimental results show that the algorithm proposed in this paper can correct the non-uniformity of IRFPA effectively and it prevails over other algorithms based on neural network.     

A novel calibration approach to structured light 3D vision inspection

Structured light 3D vision inspection is a commonly used method for various 3D surface profiling techniques. In this paper, a novel approach is proposed to generate the sufficient calibration points with high accuracy for structured light 3D vision. This approach is based on a flexible calibration target, composed of a photo-electrical aiming device and a 3D translation platform. An improved algorithm of back propagation (BP) neural network is also presented, and is successfully applied to the calibration of structured light 3D vision inspection. Finally, using the calibration points and the improved algorithm of BP neural network, the best network structure is established. The training accuracy for the best BP network structure is 0.083 mm, and its testing accuracy is 0.128 mm.     

一种适于光学实现的非线性神经网络模型及其蒙特卡洛学习算法

本文用计算机仿真研究了一种适于光学实现的非线性神经网络模型的存储客量α_c和寻址能力,提出了一个改进其触突互联矩阵的蒙特卡洛学习算法.数值研究表明,经过学习修正后的神经网络模型的寻址能力及存储容量都有较大的改进.     

一种深度学习的立体阵波达方向估计方法

针对常规波束形成主瓣宽且目标分辨能力低的问题,提出一种基于深度卷积神经网络的波达方向估计方法。算法使用常规波束形成计算二维空间功率谱,将预处理后的空间功率谱图输入深度卷积神经网络。该文利用神经网络学习解卷积映射关系,输出主瓣宽度更窄的空间功率谱图,从而实现高分辨率二维波达方向估计。该算法对阵列结构没有限制,适用于立体阵。仿真结果表明该文方法在不同目标个数、快拍数及信噪比参数下均能准确估计目标方向。该文方法目标分辨能力优于常规波束形成方法。在低快拍情况下,目标方向估计误差低于自适应波束形成方法。