共查询到20条相似文献,搜索用时 15 毫秒
1.
介绍了图像目标识别技术中的图像分割,不变性参数提取和目标分类,利用图像目标的均匀性和相应知识自适应地分割和提取图像目标,被提取的每个图像目标的不变性参数由归一化过程和Zernike矩提取,并利用MPNN模型将图像目标分类,实验结果该识别系统能识别光照不均匀或复杂背景下的图像目标。 相似文献
2.
3.
Accurate calculation of image moments. 总被引:2,自引:0,他引:2
4.
《Signal Processing Magazine, IEEE》1999,16(2):81-93
The Fourier transform has been widely used in radar signal and image processing. When the radar signals exhibit time- or frequency-varying behavior, an analysis that can represent the intensity or energy distribution of signals in the joint time-frequency (JTF) domain is most desirable. In this article, we showed that JTF analysis is a useful tool for improving radar signal and image processing for time- and frequency-varying cases. We applied JTF analysis to radar backscattering and feature extraction; we also examined its application to radar imaging of moving targets. Most methods of JTF analysis are non-parametric. However, parametric or model-based methods of time-frequency analysis, such as adaptive Gaussian and chirplets, are more suitable for radar signals and images 相似文献
5.
6.
针对利用雷达微多普勒效应的微型无人机识别问题,提出了一种基于同步压缩短时傅里叶变换(Synchrosqueezing Short-Time Fourier Transform,SSTFT)的分类识别方法.首先对无人机的微多普勒回波信号进行SSTFT从而获得信号时频谱,然后对时频谱进行多维度特征提取获得回波信号的时频特征及频率变化特征,最后将所获得联合特征输入到支持向量机(Support Vector Machine,SVM)中进而实现无人机的分类识别.基于实际雷达数据的实验结果表明,所提无人机分类方法准确率可达到97.03%. 相似文献
7.
为了提升海杂波背景下小目标探测性能,本文提出一种基于时频域深度网络的特征检测方法。首先,将观测向量转换为归一化时频图(Normalized Time-Frequency Graph, NTFG),实现海杂波抑制。在时频域,建立海杂波、含正多普勒偏移目标回波、含负多普勒偏移目标回波的三分类问题,精细化目标落在主杂波带内外的不同特性。其次,引入Inception-ResNet V2深度网络作为特征提取器,自主学习不同类别在NTFG上的深层差异性,并将差异性浓缩为一个2D特征向量。然后,在2D特征空间中,设计具有引导的三次样条曲线,获得虚警可控的判决区域,实现异常检测。最后,IPIX实测数据验证了所提算法的性能优势,能深入挖掘时频域的特性。 相似文献
8.
应用支持向量机分类的多角度目标识别技术 总被引:4,自引:1,他引:3
综合应用图像的不变矩特征和支持向量机分类方法,提出了一种对于红外图像中多角度目标的识别方法。首先通过目标分割算法求得红外图像中目标的轮廓图像,然后从轮廓图像的Hu矩、Zernike矩和Fourier-Mellin矩中选取适当阶次的矩特征组成目标在特定视角范围内的不变性特征向量;对目标的视角范围进行适当划分以解决多角度引起的目标样本多样性,并在每个划分的视角范围内分别应用支持向量机的方法进行多目标分类。测试结果表明,本文提出的方法较好地实现了红外图像中多角度目标的识别问题,是一种有效的自动目标识别算法。 相似文献
9.
10.
11.
Liang Ban Ziarani A. K. Cetinkaya C. 《Semiconductor Manufacturing, IEEE Transactions on》2006,19(4):425-431
Material source disks (sputtering targets) used in the physical vapor deposition process exhibit nonuniform erosion profiles and thickness variations in the radial direction. Noninvasive prediction of the remaining life (end-of-life) on a sputtering target during its normal service is of practical interest in semiconductor manufacturing. Based on the piezoelectric generation/detection of acoustic waves propagating and signal processing techniques, a real-time process monitoring system for predicting the end-of-life of a sputtering target is proposed and implemented. The nonstationary signal processing approaches (the time-frequency (reassigned spectrogram) analysis, the cross-spectral density estimation method, and the cross-correlation technique) were employed in analyzing the acquired ultrasonic waveforms. The advantages and shortcomings of each approach in identifying the status of erosion profiles of the sputtering target are also discussed 相似文献
12.
在探地雷达探测过程中,天线相对目标的远近变化反映在面向深度的一维时域信号(A-scan)所组成的序列的变化过程中,由此提出一种针对变化过程建模的目标识别方法。在特征提取环节,提出将时频分析与图像纹理分析相结合,首先计算A-scan信号的二维时频联合分布图像,再利用特定的图像纹理描述算子构造特征向量。识别过程根据目标与天线间距离的变化,采用无跨越单向连续隐马尔可夫模型(HMM)对序列的变化过程建模。实验表明这种基于变化过程的HMM方法比无序地利用单条A-scan特征的支持向量机方法具有更好的效果。 相似文献
13.
极化信息的有效利用可提高目标特征提取和识别的精度.针对微动目标的微多普勒提取问题,在介绍传统微多普勒提取算法的基础上,建立了微动目标的全极化回波模型,提出了一种基于全极化信息的微多普勒提取算法.该方法以时频图像的对比度作为目标函数,通过寻找一组最优极化矢量提高时频图像质量.仿真数据实验表明,本文方法比传统时频变换方法得到的时频图像的对比度更高.真实数据实验发现,本文方法得到的图像对比度高达2.56,而传统时频变换方法得到的图像对比度在0.88到1.66之间.实验结果证明了本文方法的有效性和相比传统方法的优势. 相似文献
14.
The paper introduces a new kernel for the design of a high resolution time-frequency distribution (TFD). We show that this distribution can solve problems that the Wigner-Ville distribution (WVD) or the spectrogram cannot. In particular, the proposed distribution can resolve two close signals in the time-frequency domain that the two other distributions cannot. Moreover, we show that the proposed distribution is more accurate than the WVD and the spectrogram in the estimation of the instantaneous frequency of a stepped FM signal embedded in additive Gaussian noise. Synthetic and real data collected from real-world applications are shown to validate the proposed distribution 相似文献
15.
逆合成孔径激光雷达能实现对目标的高分辨2维成像,但如果目标中包含旋转部件,旋转部件回波带来的微多普勒效应会对目标的成像造成干扰。该文提出一种含旋转部件目标微多普勒特征提取及成像方法,首先利用匹配参考信号的方法对回波信号进行一定程度的非线性补偿,然后通过二值数学形态学方法提取频率-慢时间谱图中微多普勒特征曲线的信息,并利用微多普勒特征曲线的周期性进行曲线分离,实现对目标旋转部件微动参数的快速提取。在此基础上,对主体回波信号和旋转部件回波信号进行分离,完成对目标主体的2维成像。仿真实验验证了该文算法不仅能有效剔除目标旋转部件对逆合成孔径激光雷达成像的干扰,还能通过微多普勒特征的提取为目标识别提供新的途径。 相似文献
16.
复杂电磁环境下精准检测跳频信号是实施跳频信号侦查的先决条件。针对复杂干扰下跳频信号难以检测的问题,本文提出一种名为时频语义对消的方法。该方法设计了一种具有自注意力和图注意力机制的暹罗嵌套UNet,并根据该网络提取包含跳频信号、干扰信号和噪声的时频图语义信息。将得到的结果与不包含跳频信号时频图的语义信息相消就可以得到仅包含跳频信号的时频图,实现对跳频信号的检测。仿真结果表明所提方法可以在复杂干扰下实现对跳频信号的参数估计与盲检测,在信噪比高于-5 dB和信干比高于0 dB时,虚警概率与漏警概率低于1‰。在信号时频范围检测中,对比实验表明语义对消检测方法比语义分割检测方法交并比分数提升0.31,消融实验表明注意力模块对交并比分数提升为0.022。同时分析了所提网络的复杂度,结果表明该方案具有较小的参数量以及较快的处理速度,可以运用于跳频信号的实际检测当中。 相似文献
17.
A new class of signal adaptive time-frequency representations called the adaptive constant-Q distribution (AQD) is introduced. The AQD exploits a priori knowledge about a signal's instantaneous frequency and bandwidth to perform signal-dependent smoothing of the Wigner distribution. The objective is to achieve a good tradeoff between reducing the variance and preserving the resolution by means of time-frequency dependent smoothing (specifically for use in medical Doppler ultrasound). A numerical, alias-free implementation of the AQD is presented. Deterministic, multicomponent signals as well as synthetic Doppler ultrasound signals were analyzed with the AQD. The performance of the AQD was compared with the power spectrogram, the exponential distribution, and the adaptive optimum kernel representation as well as with theory. The error was consistently lowest for the AQD. In conclusion, a new signal adaptive class of time-frequency distributions has been developed, and its potential in nonstationary signal analysis has been demonstrated 相似文献
18.
19.
In the problem of unsupervised domain adaption Extreme learning machine (ELM), the output layer parameters need to have both classification and domain adaptation functions, which often cannot be simultaneously fully utilized. In addition, traditional matching method based on data probability distribution cannot find the common subspace of source and target domains under large difference between domains. In order to alleviate the pressure of double functions of classifier parameters, the entire ELM learning process is mainly divided into two stages: feature representation and adaptive classifier learning, thus a joint feature representation and classifier learning based unsupervised domain adaption ELM model is proposed. In the feature representation stage, the source and target domain data are projected to their respective subspace while minimizing the difference in probability distribution between the two domains. In the adaptive classifier learning stage, the smooth manifold regularization term of target domain is used to improve the parameter adaptive ability. Experiments on six different types of datasets show that the proposed model has higher cross-domain classification accuracy. 相似文献