货车枕簧丢失故障动态图像识别方法

针对货车运行故障动态图像检测,提出无故障目标识别工作模式,解决货车枕簧丢失故障的自动识别问题。利用Haar特征提取枕簧特征信息,基于AdaBoost算法选取特征并构建层叠分类器,等比缩放搜索窗口检测货车图像,最终分选出无故障的枕簧图像,从而大大地减少了待识别图像的数量,显著地提高了人工识别效率。实验表明,该算法使用的特征简单,搜索策略高效,不受枕簧位置、缩放和旋转的影响,抗噪能力强,对分辨率低、局部遮挡、光照不足或过度曝光等质量较差的图像仍具有很强的适应性,所提出的方案能够满足全天候条件下的货车枕簧目标识别,为货车故障动态图像检测的工程化应用奠定了基础。     

Active polarization imaging system to discriminate adaptively with diagonal Mueller matrix

A promising method to optimize the polarization state of two-channel active polarization imaging system is presented. In this method, it is seminal that the detecting function of the imaging system is regarded as a discriminant projection of the observed objects’ polarization features (elements of the Mueller matrix). The polarization state can be seen as a physical classifier which can be obtained by training samples. The image acquired with the system that has the designed optimal polarization state become discriminative results directly. The effectiveness of the proposed method and the discriminative ability of the optimal polarization state are demonstrated by the experimental results.     

动态背景下基于光流场分析的运动目标检测算法

针对现有动态背景下运动目标检测算法的不足,提出一种基于光流场分析的运动目标检测算法.首先根据前背景在光流梯度幅值和光流矢量方向上的差异确定目标的大致边界,然后通过点在多边形内部原理获得边界内部的稀疏像素点,最后以超像素为节点,利用混合高斯模型拟合的表观信息和超像素的时空邻域关系构建马尔可夫随机场模型的能量函数,并通过使目标函数能量最小化得到最终的运动目标检测结果.该算法不需要任何先验假设,能够同时处理动态背景和静态背景两种情况.多组实验结果表明,本文算法在检测的准确性和处理速度上均优于现有算法.     

A semi-supervised learning approach for detection of phishing webpages

This paper proposes a new phishing webpage detection approach based on a kind of semi-supervised learning method-transductive support vector machine (TSVM). Firstly the features of web image are extracted for complementing the disadvantage of phishing detection only based on document object model (DOM); they include gray histogram, color histogram, and spatial relationship between subgraphs. Then the features of sensitive information are examined by using page analysis based on DOM objects. In contrast to the drawback of support vector machine (SVM) algorithm which simply trains classifier by learning little and poor representative labeled samples, this method introduces the TSVM to train classifier that it takes into account the distribution information implicitly embodied in the large quantity of the unlabeled samples, and have better performance than SVM. The experimental results show that the proposed method not only achieves better classification accuracy, but also has strong applicability as the independent method of phishing detection.     

Generalized symmetry transformation with Gaussian phase weight function

The generalized symmetry transformation (GST) is a symmetry operator detecting objects by using edge gradient directions. Conventional GST uses the cosine function to define the phase weight function (PWF), which represents the symmetry of two gradient directions. The cosine function of PWF leads to a good performance in detecting symmetrical objects. However, the weights of gradient pairs, which are considered to be asymmetrical, are relatively high, so side effects appear near the symmetry pick regions. (Note that side effects disturb the multiple object detection.) In this paper, we use the Gaussian function in calculating the symmetric weights of gradient pairs. The Gaussian function can suppress the weights of less symmetric gradient pairs. In addition, the symmetry for elliptically shaped objects can be more emphasized by controlling the width of the Gaussian function. The proposed GST is evaluated through experiments on synthetic images, which include various bright and dark plane figures, and on real images, which requires the detection of elliptical shapes.     

一种利用窗口结构提取红外弱小目标的方法

在分析红外弱小目标检测的检测概率和虚警率的基础上,引入了一种基于窗口结构的红外弱小目标检测方法.该方法无需进行图像差分,利用局部小范围内目标相对背景的灰度变化检测目标,可以在较低信噪比的情况下获得较高的检测概率和较低的虚警率.实验表明了该方法的有效性.     

基于可变形部件模型HOG特征的人形目标检测

使用单幅图像进行特定目标的检测是机器视觉领域的重要任务之一。利用机器学习的方法,使用LSVM分类器进行人形目标的检测。该方法提取图像的HOG（梯度方向直方图）特征和其对应的可变形部件来描述目标的外形特征,能够较好地解决目标由于运动而产生外形变化的问题。对常见公共区域场景进行数据采集并随机抽取了200张图像,使用所述方法对其中共1 100个人形目标进行检测,正确率识别率为78.3%。结果表明该方法具有一定的可行性和稳定性,能够较好检测出单幅图像中的人形目标并加以标注。但对于某种程度有所遮蔽的人形目标则会产生漏检的现象。     

一种目标级的遥感图像变化检测算法

传统的像素级变化检测方法对图像的配准准确度要求较高,因而在实际运用中受到很多限制.在人造目标检测的基础上,提出了一种目标级的基于局部配准误差补偿的变化检测方法.根据遥感图像中人造目标与自然目标的纹理差异,对图像中的人造目标进行检测和分割,再对分割图像采用提出的算法进行变化检测.实验表明,与传统的像素级变化检测方法相比,本算法具有较高的检测准确度,对配准准确度的要求也有所放宽,并且可以简化变化检测前的辐射校正工作和变化检测后的像素分类的工作.     

主动中子多重性法估算铀材料质量的不确定性（英文）

近年来,国际社会对核材料保护、控制和衡算日益加强。对不明材料损失量(MUF)的关注逐渐提升。铀材料质量不确定性测量在估算铀材料生产量中扮演重要角色。由于铀材料自发裂变相对较弱,主动中子多重性法被应用于估算铀材料质量。通过拟合对不同系列铀金属壳的数值模拟结果,获得了描述铀材料质量与主动中子多重性特征之间的算法和参数。得到的关系表明,可以通过分析不同重数中子多重性探测结果获得铀部件的质量。对不同探测条件下的模拟结果的定量分析,确定了探测系统设置对铀质量估算的影响,以及认知不确定性和随机不确定在估算过程中传播对质量估算的影响。对不确定度的分析获得了本文模拟采用的探测系统的最佳源强和探测时间窗设置,在此设置下,质量估算的不确定性最小。     

用黑斑的椭圆性质在二值化图像中探测眼睛

在人脸探测过程中 ,双眼的探测非常重要 ,这是因为人的眼睛是人脸上最容易探测到的特征部位。通过双眼探测还可以对人脸进行定位 ,使人脸的尺寸归一化。探讨了一种基于二值化图像中黑斑几何形状分析的双眼探测方法。因为人眼的形状比较接近椭圆 ,故在研究黑斑的几何性质时 ,重点研究黑斑的椭圆性质。所谓黑斑的椭圆性质是指按照黑斑的面积和二阶矩计算出的椭圆参数。椭圆的主要参数有椭圆的面积、质心、长短轴的长度和方向角 (椭圆长轴与水平方向的夹角 )。为了描述黑斑与椭圆的近似程度 ,定义了椭圆硬度的概念。根据眼睛黑斑椭圆参数的测试数据建立了探测眼睛的数学模型 ,并给出了根据此模型得到的双眼探测的实验结果     

低轨微小卫星及小碎片搜索/跟踪机动式大视场光电望远镜

有效反射面积小,运动速度快的空间低轨卫星和小碎片的日益增多对现有地基探测跟踪技术提出了挑战。本文分析了国外现有低轨小目标光电探测技术的发展现状,结合低轨小目标的探测需求,提出了一种用于低轨微小卫星及小碎片搜索/跟踪探测的机动式车载大视场光电望远镜设计方案。介绍了该望远镜的光学系统、跟踪架及载车,描述了它的工作模式和图像处理,讨论了系统的搜索和探测能力。结果表明,该望远镜对300 km轨道高度的目标搜索能力达到13.5星等（相当于直径5 cm目标）,可以满足搜索和跟踪低轨微小卫星及小碎片探测的实际需求。     

基于生物光学算法的海洋赤潮监测

机载海洋激光雷达具有高效搜索的能力,已成为海洋环境监测的重要手段。以机载蓝绿激光雷达监测海洋赤潮为研究应用背景,通过检测海洋中叶绿素的质量浓度来实现对海洋赤潮消长信息的获取。在分析海水固有光学性质的基础上,基于生物光学算法,构建了机载蓝绿激光雷达后向散射信号检测的的叶绿素质量浓度监测模型。仿真计算证明该模型能够有效探测出叶绿素质量浓度的大小,为机载海洋赤潮监测增加了一种方法。     

Ground-penetrating radar for soil and underground pipelines using the forward modeling simulation method

Ground-penetrating radar was simulated using the FDTD method with uniaxial anisotropic perfectly matched layers as absorbing boundary conditions. To study underground pipeline detection, the electromagnetic response properties of underground pipelines, under known soil conditions, were determined. Pipelines at different depths and with different diameters and different compositions were simulated and researched for their response. The simulation results were consistent with research and analysis, i.e., the deviation was in an acceptable range. This research is of great value to guide the detection of pipelines in cities and to explain the detection of materials underground.     

图像中近圆形对象的一种识别方法

介绍了一种利用多特征联合概率的红外图像中近圆形对象检测方法,通过计算图像中诸多小段曲线的弯拱方向来获得图像中各处存在圆的概率,并用梯度强度对这些概率进行加权,再结合对象内部为空洞的概率,共同形成对图像中近圆形对象所在位置的一个判决。经与经典的Hough变换法相比较,表现出一定的优越性。对一组不同特征的红外图像进行测试实验,均能检测出近圆形对象,表现出一定的适应性,对某些应用场景中的目标检测与识别有辅助价值。     

用频域法实现圆环形目标物的计数

研究了用频域法实现圆环形目标物计数的方法。对空域中含有圆环形目标物的二值图像，经傅里叶变换后在频域内进行滤波，再经傅里叶反变换后取一阈值进行处理，则对应于圆环形目标物即变为一个实心体，然后对实心体进行计数即可实现对圆环形目标物的计数。对该方法进行了数学描述，并分别用模拟图像和实际图像进行了实验。结果表明，该方法消除噪声的效果好，对部分封闭圆的检测同样有效，且可用于目标物有粘连的情况。该方法对圆环形目标物的尺寸相差不大的情况更为有效。     

Textual and shape-based feature extraction and neuro-fuzzy classifier for nuclear track recognition

Track counting algorithms as one of the fundamental principles of nuclear science have been emphasized in the recent years. Accurate measurement of nuclear tracks on solid-state nuclear track detectors is the aim of track counting systems. Commonly track counting systems comprise a hardware system for the task of imaging and software for analysing the track images. In this paper, a track recognition algorithm based on 12 defined textual and shape-based features and a neuro-fuzzy classifier is proposed. Features are defined so as to discern the tracks from the background and small objects. Then, according to the defined features, tracks are detected using a trained neuro-fuzzy system. Features and the classifier are finally validated via 100 Alpha track images and 40 training samples. It is shown that principle textual and shape-based features concomitantly yield a high rate of track detection compared with the single-feature based methods.     

日间恒星实时探测的视频图像处理

为了解决船载光学测量设备日间测星数量偏少的问题,研究了暗小目标电视探测技术,以保证惯导系统全天时不间断航向校准。在视频图像处理中,采用中值滤波方法提高信噪比,同时采用基于模糊理论的最大模糊熵阈值分割方法自行构建隶属度函数,自适应求取阈值,实现了星体目标检测与识别。实验结果表明：在原光学设备光学系统不变的情况下,日间可探测的恒星由大于等于3.0星等、小于等于5颗,提高到了大于等于4.5星等,大于等于250颗,满足了惯导系统全天时不间断航向校准的要求。该成果可以推广应用于电视弱小目标探测领域。     

高斯混响背景下稳健型Durbin检测器设计（英文）

This paper deals with the problem of detecting a signal whose amplitude is a scaling factor in the presence of homogeneous Gaussian noise with unknown covariance matrix.Since no uniformly most powerful test exists for the problem at hand,we devise and assess a detection strategy based on the well-known Durbin test design criteria.The closed-form expressions for the probabilities of false alarm and detection of the Durbin test are derived,which show that it bears a constant false alarm rate property against the noise covariance matrix.At the analysis stage,the performance of the new receiver is assessed,also in comparison with some classical adaptive detectors,both in matched and in mismatched signal cases.The results show that the proposed detector achieves a visible performance improvement in the presence of severe steering vector mismatch,while maintaining an acceptable detection loss for matched signal.     

利用几何特性及神经网络进行人脸探测技术的研究

在人脸识别过程中 ,首先也是最重要的一个环节是人脸探测 ,因为一旦从图像中定位并提取到了人脸 ,那么下一步的人脸识别工作就变得非常容易。眼睛是人脸图像中最容易探测的部位 ,而且通过探测双眼来发现人脸最符合人的视觉习惯。提出了一种基于几何特征分析和人工神经网络的由粗到细的两级人脸探测方法。在第一级中 ,眼睛和脸是通过测量眼睛的尺寸和眼睛与脸的位置关系探测到的 ,第一级的输出是一个尺寸归一化的人脸 ,但偶尔也伴随着一个或多个因对复杂背景中与眼睛类似的物体的误判而得到的非人脸图像 ;第二级神经网络正是用来过滤掉第一级中被误判的人脸。实验表明 ,这种由粗到细的两级人脸探测系统具有很高的稳定性和探测正确率     

面阵探测下的污染云团红外光谱仿真

研究污染云团的红外光谱仿真,对于利用仿真光谱进行光谱识别的算法研究十分重要.在单元探测器探测方式下污染云团的红外光谱仿真研究取得了一定成果,并且已经被应用于识别算法的研究工作中.针对基于成像光谱仪的污染云团识别算法研究缺乏实测数据的问题,利用具有高仿真精度的基于物理模型的污染云团扩散模型及其仿真结果网格化的特点,研究相应的云团红外光谱仿真多层模型,提出了面阵探测方式下污染云团红外光谱的仿真方法,生成了同时具有光谱维和空间维信息的数据立方体,为该研究领域提供了新的研究方法.提出的面阵探测方式下的污染云团红外光谱仿真直观地反映了污染云团的扩散,提供了完备且符合实际情况的污染云团红外光谱立方体数据,对于提高和完善红外光谱识别算法具有重要意义.