一种空间目标在轨检测图像预处理算法

通过对美国天基可见相机(SBV)在轨检测算法(Moving target indicator,MTI)算法进行改进,使其更加快速有效地检测出淹没在噪声杂波中的目标条纹.MTI改进算法使用了一种新的二维图像检测预处理算法一一依概率加窗检测算法.依概率加窗检测算依据同一大小检测窗口内,目标所在检测窗内出现的非零点比纯噪声检测窗内多的特性,通过检测窗门限滤波,在可能剔除部分目标点的同时,极大地抑制噪声.接着使用三点共线条纹检测算法,剔除可疑目标条纹,进一步降低虚警概率,提高检测概率.通过算法性能分析可知,MTI改进算法的虚警概率降低15倍,检测概率99%时所需信噪比从6降低为3,并且总体计算量降低6个数量级.MTI改进算法减少计算量的同时,降低虚警概率并提高榆测概率,在工程应用中更有利于算法的实时实现.     

基于空间距离改进的视觉显著性弱小目标检测

在人类视觉系统(HVS)中,占据重要部分的是对比度值而不是亮度。现有的基于HVS的红外弱小目标检测算法可以有效提高目标信噪比和检测率,但也存在虚警率较高、易受噪声影响等问题。针对这些问题,提出一种基于空间距离改进的视觉显著性红外弱小目标检测算法。用目标所在像素块均值与其周围像素块加权和的比值来加权目标点灰度,得到一幅显著图。在计算周围像素的灰度加权和时,将视觉注意机制中关键因素之一的空间关系考虑进去,根据空间距离加权计算,较远的地方分配较小的权值。实验表明,本文算法不仅可以降低虚警率,而且抗噪能力也比较强。     

基于稳健主成分分析和多点恒虚警的红外弱小目标检测

针对红外图像中由复杂背景和目标多形态带来的单帧检测暗弱小目标比较困难的问题,提出了一种先进行阈值分割粗提取,后进行多点信噪比精检测的算法。在粗提取阶段,提出了改进的基于稳健主成分分析(RPCA)的阈值分割算法,利用邻域稀疏度均值与整幅稀疏图像均值的比值进行阈值分割,从而进一步剔除孤立噪点和背景云层边缘的杂波。在精检测阶段,提出了基于统计特性的多点恒虚警检测算法,统计候选点在邻域内每个像元的信噪比,利用虚警率门限和统计数量阈值筛选目标点,从而克服由小目标能量弥散带来的多形态特征问题。实验结果表明,所提算法在复杂背景下的探测率达到95.6%,与利用单像元和邻域像元均值计算信噪比的方法相比,虚警率分别降低了56.1%和47.1%。     

基于光谱角累加的高光谱图像异常检测算法

针对传统方法中用作背景的像素中存在干扰像素的问题,提出一种基于光谱角累加的高光谱图像异常检测算法。通过计算测试区域待测像素光谱向量与其他光谱向量之间的夹角,并将其累加,得到图像中每个像素的异常程度;然后使用波段选择预处理方法进一步提高检测性能。HyMap高光谱数据验证表明,在虚警概率设为0.008时,检测概率达到0.73,即在提高异常检测可靠度的同时,降低了虚警概率。     

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.     

一种利用十字窗口识别红外弱小目标的新方法

为有效地识别红外图像中的弱小目标,在分析传统识别方法的基础上,提出了一种利用十字窗口识别红外弱小目标的新方法.该方法在待识别点上、下、左、右四个方向以目标大小选取四个参考点构造十字形识别窗口,根据待识别点与参考点之间的灰度梯度识别该像素点是否为目标像素.该方法无需预测背景,可以在低信噪比、强度变化剧烈的图像中获得较高的识别率与较低的虚警率.实验结果表明了该算法的有效性.     

基于小波高频距离像的红外小目标检测

针对远距离复杂背景下红外小目标检测问题,本文提出了一种基于小波高频距离像的方法。该方法首先将处理空间变换到小波域,通过分析残留背景、目标和噪声系数在高频子带的差异,定义基于邻域均值的子带系数表达形式,构造高频子带系数的中心向量,对小波高频图像进行综合形成距离像,得到红外复杂背景的抑制结果。在此基础上,利用恒虚警率算法将单帧背景抑制图像分割成候选目标、残留背景和噪声像素点。最后,在时间域基于目标运动的相关性,利用管道滤波实现红外小目标的最终检测。仿真实验结果表明,相对于经典算法,本文方法可以实现对红外复杂背景的有效抑制,增强目标信号的强度,准确稳定的从红外复杂背景中检测出小目标。     

基于SoC的海天背景下红外斑状目标检测技术

基于红外目标检测的算法在军事和民用领域发挥着重要的作用,然而,在低信噪比条件下,对于飞机、船舶等红外斑状目标的检测存在难度大,误检率、虚警率高等困难.对此,提出了 一种基于机器学习的红外斑状目标的算法,检测像素面积在3×3到100×100的红外目标.算法部分采用了形态学方法对目标进行预提取,并使用HOG特征提取与SVM...     

基于全方位形态学滤波和局部特征准则的点目标检测

对复杂背景下暗弱点目标和背景杂波特性进行了分析,提出了一种基于全方位多尺度的形态学滤波和局部特征准则的点目标检测方法。实验结果表明,在复杂背景和低信噪比条件下,所提算法的目标检测概率达到99.8%,虚警率为0.1%。与最大中值滤波法、高斯差分尺度空间法、高斯混合模型法进行对比,结果表明,所提算法对复杂背景的抑制作用较好,且算法复杂度不高,易于实时实现。     

基于Bertrand曲面模型的边缘检测算法

针对图像边缘检测过程中噪声抑制与细节保留不能兼顾的问题,提出一种基于Bertrand曲面模型的边缘检测算法.在确定像素级边缘的基础上,选取沿边缘方向的带状域为拟合区域,利用Bertrand曲面具有沿母线各点的法线与母线共面的性质,将拟合曲面区域内的像素点信息转化为边缘曲线的活动坐标,并对转化后的像素点坐标和归一化灰度值进行拟合,求得亚像素边缘到像素级边缘的法向距离,实现图像亚像素边缘的检测.用视觉测量系统对量块直线边缘进行实验,并与改进Facet曲面拟合亚像素边缘检测算法比较,说明基于Bertrand曲面模型的边缘检测算法具有较高的定位精度,测得一等量块的直线度误差在1μm以内,多次测量的误差平均值为-0.811μm,可靠性高.通过机油泵泵体测量实例,说明本文算法可以应用于机械零件的精密测量,尤其适用于中心距、孔径等的测量.     

Target detection algorithm in side-scan sonar image based on pixel importance value measurement

Side-scan sonar detection application always combines with unstable results.A two-stage novel pixel importance value measurement algorithm is proposed to stabilize the detection ability and false alarm probability simultaneously.In first stage of the algorithm,a new feature defined as pixel importance value(PIV) is proposed in terms of distances between the target pixel and each other pixels.PIV measurement of current pixel is defined as the weighted sum of all remaining segmented pixels.The weighted part refers to Gaussian kernel,which means closer pixels gets higher weight.Thus,targets with higher PIV can be located.In the second stage,we use convolutional neural network as classifier to eliminate the dot-like false targets.Our experiment data is obtained by autonomous underwater vehicle,where we demonstrate superior performance of our algorithm over the state-of-the-art sonar detection algorithms in terms of 90.39% recall rate and 2.39% false alarm probability.     

结合区域提取和改进卷积神经网络的水下小目标检测

针对水下小目标信息量有限而难以提取有效特征导致的检测性能不佳问题,提出了一种结合区域提取和融合Hu矩特征的改进卷积神经网络水下小目标检测方法。该方法包含区域提取和分类两个步骤。首先以马尔可夫随机场分割算法为基础进行区域提取,对潜在目标定位的同时降低伪目标对后续分类的干扰;然后提取潜在目标区域的Hu矩特征并融入卷积神经网络,形成一种形状特征表征能力更强的改进卷积神经网络用于分类。声呐实测数据处理结果表明,该方法可以有效提升对水下小目标的发现概率和正确报警率,与其他目标检测方法相比,该方法具有更好的检测性能和泛化性。     

Small target detection using bilateral filter and temporal cross product in infrared images

We introduce a spatial and temporal target detection method using spatial bilateral filter (BF) and temporal cross product (TCP) of temporal pixels in infrared (IR) image sequences. At first, the TCP is presented to extract the characteristics of temporal pixels by using temporal profile in respective spatial coordinates of pixels. The TCP represents the cross product values by the gray level distance vector of a current temporal pixel and the adjacent temporal pixel, as well as the horizontal distance vector of the current temporal pixel and a temporal pixel corresponding to potential target center. The summation of TCP values of temporal pixels in spatial coordinates makes the temporal target image (TTI), which represents the temporal target information of temporal pixels in spatial coordinates. And then the proposed BF filter is used to extract the spatial target information. In order to predict background without targets, the proposed BF filter uses standard deviations obtained by an exponential mapping of the TCP value corresponding to the coordinate of a pixel processed spatially. The spatial target image (STI) is made by subtracting the predicted image from the original image. Thus, the spatial and temporal target image (STTI) is achieved by multiplying the STI and the TTI, and then targets finally are detected in STTI. In experimental result, the receiver operating characteristics (ROC) curves were computed experimentally to compare the objective performance. From the results, the proposed algorithm shows better discrimination of target and clutters and lower false alarm rates than the existing target detection methods.     

浅海动态环境下蛙人探测声呐布放深度优化技术

蛙人探测声呐的探测性能及作用距离不仅与目标海区的声速剖面和自身性能参数有关,还与其布放深度有关,其布放深度的选择对于水下目标探测具有重要意义。以射线跟踪理论为基础,建立了基于探测概率感知的声呐性能评估模型;提出了一种声呐布放深度自适应优化算法,该算法以不同准则进行两次筛选优化后可得到声呐最佳布放深度。仿真结果表明：在不同水文情况下,所提优化算法可以有效选择最优布放深度,进而提高声呐对目标的探测性能。     

噪声中的线谱检测及自动提取研究

线谱是水中目标检测与识别的重要特征,其自动提取对于水中无人探测系统至关重要。采用统计信号分析方法得出了线谱检测的接收机工作特性曲线,理论分析了线谱自动提取时一定虚警概率下的峰高门限,并提出一种自动门限以适应不同背景方差的变化。实测数据分析表明,本文方法在不同噪声背景下自动提取线谱的效果较好。     

Prediction of the performance of an algorithm for the detection of small targets in infrared images

Detection of small targets in infrared (IR) images is important in IR image processing. For the prediction of performance of a detection algorithm, it is necessary to calculate the probability of detection and probability of false alarm. A method is developed to calculate the probabilities in this paper. The detection is divided into two parts: the first part, which is called pre-detection, is to find out candidates for targets in a single frame of an image; and the second part is to localize the target in multiple frames of the image. Under some assumptions, the pre-detection probability, the false detection probability of single frame, detection probability and false alarm probability are derived. The algorithm for the detection of small target in IR image, which is used for the derivation of the probabilities, is contrast threshold detection based on background prediction, and a pipeline filter is used for multiframe image processing. The results show the relationship of the probabilities to the contrast of target to background, SNR, and contrast threshold.     

基于多特征的快速红外弱小目标检测算法

为了从全向红外搜索和跟踪系统采集的海量大视场高分辨率红外图像中快速准确地检测出红外弱小目标,本文提出了一种基于由粗到细的分阶段检测策略和时空域特征融合的红外弱小目标检测算法.首先,通过引入基于频域的快速显著性检测算法预先检测出目标可能存在的候选区域;其次,对候选区域进行角点检测以判定是否存在候选目标;最后,通过结合帧间时空域特征对候选目标进行进一步判定,以提取真实目标、删除虚假目标.多种实际场景的实验结果表明,该目标检测算法不仅运算量小而且探测概率高、虚警率低,是一种工程实用性能很好的红外弱小目标检测算法.     

一种被动声呐信号检测的低虚警率方法

针对被动声呐多目标信号检测中的噪声背景归一化问题,提出了一种基于数学形态学滤波的噪声背景归一化新方法。该方法利用数学形态学处理中的膨胀和腐蚀算子,以及基于多项式拟合的数据均值估计方法,构造出了一种能够较为准确的估计噪声门限的方法,并以之进行噪声背景归一化,在较好保留信号信息的前提下较大程度的抑制了噪声,有效降低了多目标信号检测的虚警概率。通过计算机仿真对比了该算法与S3PM算法、OTA算法的性能,结果表明该噪声背景归一化算法能够在检测概率损失较小的情况下较大幅度地降低检测的虚警率。实际被动声呐数据处理的对比结果同样验证了该算法的有效性。      

A new super-resolution PIV accelerated by characteristic pixel selection

A new high-speed super-resolution PIV was proposed using characteristic pixel selection to accelerate the successive abandonment (SA) with recursive window subdivision. The performance and applicability of the proposed PIV were evaluated. In the SA calculation with the characteristic pixel selection, 1000 candidates are narrowed down to only one at over 50 % of the measurement points, and the number of error vectors is reduced because the difference between the cumulative intensities of a correct candidate and of other ones becomes clear due to the characteristics of selected pixels. In all recursive processes, error checks are carefully performed using the summation of the distribution of the cumulative intensity difference distribution, which is suitable for the SA method. In a comparison of the time per velocity vector, the present super-resolution PIV was shown to be 10 times faster than the former ordinary resolution PIV. Another feature of the present super-resolution PIV is that the velocity vectors are obtained in the region very close to the image boundaries and masked regions by using the recursive algorithm.