红外车辆目标的自动模糊分割

针对红外图像的特点,提出了一种基于遗传算法的自动模糊分割红外车辆目标图像的方法.首先选取图像的感兴趣区域以加快运算速度;然后对感兴趣区域图像进行模糊增强,借助于二维OTSU方法对增强后的感兴趣区域进行阈值分割,为了加快分割算法的速度,先限定一个最佳阈值范围,再利用遗传算法在此阈值范围内自动搜索最佳分割阈值;为了弥补单独利用二维OTSU方法分割的不足,采用缩短模糊边缘宽度的方法来提取感兴趣区域红外车辆目标图像的边缘.最后把二维OTSU方法分割的图像与模糊边缘提取得到的边缘图像进行或运算后进行填充以得到最终的车辆目标分割图像.实验结果表明,对于红外车辆目标图像,一维OTSU和二维OTSU算法只是基本分割出了红外车辆目标的主体,而本文提出的自动模糊分割技术不仅准确分割出了红外车辆目标的主体,而且对于坦克的模糊炮塔亦得到了完整的分割.     

时空联合的红外运动目标提取算法

针对红外图像对比度差、边缘模糊的特点,提出了一种基于时空联合的红外序列图像目标提取的新方法.算法充分利用了红外目标的亮度特征、背景信息以及运动信息.时域分割中通过建立帧差图像背景的高斯分布模型,采用变化检测模板来确定红外目标约束区域.然后,构造图像像素与区域之间的空间关系隶属度矩阵并约束到传统的模糊聚类算法中,空域分割则利用该模糊聚类来对目标约束区域进行有效分割.最后将时空分割结果融合便能实现最终的红外目标提取.实验结果表明,该方法简单有效,能准确提取动态场景中的红外目标.     

改进的模糊阈值图像分割方法

提出了一种自适应的模糊阈值图像分割方法,通过预分割和直方图信息相结合的方法,解决了传统的模糊闽值图像分割法难以自动获取窗宽的困难;并针对模糊闽值图像分割方法不能适用于直方图呈单峰分布的图像的缺陷,提出了一个新的平滑迭代公式。该平滑迭代公式利用像素点的邻域信息使图像增强,再使用自适应的模糊阈值图像分割方法进行分割,可以拓宽模糊阈值图像分割方法的适用范围。实验结果表明,使用该方法的目标分割正确率达97.3％,显示了较高的分割精度和较强的鲁棒性。     

利用模糊综合评判技术提取红外图像海天线

海天线检测是红外图像自动目标识别技术的一项重要内容.本文深入分析了海天背景红外图像的特点,提出了一种基于四项特征评分因子的海天线模糊识别算法.首先对梯度图进行Radon变换以提取候选海天线,然后联合候选海天线四种特征信息进行模糊综合评判,最后根据评判值大小给出海天线识别结果.算法对比和实测数据结果表明,该方法适用范围广,抗噪声和抗干扰能力强,识别率高,可以有效地检测出各种复杂海天背景下的海天线,为进一步应用(如目标识别)奠定了良好的基础.     

基于空间邻域信息的二维模糊聚类图像分割

传统模糊C均值聚类(FCM)算法进行图像分割时仅利用了像素的灰度信息,并且使用对噪声较敏感的欧氏距离作为像素与聚类中心距离度量的标准,因此抗噪性能较差.为了克服传统FCM算法的局限性,本文提出了一种基于空间邻域信息的二维模糊聚类图像分割方法(2DFCM).该方法利用二维直方图描述的像素邻域关系属性,一方面为聚类提供较准确的初始聚类中心,从而避免聚类中的死点问题;另一方面通过提出聚类中心同时在像素值、像素邻域值二维方向上进行更新的思想,建立了包含邻域信息的新的聚类目标函数,实现了图像的分割.实验结果表明,这种方法抗噪能力强、收敛速度快,是一种有效的模糊聚类图像分割方法.     

多阈值S-F的光照不均图像分割

针对钢管识别统计系统开发中,图像分割环节易受光照不均匀影响的问题,以及对图像增强处理后再分割导致目标错分的不足,本文提出一种多阈值S-F(分割-融合)的图像分割方法。该方法根据改进的Otsu多阈值法,采用形态学操作与图像融合技术,实现堆垛钢管目标的提取。实验结果表明,在光照不均匀情况下,该方法对钢管图像的分割效果明显优于传统方法,具有不受光照优劣程度影响、适应性强的特点,可应用于机器视觉领域的目标识别。     

一种提高人耳特征点识别度的目标区域提取方法

对人耳进行特征识别多采用SURF算法,但该算法应用时极易受到图像中非目标区域的干扰,进而影响人耳特征点的检测和匹配准确度.基于目标区域的人耳特征识别算法可以突出目标区,而尽可能地抑制背景区域的影响.针对此问题,提出一种复合图像分割算法—KRM法作为人耳识别的预处理方法,将图像中人耳所在目标区域提取出来.该KRM法分为3步:首先利用k-means聚类算法将图像初步分割为前景目标区域和背景两类;再通过区域生长算法对过度分割的区域进行合并;最后应用形态学腐蚀的方法进行滤波得到人耳所在的目标区域.将KRM目标区域提取和SURF方法联用(简称KRM-SURF算法)应用于50组人耳图像,进行人耳特征点的检测与匹配,实验结果表明,特征点识别度(RD)均值达到0.924,KRM法的使用能极大地提高基于SURF算法的人耳特征识别的准确性.     

SAR图像目标边界流和边沿流特征提取

本文将目标边界流特征和目标边沿流特征的概念和边缘特征提取的方法用于SAR图像的目标检测和识别这个方法先采用中值滤波、自动图像分割、数学形态学二值滤波、十字模板边缘提取方法,获得目标单像素边缘再对所得边缘采用边界跟踪、边界流信息熵和边沿流信息熵运算,得到边界流特征和边沿流特征.对测试图像和真实SAR图像进行了实验,结果表明,该方法能很好地描述目标的特征,编程简单、运算速度较快.     

基于 NSCT 域特征和 PCNN 的SAR 图像目标分割

针对 SAR 图像的目标自动分割问题,在分析非下采样轮廓波变换和脉冲耦合神经网络的基础上,提出了一种基于非下采样轮廓波域特征图和 PCNN 的 SAR 图像目标分割算法.对 SAR 图像经过 NSCT 分解后的高、低频图像分别运用不同方式进行处理.对低频图用 PCNN 进行分割以获取目标所在的区域,对高频子带构造了特征图,对特征图利用 PCNN 进行分割以获取目标的精细结构.利用 MSTAR 数据进行了仿真实验,并与基于模糊 C 均值的分割算法、基于马尔可夫随机场的分割算法进行了对比.实验结果表明,所提出算法对 SAR 图像目标的分割结果更为准确,同时较其它算法具有更强的抗噪性能,是一种有效可行的 SAR 目标分割算法.     

基于边缘轮廓特征匹配的飞机目标识别

为快速准确识别可见光机场区域遥感图像中飞机目标的机型,提出一种基于边缘轮廓特征匹配进行识别的方法.对图像进行各向异性扩散处理滤除噪声,针对飞机停机位置是否有阴影存在,选择不同策略分割飞机目标.采用一种改进的区域分割方法提取目标,通过canny算子提取边缘轮廓,用主成份分析法获取飞机主轴,沿主轴垂直方向用相等间隔采样,提取主轴两侧边缘点间距作为特征参数.使用等级差数法对特征进行匹配,实现机型识别.实验结果表明,该方法能准确检测、识别飞机目标,统计得目标检测准确率96.44%,识别率94.07%,验证了算法的有效性.可将算法应用在对机场区域军事目标侦查识别中,能较为快速准确地识别飞机目标,体现了算法的应用价值.     

基于 SVM 的模糊图像识别

目的研究如何准确识别清晰图像与不同程度的模糊失真图像。方法首先对图像进行特征提取,主要从离散余弦变换域内的频率系数统计特征、峰度值、颜色饱和度三方面进行。然后在不同程度的模糊图像库中,利用支持向量机分辨出模糊图像。结果基于上述3种图像特征的组合,非常适合用于描述图像模糊现象,并且运用支持向量机分类器可以较为准确快速地区分出高斯模糊图像和清晰图像。结论提取模糊图像具有表征性的特征,可应用于不同程度模糊图像的识别,且运用支持向量机分类结果准确度也较高。此方法可应用于图像处理前期,剔除有碍信息表达的模糊图像。     

Depth extraction of three-dimensional objects using block matching for slice images in synthetic aperture integral imaging

We describe a computational method for depth extraction of three-dimensional (3D) objects using block matching for slice images in synthetic aperture integral imaging (SAII). SAII is capable of providing high-resolution 3D slice images for 3D objects because the picked-up elemental images are high-resolution ones. In the proposed method, the high-resolution elemental images are recorded by moving a camera; a computational reconstruction algorithm based on ray backprojection generates a set of 3D slice images from the recorded elemental images. To extract depth information of the 3D objects, we propose a new block-matching algorithm between a reference elemental image and a set of 3D slice images. The property of the slices images is that the focused areas are the right location for an object, whereas the blurred areas are considered to be empty space; thus, this can extract robust and accurate depth information of the 3D objects. To demonstrate our method, we carry out the preliminary experiments of 3D objects; the results indicate that our method is superior to a conventional method in terms of depth-map quality.     

Magnetic resonance imaging reconstruction via non‐convex total variation regularization

Magnetic resonance imaging (MRI) reconstruction model based on total variation (TV) regularization can deal with problems such as incomplete reconstruction, blurred boundary, and residual noise. In this article, a non‐convex isotropic TV regularization reconstruction model is proposed to overcome the drawback. Moreau envelope and minmax‐concave penalty are firstly used to construct the non‐convex regularization of L₂ norm, then it is applied into the TV regularization to construct the sparse reconstruction model. The proposed model can extract the edge contour of the target effectively since it can avoid the underestimation of larger nonzero elements in convex regularization. In addition, the global convexity of the cost function can be guaranteed under certain conditions. Then, an efficient algorithm such as alternating direction method of multipliers is proposed to solve the new cost function. Experimental results show that, compared with several typical image reconstruction methods, the proposed model performs better. Both the relative error and the peak signal‐to‐noise ratio are significantly improved, and the reconstructed images also show better visual effects. The competitive experimental results indicate that the proposed approach is not limited to MRI reconstruction, but it is general enough to be used in other fields with natural images.     

Blur-metric-based resolution enhancement of computationally reconstructed integral images

A computational integral imaging reconstruction technique can reconstruct a set of plane images of three-dimensional (3-D) objects along the output plane, in which only the plane object image (POI) reconstructed on the right planes where the objects were positioned is highly focused, whereas the other POIs reconstructed away from these planes are unfocused and blurred. In fact, these blurred POIs act as additional noises to other object images reconstructed on different output planes, so that the resolution of reconstructed object images should be considerably deteriorated. In this paper, a novel approach is proposed to effectively reduce the blurred images occurring in the focused POIs by employing a blur metric. From the estimated blur metric of each reconstructed POI, the right output planes where the objects were located can be detected. In addition, with an estimated blur metric, focused POIs can be adaptively eroded by a simple gray level erosion operation because it reduces regional expansion caused by the blur effect. The gray values of the eroded POIs are then finally remapped by referencing the original POIs. Some experiments revealed an average increase of 1.95 dB in the peak signal-to-noise ratio in the remapped POIs compared with that of the originally reconstructed POIs, and that the original forms of the object images in the remapped POIs could be preserved even after they had gone through an erosion operation. This feasibility test of the proposed scheme finally suggests a possibility of its application to robust detection and recognition of 3-D objects in a scene.     

离焦模糊图像的自适应滤波及逆滤波器复原

在微操作中,显微视觉系统获取的图像通常是离焦模糊图像.根据最小二乘原理和回归模型设计自适应滤波器,用于消除图像噪声,提高图像的信噪比;离焦模糊图像的退化模型可用圆盘函数描述,利用模糊图像频域的零点位置来估计圆盘函数的模糊参数;采用基于简化Wiener滤波的逆滤波器方法对模糊图像进行复原.对算法进行了仿真和实验分析,结果表明,该方法能够以较少的运算时间代价获取较好的复原效果.     

Recovery of blurred images obtained through turbulent media

Method for recovering blurred images taken through turbulent media, such as the atmosphere is presented. In this method amplitude and phase of the object's image are separately determined by special techniques of data processing, and finally the original image can be recovered. Principle and algorithm are described and the techniques to determine amplitude and phase are introduced by use of the results of computer simulations as well as light propagation experiments. As a demonstration to verify the utility of this method, images recovered by this method by use of data taken through a telescope of 1.5 m diameter, and the results of a computer simulation with atmospheric turbulence are shown. The results suggest that the presented method is well suited for the retrieval of blurred images.     

A study on restoration of iris images with motion‐and‐optical blur on mobile iris recognition devices

Iris recognition is a form of biometric technology that authenticates individuals by using the unique iris patterns between the pupil and the sclera. To solve security problems in mobile environments, mobile iris recognition devices have been commercialized recently. A motion‐and‐optical blurred image can be sometimes captured because users capture the iris images of a testee by holding the recognition devices. Motion‐and‐optical blurred images reduce iris recognition accuracy. Previous researches of restoring iris image only dealt with optical or motion blurred image. To overcome these problems, we propose a new method of restoring motion‐and‐optical blurred iris images at the same time. This article presents three contributions over previous research. (1) A new focus assessment method is proposed to measure accurate focus scores regardless of motion blurring. (2) Previous research restored coexisting motion‐and‐optical blurred images in terms of visibility, but in this article, we restored them in terms of recognition. (3) We used a modified CLS (Constrained Least Square) filter to prevent the zero‐crossing of the PSF (Point Spread Function) of motion blurring with a uniform shape. So, the iris recognition accuracy was better than when using a conventional CLS filter. Experimental results showed that the EER was 0.796% when using the proposed method and it was 1.431% when not using the proposed method. Consequently, the EER was reduced as much as 0.635% (1.431–0.796%) when using the proposed method. © 2009 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 19, 323–331, 2009     

Holographic Image Restoration by Using an Unconstrained Single Deblurring Filter

The unconstrained single deblurring filter for coherent optical restoration of blurred image is produced in a modified Rayleigh interferometer with blurred point spread function (PSF) h(x, y) and doubly blurred PSF h(x, y)=h(x, y)?h(x, y),? denoting convolution. The linear-motion blurred images and the defocusing blurred images are corrected with the present holographic filter, and it is shown that the restored images are significantly improved.     

基于颜色空间变换和CNN的自适应去模糊方法

目的研究数字图像中的去模糊问题,从受损的模糊图像中恢复出清晰图像。方法针对现有图像去模糊算法无法保留图像高频信息及容易产生振铃效应等问题,提出一种基于Y通道反卷积和卷积神经网络的两阶段自适应去模糊算法(SDYCNN)。在第1阶段,将数字图像转换至YUV颜色空间,根据图像无参考质量评价分数与模糊核尺寸之间的对应关系,在Y通道内自适应确定模糊核尺寸并进行反卷积增强;第2阶段将第1阶段中的反卷积增强作为预处理方式,通过4层卷积神经网络建立反卷积增强后的图像与清晰图像之间的映射关系,实现图像去模糊。结果轻微模糊图像在第1阶段便能够得到较好的去模糊效果,严重模糊图像经过第1阶段的反卷积增强,也有助于神经网络中特征的快速提取。结论实验结果表明,该算法不仅对于模糊图像具有良好的恢复效果,运算效率也有显著提升。     

随机图像选取与自适应背景更新的运动物体检测

为克服运动目标检测中易出现的光照变化、遮挡、虚假目标等现象,提出了一种随机图像选取与自适应背景更新的运动物体检测方法.该方法从视频序列中随机选取一帧图像作为初始背景,根据变化标记矩阵对背景进行自适应迭代更新,以提取可靠的背景图像,实现运动物体的检测.实验结果表明,采用该算法提取的背景不存在混合现象,且在光照变化较大以及运动物体之间存在遮挡的情况下,能够构造出可靠的背景,检测出的目标物体清晰可见.