多神经网络图像分割算法

为了提高机器视觉系统的图像分割精度,提出了一种以小波变换为分类特征的多神经网络(Multi-Neural Network,MNN)图像分割算法,该算法包括小波特征提取、MNN区域划分和MNN分类3个阶段。其中,小波特征提取根据小波变换的各层图像扩展得到分类特征;MNN区域划分将初分割边界附近区域分为训练样本区域和待分类区域(待分区域),并用多边形拟合算法将待分区域划分为多个局部待分区域;MNN分类将每个局部待分区域的像素用区域内的神经网络分类器进行分类,确定目标像素和背景像素,将目标像素合并后再进行一定的后处理即可得到分割结果。以轴承表面缺陷检测系统采集的轴承缺陷图像为对象,对MNN算法和阈值分割算法进行了对比试验,结果显示MNN算法的像素数量误差(Pixel Error,PE)相比阈值分割算法降低了75%,分割精度显著提高。     

气泡高干扰在线铁谱视频图像的磨粒快速分割算法研究

针对在线铁谱视频图像气泡高干扰所面临的磨粒分割困难问题,提出一种气泡高干扰在线铁谱视频图像的磨粒快速分割算法。首先运用运动检测的方法确定视频中气泡的位置,并用相邻帧相同位置的图像信息对气泡区域进行处理,再使用双边滤波对处理后的图像进行平滑去噪,实现气泡干扰的初步抑制;最后基于抑制气泡图像的灰度直方图,对每一帧图像选取其自适应的阈值,实现在线铁谱视频图像中磨粒的快速分割。该研究为在线铁谱的磨粒分割与后续对磨粒特征的智能提取和分析奠定了基础。     

海上监控图像的海天线及其目标区域提取

根据海上监控采集海面图像的特点,提出一种适用于海天线及其目标区域提取的方法。该方法具体是对海面图像灰度化和高斯滤波后,首先根据灰度差异采用OSTU自适应阈值分割方法对海面图像进行分割;接着利用Canny边缘检测得到图像的所有边缘点集,从而获得主要轮廓;然后,根据梯度方向,减少无关边缘点,来改进Hough变换,连接相邻斜率误差小的直线段,减少直线段数目,选取有效程度最大的直线为海天线;最后,依据海天线的位置确定海天线区域。海天线区域的确定,为运动目标的检测、跟踪和避障节约了时间,并减少了误检概率。     

一种对ICT复杂图像进行二值化处理的方法

比较、评价了各种经典的图像二值化方法对ICT图像二值化处理的效果。综合分析及实践发现，实验法、大津法不适用于对ICT图像进行二值化分割；而迭代法、边缘检测阈值法、加权边缘检测阈值法对ICT图像二值化分割效果不好。直方图法处理的效果稍好，但它仍将非常接近背景灰度的目标物部分错误划分为背景。提出“背景法”、直方图法与背景法结合的改进方法，可对ICT断层图像进行较理想的二值化分割。     

基于背景差分的高铁钢轨表面缺陷图像分割

高铁钢轨表面图像具有光照变化、反射不均、特征少等特点,使得缺陷自动检测极为困难。为了在高速运动过程中,从复杂的钢轨表面图像中分割出缺陷,根据钢轨表面图像具有沿钢轨方向像素值基本不变的特征,建立钢轨表面图像背景模型,提出了基于背景差分的钢轨表面缺陷检测算法,主要包括钢轨区域提取、背景建模差分、阈值分割和图像滤波4个步骤,其主要特点是将视频监控中的背景差分法推广到缺陷图像分割领域,同时借助自适应阈值分割和滤波技术,在一定程度上,解决了铁轨表面缺陷分割过程中图像光照变化、反射不均、特征少等不利因素的影响。实验仿真和现场测试结果均表明,该方法对块状缺陷能很好地识别,召回率和准确率分别达96%和80.1%。     

基于数学形态学的图像分割研究及应用

给出了一种基于图像灰度分布统计特征期望值为阈值的二值化方法．并在此基础上运用数学形态学方法进行图像目标定位分割。应用于有噪声的车牌图像二值化和分割、实验表明本文提出的二值化算法的计算复杂度．执行时间、二值化效果以及适用范围等方面均有较佳表现．形态学分割算法定位效果好、分割精度高。     

图像分割技术及其在模拟指针式仪表自动检定系统中的应用

图像分割技术是智能模拟指针式仪表检定系统中的用的图像分割方法是根据系统自身的实际特点,将多种图像分割方法如基于阈值的分割和基于区域的分割综合使用,并根据具体的要求,对基于边缘的分割方法做了适当的改进,最终得到了适合系统自身要求和特点的图像分割方法.在实际应用系统中,该算法的分割的效果好于其他的单一的图像分割算法,分割出的图像清晰,噪声小,为后续的处理程序提供了良好的基础.     

Hartmann-Shack波前传感器图像自动阈值选取方法的研究

Hartman-Shack传感器图像阈值的选取对质心探测精度有较大的影响,而质心探测精度将直接影响探测波前的复原。通过介绍几种阈值分割的原理和方法,在比较多幅基点图像质心探测精度的基础上,提出了基于灰度拉伸的大津法是阈值分割的较好方法。     

基于混沌和声搜索算法的图像分割研究

为克服图像分割中图像目标与背景之间存在的对比度低、亮度分布不均匀以及在搜索最佳阈值的过程中有可能陷入局部最优解等问题,运用混沌和声搜索算法搜寻最佳阈值,将模糊Renyi熵与混沌和声搜索算法相结合进行图像分割。实验结果显示了该方法对红外人体图像以及单峰、双峰、多峰图像都有较好的分割效果。     

水下海参图像处理方法研究

由于海参养殖的水下环境相对复杂,导致水下拍摄得到的图像容易受到水源浑浊度、光照折射或水流速度等环境因素影响,造成图像质量低,更有海草、水下生物、人类垃圾等背景因素干扰,给图像中海参的检测造成了很大的困难。结合水下海参图像的实际情况,首先对采集到的图像进行灰度值范围调整使图像更加明亮,突出图像中的重要信息;选择高斯高通滤波器使图像得到锐化处理去除图像的低频部分,突出图像边缘;之后采用全局阈值分割方法将目标对象和背景较为明显的区分出来。通过上述方法实现海参图像的增强、分割处理,为水下海参图像的检测识别提供一种有效的图像处理方法。     

A novel segmentation approach for noisy medical images using Intuitionistic fuzzy divergence with neighbourhood‐based membership function

Medical image segmentation demands higher segmentation accuracy especially when the images are affected by noise. This paper proposes a novel technique to segment medical images efficiently using an intuitionistic fuzzy divergence–based thresholding. A neighbourhood‐based membership function is defined here. The intuitionistic fuzzy divergence–based image thresholding technique using the neighbourhood‐based membership functions yield lesser degradation of segmentation performance in noisy environment. Its ability in handling noisy images has been validated. The algorithm is independent of any parameter selection. Moreover, it provides robustness to both additive and multiplicative noise. The proposed scheme has been applied on three types of medical image datasets in order to establish its novelty and generality. The performance of the proposed algorithm has been compared with other standard algorithms viz. Otsu's method, fuzzy C‐means clustering, and fuzzy divergence–based thresholding with respect to (1) noise‐free images and (2) ground truth images labelled by experts/clinicians. Experiments show that the proposed methodology is effective, more accurate and efficient for segmenting noisy images.     

基于线性插值的图像分块分割算法

图像分割是图像处理中的一个重要问题,全局阈值法和局部阈值法是图像分割中广泛采用的两类方法。全局阈值法效率高,但是对局部信息不敏感;局部阈值法对局部信息保存较好,但是计算效率不高。针对这一问题,该文提出一种基于分块的图像分割算法。首先将图像分块,然后,每块内采用最大类间方差法计算阈值,最后,块与块之间采用线性插值的方式平滑分割图像。实验表明,该文提出的算法和全局阈值法相比能较好地保存局部信息,和局部阈值法相比有更高的运行效率,在实际中推荐使用。     

Grain‐oriented segmentation of images of porous structures using ray casting and curvature energy minimization

We segment an image of a porous structure by successively identifying individual grains, using a process that requires no manual initialization. Adaptive thresholding is used to extract an incomplete edge map from the image. Then, seed points are created on a rectangular grid. Rays are cast from each point to identify the local grain. The grain with the best shape is selected by energy minimization, and the grain is used to update the edge map. This is repeated until all the grains have been recognized. Tests on scanning electron microscope images of titanium oxide and aluminium oxide show that their process achieves better results than five other contour detection techniques.     

基于改进多尺度分形特征的红外图像弱小目标检测

为提高红外图像弱小目标检测的准确率和实时性,在分析用于红外图像增强的分形参数K相关的多尺度分形特征(MFFK)基础上,提出了一种基于改进多尺度分形特征(IMFFK)的红外图像弱小目标检测算法。首先,将基于地毯覆盖法的分形维数计算公式代入MFFK计算公式,提出了一种改进多尺度分形特征(IMFFK)用于图像增强。其次,对IMFFK特征计算进行简化,采用自适应阈值分割得到感兴趣目标区域,提出了一种具有较高计算效率的红外图像弱小目标检测算法。最后,通过仿真图像分析了主要参数对图像增强和算法耗时的影响,采用红外真实图像进行了算法检测性能测试,并与当前基于局部对比度测度的目标检测算法进行了对比。实验结果表明,提出的算法虽然在一些检测场景具有较多虚警,但能同时适用于弱小目标和较大目标检测,且无论目标为亮目标或暗目标。提出算法对于低分辨率红外图像(320×240)检测接近30 frame/s。提出算法具有较强的适用性,能够检测出红外图像中具有较高局部对比度的目标。     

Automatic thresholding from the gradients of region boundaries

We present an approach for automatic threshold segmentation of greyscale images. The procedure is inspired by a reinterpretation of the strategy observed in human operators when adjusting thresholds manually and interactively by means of ‘slider’ controls. The approach translates into two methods. The first one is suitable for single or multiple global thresholds to be applied globally to images and consists of searching for a threshold value that generates a phase whose boundary coincides with the largest gradients in the original image. The second method is a variation, implemented to operate on the discrete connected components of the thresholded phase (i.e. the binary regions) independently. Consequently, this becomes an adaptive local threshold procedure, which operates relative to regions, rather than to local image subsets as is the case in most local thresholding methods previously published. Adding constraints for specifying certain classes of expected objects in the images can improve the output of the method over the traditional ‘segmenting first, then classify’ approach.     

基于视频图像的工业仪表高速自动识别

针对工业现场高速变化的灯板仪表全自动检测的问题,提出了一种基于图像识别的检测方法。本方法采用高速工业相机获取仪表视频图像,并采用数字图像处理技术,对单帧图像用OSTU算法进行二值化分割,运用邻域平均法进行平滑处理,通过闭运算以完成去噪,运用Hough变换以实现对灯的逐一定位,并用灰度检测法对灯状态进行判读,最后保存视频。研究结果表明,该算法可靠有效,系统能够高速准确地应对仪表位置的变化,为进一步的一般化视频检测奠定了基础。     

Automatic setae segmentation from Chaetoceros microscopic images

A novel image processing model Grayscale Surface Direction Angle Model (GSDAM) is presented and the algorithm based on GSDAM is developed to segment setae from Chaetoceros microscopic images. The proposed model combines the setae characteristics of the microscopic images with the spatial analysis of image grayscale surface to detect and segment the direction thin and long setae from the low contrast background as well as noise which may make the commonly used segmentation methods invalid. The experimental results show that our algorithm based on GSDAM outperforms the boundary‐based and region‐based segmentation methods Canny edge detector, iterative threshold selection, Otsu's thresholding, minimum error thresholding, K‐means clustering, and marker‐controlled watershed on the setae segmentation more accurately and completely. Microsc. Res. Tech. 77:684–690, 2014. © 2014 Wiley Periodicals, Inc.     

基于二维对称Tsallis交叉熵的小目标图像阈值分割

现有的阈值分割方法应用于目标与背景面积相差悬殊的小目标图像时,几乎都失效.为此,提出了基于对称Tsallis交叉熵及背景与目标面积差的小目标图像阈值分割方法.对称Tsallis交叉熵准则能确保准确分割时目标和背景内部的灰度均匀,而背景与目标面积差可抑制均等分割的趋势,二者综合构成了更为合理的阈值选取准则函数.首先导出了一维阈值选取公式;然后经推广得到基于二维斜分对称Tsallis交叉熵及背景与目标面积差的阈值选取公式,给出了其快速递推算法及相应的简化方法.大量实验结果表明:与目前性能较优越的二维斜分Otsu、最大熵、非对称交叉熵阈值分割方法相比,所提出的方法在小目标图像分割效果上具有极为明显的优势.     

Application of Dynamic Thresholding of Video Images for Measuring the Interfacial Tension of Liquids and Contact Angles

A gradient image-segmentation method for analyzing fields of images containing several objects is developed. This method uses the dynamic thresholding and is applied to the analysis of video images of pendant and sessile drops for determining the interfacial tension of liquids and contact angles.     

In-situ water level measurement using NIR-imaging video camera

Water level is a critical component for observation and management of water resources. Since video surveillance is becoming a standard configuration of gauging stations, more attentions have been paid on the image-based water level measurement techniques in recent years. Instead of human eyes, images of staff gauge can be captured by a camera and automatically processed to detect the readings of water line. However, it is quite unreliable to identify the readings on staff gauges due to low imaging resolution, tilt viewing angle and complex illumination in field conditions. For that reason, most of the existing methods failed to make effective automatic measurements in practice. To resolve difficulties of poor visibility, image distortions and ambient noises in in-situ water level measurement with standard bicolor staff gauges, an all-weather, real-time and automatic flow measurement system using single near infrared (NIR)-imaging video camera is developed. Lab and in-situ experiments demonstrate that NIR-imaging is efficient to enhance image contrast and suppress reflection noises on the water surface, which successfully overcomes the limitation of water line detection with current visible light (VIS)-imaging systems in clear water and low velocity conditions. The proposed water level conversion method is based on image ortho-rectification. It requires no on-site calibration by utilizing natural corresponding points to build the perspective transformation between the staff gauge Region of Interest (ROI) and an orthographic template image. The measurement resolution of water level (1 mm) is controlled by the physical resolution (1 mm/pixel) of the template. The proposed water line detection approach is integrated by a series of algorithms: the Order-Statistic Filtering for adaptive thresholding, the Morphological Opening Operation for local noise suppression, the Multi-points Continuity Criterion for water line locating and the Median Filtering for random noise elimination. It overcomes the deficiency of traditional Otsu method for images with uneven illumination. The reliability of system under various weather and illumination conditions is proved with water level accuracy up to 1 cm and Effective Data Ratios up to 95%.