Inherent bias in correlation averaged images

The correlation averaging algorithm frequently used to enhance micrographs of repeating structures contains an inherent bias that favours images with larger pixel values or positive noise levels. This bias not only skews the composite image toward higher pixel values, but also distorts the image by increasing the value of high-valued pixels more than that of low-valued pixels. These errors are especially important in scanning probe microscopy images where the pixel value reflects a distinct height. A similar algorithm that uses a structure function in place of the correlation function eliminates this bias.     

Segmenting time‐lapse phase contrast images of adjacent NIH 3T3 cells

We present a new method for segmenting phase contrast images of NIH 3T3 fibroblast cells that is accurate even when cells are physically in contact with each other. The problem of segmentation, when cells are in contact, poses a challenge to the accurate automation of cell counting, tracking and lineage modelling in cell biology. The segmentation method presented in this paper consists of (1) background reconstruction to obtain noise‐free foreground pixels and (2) incorporation of biological insight about dividing and nondividing cells into the segmentation process to achieve reliable separation of foreground pixels defined as pixels associated with individual cells. The segmentation results for a time‐lapse image stack were compared against 238 manually segmented images (8219 cells) provided by experts, which we consider as reference data. We chose two metrics to measure the accuracy of segmentation: the ‘Adjusted Rand Index’ which compares similarities at a pixel level between masks resulting from manual and automated segmentation, and the ‘Number of Cells per Field’ (NCF) which compares the number of cells identified in the field by manual versus automated analysis. Our results show that the automated segmentation compared to manual segmentation has an average adjusted rand index of 0.96 (1 being a perfect match), with a standard deviation of 0.03, and an average difference of the two numbers of cells per field equal to 5.39% with a standard deviation of 4.6%.     

Signal‐to‐noise ratio enhancement on SEM images using a cubic spline interpolation with Savitzky–Golay filters and weighted least squares error

A new technique based on cubic spline interpolation with Savitzky–Golay smoothing using weighted least squares error filter is enhanced for scanning electron microscope (SEM) images. A diversity of sample images is captured and the performance is found to be better when compared with the moving average and the standard median filters, with respect to eliminating noise. This technique can be implemented efficiently on real‐time SEM images, with all mandatory data for processing obtained from a single image. Noise in images, and particularly in SEM images, are undesirable. A new noise reduction technique, based on cubic spline interpolation with Savitzky–Golay and weighted least squares error method, is developed. We apply the combined technique to single image signal‐to‐noise ratio estimation and noise reduction for SEM imaging system. This autocorrelation‐based technique requires image details to be correlated over a few pixels, whereas the noise is assumed to be uncorrelated from pixel to pixel. The noise component is derived from the difference between the image autocorrelation at zero offset, and the estimation of the corresponding original autocorrelation. In the few test cases involving different images, the efficiency of the developed noise reduction filter is proved to be significantly better than those obtained from the other methods. Noise can be reduced efficiently with appropriate choice of scan rate from real‐time SEM images, without generating corruption or increasing scanning time.     

Single-image signal-to-noise ratio estimation

A method for estimating the signal-to-noise ratio from a single image is presented in this paper. The autocorrelation-based technique requires that image details be correlated over distances of a few pixels, while the noise is assumed to be uncorrelated from pixel to pixel. The latter is shown to be a good approximation in the case of scanning electron microscope (SEM) images provided that the video signal is not band limited. The noise component is derived from the difference between the image autocorrelation at zero offset and an estimate of the corresponding noise-free autocorrelation. Nonlinear effects introduced by intensity saturation and their implications on the image signal-to-noise ratio are also discussed.     

Visibility of objects in computer simulations of noisy micrographs

Thresholds of visibility for objects in images with random pixel noise are predicted in terms of the signal-to-noise ratio. From trials with volunteers marking test images, we determined visibility thresholds of objects obscured by random pixel noise. The test images had objects with a variety of simple shapes and relatively little internal structure. Aside from the noise, the background of the test images was smooth and featureless. We extend the threshold signal-to-noise ratio measurements of Rose and others to a variety of object sizes and shapes. For objects with areas less than a disc subtending 2° at the eye, visibility depends on the averaged difference in intensity from background, the noise level and the number of pixels in the object. Visibility does not seem to depend on object shape.     

High-resolution large-strain measurement of plastically deformed specimen by Fourier phase correlation

The Fourier phase correlation method is applied to position circular marks one by one printed on the specimen before and after deformation in order to measure the large plastic strain. This method is extremely sensitive to the differences between the profile of a mark and noise. Therefore, it detects marks easily even under non-uniform illumination without any processing such as flattening illumination, noise exclusion and boundary enhancement in image transformation into binary codes. In addition, a new method of moving reference images (MRI) to position the observed marks in the sub-pixel range is proposed. The MRI technique introduces the theoretical resolution 1/n (n: number of divisions of one pixel). The MRI method enables the positions of the deformed marks to be determined with a resolution of ±0.1 pixel (standard deviation σ_p is 0.042 pixel). Strain of the tensile specimen can be measured within an error of ±0.0015 (standard deviation σ_ε=0.00055).     

An expanded approach to noise reduction from high-resolution STEM images based on the maximum entropy method

An expanded use of the maximum entropy method (MEM) is suggested to reduce noise from an experimental high-angle annular dark-field (HAADF) scanning transmission electron microscope (STEM) image. The MEM is combined with an estimate of the standard deviation of noise from an experimental HAADF STEM image and low-pass filtering using the information limit for an incoherent STEM image. Consequently, the present method has just one parameter of a Lagrange multiplier. It is demonstrated that the present method can reduce noise efficiently in high-resolution HAADF STEM images.     

A microcomputer based digital imaging system for ion microanalysis

A microcomputer based digital imaging system was developed for a Cameca IMS-3f ion microscope permitting real-time digital acquisition of secondary ion images. Image signal-to-noise enhancement results from random noise reduction by real-time ensemble averaging and from a reduction of pattern noise in the charge injection device (CID) array by subtraction of blank frames. Acquired images comprise 244times248 pixel arrays with 8-bit intensity resolution. Images are displayed on a colour monitor in a grey scale or pseudo-colour using one of four programmable lookup tables. Image processing software permits off-line ion image enhancement and manipulation as well as multitechnique digital image correlations. System capability is illustrated by a biological example involving digital imaging studies of Al distribution in osteomalacic bone tissue, including correlative light microscopy and ion microanalysis.     

针对重复图案晶片自动检测的新方法

具有重复图案阵列的IC在整个芯片中占有较大的比例,针对此类IC的结构特征,提出一种基于旋转不变子空间技术（ESPRIT）算法的自比较模板匹配缺陷检测方法。首先应用ESPRIT算法精确计算出重复图案中结构块的大小,然后将图像中所有结构块对应位置的像素值平均,计算出标准结构块。再根据显微镜视野大小将标准结构块进行扩展形成标准模板,通过比对实现缺陷检测。试验结果表明：采用ESPRIT算法求取结构块大小,具有较高的速度和精度,能够满足IC检测实时性要求。算法的计算复杂度为O(N的3/2次方) ,精度可以达到0.04个像素。     

强噪声下自适应Canny算子边缘检测

针对传统Canny算子不能有效滤除图像在解码处理和传输过程产生的椒盐噪声、无法保留边缘细节的问题,提出强噪声下Canny算子图像边缘检测算法.依据椒盐噪声的极值性、灰度差值性,将像元点划分为噪声点、疑似噪声点;根据分类之后的像元点自适应地改变滤波器窗口的大小和权值,在降低噪声影响的同时能较好地保留图像细节.引入8个方向...     

遥感相机焦面CCD机械拼接中重叠像元数的确定

针对推扫式遥感相机前后摆成像中焦面拼接的CCD其像片间出现视场漏缝的问题,提出了一种焦面CCD的非均匀重叠像元数机械拼接方法.首先,构造了相机前后摆的成像原理模型;接着,分析了CCD间产生地面视场漏缝的原因,并依据模型推导出相机前摆和后摆时各CCD视场间漏缝大小的数学公式,从而得到各片间容许的最少重叠像元数量公式.在此基础上,进一步给出了后续电子学图像拼接与配准中片间重复像元数的函数形式,为实现图像无缝拼接提供了软件实现依据.分析表明,该方法的理论误差ε小于一个像元尺寸,即|ε/D|＜1.最后,给出了工程应用实例.     

Estimation of optimum number of cycles for combustion analysis using measured in-cylinder pressure signal in conventional CI engine

Analysis of measured in-cylinder pressure data provides various parameters that characterize engine combustion process. Advanced engine control technologies use cylinder pressure based combustion parameters for closed loop control. Four step signal processing (i) absolute pressure correction, (ii) crank angle position referencing, (iii) cycle averaging and (iv) filtering is typically applied to get sufficiently accurate cylinder pressure data for an engine cycle. This paper focuses on cycle averaging and filtering of in-cylinder pressure signal from a conventional compression ignition (CI) engine. Experiments are conducted at different engine load and compression ratios at 1500 rpm. The in-cylinder pressure trace of 2500 consecutive engine cycles is recorded and analyzed. Effect of in-cylinder pressure signal noise and cyclic variation on combustion analysis is investigated. A method based on standard deviation of pressure and pressure rise rate is used to find sufficient minimum number of engine cycles to be recorded for averaging to get reasonably accurate pressure data independent of cyclic variability.     

基于边缘像元投影的微小轴承亚像素边缘检测

对于微小轴承的内径测量,基于机器视觉的非接触测量技术具有广阔的应用前景。针对轴承的边缘像素点,投影将感光像元划分成感光值不同的两个部分,像素单元的最终灰度值为投影两侧局部灰度统计值的面积加权平均值,取该像素的矩形邻域,并对其邻域灰度值进行高斯加权处理,将边缘投影按泰勒公式展开为二次曲线得到高精度亚像素边缘。针对微小轴承进行实验,将本文算法与基于直线拟合的方法进行比较,每毫米包括约217.18个像素,检测的圆度提高0.04%以上,标准差减少不少于7.27%,因此可更准确获取微小轴承内径的亚像素边缘。     

基于双线性滤波器的安防相机低照度去噪研究

针对安防监控摄像机在夜间拍摄时易产生噪声,以及所使用的图像去噪方法易产生图像边缘特征退化的问题,提出了一种基于图像像素值差的监控摄像机低照度去噪方法。通过计算像素值差,再下采样得到图像金字塔,以获得整体和局部像素值差的索引,由此设计了一种适用于低照度的去噪卷积核,以增加噪声对中间像素值的影响,实现低照度环境下图像噪点的消除并保留其边缘特征。实验表明,该方法在低照度场景实际应用中既去除了图像的噪点,又保留了边缘信息。与基于小波变换的算法相比,该方法明显提升了图像的主观评价质量,并在Imatest信噪比测试中提升了图像的亮度信噪比。     

Method of calibration of a fluorescence microscope for quantitative studies

Confocal microscopy is based on measurement of intensity of fluorescence originating from a limited volume in the imaged specimen. The intensity is quantized in absolute (albeit arbitrary) units, producing a digital 3D micrograph. Thus, one may obtain quantitative information on local concentration of biomolecules in cells and tissues. This approach requires estimation of precision of light measurement (limited by noise) and conversion of the digital intensity units to absolute values of concentration (or number) of molecules of interest. To meet the first prerequisite we propose a technique for measurement of signal and noise. This method involves registration of a time series of images of any stationary microscope specimen. The analysis is a multistep process, which separates monotonic, periodic and random components of pixel intensity change. This approach permits simultaneous determination of dark and photonic components of noise. Consequently, confidence interval (total noise estimation) is obtained for every level of signal. The algorithm can also be applied to detect mechanical instability of a microscope and instability of illumination source. The presented technique is combined with a simple intensity standard to provide conversion of relative intensity units into their absolute counterparts (the second prerequisite of quantitative imaging). Moreover, photobleaching kinetics of the standard is used to estimate the power of light delivered to a microscope specimen. Thus, the proposed method provides in one step an absolute intensity calibration, estimate of precision and sensitivity of a microscope system.     

工业CT含噪图像中的裂纹探测

提出一种基于快速Beamlet变换的工业CT含噪图像的裂纹探测方法。首先,分析图像在单尺度下的Beamlet组成以及Beamlet间的相互关系,设计出一种快速Beamlet变换。基于快速Beamlet变换,引入一个关于吻合度的控制量,并结合Beamlet自身的多尺度树型结构,采取"自上向下"寻找目标函数最优值的思路,得到裂纹的探测结果。最后,结合探测结果的相邻区域的像素特性,提取出含噪裂纹的区域边界。分别对含有裂纹的CT图像,以及叠加方差为0.1的高斯白噪声图像,叠加强度为0.1的椒盐噪声图像进行探测实验。结果表明,与基于Laplace、Canny或小波的探测方法相比,该方法能有效探测到工业CT含噪图像中的裂纹。因为Beamlet是以线基的方式分析图像数据,所以该方法对噪声干扰具有很好的抑制能力,成功实现了工业CT含噪图像中的裂纹探测。     

基于改进Census变换的图像立体匹配算法研究

针对传统的 Census 区域匹配算法过分依赖窗口中心像素信息,导致算法受到噪声干扰时匹配精度降低的问题,提出一种基于改进 Census 变换的匹配算法.采用局部像素反差值为中心像素选择的评判标准,对传统的 Census 变换进行改进,增强了窗口像素信息的利用,提高了算法对像素值突变的适应性,使算法有更好的鲁棒性;代价聚合阶段采用引导图滤波算法并结合多尺度聚合模型,增强平坦区域像素间的区分度;采用 win-take-all 算法选取最优视差值,完成视差计算;采用区域投票策略和中值滤波算法完成视差精化.利用该改进算法对 Middlebury 平台提供的标准图像进行实验,实验结果表明该算法较传统Census 算法有较好的抗噪能力和立体匹配精度.     

A comparison of different intensity gradient formulae for orientation analysis of electron micrographs

A detailed study of different intensity gradient formulae was conducted using 91 separate images covering a wide range of microfabric. It was found that a formula based on the 3 × 3 local pixel array gives almost identical results over all images with the standard formula based on 20 pixels forming a circular grouping within the 5 × 5 local array. Other formulae based on other groupings of pixels within the same pixel arrays give markedly different results. Computation times are reduced up to 45% when the 8,5 formula is used in place of the 20,14 one, and would thus seem to have advantages when large numbers of images are to be processed. While all formulae give somewhat similar domain-segmented images when viewed visually, differences are noted both in the number of segmented domains as well as in the area of domains in each orientation sector. It is possible that the radius of the modal filter selected for the domain segmentation should be varied according to the formula type used.     

Retrieving true images through fine grid steps for enhancing the resolution beyond the classical limits: theory and simulations

We developed the theory of, and tested by extended simulations, a novel method for retrieving true images in a grid step much finer than both the acquisition and the optical microscope limits. We believe that the method is promising in view of avoiding the limitations on the resolution improvement in direct imaging mode systems. Two basic concepts are involved: (i) random (up to 3D) relative displacements of objects with respect to the receiving matrix and (ii) the use of a reference object firmly fixed to small signal objects for avoiding the displacement measurements. The retrieved images are created by rearranging a set of true images acquired with a lower resolution equal to the matrix pixel size. We demonstrate the good quality of the retrieved images and the possibility to visualize and detect small (convolved) objects not observed into the captured images. The method provides good opportunities for effective applications of different inverse algorithms for improving the resolution requiring, as a rule, more precisely sampled images, but at arbitrary relations between the pixel size and the optical diffraction limit. We further demonstrated the application of some deconvolution procedures for extracting highly resolved images in the object and image planes in the presence of noise. The possibility to resolve small objects beyond the two classical limits is shown by means of simulations. The estimates for the method's limiting resolution, combined with proper deconvolution processing, show that resolution in the lower nano‐dimension scale (below 10 nm) could be achieved. The requirements to the implementation of the novel method are commented as well.     

高帧频CMOS相机对光通信精跟踪系统影响分析

分析了互补金属氧化物半导体(CMOS)相机作为机载光通信精跟踪系统的位置传感器的优势,采用基于CMOS相机与FPGA研制的精跟踪系统,可实现帧频8 800 Hz的图像采集、数据传输、实时图像处理、光斑质心计算以及伺服控制等功能,分析了CMOS相机的噪声、帧频以及光斑质心算法对精跟踪系统跟踪精度的影响,给出了对应的数值分析与实验结果。实验结果表明:CMOS相机噪声对光斑坐标x轴、y轴引起坐标的标准差分别为0.291 9像素,0.120 2像素;当相机帧频从2 200 Hz提高到8 800 Hz时跟踪标准差从1.8降低到0.5个像素;外场光斑质心算法偏差比理想高斯光斑计算结果大8.34-0.71像素。光斑中心算法对跟踪精度影响明显,设计更具针对性的自适应光斑中心算法可以大大改善精跟踪的效果。