一种高效的油液磨粒图像自适应分割方法

为了提高油液磨粒图像的分割效果,优化磨粒自动识别工作的重要环节,提出了一种结合分水岭算法及区域相似度合并的油液磨粒图像自适应分割方法。对于待分割图像,首先通过形态学重构和基于形态学的自适应H-minima技术对梯度图像进行修正,利用分水岭算法完成磨粒图像的一次分割;其次提取分水岭分割后同质区域的Lab颜色特征、局部二值模式（LBP）纹理特征作为区域的量化指标,基于Bhattacharyya系数分别计算区域间的颜色、纹理相似度,设计可以实现权重自适应调整的颜色、纹理特征融合规则,以此来获取同质区域的综合相似度矩阵,实现过分割区域的合并;最后基于统一的后处理算法完成磨粒图像的完整分割。选择60幅磨粒图像对所提方法的分割效果进行测试,结果表明,单幅图像的平均分割速度在12 s左右,分割正确率稳定在90%以上,所提方法避免了用户在分割过程中对图像的交互式处理,较好地平衡了分割效率和分割精度,自适应程度明显提高。      

基于区域小波变换的序列显微图像融合

利用图像处理技术实现序列显微图像的融合显示是解决显微成像焦深有限的有效方法.针对空域融合算法在处理离焦步长较大和透明显微图像时存在的块效应和重影现象,基于图像像素间高度的相关性和小波变换多分辨率分析的特点,提出了一种基于区域小波变换的序列显微图像融合算法:在小波变换域内进行序列显微图像融合时,不同频段选择不同的融合算子;在确定融合的小波系数时,不仅考虑到相应位置的小波系数,还考虑了与它相邻的小波系数,使得融合后图像克服了空域算法存在的块效应和重影现象.同时,研究了不同的小波滤波器对图像融合质量的影响.实验证明,该方法具有很好的融合效果.      

基于ICA和BP神经网络相结合的掌纹识别

提出了一种基于独立成分分析(ICA,Independent Comment Analysis)和多层前馈(BP,Back Propagation)神经网络相结合的方法对掌纹进行识别.首先采用一种新的方法检测角点,得到掌纹图像的不变特征点,根据这些点校正掌纹方向并得到掌纹的感兴趣区域.对该区域采用定点快速ICA算法(FastICA),得到掌纹特征子空间,然后构建BP神经网络,并采用训练样本得到的掌纹特征进行训练,得到合适的权值.对香港理工大学掌纹数据库进行测试,与主成分分析(PCA,Principal Components Analysis)提取特征的方法进行比较,取得了较高的识别率.      

基于混合稀疏表示的二维压缩感知SAR成像

压缩感知(CS)理论在合成孔径雷达(SAR)成像中应用广泛。针对包含城市、河流等区域的非稀疏场景压缩感知SAR成像,提出基于近似观测模型的混合稀疏表示(MSR)压缩感知SAR成像方法。该方法将复杂的SAR图像分解成点、线、面,并将线、面分别通过离散余弦变换和曲波变换转换到稀疏域,使压缩感知的稀疏性条件得以满足,通过求解基于近似观测模型的二维压缩感知优化问题重建非稀疏场景的SAR图像。所提方法能够实现降采样率条件下对包含城市、河流等非稀疏场景区域的成像,仿真场景和实测场景成像结果表明了所提方法的有效性。     

基于小波域清晰度评价的光场全聚焦图像融合

传统全聚焦图像融合以相机多次曝光拍摄的多聚焦图像为基础,光场相机在单次曝光后可计算空间任意深度的重聚焦图像,为后期全聚焦图像的获取提供便利。提出了一种基于小波变换的光场全聚焦图像获取算法,可有效避免传统空域图像融合算法的块效应,获得较高质量的全聚焦图像。该算法通过对微透镜阵列光场相机获得的4D光场数据进行空间变换与投影,得到用于全聚焦图像融合的重聚焦图像,对各帧重聚焦图像进行小波分解提取高、低频子图像集,提出区域均衡拉普拉斯算子、像素可见度函数分别构建融合图像的高、低频小波系数实现图像融合,其性能优于传统的区域清晰度评价函数。实验验证了所提算法的正确性和有效性,采用Lytro光场相机的原始数据计算了融合全聚焦图像,与传统图像融合算法相比,人眼视觉效果更好,客观图像指标也得到了提高。      

基于长椭球面波函数去卷积方法的高动态范围太阳射电成像

利用综合孔径射电望远镜对太阳进行观测时,通过对图像中存在的明亮扩展源进行准确建模并移除,可以更好地观测视场内的微弱源并提高图像的动态范围。在射电天文领域,主要利用CLEAN算法对图像中的明亮源进行移除,以显示微弱的背景。然而,使用图像像素作为基函数的CLEAN算法的固有限制导致其对扩展源的建模效果较差。为了克服这种限制,将基于长椭球面波函数（Prolate Spheroidal Wave Functions,PSWF）的去卷积方法应用于太阳射电成像。PSWF最优正交基由脏图中的感兴趣区域（Region of Interest,ROI）和UV覆盖共同决定。为了验证该方法的有效性,基于PSWF正交基对均匀圆环阵观测的太阳射电图像进行去卷积,并从动态范围和保真度两个方面定量化对比了CLEAN算法和基于PSWF正交基方法的性能。基于PSWF正交基去卷积方法剩余脏图中的微弱源更接近真实情况且动态范围更高。     

基于多特征的高分遥感图像分割算法研究

针对传统的图像分割算法不能完全适用具有多种特征〖BF〗（〖BFQ〗光谱特征、纹理特征和几何特征〖BF〗）〖BFQ〗的高分辨率遥感图像的问题,提出了一种基于多特征的遥感图像分割算法。算法基于改进的均值漂移滤波和自动标记分水岭分割方法来实现最终分割。首先利用自动标记分水岭分割方法对遥感图像进行分割,进而采用仿射不变矩形状特征算子提取图像几何特征;其次对图像进行主成分分析,计算第一主成分灰度共生矩阵,分析矩阵特性得出纹理特征;然后结合光谱特征通过改进的均值漂移方法得到多特征滤波结果;最后利用分水岭分割方法实现高分辨率遥感图像分割。为了表明算法的分割效果,利用基于多光谱信息熵方法对算法和单一的分水岭分割方法进行非监督评价。研究结果表明,算法可较好地改善遥感图像的过分割问题,是一种适合高空间分辨率多光谱遥感图像的分割算法。     

An approach to multi-temporal MODIS image analysis using image classification and segmentation

This paper discusses an approach for river mapping and flood evaluation based on multi-temporal time series analysis of satellite images utilizing pixel spectral information for image classification and region-based segmentation for extracting water-covered regions. Analysis of MODIS satellite images is applied in three stages: before flood, during flood and after flood. Water regions are extracted from the MODIS images using image classification (based on spectral information) and image segmentation (based on spatial information). Multi-temporal MODIS images from “normal” (non-flood) and flood time-periods are processed in two steps. In the first step, image classifiers such as Support Vector Machines (SVM) and Artificial Neural Networks (ANN) separate the image pixels into water and non-water groups based on their spectral features. The classified image is then segmented using spatial features of the water pixels to remove the misclassified water. From the results obtained, we evaluate the performance of the method and conclude that the use of image classification (SVM and ANN) and region-based image segmentation is an accurate and reliable approach for the extraction of water-covered regions.     

基于区域灰度统计信号处理的图像融合方法

针对多源图像融合问题,提出了一种在多分辨率框架下基于区域内灰度特征统计信号的融合算法.利用图像灰度特征的区域生长法对源图像进行区域分割,并以裂缝边缘作为特征区域的闭合边界,对源图像与分割结果的区域映射图作多分辨率变换.在图像低频部分,以联合区域映射图为指导,在区域内建立信号与噪声的高斯混合分布模型,利用期望极大化(EM,Expectation Maximization)算法迭代估计噪声模型分布参数,获得低频融合结果;在图像高频部分,根据系数在区域映射图上的位置差异分别采用窗口系数加权平均法和系数绝对值选大法进行融合,将低频和高频融合结果反变换得到最终融合图像.融合结果表明:该方法是可行和高效的,且比其他图像融合方法具有更好的性能.     

利用多尺度区域权重的红外可见光图像融合

针对红外、可见光图像融合，提出了一种利用多尺度区域权重的方法，获得既具有良好视觉效果又能保留和继承特征细节的融合结果。首先，通过改变平滑滤波的参数构建多尺度分解，获得一系列具有不同细节的双波段图层；其次，基于每一图层生成区域权重图，实现双波段图层融合;最后，利用适当的系数对融合后的图层进行合成，获得最终的融合结果。通过对比实验结果的主客观评价分析，证明本文方法能够产生最佳视觉效果的融合图像，能保留和突出原始双波段图像中的细节与特征，并能获得最佳的客观评价指标。     

Rainfall estimation over a Mediterranean region using a method based on various spectral parameters of SEVIRI-MSG

The ultimate objective of this paper is the estimation of rainfall over an area in Algeria using data from the SEVIRI radiometer (Spinning Enhanced Visible and Infrared Imager). To achieve this aim, we use a new Convective/Stratiform Rain Area Delineation Technique (CS-RADT). The satellite rainfall retrieval technique is based on various spectral parameters of SEVIRI that express microphysical and optical cloud properties. It uses a multispectral thresholding technique to distinguish between stratiform and convective clouds. This technique (CS-RADT) is applied to the complex situation of the Mediterranean climate of this region. The tests have been conducted during the rainy seasons of 2006/2007 and 2010/2011 where stratiform and convective precipitation is recorded. The developed scheme (CS-RADT) is calibrated by instantaneous meteorological radar data to determine thresholds, and then rain rates are assigned to each cloud type by using radar and rain gauge data. These calibration data are collocated with SEVIRI data in time and space.     

实时棒材图像识别与跟踪方法研究

多目标识别跟踪的关键问题是特征提取和目标匹配.为了提取生产线上堆积棒材的特征,提出粘连目标分割和多目标识别的方法.采用中值滤波和形态学滤波去除噪声,自适应阈值化和分水岭变换分割粘连目标;然后采用区域统计、参数识别、噪声区域去除以及聚类分析等手段进行目标特征识别,提取出棒材的质心点坐标作为特征;对棒材图像序列提出采用模板匹配、相近位移匹配和Kalman滤波的方法建立跟踪链,通过插入、删除、更新链节点进行目标跟踪;对于图像处理中可能出现的漏检目标和虚增目标,进行了计数结果校正.在现场采集了100帧连续图像后,采用此方法跟踪计数的精度为96.2%.      

Adjustment of systematic errors in ALS data through surface matching

Surface matching is a well researched topic in both Computer Vision (CV) and terrestrial laser scanning (TLS) or ground based light detection and ranging (LiDAR), but the extent of the range images derived from these technologies is typically orders of magnitude smaller than those derived from airborne laser scanning (ALS), also known as airborne LiDAR. Iterative closest point (ICP) and its variants have been successfully used to align and register multiple overlapping views of the range images for CV and TLS applications. However, many challenges are encountered in applying the ICP approach to ALS data sets. In this paper, we address these issues, explore the possibility of automating the algorithm, and present a technique to adjust systematic discrepancies in overlapping strips, using geometrical attributes in a given terrain. In this method, the ALS point samples used in the algorithm are selected depending on their ability to constrain the relative movement between the overlapping laser strips. The points from overlapping strips are matched through modified point to plane based on the ICP method.     

New method of enhancement using wavelet transforms applied to SODISM telescope

PICARD is a space-based observatory hosting the Solar Diameter Imager and Surface Mapper (SODISM) telescope, which has continuously observed the Sun from July 2010 and up to March 2014. In order to study the fine structure of the solar surface, it is helpful to apply techniques that enhance the images so as to improve the visibility of solar features such as sunspots or faculae. The objective of this work is to develop an innovative technique to enhance the quality of the SODISM images in the five wavelengths monitored by the telescope at 215.0?nm, 393.37?nm, 535.7?nm, 607.1?nm and 782.2?nm. An enhancement technique using interpolation of the high-frequency sub-bands obtained by Discrete Wavelet Transforms (DWT) and the input image is applied to the SODISM images. The input images are decomposed by the DWT as well as Stationary Wavelet Transform (SWT) into four separate sub-bands in horizontal and vertical directions namely, low-low (LL), low–high (LH), high-low (HL) and high–high (HH) frequencies. The DWT high frequency sub-bands are interpolated by a factor 2. The estimated high frequency sub-bands (edges) are enhanced by introducing an intermediate stage using a Stationary Wavelet Transform (SWT), and then all these sub-bands and input image are combined and interpolated with half of the interpolation factor α/2, used to interpolate the high-frequency sub-bands, in order to reach the required size for IDWT processing. Quantitative and visual results show the superiority of the proposed technique over a bicubic image resolution enhancement technique. In addition, filling factors for sunspots are calculated from SODISM images and results are presented in this work.     

Application of CCD drift-scan photoelectric technique on monitoring GEO satellites

Geosynchronous Earth Orbit (GEO) satellites are widely used because of their unique characteristics of high-orbit and remaining permanently in the same area of the sky. Precise monitoring of GEO satellites can provide a key reference for the judgment of satellite operation status, the capture and identification of targets, and the analysis of collision warning. The observation using ground-based optical telescopes plays an important role in the field of monitoring GEO targets. Different from distant celestial bodies, there is a relative movement between the GEO target and the background reference stars, which makes the conventional observation method limited for long focal length telescopes. CCD drift-scan photoelectric technique is applied on monitoring GEO targets. In the case of parking the telescope, the good round images of the background reference stars and the GEO target at the same sky region can be obtained through the alternating observation of CCD drift-scan mode and CCD stare mode, so as to improve the precision of celestial positioning for the GEO target. Observation experiments of GEO targets were carried out with 1.56-meter telescope of Shanghai Astronomical Observatory. The results show that the application of CCD drift-scan photoelectric technique makes the precision of observing the GEO target reach the level of 0.2″, which gives full play to the advantage of the long focal length of the telescope. The effect of orbit improvement based on multi-pass of observations is obvious and the prediction precision of extrapolating to 72-h is in the order of several arc seconds in azimuth and elevation.     

基于改进SIFT的图像配准算法

为解决存在较大程度旋转和缩放的图像配准问题,提出了一种基于尺度不变特征变换(SIFT,Scale Invariant Features Transform)的图像配准算法.采用对数极坐标变换(LPT,Log-Polar Transform)进行图像粗匹配,对图像旋转角度和缩放尺度变化量进行估计,并对图像加以校正;在粗匹配的基础上对图像进行分块,根据信息熵原理提取子块的SIFT特征和不变矩特征,构造新型的特征描述符;结合欧氏距离和Procrustes迭代算法获得图像的同名点对,并估计图像形变参数,完成图像配准.实验结果表明:该算法速度快、稳定性强,并能达到亚像素级的匹配精度.     

基于改进PCNN的轴尖表面缺陷检测

脉冲耦合神经网络(PCNN,Pulse Coupled Neural Network)与传统神经网络不同,不经过训练即可用于图像处理.针对PCNN模型中结构参数较多,且需要人工反复试验进行设置的困难,改进模型结构,简化了馈送输入和连接输入,减少了待定参数;根据邻域灰度动态地计算内部连接系数,由邻域的欧氏距离计算权值矩阵,再由图像的灰度特征计算动态阈值.将改进的PCNN用于陀螺轴尖表面缺陷图像的分割,用基于完整性与正确性指标的缓冲区匹配方法评价所提方法、最大熵法及Canny方法.针对不同缺陷图像的实验表明:所提算法的完整性与正确性都高于0.9,证明所提方法更有效.     

基于图像特征分析的物体轮廓提取

对物体的轮廓进行分析提取,是计算机视觉方向的基础问题之一,对其进行研究对于复杂场景的分析理解至关重要。本文对室内场景图像进行研究,基于图像特征进行图像分割,提取物体轮廓。在彩色场景图像全局轮廓后验边界概率（gPb）提取算法的基础上,加入深度图像信息,对室内场景的彩色、深度（RGB-D）图像中的物体轮廓进行分析。通过多尺度信息融合,计算得到多尺度轮廓后验概率（mPb）和谱后验概率（sPb）,两后验概率加权综合得到gPb。而后结合超度量轮廓图与分水岭算法,对基于方向特征变化的gPb图像融合处理,最终得到清晰的物体轮廓。本文所提方法在通用的RGB-D数据库基础上进行实验。实验结果表明,本文所提出的方法能提取出清晰的室内物体轮廓图。      

The automated prediction of solar flares from SDO images using deep learning

In the last few years, there has been growing interest in near-real-time solar data processing, especially for space weather applications. This is due to space weather impacts on both space-borne and ground-based systems, and industries, which subsequently impacts our lives. In the current study, the deep learning approach is used to establish an automated hybrid computer system for a short-term forecast; it is achieved by using the complexity level of the sunspot group on SDO/HMI Intensitygram images. Furthermore, this suggested system can generate the forecast for solar flare occurrences within the following 24 h. The input data for the proposed system are SDO/HMI full-disk Intensitygram images and SDO/HMI full-disk magnetogram images. System outputs are the “Flare or Non-Flare” of daily flare occurrences (C, M, and X classes). This system integrates an image processing system to automatically detect sunspot groups on SDO/HMI Intensitygram images using active-region data extracted from SDO/HMI magnetogram images (presented by Colak and Qahwaji, 2008) and deep learning to generate these forecasts. Our deep learning-based system is designed to analyze sunspot groups on the solar disk to predict whether this sunspot group is capable of releasing a significant flare or not. Our system introduced in this work is called ASAP_Deep. The deep learning model used in our system is based on the integration of the Convolutional Neural Network (CNN) and Softmax classifier to extract special features from the sunspot group images detected from SDO/HMI (Intensitygram and magnetogram) images. Furthermore, a CNN training scheme based on the integration of a back-propagation algorithm and a mini-batch AdaGrad optimization method is suggested for weight updates and to modify learning rates, respectively. The images of the sunspot regions are cropped automatically by the imaging system and processed using deep learning rules to provide near real-time predictions. The major results of this study are as follows. Firstly, the ASAP_Deep system builds on the ASAP system introduced in Colak and Qahwaji (2009) but improves the system with an updated deep learning-based prediction capability. Secondly, we successfully apply CNN to the sunspot group image without any pre-processing or feature extraction. Thirdly, our system results are considerably better, especially for the false alarm ratio (FAR); this reduces the losses resulting from the protection measures applied by companies. Also, the proposed system achieves a relatively high scores for True Skill Statistics (TSS) and Heidke Skill Score (HSS).