Multi-resolution image segmentation based on Gaussian mixture model

Mixture model based image segmentation method, which assumes that image pixels are independent and do not consider the position relationship between pixels, is not robust to noise and usually leads to misclassification. A new segmentation method, called multi-resolution Ganssian mixture model method, is proposed. First, an image pyramid is constructed and son-father link relationship is built between each level of pyramid. Then the mixture model segmentation method is applied to the top level. The segmentation result on the top level is passed top-down to the bottom level according to the son-father link relationship between levels. The proposed method considers not only local but also global information of image, it overcomes the effect of noise and can obtain better segmentation result. Experimental result demonstrates its effectiveness.

Geographic,Geometrical and Semantic Reconstruction of Urban Scene from High Resolution Oblique Aerial Images

An effective approach is proposed for 3D urban scene reconstruction in the form of point cloud with semantic labeling. Starting from high resolution oblique aerial images,our approach proceeds through three main stages: geographic reconstruction, geometrical reconstruction and semantic reconstruction. The absolute position and orientation of all the cameras relative to the real world are recovered in the geographic reconstruction stage. Then, in the geometrical reconstruction stage,an improved multi-view stereo matching method is employed to produce 3D dense points with color and normal information by taking into account the prior knowledge of aerial imagery.Finally the point cloud is classified into three classes(building,vegetation, and ground) by a rule-based hierarchical approach in the semantic reconstruction step. Experiments on complex urban scene show that our proposed 3-stage approach could generate reasonable reconstruction result robustly and efficiently.By comparing our final semantic reconstruction result with the manually labeled ground truth, classification accuracies from86.75% to 93.02% are obtained.

Multi-sensor image registration using multi-resolution shape analysis

Multi-sensor image registration has been widely used in remote sensing and medical image field, but registration performance is degenerated when heterogeneous images are involved. An image registration method based on multi-resolution shape analysis is proposed in this paper, to deal with the problem that the shape of similar objects is always invariant. The contours of shapes are first detected as visual features using an extended contour search algorithm in order to reduce effects of noise, and the multi-resolution shape descriptor is constructed through Fourier curvature representation of the contour’s chain code. Then a minimum distance function is used to judge the similarity between two contours. To avoid the effect of different resolution and intensity distribution, suitable resolution of each image is selected by maximizing the consistency of its pyramid shapes. Finally, the transformation parameters are estimated based on the matched control-point pairs which are the centers of gravity of the closed contours. Multi-sensor Landsat TM imagery and infrared imagery have been used as experimental data for comparison with the classical contour-based registration. Our results have been shown to be superior to the classical ones.

Fast image super-resolution algorithm based on multi-resolution dictionary learning and sparse representation

Sparse representation has attracted extensive attention and performed well on image super-resolution（SR） in the last decade. However, many current image SR methods face the contradiction of detail recovery and artifact suppression. We propose a multi-resolution dictionary learning（MRDL） model to solve this contradiction, and give a fast single image SR method based on the MRDL model. To obtain the MRDL model, we first extract multi-scale patches by using our proposed adaptive patch partition method（APPM）. The APPM divides images into patches of different sizes according to their detail richness. Then, the multiresolution dictionary pairs, which contain structural primitives of various resolutions, can be trained from these multi-scale patches.Owing to the MRDL strategy, our SR algorithm not only recovers details well, with less jag and noise, but also significantly improves the computational efficiency. Experimental results validate that our algorithm performs better than other SR methods in evaluation metrics and visual perception.

Magnetic-resonance image segmentation based on improved variable weight multi-resolution Markov random field in undecimated complex wavelet domain

To solve the problem that the magnetic resonance(MR)image has weak boundaries,large amount of information,and low signal-to-noise ratio,we propose an image segmentation method based on the multi-resolution Markov random field(MRMRF)model.The algorithm uses undecimated dual-tree complex wavelet transformation to transform the image into multiple scales.The transformed low-frequency scale histogram is used to improve the initial clustering center of the K-means algorithm,and then other cluster centers are selected according to the maximum distance rule to obtain the coarse-scale segmentation.The results are then segmented by the improved MRMRF model.In order to solve the problem of fuzzy edge segmentation caused by the gray level inhomogeneity of MR image segmentation under the MRMRF model,it is proposed to introduce variable weight parameters in the segmentation process of each scale.Furthermore,the final segmentation results are optimized.We name this algorithm the variable-weight multi-resolution Markov random field(VWMRMRF).The simulation and clinical MR image segmentation verification show that the VWMRMRF algorithm has high segmentation accuracy and robustness,and can accurately and stably achieve low signal-to-noise ratio,weak boundary MR image segmentation.

Study on threshold segmentation of multi-resolution 3D human brain CT image

In order to effectively improve the pathological diagnosis capability and feature resolution of 3D human brain CT images,a threshold segmentation method of multi-resolution 3D human brain CT image based on edge pixel grayscale feature decomposition is proposed in this paper.In this method,first,original 3D human brain image information is collected,and CT image filtering is performed to the collected information through the gradient value decomposition method,and edge contour features of the 3D human brain CT image are extracted.Then,the threshold segmentation method is adopted to segment the regional pixel feature block of the 3D human brain CT image to segment the image into block vectors with high-resolution feature points,and the 3D human brain CT image is reconstructed with the salient feature point as center.Simulation results show that the method proposed in this paper can provide accuracy up to 100%when the signal-to-noise ratio is 0,and with the increase of signal-to-noise ratio,the accuracy provided by this method is stable at 100%.Comparison results show that the threshold segmentation method of multi-resolution 3D human brain CT image based on edge pixel grayscale feature decomposition is signicantly better than traditional methods in pathological feature estimation accuracy,and it effectively improves the rapid pathological diagnosis and positioning recognition abilities to CT images.

An image registration method based on multi-resolution morphology contour detection

Combined with the printing application,an image registration method based on the multi-resolution morphology contour detection was proposed.First,a direction based multi-resolution gray morphology in the scheme was proposed to realize the contour extraction.Then,based on the contour features,the subspace image registration was proposed to deal with issues of the computing complexity appeared in the traditional image registration methods.The proposed image registration was efficiently applied in the defect inspection of printing images.

Wavelet Transform for Image Compression Using Multi-Resolution Analytics: Application to Wireless Sensors Data

The aggregation of data in recent years has been expanding at an exponential rate. There are various data generating sources that are responsible for such a tremendous data growth rate. Some of the data origins include data from the various social media, footages from video cameras, wireless and wired sensor network measurements, data from the stock markets and other financial transaction data, supermarket transaction data and so on. The aforementioned data may be high dimensional and big in Volume, Value, Velocity, Variety, and Veracity. Hence one of the crucial challenges is the storage, processing and extraction of relevant information from the data. In the special case of image data, the technique of image compressions may be employed in reducing the dimension and volume of the data to ensure it is convenient for processing and analysis. In this work, we examine a proof-of-concept multiresolution analytics that uses wavelet transforms, that is one popular mathematical and analytical framework employed in signal processing and representations, and we study its applications to the area of compressing image data in wireless sensor networks. The proposed approach consists of the applications of wavelet transforms, threshold detections, quantization data encoding and ultimately apply the inverse transforms. The work specifically focuses on multi-resolution analysis with wavelet transforms by comparing 3 wavelets at the 5 decomposition levels. Simulation results are provided to demonstrate the effectiveness of the methodology.

基于Oblique显示技术的钝性外伤患者肋骨隐匿性骨折的影像诊断及其分型研究

目的:应用Oblique显示技术对肋骨隐匿性骨折进行诊断及分型。方法:选取2014年3月-2018年12月就诊于本院106例明确钝性外伤史疑似肋骨隐匿性骨折患者,所有患者经常规显示技术均未发现肋骨骨折。所有患者应用Oblique显示技术检查进行肋骨隐匿性骨折诊断,并根据骨折累及程度及表现形式进行肋骨隐匿性骨折分型。结果:106例患者共有189根肋骨存在隐匿性骨折,分为单皮质缘骨折、双皮质缘骨折及钝性骨挫伤。其中单皮质缘骨折105根,包括单纯外缘骨皮质褶皱成角64根,内缘骨皮质褶皱成角41根;内缘及外缘同时褶皱成角的双皮质缘骨折55根;骨皮质缘光滑无褶皱,仅表现为髓腔内骨髓水肿的钝性骨挫伤29根。结论:Oblique显示技术可以提高肋骨隐匿性骨折的诊断率,并能依据累及程度及表现形式进行合理分型。

Multi-Focus Image Fusion Based on Wavelet Transformation

In the fusion of image,how to measure the local character and clarity is called activity measurement. According to the problem,the traditional measurement is decided only by the high-frequency detail coefficients, which will make the energy expression insufficient to reflect the local clarity. Therefore,in this paper,a novel construction method for activity measurement is proposed. Firstly,it uses the wavelet decomposition for the fusion resource image, and then utilizes the high and low frequency wavelet coefficients synthetically. Meantime,it takes the normalized variance as the weight of high-frequency energy. Secondly,it calculates the measurement by the weighted energy,which can be used to measure the local character. Finally,the fusion coefficients can be got. In order to illustrate the superiority of this new method,three kinds of assessing indicators are provided. The experiment results show that,comparing with the traditional methods,this new method weakens the fuzzy and promotes the indicator value. Therefore,it has much more advantages for practical application.

An Effective Method of Threshold Selection for Small Object Image

The image segmentation difficulties of small objects which are much smaller than their background often occur in target detection and recognition. The existing threshold segmentation methods almost fail under the circumstances. Thus, a threshold selection method is proposed on the basis of area difference between background and object and intra-class variance. The threshold selection formulae based on one-dimensional (1-D) histogram, two-dimensional (2-D) histogram vertical segmentation and 2-D histogram oblique segmentation are given. A fast recursive algorithm of threshold selection in 2-D histogram oblique segmentation is derived. The segmented images and processing time of the proposed method are given in experiments. It is compared with some fast algorithms, such as Otsu, maximum entropy and Fisher threshold selection methods. The experimental results show that the proposed method can effectively segment the small object images and has better anti-noise property.

Analysis of color distortion and optimum fusion for remote sensing images using the statistical property of wavelet decomposition

IHS （Intensity, Hue and Saturation） transform is one of the most commonly used tusion algonthm. But the matching error causes spectral distortion and degradation in processing of image fusion with IHS method. A study on IHS fusion indicates that the color distortion can＇t be avoided. Meanwhile, the statistical property of wavelet coefficient with wavelet decomposition reflects those significant features, such as edges, lines and regions. So, a united optimal fusion method, which uses the statistical property and IHS transform on pixel and feature levels, is proposed. That is, the high frequency of intensity component Ⅰ is fused on feature level with multi-resolution wavelet in IHS space. And the low frequency of intensity component Ⅰ is fused on pixel level with optimal weight coefficients. Spectral information and spatial resolution are two performance indexes of optimal weight coefficients. Experiment results with QuickBird data of Shanghai show that it is a practical and effective method.

Multi-resolution network based image steganalysis model

Recently, many steganalysis approaches improve their feature extraction ability through addingconvolutional layers. However, it often leads to a decrease of resolution in the feature map during downsampling,which makes it challenging to extract weak steganographic signals accurately. To address this issue, this paperproposes a multi-resolution steganalysis net (MRS-Net). MRS-Net adopts a multi-resolution network to extract globalimage information, fusing the output feature map to ensure high-dimensional semantic information andsupplementing low-level detail information. Furthermore, the model incorporates an attention module which cananalyze image sensitivity based on different channel and spatial information, thus effectively focusing on areas withrich steganographic signals. Multiple benchmark experiments on the BOSSBase 1.01 dataset demonstrate that theaccuracy of MRS-Net significantly improves by 9.9% and 3.3% compared with YeNet and SRNet, respectively,demonstrating its exceptional steganalysis capability.

基于高速摄像法的气液段塞流特性参数测量

在气-液两相段塞流中,液膜段的湿壁分数和弹头倾角是研究气弹段持液率和气弹特性的关键.文章基于光学成像技术设计了高速摄像法的视觉传感器测试系统,实现了气液两相段塞流气弹特性参数测量.基于图像分析技术,优化边缘检测算子,精确定位边缘位置,高质量提取了轮廓分布,将机器学习模型作为预测多相流特征参数的一种新方法,对气弹弹头倾角进行数据建模研究,分析了适用于分层平滑流、波浪和环状流的表观粗糙表面(modified apparent rough surface,MARS)模型以及段塞流气弹区和液膜区流动状态机理,在MARS模型基础上进行了参数优化,使其适用于段塞流液膜处的湿壁分数和持液率求解.结果表明:通过数据拟合的多项式模型弹头倾角α的模型预测结果的平均绝对百分比误差(MAPE)为8.87%,94.6%预测结果处于相对误差±20%的范围内.通过修正MARS模型提出的水平管段塞流液膜处湿壁分数预测结果的MAPE为9.42%,96.1%预测结果处于相对误差±20%的范围内.液膜持液率预测模型结果的MAPE约为8.04%,93.4%预测结果处于相对误差±20%的范围内.

腰椎滑脱对椎体节段血管及OLIF手术通道安全区的影响

目的通过MRI研究Ⅰ~Ⅱ度腰椎滑脱对腰椎左侧节段血管及OLIF手术通道安全区的影响。方法回顾2019年10月~2022年11月收治的221例符合选择标准的腰腿痛患者腰椎MRI资料,将椎间隙高度无明显塌陷且椎体间无移位的腰椎间盘突出、椎间盘源性腰痛、退行性腰椎管狭窄症患者归为对照组(n=107),Ⅰ~Ⅱ度腰椎滑脱的患者归为观察组(n=114)。测量并比较左侧椎体侧方节段血管发出角度、血管走行角度、血管长度、椎体不同Moro区域节段血管间距及其与相邻椎间盘边缘的距离,分析血管间面积和22 mm模拟通道内出现血管比例的差异。结果①与对照组相比,观察组L_(4)血管发出角度、L_(1)/L_(3)/L_(4)/L_(5)血管走行角度、L_(2)/L_(4)血管长度均减小;在Ⅱ~Ⅲ、Ⅲ~Ⅳ区交界,L_(1,2)、L_(3,4)节段血管间距减小,二者血管间面积也相应减小。②两组L_(4,5)椎体节段血管间距在Ⅲ~Ⅳ区交界均<22 mm,除L_(3~5)在I~Ⅱ区交界节段血管尾侧与同序列椎体下缘的距离>10 mm外,两组患者其它部位节段血管与同序列椎体上/下缘的距离均<10 mm。节段血管与同序列椎体下缘距离:观察组L_(1) Ⅱ~Ⅳ区交界的距离减小,L_(4) Ⅱ~Ⅳ区交界的距离增大;节段血管与同序列椎体上缘距离:观察组L_(2) I~Ⅱ区交界、L_(2)~5 Ⅱ~Ⅲ区交界、L_(2)/L_(4) Ⅲ~Ⅳ区交界的距离减小。③观察组模拟通道内发现血管的比例大于对照组,其中以L_(4,5) I区、Ⅱ区发现L_(5)节段血管的差异最显著。结论I~Ⅱ度腰椎滑脱影响腰椎左侧节段血管的走形,缩小L_(1,2)、L_(3,4)节段血管间距,缩小Ⅱ~Ⅲ区交界L_(2)~L_(5)节段血管至同序列椎体上缘的距离,增加了OLIF通道建立的难度和风险。尤其是在L_(4,5)间隙进行操作时,需避免在距椎间盘边缘10 mm的椎体及Ⅲ~Ⅳ区交界放置OLIF扩张通道,可适当将通道向Ⅱ区头侧调整,密切注意L_(5)节段血管在I区、Ⅱ区的走形分布以避免损伤。

基于几何约束的倾斜影像特征匹配方法

针对倾斜影像视角变换较大、重复纹理导致匹配数量少、匹配精度不高的问题,提出一种适用于倾斜影像的特征点、线分级匹配方法.首先,用直线提取(检测)算法(LSD)获取影像直线特征,并将直线特征以一定约束进行直线组对,构建直线对区域与改进的SIFT(scale-invariant feature transform)特征描述符进行匹配,使用RANSAC算法剔除误匹配,获得初始匹配结果后再进行核线约束;然后,利用已获得直线对区域进行影像局部纠正,在纠正后的局部影像上采用SIFT匹配并反算回原始影像,利用得到的同名点全局纠正倾斜影像,并进行特征点匹配与采用基于方格的运动统计算法(GMS)剔除误匹配,仍将匹配结果反算回原始影像上;最后,将仿射尺度不变特征变化结果与点拓展匹配结果进行合并,得到最终匹配结果.试验结果表明:本文方法匹配正确率与经典的仿射不变匹算法(ASIFT)的正确率相差不大,但匹配数量却是ASIFT算法的1倍~3倍.

基于Pix4Dmapper软件的大疆Phantom4 RTK正射影像图精度分析

针对无人机正射影像图精度分析,提出一种基于Pix4Dmapper软件的大疆Phantom4 RTK正射影像图精度分析方法。该方法以大疆精灵4 RTK为例,利用Pix4Dmapper软件对其飞行的照片数据进行解算,针对同一作业区域比较了3D倾斜摄影飞行模式和2D正射飞行模式所产生的正射影像在平面精度和高程精度上的差异;分析了低空飞行模式下,采用像控点参与解算对成果精度的影响。实验结果表明,Phantom4 RTK在低空2D正射影像飞行模式下可以达到免像控生产正射影像图和地形图的要求,为无人机正射影像图精度分析提供了一种新思路。

基于参数自适应CNN的倾斜摄影影像融合方法

针对目前遥感影像融合的精度较差的问题,搭建一种基于参数自适应卷积神经网络(CNN)的倾斜摄影影像融合方法。首先,搭建一种基于深度卷积神经网络的影像融合模型。然后,引入通道相似度注意力,以实现网络特征的自适应学习。结果表明,所提算法在空间相关系数方面的表现较好,指标为0.758,通用影像质量为0.57。此外,算法的影像失真较少,峰值信噪比为23.75 dB。在Pavia Center数据集中,所提算法展示出较好的空间细节信息保持能力、影像融合精度、匹配效果以及影像融合效果。在CAVE数据集上也具有较好的影像融合效果,相关系数最大,为0.997,光谱角反射和相对全局合成误差分别为4.440和2.617。实验结果证明,所提模型在影像融合方面具有较好的细节信息保持能力。

基于机载激光扫描和倾斜影像匹配的点云构建及差异性分析

点云是三维空间数据的重要组成部分,机载激光扫描和倾斜影像匹配是两种主要点云构建技术,具有共性与差异性。本文旨在分析机载激光扫描和倾斜影像匹配技术构建点云的差异性,首先阐述了机载激光扫描和倾斜影像匹配点云构建的具体方法,然后结合案例,从点云构建效果、数据完整性、密度、精度和植被穿透性等方面进行比较和分析。结果表明,两种技术构建点云都具有较好的数据完整性,均能生产高密集点云,远超规范要求,精度相当且较高,但是激光扫描在遮挡区有漏洞,影像匹配局部盲区细节失真;激光点云具有较好的穿透性,能实现不同植被覆盖区域地面点云构建,而影像匹配点穿透性差,在密集植被区域存在缺失甚至不能构建地面点。研究结果为后续点云构建方法的选用和优化提供了参考,具有一定的研究与应用价值。

卫星斜视影像地形遮蔽区的自动检测研究

针对卫星斜视影像普遍存在的地形遮挡问题,文章提出了基于有理多项式系数(Rational Polynomial Coefficients,RPC)模型的高分辨率卫星影像遮挡检测方法。利用不规则三角网(Triangulated Irregular Network,TIN)表达三维地形表面,并采用光线追踪法精确探测入射光线与三角网的相交情况以确定交点间的遮挡关系。为提高遮挡检测效率,设计了一种极小化地形搜索区的方法,并利用斜视成像的真实卫星影像及模拟卫星影像进行了实验验证,结果表明该方法对遮挡区域的检测精度高于97%,能有效地识别出受地形遮挡影响的区域,自动生成准确的地形遮蔽区域掩膜影像。