基于微透镜阵列的实时三维物体识别

提出一种基于微透镜阵列多视角成像特点,将三维物体的深度信息转化为二维透射像阵列的角度信息,利用光学二维图像识别技术,实现对三维物体识别的方法.对识别过程进行了理论分析和计算,用匹配滤波的方法实现了对三维物体骰子的实时识别.实验结果表明,本方法的相关识别能力较高,并且具有很强的灵活性,对于有微小旋转、微小平移的三维物体也可进行识别.     

基于二维直方图和模糊熵准则的阈值化方法

图像分割是图像分析、识别和理解的基础。图像分割主要是指将图像分成各具特性的区域并提取出感兴趣目标的技术,其研究多年来一直受到人们的高度重视;阈值化法是图像分割的一种重要方法,在图像处理与识别中广为应用;针对图像分割中细节往往被忽略导致后续处理困难的问题,基于模糊关系和最大模糊熵原理提出了一种阈值化方法,对二维直方图进行模糊分割;为了获得图像分割中的细节,提出的方法根据最大熵原则自动确定模糊区域和门限,进而获得二维模糊熵和遗传算法最优解,最后获得图像细节;通过对不同灰度水平和颜色类型图像进行实验比较,实验结果表明提出的方法优于二维非模糊方法和一维模糊熵分割法,得到该方法在图像分割中获得细节的结论。     

Improving the Accuracy of Nearest-Neighbor Classification Using Principled Construction and Stochastic Sampling of Training-Set Centroids

A conceptually simple way to classify images is to directly compare test-set data and training-set data. The accuracy of this approach is limited by the method of comparison used, and by the extent to which the training-set data cover configuration space. Here we show that this coverage can be substantially increased using coarse-graining (replacing groups of images by their centroids) and stochastic sampling (using distinct sets of centroids in combination). We use the MNIST and Fashion-MNIST data sets to show that a principled coarse-graining algorithm can convert training images into fewer image centroids without loss of accuracy of classification of test-set images by nearest-neighbor classification. Distinct batches of centroids can be used in combination as a means of stochastically sampling configuration space, and can classify test-set data more accurately than can the unaltered training set. On the MNIST and Fashion-MNIST data sets this approach converts nearest-neighbor classification from a mid-ranking- to an upper-ranking member of the set of classical machine-learning techniques.     

Preliminary result of an advanced tangential X-ray pinhole camera system with a duplex MWPC on KSTAR plasma

An advanced tangential X-ray pinhole camera (TXPC) has been developed for KSTAR by utilizing a 2-D duplex multi-wire proportional counter (MWPC) detector. The KSTAR MWPC employs a 2-D parallel type readout system for high temporal resolution and adopts a duplex type for the capability of electron temperature measurement via the multi-color method. This paper presents the performance test result of the developed MWPC system utilizing a Fe-55 X-ray source. As a preliminary experimental result from the 2012 KSTAR campaign, the clear presentation of sawtooth activities and its frequency change, and 2-D plasma images during the vertical disruption event are given.     

Region-based SIFT approach to iris recognition

Traditional iris recognition systems transfer iris images to polar (or log-polar) coordinates and have performed very well on data that tends to have a centered gaze. The patterns of an iris are part of a 3-D structure that is captured as a two-dimensional (2-D) image and cooperative iris recognition systems are capable of correctly matching these 2-D representations of iris features. However, when the gaze of an eye changes with respect to the camera lens, many times the size, shape, and detail of iris patterns will change as well and cannot be matched to enrolled images using traditional methods. Additionally, the transformation of off-angle eyes to polar coordinates becomes much more challenging and noncooperative iris algorithms will require a different approach. The direct application of the scale-invariant feature transform (SIFT) method would not work well for iris recognition because it does not take advantage of the characteristics of iris patterns. We propose the region-based SIFT approach to iris recognition. This new method does not require polar transformation, affine transformation or highly accurate segmentation to perform iris recognition and is scale invariant. This method was tested on the iris challenge evaluation (ICE), WVU and IUPUI noncooperative databases and results show that the method is capable of cooperative and noncooperative iris recognition.     

Spin-image surface matching based target recognition in laser radar range imagery

We explore the problem of in-plane rotation-invariance existing in the vertical detection of laser radar (Ladar) using the algorithm of spin-image surface matching. The method used to recognize the target in the range imagery of Ladar is time-consuming, owing to its complicated procedure, which violates the requirement of real-time target recognition in practical applications. To simplify the troublesome procedures, we improve the spin-image algorithm by introducing a statistical correlated coefficient into target recognition in range imagery of Ladar. The system performance is demonstrated on sixteen simulated noise range images with targets rotated through an arbitrary angle in plane. A high efficiency and an acceptable recognition rate obtained herein testify the validity of the improved algorithm for practical applications. The proposed algorithm not only solves the problem of in-plane rotation-invariance rationally, but also meets the real-time requirement. This paper ends with a comparison of the proposed method and the previous one.     

基于支持向量机的舰船图像识别

支持向量机(SVM)是一种基于超平面分类的新的学习方法,具有很强的泛化能力。研究了支持向量机的学习机理,以及实现支持向量机的序贯最小优化算法(SMO),并用来对舰船图像进行识别。首先将待识别目标进行二维小波分解,获取不同尺度下的小波系数,然后对其进行主元分析,得到的主元分量作为支持向量机的特征量输入。实验结果表明,该方法具有良好的分类性能。     

Three-dimensional-object recognition by use of single-exposure on-axis digital holography

On-axis phase-shifting digital holography requires recording of multiple holograms. We describe a novel real-time three-dimensional- (3-D-) object recognition system that uses single-exposure on-axis digital holography. In contrast to 3-D-object recognition by means of a conventional phase-shifting scheme that requires multiple exposures, our proposed method requires only a single digital hologram to be synthesized and used to recognize 3-D objects. A benefit of the proposed 3-D recognition method is enhanced practicality of digital holography for 3-D recognition in terms of its simplicity and greater robustness to external scene parameters such as moving targets and environmental noise factors. We show experimentally the utility of the single-exposure on-axis digital holography-based 3-D-object recognition method.     

工业CT图像轮廓提取与优化系统设计

为了有效地提高面向逆向工程的工业CT图像轮廓处理的精准性和计算效率,开发了工业CT图像轮廓提取与优化系统。该系统主要包括图像处理模块、轮廓识别与提取模块、轮廓拟合与优化模块、图像处理经典算法库和数据接口等功能模块,实现了从工业CT图像到轮廓数据的三维点云与轮廓线的标准格式输出,并与逆向工程软件对接。同时,在轮廓拟合与优化模块中,提出了基于全局约束模型的轮廓拟合策略,并采用有向图分解和数值代数相结合的计算方法,提高了轮廓拟合的计算精度和效率。最后通过实例验证了系统的有效性和正确性。     

Human eye localization using the modified Hough transform

The precise localization of parts of a human face such as mouth, nose or eyes is important for their image understanding and recognition. The developed successful computer method of eyes and eyelids localization using the modified Hough transform is presented in this paper. The efficiency of this method was tested on two publicly available face images databases and one private face images database with the location correctness better than 96% for a single eye or eyelid and 92% for eye and eyelid couples.     

Optical display of magnified, real and orthoscopic 3-D object images by moving-direct-pixel-mapping in the scalable integral-imaging system

In this paper, we proposed a novel approach for reconstruction of the magnified, real and orthoscopic three-dimensional (3-D) object images by using the moving-direct-pixel-mapping (MDPM) method in the MALT(moving-array-lenslet-technique)-based scalable integral-imaging system. In the proposed system, multiple sets of elemental image arrays (EIAs) are captured with the MALT, and these picked-up EIAs are computationally transformed into the depth-converted ones by using the proposed MDPM method. Then, these depth-converted EIAs are combined and interlaced together to form an enlarged EIA, from which a magnified, real and orthoscopic 3-D object images can be optically displayed without any degradation of resolution. Good experimental results finally confirmed the feasibility of the proposed method.     

Fluorescence multicolor hologram recorded by using a macrolens array

We present a new efficient method for obtaining multicolor digital Fresnel holograms of three-dimensional (3-D) objects emitting incoherent fluorescent light. For each emitted fluorescent light wavelength, a single monochromatic digital modified Fresnel hologram of the 3-D scene is generated. Using a macrolens array, only nine projections of the 3-D scene are captured in a single camera shot and are then digitally processed to yield the hologram. The 3-D reconstructed images from all the monochromatic holograms are composed into a single multicolor image of the 3-D scene. The proposed holographic technique is demonstrated by experiments.     

Visualization of three-dimensional temperature distributions in a large-scale furnace via regularized reconstruction from radiative energy images: numerical studies

In this paper, the possibility of visualization of three-dimensional (3-D) temperature distributions in large-scale boiler furnaces from radiative energy images captured by multiple charge-coupled device (CCD) cameras mounted around the furnace is studied numerically. For the calculation of the radiative energy image formation, a fast algorithm proposed by the authors for pinhole imaging is used in this paper, which is based on the Monte Carlo method and combined with a concept of angular factor effective for image formation. This algorithm is applicable for the emitting, absorbing, and isotropic scattering medium. For the inversion of the 3-D temperature distributions which is an ill-posed problem, a modified Tikhonov regularization method is improved, where the finite difference regularizer is defined and can be used in 3-D cases, and the optimal regularization parameter is suggested to be selected by using a post-treatment method. For a 3-D unimodal temperature distribution, the numerical simulation results show that the reconstruction errors for the 3-D temperature distribution can be maintained at levels similar to the measurement error and the visualization quality of the temperature distribution is satisfactory. For a kind of bimodal temperature distribution, the reconstruction errors are higher than those for the unimodal distribution, but the bimodal feature of the temperature distribution can also be reproduced clearly.     

3-D imaging and modeling – Part I: acquisition and registration

The key modules in a typical reverse engineering system consist of the acquisition, registration, and the integration of range images of three-dimensional objects with complex geometry. We first present a self-made optical digitizer that is employed for quickly acquiring range data from multiple views, then introduce a visualized registration procedure that can provide a reliable estimate for the fine registration of multiple range images in a unified coordinate system. Experiment results are given to show the effectiveness of this approach in 3-D imaging and multiple range image registration for the applications of reverse engineering.     

Fuzzy-Fusion Classifier for MMV Full-Polarization Radar Target Recognition Using One-Dimensional Scattering Centers

One-dimensional (1-D) scattering centers on a target can be used for radar target recognition purpose. In this paper, a new classifier based on fuzzy techniques is proposed, which is very much suitable for target recognition using 1-D scattering centers. Furthermore, information fusion techniques for full polarimetric receiver data are utilized at the decision level. The performance of the proposed fuzzy-fusion classifier is evaluated using a dataset of five aircraft models measured by millimeter-wave (MMW) full-polarization stepped-frequency radar in the compact range facility.     

3-D imaging and modeling – Part I: acquisition and registration

The key modules in a typical reverse engineering system consist of the acquisition, registration, and the integration of range images of three-dimensional objects with complex geometry. We first present a self-made optical digitizer that is employed for quickly acquiring range data from multiple views, then introduce a visualized registration procedure that can provide a reliable estimate for the fine registration of multiple range images in a unified coordinate system. Experiment results are given to show the effectiveness of this approach in 3-D imaging and multiple range image registration for the applications of reverse engineering.     

A miniature high resolution 3-D imaging sonar

This paper discusses the design and development of a miniature, high resolution 3-D imaging sonar. The design utilizes frequency steered phased arrays (FSPA) technology. FSPAs present a small, low-power solution to the problem of underwater imaging sonars. The technology provides a method to build sonars with a large number of beams without the proportional power, circuitry and processing complexity. The design differs from previous methods in that the array elements are manufactured from a monolithic material. With this technique the arrays are flat and considerably smaller element dimensions are achievable which allows for higher frequency ranges and smaller array sizes. In the current frequency range, the demonstrated array has ultra high image resolution (1″ range × 1° azimuth × 1° elevation) and small size (<3″ × 3″). The design of the FSPA utilizes the phasing-induced frequency-dependent directionality of a linear phased array to produce multiple beams in a forward sector. The FSPA requires only two hardware channels per array and can be arranged in single and multiple array configurations that deliver wide sector 2-D images. 3-D images can be obtained by scanning the array in a direction perpendicular to the 2-D image field and applying suitable image processing to the multiple scanned 2-D images. This paper introduces the 3-D FSPA concept, theory and design methodology. Finally, results from a prototype array are presented and discussed.     

Performance-enhanced recognition of a far and partially occluded 3-D object by use of direct pixel-mapping in computational curving-effective integral imaging

In this paper, we propose a novel approach to enhance the recognition performance of a far and partially occluded three-dimensional (3-D) target in computational curving-effective integral imaging by using the direct pixel-mapping (DPM) method. With this scheme, the elemental image array (EIA) originally picked up from a far and partially occluded 3-D target can be converted into a new EIA just like the one virtually picked up from a target located close to the lenslet array. Due to this characteristic of DPM, resolution and quality of the reconstructed target image can be highly enhanced, which results in a significant improvement of recognition performance of a far 3-D object. In addition, the computational time required for reconstruction of a far 3-D target could be also reduced because the distance between the lenslet array and image plane is virtually shortened in the new EIA transformed by DPM. Experimental results reveal that image quality of the reconstructed target image and object recognition performance of the proposed system have been improved by 1.75 dB and 4.56% on the average in PSNR (peak-to-peak signal-to-noise ratio) and NCC (normalized correlation coefficient), respectively, compared to the conventional system.     

4-D image reconstruction for Streak Tube Imaging Lidar

The Streak Tube Imaging Lidar (STIL) which is the flash laser radar is developed in recently years. It can output the 4-D image (3-D range image + 1-D intensity image). The core part of STIL is the streak tube, which is a line array detector. STIL can directly collect the image of returned light pulse, and consequently, it can’t output the 4-D target image. In the paper, the peak detection as the reconstruction algorithm is used to reconstruct the multiple streak tube images, and the clear 4-D target images are obtained. The outdoor imaging experiments are completed for the far distance buildings. The reconstruction for the multiple streak tube images is finished, and the results state that the peak detection algorithm can output the high-resolution 4-D target images. Original Text ? Astro, Ltd., 2009. The article is published in the original.     

Face recognition based on PCA image reconstruction and LDA

Face recognition has become a research hotspot in the field of pattern recognition and artificial intelligence. Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) are two traditional methods in pattern recognition. In this paper, we propose a novel method based on PCA image reconstruction and LDA for face recognition. First, the inner-classes covariance matrix for feature extraction is used as generating matrix and then eigenvectors from each person is obtained, then we obtain the reconstructed images. Moreover, the residual images are computed by subtracting reconstructed images from original face images. Furthermore, the residual images are applied by LDA to obtain the coefficient matrices. Finally, the features are utilized to train and test SVMs for face recognition. The simulation experiments illustrate the effectivity of this method on the ORL face database.