基于GPU的球面深度图实时绘制

提出了一种GPU加速的实时基于图像的绘制算法.该算法利用极坐标系生成对物体全方位均匀采样的球面深度图像;然后根据推导的两个预变换公式将单幅球面深度图像预变换到物体包围球的一个与视点相关的切平面上,以生成中间图像;再利用纹理映射生成最终目标图像.利用现代图形硬件的可编程性和并行性,将预变换移植到Vertex Shader来加快绘制速度;利用硬件的光栅化功能来完成图像的插值,以得到连续无洞的结果图像.此外,还在Pixel Shader上进行逐像素的光照以及环境映射的计算,生成高质量的光照效果.最终,文章解决了算法的视点受限问题,并设计了一种动态LOD(Level of Details)算法,实现了一个实时漫游系统,保持了物体间正确的遮挡关系.

Improved digital processing method used for image motion measurement based on hybrid opto-digital joint transform correlator

Hybrid opto-digital joint transform correlator (HODJTC) is effective for image motion measurement, but it is different from the traditional joint transform correlator because it only has one optical transform and the joint power spectrum is directly input into a digital processing unit to compute the image shift. The local cross-correlation image can be directly obtained by adopting a local Fourier transform operator. After the pixel-level location of cross-correlation peak is initially obtained, the up-sampling technique is introduced to relocate the peak in even higher accuracy. With signal-to-noise ratio >= 20 dB, up-sampling factor k >= 10 and the maximum image shift <= 60 pixels, the root-mean-square error of motion measurement accuracy can be controlled below 0.05 pixels.

IMP 重离子治癌中的剂量计算方法

Basic algorithms of biological effective dose optimization and dose distribution on CT image for the heavy ion therapy project at the Institute of Modern Physics(IMP),Chinese Academy of Sciences(CAS) are reported in this paper.Firstly,biological effective dose optimization is conducted in water.According to the relationship between CT number and water equivalent path length,an integral algorithm is used to calculate the average dose within a pixel and then the dose distribution in tissue is derived.Secondly...中文文摘：针对深部肿瘤重离子治疗临床试验的需求,首先在水介质中进行生物有效剂量的优化计算,然后根据CT图像中像素CT值与水等效长度转换系数之间的关系,结合水中的深度剂量分布曲线对每个像素进行积分得到CT图像上的生物有效剂量分布。同时介绍了基于被动式束流配送系统适形照射时的剂量确定方式,并提出二维适形放疗也应使用分层照射方式以适应治疗时的不同要求。这些方法适合目前及今后在IMP进行的重离子治癌临床试验研究中治疗计划系统的需要。

A TOF-PET prototype with position sensitive PMT readout

A prototype of time-of-flight positron emission computed tomography (TOF-PET) has been developed for acquiring the coincident detection of 511 keV gamma-rays produced from positron annihilation. It consists of two 80.5 mmx80.5 mm LYSO scintillator arrays (composed of 35 x35 pixel finger crystals) with the position sensitive photomultiplier tubes R2487 as the readout. Each array is composed of 2 mm x2 mm x 15 mm finger crystals and the average pixel pitch is 2.30 mm. The measured results indicate that the TOF information has the potential to significantly enhance the image quality by improving the noise variance in the image reconstruction. The best spatial resolution (FWHM) of the prototype for the pairs of 511 keV gamma-rays is 1.98 mm and 2.16 mm in the x and y directions, respectively, which are smaller than the average pixel pitch of 2.30 mm.

Local histogram based geometric invariant image watermarking

Compared with other existing methods, the feature point-based image watermarking schemes can resist to global geometric attacks and local geometric attacks, especially cropping and random bending attacks (RBAs), by binding watermark synchronization with salient image characteristics. However, the watermark detection rate remains low in the current feature point-based watermarking schemes. The main reason is that both of feature point extraction and watermark embedding are more or less related to the pixel position, which is seriously distorted by the interpolation error and the shift problem during geometric attacks. In view of these facts, this paper proposes a geometrically robust image watermarking scheme based on local histogram. Our scheme mainly consists of three components: (1) feature points extraction and local circular regions (LCRs) construction are conducted by using Harris-Laplace detector; (2) a mechanism of grapy theoretical clustering-based feature selection is used to choose a set of non-overlapped LCRs, then geometrically invariant LCRs are completely formed through dominant orientation normalization; and (3) the histogram and mean statistically independent of the pixel position are calculated over the selected LCRs and utilized to embed watermarks. Experimental results demonstrate that the proposed scheme can provide sufficient robustness against geometric attacks as well as common image processing operations. (C) 2010 Elsevier B.V. All rights reserved.

A Unified Tensor Level Set for Image Segmentation

This paper presents a new region-based unified tensor level set model for image segmentation. This model introduces a three-order tensor to comprehensively depict features of pixels, e.g., gray value and the local geometrical features, such as orientation and gradient, and then, by defining a weighted distance, we generalized the representative region-based level set method from scalar to tensor. The proposed model has four main advantages compared with the traditional representative method as follows. First, involving the Gaussian filter bank, the model is robust against noise, particularly the salt-and pepper-type noise. Second, considering the local geometrical features, e. g., orientation and gradient, the model pays more attention to boundaries and makes the evolving curve stop more easily at the boundary location. Third, due to the unified tensor pixel representation representing the pixels, the model segments images more accurately and naturally. Fourth, based on a weighted distance definition, the model possesses the capacity to cope with data varying from scalar to vector, then to high-order tensor. We apply the proposed method to synthetic, medical, and natural images, and the result suggests that the proposed method is superior to the available representative region-based level set method.

Organic integrated pixels fabricated by a full evaporation method

An organic integrated pixel consisting of an organic light-emitting diode driven by an organic thin-film field-effect transistor (OTFT) was fabricated by a full evaporation method oil a transparent glass substrate. The OTFT was designed as a top-gate Structure, and the insulator is composed of a double-layer polymer of Nylon 6 and Teflon to lower the operation voltage and the gate-leakage current, and improve the device stability. The field-effect mobility of the OTFT is more than 0.5 cm(2) V-1 s(-1), and the on/off ratio is larger than 10(3). The brightness of the pixel reached as large as 300 cd m(-2) at a driving current of 50 mu A.

Effect of spatial variation on areal evapotranspiration simulation in Haibei, Tibet plateau, China

Quantification of areal evapotranspiration from remote sensing data requires the determination of surface energy balance components with support of field observations. Much attention should be given to spatial resolution sensitivity to the physics of surface heterogeneity. Using the Priestley-Taylor model, we generated evapotranspiration maps at several spatial resolutions for a heterogeneous area at Haibei, and validated the evapotranspiration maps with the flux tower data. The results suggested that the mean values for all evapotranspiration maps were quite similar but their standard deviations decreased with the coarsening of spatial resolution. When the resolution transcended about 480 m, the standard deviations drastically decreased, indicating a loss of spatial structure information of the original resolution evapotranspiration map. The absolute values of relative errors of the points for evapotranspiration maps showed a fluctuant trend as spatial resolution of input parameter data layers coarsening, and the absolute value of relative errors reached minimum when pixel size of map matched up to measuring scale of eddy covariance system. Finally, based on the analyses of the semi-variogram of the original resolution evapotranspiration map and the shapes of spatial autocorrelation indices of Moran and Geary for evapotranspiration maps at different resolutions, an appropriate resolution was suggested for the areal evapotranspiration simulation in this study area.

基于图像统计信息的去椒盐噪声算法

提出了一种基于图像统计信息的去椒盐噪声算法。该方法利用图像中心像素邻域的均值和方差来消除图像中椒盐噪声的影响。首先给出了这种算法的基本原理和实现,然后分别应用中值滤波算法、自适应中值滤波算法以及该算法对有椒盐噪声污染的图像进行滤波,并对实验结果进行比较和分析。实验结果表明,该方法不仅能够消除椒盐噪声,而且能够保持图像的特征信息。最后,对这种算法复杂度也进行了计算分析,并将其和中值滤波算法以及自适应中值滤波算法的复杂度进行了比较,证明该算法比中值滤波算法以及自适应中值滤波算法的复杂度低。

双向联想记忆神经网络的一种编码策略

本文提出一种双向联想记忆神经网络的按‘位’加权编码策略，并给出了求取权值的速推算法．它将Ｋｏｓｋｏ双向联想记忆神经网络按海明距离进行模式匹配的原则，修正为按加权海明距离进行模式匹配，从而可以使得对不满足连续性的所谓“病态结构”的一类样本模式集，同样具有良好的联想能力．对二值图象模式存贮、联想的计算机模拟实验表明，此方法具有优良的性能和实用价值。

一种高智能化的通用自动胶片判读仪

自动通用胶片判读仪是一种高智能化的精密光学测试设备,采用了计算机控制飞点扫描技术、精密光学测量技术、图象跟踪测量与信息处理技术。飞点管分辨率达4096×4096象元,通过光学系统胶片上获得6.55μm高分辨率,飞点扫描方式灵活多样且可随意控制,通用于目前我国靶场所有的电影经纬仪和高速摄影机35mm胶片的数据判读。具有自动判读和半自动判读两种工作模式,自动判读的速度为5帧/秒,自动判读的精度为σ=±0.011mm,半自动判读的精度为σ=±0.009mm,测量数据可以记录、打印和显示。

用光学阵列传感器的机器人物体分类系统

本文介绍用光学阵列传感器的机器人物体分类系统。传感器直接安装在机器人的两个手指上。被抓物体的阴影通过光导纤维传到安放在“安全区”的光敏元件上。计算机识别物体的轮廓后命令机器人抓握物体,并把它运送到指定的地点从而达到物体分类的目的。

图象噪声的模糊聚类平滑方法

模糊C-means算法在聚类分析中已得到了成功的应用,本文提出一种利用模糊C-means算法消除噪声的新方法。一般来说,图象中的噪声点就是其灰度值与其周围象素的灰度值之差超过某个门限值的点。根据这个事实,首先利用模糊C-means算法分类,再利用标准核函数检测出噪声点,然后将噪声点去掉。由于只修改噪声点处的象素灰度值,而对于其它象素的灰度值不予改变,所以本算法能够很好地保护细节和边缘。本方法每次处理3×3个点,而以往的方法只能每次处理一个点,所以本方法能提高运算速度。文中给出了利用本方法对实际图象处理的结果,并与梯度倒数权值法进行了定量的比较。

基于Zernike正交矩的图像亚像素边缘检测算法改进

介绍了Zernike矩及基于Zernike矩的图像亚像素边缘检测原理,针对Ghosal提出的基于Zernike矩的亚像素图像边缘检测算法检测出的图像存在边缘较粗及边缘亚像素定位精度低等不足,提出了一种改进算法.推导了7×7 Zernike矩模板系数,提出一种新的边缘判断依据.改进的算法能较好检测图像边缘并实现了较高的边缘定位.最后,设计了3组不同的实验.实验结果同Canny算子及Ghosal算法相比,证明了改进算法的优越性.

乡镇级人口统计数据空间化研究

Population data which collected and saved according to administrative region is a kind of statistical data. As a traditional method of spatial data expression, average distribution in every administrative region brings population data on a low spatial and temporal precision. Now, an accurate population data with high spatial resolution is becoming more and more important in regional planning, environment protection, policy making and rural-urban development. Spatial distribution of population data is becoming more important in GIS study area. In this article, the author reviewed the progress of research on spatial distribution of population. Under the support of GIS, correlative geographical theories and Grid data model, Remote Sensing data, terrain data, traffic data, river data, resident data, and social economic statistic were applied to calculate the spatial distribution of population in Fujian province, which includes following parts: (1) Simulating of boundary at township level. Based on access cost index, land use data, traffic data, river data, DEM, and correlative social economic statistic data, the access cost surface in study area was constructed. Supported by the lowest cost path query and weighted Voronoi diagram, DVT model (Demarcation of Villages and Towns) was established to simulate the boundary at township level in Fujian province. (2) Modeling of population spatial distribution. Based on the knowledge in geography, seven impact factors, such as land use, altitude, slope, residential area, railway, road, and river were chosen as the parameters in this study. Under the support of GIS, the relations of population distribution to these impact factors were analyzed quantificationally, and the coefficients of population density on pixel scale were calculated. Last, the model of population spatial distribution at township level was established through multiplicative fusion of population density coefficients and simulated boundary of towns. (3) Error test and analysis of population spatial distribution base on modeling. The author not only analyzed the numerical character of modeling error, but also its spatial distribution. The reasons of error were discussed.