Description: ************************************************************************ * * * * * THIS IS THE H Y P L A S 2.0 README FILE * * ----------------- * * * * HYPLAS is a finite element program for implicit small and large * * strain analisys of hyperelastic and elasto-plastic two-dimensional * * and axisymmetric solids * * * * HYPLAS v2.0 is the companion software to the textbook: * * EA de Souza Neto, D Peric & DRJ Owen. Computational Methods for * * Plasticity: Theory and Applications. Wiley, Chichester, 2008. * * (www.wiley.com/go/desouzaneto) * * * * Copyright (c) 1998-2008 EA de Souza Neto, D Peric, D.R.J. Owen * *----------------------------------------------------------------------* * File last updated: 18 October 2008 * * * * This file belongs in the directory ../HYPLAS_v2.0 * ************************************************************************ * * * I M P O R T A N T * * * * READ SECTIONS 0 TO 3 OF THIS FILE CAREFULLY BEFORE ATTEMPTING * * TO COMPILE AND RUN THE PROGRAM HYPLAS ON YOUR COMPUTER !! * * * * THE AUTHORS DO NOT GUARANTEE THAT ANY SUGGESTIONS/INSTRUCTIONS * * GIVEN IN THIS README FILE WILL WORK ON ANY PARTICULAR OPERATING * * SYSTEM. IF YOU DECIDE TO FOLLOW ANY SUGGESTIONS/INSTRUCTIONS * * GIVEN HERE YOU MUST DO SO AT YOUR OWN RISK. * * * * * * BUG REPORTS: Please send bug reports to * * * * hyplas_v2.0@live.co.uk * * * * Messages sent to the authors' personal email addresses * * will NOT be answered. * ************************************************************************ This file contains the following sections: 0. Copyright statement and disclaimer 0.(a) Copyright statement 0.(b) Disclaimer 0.(c) Conditions of use 1. Introduction 1.(a) Note on portability 2. Compiling and running HYPLAS 2.(a) Memory requirements 2.(b) Testing a newly compiled executable 3. The HYPLAS directory tree 4. Cross-referencing between the source code and the textbook 5. HYPLAS error messaging 6. Further remarks on HYPLAS ************************************************************************ 0. COPYRIGHT STATEMENT AND DISCLAIMER ================================== 0.(a) Copyright statement ------------------- You may only use this program for your own private purposes. You are not allowed, in any circumstances, to distribute this program (including its source code, executable and any other files related to it, either in their original version or any modifications introduced by you, the authors or any other party) in whole or in part, either freely or otherwise, in any medium, without the prior written consent of the copyright holders. 0.(b) Disclaimer ---------- This program (including its source code, executable and any other files related to it) is provided "as is" without warranty of any kind, either expressed or implied, including, but not limited to, any implied warranties of fitness for purpose. In particular, THIS PROGRAM IS BY NO MEANS GUARANTEED TO BE FREE FROM ERRORS. This program (or any modification incorporated to it by you, the authors or any other party) will run entirely at your risk. The results produced by this program are in no way guaranteed to be fit for any purpose. Under no circumstances will the authors/copyright holders be liable to anyone for damages, including any general, special, incidental or consequential damages arising from the use or inability to use the program (including, but not limited to, loss or corruption of data, failure of the program to operate in any particular way as well as damages arising from the use of any results produced by the program for any purpose). 0.(c) Conditions of use ----------------- You may only use this program if you fully understand and agree with the terms of the above disclaimer. You must not use this program if you do not agree with or do not understand (fully or in part) these conditions of use. 1. INTRODUCTION ============ HYPLAS is a finite element code for small and large strain analysis of hyperelastic and elasto-plastic solids. Most procedures implemented in HYPLAS are described in detail in its companion textbook: EA de Souza Neto, D Peric & DRJ Owen. Computational Methods for Plasticity: Theory and Applications. Wiley, Chichester, 2008 (www.wiley.com/go/desouzaneto). 1.(a) Note on Portability ------------------- HYPLAS has been written in standard ANSI FORTRAN 77. Currently, the only known (and deliberate) exceptions to the FORTRAN 77 ANSI standard are the instructions: INCLUDE '' used in many routines to include the HYPLAS database files (common blocks and global variables), and; CALL GETENV('HYPLASHOME',HYPLASHOME) used in subroutine "ERRPRT" (file ../HYPLAS_v2.0/src/GENERAL/errprt.f). This instruction inquires the name of the system environment variable HYPLASHOME and writes it on the character string HYPLASHOME. This instruction is NOT part of the ANSI FORTRAN 77 standard, but seems to work in most currently available FORTRAN 77 compilers. 2. COMPILING AND RUNNING H Y P L A S ================================== The HYPLAS source code is stored in directory ../HYPLAS_v2.0/src/ (../HYPLAS_v2.0/ being the current directory) and all its subdirectories. To generate an executable file, you just need to compile the FORTRAN source files: ../HYPLAS_v2.0/src/hyplas.f and ../HYPLAS_v2.0/src/*/*.f together. We recommend that the executable HYPLAS be stored in the directory ../HYPLAS_v2.0/bin to which the environment variable HYPLASHOME should be set (see below how to set a system environmental variable). WINDOWS (R) systems ------------------- On Microsoft Windows(R) systems, HYPLAS has been successfully compiled using Intel Visual Fortran Compiler(R) integrated with Microsoft Visual Studio(R). Here you only need to create a project that contains all Fortran source files mentioned above as well as the include files ..\HYPLAS_v2.0\src\*.INC On a Windows XP system, the system environment variable HYPLASHOME can be set as follows: 1. Open a File Manager 2. Right-click on the "My Computer" icon 3. Select "Properties" on the drop-down menu 4. A new window named "System Properties" will pop-up. Here select the "Advanced" tab. 5. On the "Advanced" tab, click the "Environment Variables" button. 6. A new window titled "Environment Variables" will pop-up. Here click the button "New" in the "System Variables" section of the window. 7. A new window will pop-up titled "New System Variable". Here you should fill the fields "Variable name" and "Variable Value", respectively, with HYPLASHOME and the path name (in full) of the directory ..\HYPLAS_v2.0\bin. 8. Press "OK" on the relevant pop-up windows. 9. The next time the computer is REBOOTED, this variable will be set to the correct path and HYPLAS should be able to find the error messages file ERROR.RUN if required. UNIX/LINUX systems ------------------ In a UNIX/LINUX operating system using a C-shell, for instance, the HYPLASHOME environment variable should be set with the command: setenv HYPLASHOME where here denotes the full path to the directory ../HYPLAS_v2.0/bin. To compile HYPLAS (from directory ../HYPLAS_v2.0/src) with a FORTRAN 77 compiler such as g77, you can use the command: g77 -o ../bin/hyplas hyplas.f */*.f Note that the executable file "hyplas" will be stored in the directory ../HYPLAS_2.0/bin (i.e. the directory set in the HYPLASHOME environment variable). Alternatively, you may use the Makefile provided (with suitable modifications, if needed) to create the HYPLAS executable. IMPORTANT: Before generating a HYPLAS executable, read Sections 2.(a) and 2.(b) below. 2.(a) Memory Requirements ------------------- HYPLAS memory requirements depend on the array dimensioning parameters set in files: ../HYPLAS_v2.0/src/ ELEMENTS.INC GLBDBASE.INC MATERIAL.INC MAXDIM.INC Files ELEMENTS.INC, GLBDBASE.INC and MATERIAL.INC contain parameters which are associated with the currently implemented finite elements and materials. DO NOT MODIFY THEM ! unless you are absolutely sure of what you are doing (only developers coding new elements or new material models/analysis types may need to modify them by changing the existing dimensioning parameters and/or including new parameters). The ONLY dimensioning file that can be safely modified by the average user is the file MAXDIM.INC This file contains the array dimensioning parameters related to the maximum permissible dimension of problems to be analysed by HYPLAS. These parameters include the maximum number of nodes, elements, element groups, etc. If necessary, CHANGE THESE PARAMETERS TO SUIT YOUR PROBLEM SIZE/MEMORY REQUIREMENTS before compiling HYPLAS. 2.(b) Testing a newly compiled executable ----------------------------------- After you have successfully compiled the HYPLAS source code and created an executable file, the next step is to run some tests to verify that HYPLAS is working well. To do this, proceed as follows: The directory ../HYPLAS_v2.0/book_examples/data_files contains a series of data files named .dat of benchmarked examples described in the companion textbook. The corresponding (benchmarked) result files are in the directory ../HYPLAS_v2.0/book_examples/result_files This directory contains a series of result files named .res generated with the current version of HYPLAS on a tested platform. All these files have been named such that their names start with the textbook section number where the corresponding example is described. For instance, files 14_9_2_tresca.dat and 14_9_2_tresca.res refer to a problem described in section 14.9.2 of the textbook, and so on. To check that HYPLAS is working well on your platform, after compiling HYPLAS, run the program HYPLAS for the examples of files .dat and compare the newly generated results .res with their benchmarked counterparts (of the same filename) in the result_files directory. To run an example, execute HYPLAS and use the keyboard to enter the name of the corresponding data file in full (including the extension .dat). To compare the benchmarked .res files against their newly generated you may proceed as follows: 1. On MICROSOFT WINDOWS systems - Here we have successfully used the software "ExamDiff" (the task was made particularly easy by selecting "View" and then the "Show Differences Only" option - this refers to version 1.8 of this software). 2. On UNIX/LINUX systems - Here we use the "diff" command from a shell window (and set the option to ignore blank spaces). A shell script may be used to perform this task automatically (including running HYPLAS and checking for result file differences) for all benchmarked examples provided. IMPORTANT: THE ONLY ACCEPTABLE DIFFERENCES BETWEEN A THE NEWLY GENERATED RESULT FILES AND THEIR BENCHMARKED COUNTERPARTS ARE THE DIMENSIONING PARAMETERS (FROM FILE MAXDIM.INC) USED TO COMPILE THE NEW EXECUTABLE (THESE PARAMETERS ARE PRINTED RIGHT AT THE BEGINNING OF THE RESULT FILES) AND NUMERICAL DIFFERENCES IN RESULTS DUE TO NUMERICAL "ROUNDING-OFF" (THESE ARE VERY SMALL DIFFERENCES THAT DEPEND ON THE PRECISION OF ARITHMETIC OPERATIONS IN THE PLATFORM USED). ALSO NOTE THAT THE EXAMPLES OF THE COMPANION TEXTBOOK DO NOT COVER ALL FEATURES OF HYPLAS. HENCE THIS TEST DOES NOT GUARANTEE THAT EVERYTHING IS WORKING PROPERLY. 3. THE H Y P L A S DIRECTORY TREE ================================ 3.(a) Summary ------- ../ HYPLAS_v2.0/ bin/ book_examples/ data_files/ result_files/ man/ html/ src/ CRYSTAL/ DAMAGE/ DAMAGED_ELASTIC/ DRUCKER_PRAGER/ ELASTIC/ ELEMENTS/ GENERAL/ MATERIALS/ MATHS/ MOHR_COULOMB/ OGDEN/ TRESCA/ VON_MISES/ VON_MISES_MIXED/ 3.(b) Description ----------- The HYPLAS program directory tree is organised as follows: ../HYPLAS_v2.0/ (this directory) This is the HYPLAS root directory, where the HYPLAS directory tree starts. ../HYPLAS_v2.0/bin/ This directory contains the file ERROR.RUN where most HYPLAS error/warning messages are. IMPORTANT: the environment variable HYPLASHOME should be set to this directory. Otherwise, HYPLAS will not find its error/warning messages when required. We also recommend that the EXECUTABLE of HYPLAS be stored in this directory. ../HYPLAS_v2.0/book_examples/ This directory has the following subdirectories: ../HYPLAS_v2.0/book_examples/data_files ../HYPLAS_v2.0/book_examples/result_files Refer to Section 2.(b) above for further details. ../HYPLAS_v2.0/man/ This is the HYPLAS documentation/manuals directory. It contains the following files: input_man.txt - A concise input data manual for HYPLAS in ASCII format; hyplas_calltree.txt - Contains a flowgraph (shows the call tree) of HYPLAS in ASCII-format. Note: calls to function subprograms are not included in this flowgraph; and the subdirectory: ../HYPLAS_v2.0/man/html This directory contains the hypertext (HTML) format Fortran source code and of manual pages of the entire HYPLAS program. Manual pages with descriptions of each function/subprogram including their argument list are linked to their corresponding HTML-format source code. This allows the user the navigate through the HYPLAS source code using a web browser. To start at the main program, use your web browser to open the file hyplas.html. This facility should be helpful to those trying to understand the flow of program HYPLAS. ../HYPLAS_v2.0/src/ This directory (and its subdirectories) contains the Fortran source code of HYPLAS. The files containing the sources are named following the standard practice: .f where is the name of the FORTRAN procedure (subroutine, function subprogram, etc.) whose source code is in file .f. The source code of the HYPLAS main program is in file hyplas.f and the HYPLAS database (COMMON blocks, array dimensioning parameters and other global parameters) is coded in the "include files" ELEMENTS.INC GLDBASE.INC MATERIAL.INC MAXDIM.INC in this directory. In addition, this directory contains a file named "Makefile" (UNIX-LINUX Release only) which may be used for compiling and linking HYPLAS in UNIX/LINUX systems. The subdirectories of ../HYPLAS_v2.0/src are as follows: ../HYPLAS_v2.0/src/CRYSTAL Contains the source code of all procedures related to the finite strain single crystal plasticity model implemented in HYPLAS. ../HYPLAS_v2.0/src/DAMAGE Source files of the procedures related to the Lemaitre ductile damage model implementation. ../HYPLAS_v2.0/src/DAMAGED_ELASTIC Source files of the procedures related to the damaged elasticity model with crack closure effect. ../HYPLAS_v2.0/src/DRUCKER_PRAGER Source files of the procedures related to the implemented Drucker-Prager plasticity model. ../HYPLAS_v2.0/src/ELASTIC Source files of the procedures related to the linear elasticity model (Hencky model under large strains) implemented. ../HYPLAS_v2.0/src/ELEMENTS Source files of the element interfaces and element-related procedures. ../HYPLAS_v2.0/src/GENERAL Source files of general procedures. ../HYPLAS_v2.0/src/MATERIALS Source files of the material interfaces. ../HYPLAS_v2.0/src/MATHS Source files of the mathematical procedures. ../HYPLAS_v2.0/src/MOHR_COULOMB Source files of the procedures related to the implemented Mohr-Coulomb plasticity model. ../HYPLAS_v2.0/src/OGDEN Source files of the procedures related to the implemented Ogden hyperelasticity model. ../HYPLAS_v2.0/src/TRESCA Source files of the procedures related to the implemented Tresca plasticity model. ../HYPLAS_v2.0/src/VON_MISES Source files of the procedures related to the implemented von Mises plasticity model with isotropic hardening. ../HYPLAS_v2.0/src/VON_MISES_MIXED Source files of the procedures related to the implemented von Mises plasticity model with mixed isotropic/kinematic hardening. 4. CROSS-REFERENCING BETWEEN THE SOURCE CODE AND THE TEXTBOOK ========================================================== Many references are made in the textbook to various subprograms of HYPLAS. These are usually made when a particular procedure described in the text is implemented in the program. The reader should refer to the textbook index. Also, a substantial number of comment lines have been added to the source code of HYPLAS with reference to sections, figures, boxes, etc of the textbook related to the part of the code in question. Such references are usually displayed after the word "REFERENCE:" (in capitals) on commented lines. Searching for this word will take you to the line of code where the particular routine has a reference to the textbook. NOTE: Occasional references to other textbooks/journal papers are also made following the word "REFERENCE:" on commented lines. 5. HYPLAS ERROR MESSAGING ====================== Most error/warning messages issued by HYPLAS are in the ASCII-format file ERROR.RUN (kept in the HYPLASHOME directory - ../HYPLAS_v2.0/bin). All such error/warning messages have an identification code (e.g. ED0015) which is printed both to the standard output (this is usually the computer screen) and to the relevant results file. If you wish to find where in the source code a particular message is being issued, then perform a search for the corresponding message identification code in the entire source code of HYPLAS. 6. FURTHER REMARKS ON HYPLAS ========================= 6.(a) Program efficiency THIS SECTION IS OF INTEREST ONLY TO THOSE WANTING TO MAKE HYPLAS RUN FASTER. It is particularly stressed in the textbook that this program has not been designed having efficiency in mind (refer to Section 5.1.2 of the textbook). Its structure has been designed mainly to illustrate in a relatively clear manner the computer implementation of the techniques and algorithms described in the text, with a particular view to the implementation of solid constitutive models and finite elements. For those who are especially interested in the speed of the code, there are a few tips that could help in this direction. Unfortunately, these involve modifications to the source code which is probably most appropriate to readers with a good level of experience in finite element programming. To those with this particular interest, we can suggest the following: (i) The use of faster linear solvers This is probably the change that would result in a greater gain in efficiency. The Frontal Method adopted in subroutine FRONT (file ../HYPLAS_v2.0/src/GENERAL/front.f) has been designed originally to save memory (back in the days when computer memory was severely limited). There are currently a vast number of methodologies which focus on speeding up the linear solution, in addition to reducing memory storage requirements (which is a particularly important issue in the solution of large scale problems). Some of these are extensions/refinements of the original Frontal solver. We remark that a number of such procedures (with their respective source codes) are available (conditions may apply) from the LAPACK (Linear Algebra PACKage - http://www.netlib.org/lapack) repository or from the HSL Library (http://www.cse.cse.scitech.ac.uk/nag/hsl). For the reader interested in gaining speed, we would recommend the replacement of the existing solver of FRONT by a faster one. We remark though that this is a substantial programming task. Another aspect here is the fact that computing times in FRONT are directly linked to the frontwidth of the system which, in the present version of HYPLAS is fixed and depends, for a given mesh, on how the degrees of freedom are numbered (node numbering). The incorporation of a frontwidth optimiser (which re-numbers the degrees of freedom in order to minimise the frontwidth) in FRONT could produce some good savings in computing times. Such savings become particularly noticeable in larger problems where the original node numbering produces an excessively large frontwidth. (ii) Material-specific computations The issues pointed out here affect only the computing times for specific material models and are expected to have a much lower impact in overall speed than the linear solver issue discussed above. Some of the material model-specific computations carried out in HYPLAS could be made a bit faster. For example, for isotropic models whose stress update is carried out in the principal stress space (such as the Tresca and Mohr-Coulomb models - see routines SUTR and SUMC, files ../HYPLAS_v2.0/src/TRESCA/sutr.f and ../HYPLAS_v2.0/MOHR_COULOMB/sumc.f, respectively) the spectral decomposition of the stress in carried out in the state update update routine and then repeated in the corresponding routine for computation of the consistent tangent operator (refer to files ../HYPLAS_v2.0/src/TRESCA/cttr.f and ../HYPLAS_v2.0/src/MOHR_COULOMB/ctmc.f, respectively, for the Tresca and Mohr-Coulomb plasticity models). Some savings in computing time can be achieved here by storing the stress eigenprojection tensors (these can be stored as state variables) during the execution of the state updating and then retrieving them later for use in the computation of the consistent tangent operator. This change can be incorporated to the code relatively easily. The computation of the exponential map and is derivative for the single crystal plasticity model (routines EXPMAP, file ../HYPLAS_v2.0/src/CRYSTAL/expmap.f and DEXPMP, file ../HYPLAS_v2.0/src/CRYSTAL/dexpmp.f) is carried out in three dimensions (these routines have been adapted from an earlier three-dimensional code). To improve efficiency, these can be adapted to work only in two-dimensional problems by removing the unnecessary operations related to the third dimension. 6.(b) Output of nodal averaged values The reader should be aware that the way in which nodal averaged values of stresses and other variables are calculated in HYPLAS is very basic (and rudimentary). This feature of the program is made available only to help those interested in producing contour plots, etc from results presented in HYPLAS result files and should be useful in many circumstances of interest. This facility has in fact been used in producing many of the figures presented in the textbook. But note, for example, that the values of incremental plastic multipliers for plasticity models may take (inadmissible) negative values when extrapolated from Gauss-point to nodes and averaged. We remark that more sophisticated and refined techniques of transferring Gauss point values of variables to nodal points and obtaining the corresponding smoothed field are available in the current literature. These fall outside the scope of the companion textbook of HYPLAS. Platform: |
Size: 11008084 |
Author:gtcewli3 |
Hits:
Description: 科恩著“时_频分析:理论与应用”该书全面系统地阐述了时-频分析的理论与应用。全书共分19章，主要内容有：时间分析、频率分析、尺度分析、时间-带宽关系；瞬时频率；密度和局部量；短-时傅里叶变换；时-频分析；魏格纳分布；时-频表示；计算方法；综合问题；空间/空间-频率表示；时间尺度表示；算子；一般联合表示；随机信号和高阶时-频分布。每一个概念都有举例说明，而且还给出了这些方法如何推广到其它的变量，如尺度等。
西安交通大学出版社-Cohen book when _-frequency analysis: Theory and Applications The book comprehensively and systematically elaborated on when- frequency analysis of the theory and application. Book is divided into 19 chapters, the main elements are: time analysis, frequency analysis, scale analysis, the time- bandwidth relations instantaneous frequency density and local capacity short- when the Fourier transform time- frequency analysis Wigner distribution time- frequency representation calculation method integrated problem space/space- the frequency of express time scales express operator general joint express random signal and high-order time- frequency distribution. Each have examples of a concept, but also how these methods are given to other variables, such as the scale and so on. Xi an Jiaotong University Press Platform: |
Size: 6445056 |
Author:宋知用 |
Hits:
Description: 在使用扩散过程平滑噪声之后引入反扩散过程来恢复边缘，结合尺度空间理论和反扩散函数对图像进行去噪处理。该方法使用最小描述长度(MDI )准则自适应地选择图像中每一点处的最优尺度对图像进行滤波。加入尺度范围限制降低了过平滑和欠平滑的影响。改进了反扩散函数模型，对降质图像中的边缘进行恢复。与经典的滤波方法以及各向异性扩散方程的结果相比。本文方法取得了较好的效果。-Diffusion process in the use of noise after smoothing the introduction of anti-diffusion process to restore the edge, combined with scale space theory and the anti-spread function of image denoising treatment. This method is the use of minimum description length (MDI) criteria for adaptive selection of images every point in the Agency Platform: |
Size: 225280 |
Author:liqiangqiang |
Hits:
Description: IOI国家集训队2000论文集
包括：
陈彧：《信息学竞赛中的思维方法》
方奇：《动态规划》
高寒蕊：《递推关系的建立及在信息学竞赛中的应用》
郭一：《数学模型及其在信息学竞赛中的应用》
江鹏：《探索构造法解题模式》
李刚：《动态规划的深入讨论》
龙翀：《解决空间规模问题的几种常用的存储结构》
骆骥：《数学模型的建立和选择》
施遥：《人工智能在围棋程序中的应用》
肖洲：《数据结构的在程序设计中的应用》
谢婧：《规模化问题的解题策略》
徐串：《论程序的调试技巧》
徐静：《图论模型的建立与转化》
杨江明：《论数学策略在信息学问题中的应用》
杨培：《非最优化算法初探》
张辰：《动态规划的特点及其应用》
张力：《类比思想在解题中的应用》
张一飞：《冗繁削尽留清瘦——浅谈信息的充分利用》
-IOI National Team 2000 Proceedings
Include:
Chen Yu: "Information science contest in the way of thinking"
Qi Fang: "dynamic programming"
Alpine Rui: "The establishment of recurrence relations and the application of informatics competition"
Guo: "Mathematical Model and Its Application in the application of informatics competition"
Chiang Peng: "explore the construction method for problem solving mode"
Li Gang: "Dynamic Programming in-depth discussion"
Long Chong: "to solve the problem of space the size of several commonly used storage structure"
Luo Ji: "Mathematical model and select"
Shi Yao: "Artificial Intelligence Go program in the application"
Zhou Xiao: "data structures in the program design of the application"
Xie Jing: "Solving the problem of scale strategy"
Xu string: "On Debugging Skills"
Xu: "Graph Theory model and Transformation"
YANG Jiang-ming: "On the mathematical strategy issues in the information in the application"
YANG Pei: "Non-optimization algorith Platform: |
Size: 2275328 |
Author:Jesse_Luo |
Hits:
Description: 本文运用尺度空间理论检测人体，通过集成
面向梯度与histogramof尺度空间理论
-Human detection is the task of finding presence and position
of human beings in images. In this paper, we apply
scale space theory to detecting human in still images. By integrating
scale space theory with histogramof oriented gradients
(HOG), we designed a new feature descriptor called
scale space histogram of oriented gradients (SS-HOG). SSHOG
focus on the multiple scale property of describe an
object. Using HOGs at multiple scale, SS-HOG encodes
more information to discriminate human bodies from other
object types than traditional uni-scale HOGs. Experiments
on INRIA person dataset demonstrate the effectiveness of
our method. Platform: |
Size: 142336 |
Author:谷川 |
Hits:
Description: 包含了5篇Tony Lindeberg 尺度空间理论经典著作，希望大家能够受益-Contains five Tony Lindeberg Scale Space Theory classics, I hope we can benefit from Platform: |
Size: 4631552 |
Author:高明亮 |
Hits:
Description: 本项目将多尺度理论、动态规划法、模式识别和区域划分等技术应用到图像的立体匹配之中，减少立体匹配过程中的搜索空间、计算时间和匹配错误；-This project will multi-scale theory, the dynamic programming, pattern recognition and regional division technology application of image matching of three-dimensional, reduce the search space stereo matching process, and computing time and matching error
Platform: |
Size: 73728 |
Author:蒋星星 |
Hits:
Description: 在基于Mean-Shift的目标跟踪算法中，尺度自适应机制是算法研究的一个重要方向。一种典型的方法采
用Lindeberg的尺度空间理论以获取目标尺度信息。但现有算法中将尺度由2．D矢量压缩为1一D量，未能精细地
刻画目标仿射变换时的尺度变化，从而限制了算法的适用范围。为此，该文将尺度维1．D滤波推广至2一D，构造得
到了相应的垂D尺度空间，并利用窄间维和尺度维的Mean-Shift交替迭代，实现了同时在空间位置和尺度方向对
目标的有效跟踪，提高了算法在目标尺度变化时的自适应性，并扩大了算法的适用范围。
关键-Scale adaptive mechanism in the target tracking algorithm based on Mean-Shift algorithm is an important direction of research. A typical method using Lindeberg' s scale space theory to obtain the target scale. However, the existing algorithms will scale from 2. D vector compression for 1-D volume, not fine to portray the target affine transformation of scale changes, thus limiting the scope of application of the algorithm. To this end, the paper will scale dimension. D filter to the 2-D structure has been the vertical D scale space and the use of narrow inter-dimension and scale dimension of the Mean-Shift alternating iteration, effective tracking of the target position in space and scale direction to improve the self-adaptive algorithm when the target scale changes, and to expand the scope of application of the algorithm. Key Platform: |
Size: 536576 |
Author:张娜 |
Hits:
Description: 根据linderburg的尺度空间理论提取了尺度不变特征，包含图片，能够运行。-Extracted under linderburg scale space theory scale invariant feature, including pictures, be able to run. Platform: |
Size: 1391616 |
Author:黄少 |
Hits:
Description: SIFT 由特征提取，特征描述符描述和特征匹配 3 部分构成，该算子特征提取数目庞大，建立特征描述符运算
量高，导致算法效率低。提出了一种 SEC( SIFT-Edge-Corner) 算法，在图像尺度空间提取角点代替 SIFT 特征点，并根
据角点是边缘曲率极值理论，预先采用 Canny 算子得到高斯边缘图像金字塔，再提取角点并进行尺度选择。实验结
果表明: 该算法在保障高准确率的前提下大幅度提高特征提取效率-By the SIFT feature extraction, feature descriptions and feature matching descriptors 3 parts, the large number of feature extraction operator established feature descriptor computation high, resulting in low efficiency of the algorithm. Presents a SEC (SIFT-Edge-Corner) algorithm, the image scale space instead of SIFT feature extraction corner points and corner points based on extreme value theory is an edge curvature in advance using Canny operator edge image obtained Gaussian pyramid, and then extract corner point and scale selection. Experimental results show that: the algorithm protect high accuracy under the premise of feature extraction efficiency greatly improved Platform: |
Size: 231424 |
Author:焦婷 |
Hits:
Description: 尺度空间相关资料及代码(OpenCV编程中的资料和代码）-Scale space-related data and codeScale space-related data and code (OpenCV programming data and code) Platform: |
Size: 1351680 |
Author:梅栋 |
Hits:
Description: 提出一种多尺度理论与无味卡尔曼滤波
器(UKF)相结合的视频跟踪改进算法。利用多尺度理论统计跟踪窗内的信息量，使用 UKF 对得到的信息量进行预测，通过修正后的信息量
计算窗口变化比例系数，对尺度任意变化的目标进行跟踪。-An novel
method is proposed that is multi-scale space theory combined with Unscented Kalman Filter(UKF). UKF filter is introduced to predict the
information in the tracking window which is calculated by the multi-scale space theory. The proportion of the target image area is got by the
modified information. It is implemented by the combination of the Mean-Shift tracking algorithm and UKF to track targets. Platform: |
Size: 1158144 |
Author:小于 |
Hits:
Description: 为在图像处理、计算器视觉、信号处理上所使用的一项技术。 高斯金字塔本质上为信号的多尺度表示法，亦即将同一信号或图片多次的进行高斯模糊，并且向下取样， 藉以产生不同尺度下的多组信号或图片以进行后续的处理，例如在影像辨识上，可以藉由比对不同尺度下的图片，以防止要寻找的内容可能在图片上有不同的大小。 高斯金字塔的理论基础为尺度空间理论，而后续也衍生出了多分辨率分析。-
As a technique in image processing, computer vision, signal processing is used. Gaussian pyramid is essentially a multi-scale representation of the signal, that signal will be the same picture multiple times or Gaussian blur, and down sampling, in order to produce multiple sets of different scales for subsequent signal or image processing, for example in the image the identification, can be compared to the picture at different scales, in order to prevent the contents may have to look for a different size in the picture. The theoretical basis for the scale-space Gaussian pyramid theory, but also spawned a follow-up multi-resolution analysis. Platform: |
Size: 244736 |
Author:robin lee |
Hits:
Description: Meanshift跟踪算法针对传统的MeanShift跟踪算法在目标发生遮挡时容易导致目标丢失的情况,提出了一种改进的MeanShift跟踪算法.将多尺度空间理论、Kalman滤波器与遮挡算法相结合,当目标发生遮挡后,利用Kalman估计目标信息量,能对目标尺寸有后续跟踪能力.实验结果表明,当目标发生遮挡后,改进的跟踪算法对目标无论增大或减小都能连续地、自动地选择大小合适的跟踪窗口.-Meanshift tracking algorithmMeanshift tracking algorithm for traditional Meanshift in target tracking algorithm when the shade is easy to cause the condition of the missing target, an improved Meanshift tracking algorithm is proposed. The multi-scale space theory, combined with block algorithm, Kalman filter the shade when the target, the target information is estimated based on the Kalman, follow-up tracking ability to target size. The experiment results show that when the target after shade, the improved algorithm of target tracking can both increase/decrease continuously and automatically appropriate to the size of the tracking window. Platform: |
Size: 188416 |
Author:樊梦丹 |
Hits:
Description: 被誉为大自然的几何学的分形( fractal) 理论是现代数学的一个新分支。分形的概念最早是由美籍数学家曼德布罗特( B. B. Mandelbort) 提出 ， 分形是指部分与整体以某种方式相似的形体。自相似原则和迭代生成原则是分形理论的重要原则。它表征分形在通常的几何变换下具有不变性， 即标度无关性。分形维数是分形理论中的主要参数， 分形维数反映了复杂形体占有空间的有效性， 它是复杂形体不规则性的量度。(The fractal (fractal) theory, which is known as the geometry of nature, is a new branch of modern mathematics. The concept of fractal was first proposed by the American mathematician, Mandelbrot (B. B. Mandelbort), which refers to a form that is similar in some way to the whole. The principle of self similarity and the principle of iterative generation are the important principles of the fractal theory. It characterizes the invariance of the fractal under the usual geometric transformation, namely the scale independence. Fractal dimension is the main parameter in fractal theory. Fractal dimension reflects the validity of the occupied space of complex form, and it is the measurement of the irregularity of complex shape.) Platform: |
Size: 7168 |
Author:张子晨55 |
Hits:
Description: 基于计算机视觉所提出的尺度空间融合理论的空间聚类方法(Spatial clustering method based on the theory of scale space fusion proposed by computer vision) Platform: |
Size: 38912 |
Author:cy0225 |
Hits:
Search - Scale Space Theory