Categories
BLOG

dice strain

Introduction

DICe is an open source digital image correlation (DIC) tool intended for use as a module in an external application or as a standalone analysis code. Its primary capability is computing full-field displacements and strains from sequences of digital images. These images are typically of a material sample undergoing a characterization experiment, but DICe is also useful for other applications (for example, trajectory tracking and object classification). DICe is machine portable (Windows, Linux and Mac) and can be effectively deployed on high performance computing platforms (DICe uses MPI parallelism as well as threaded on-core parallelism). Capabilities from DICe can be invoked through a customized library interface, via source code integration of DICe classes or through a standalone executable from the command line.

DIC, in general, has become a popular means of determining full-field displacements from digital images, it has also become a vital component in material characterization applications that use full-field information as part of a parameter inversion process. DIC is also used extensively for constitutive model development and validation as well as physics code validation. DICe aims to enable more seamless integration of DIC in these types of applications by providing a DIC tool that can be directly incorporated in an external application.

DICe is different than other DIC codes in the following ways: First, subsets can be of arbitrary shape. This enables tracking of oblong objects that otherwise would not be trackable with a square subset. DICe also incudes a robust simplex optimization method that does not use image gradients (this method is useful for data sets that are impossible to analyze with the traditional Lucas-Kanade-type algorithms, for example, objects without speckles, images with low contrast, and small subset sizes [email protected] , phone (505) 845-7446.

Getting Started

There are three modes in which DICe can be used.

  • As a standalone executable
  • As integrated code in an external application by static linking to DICe
  • As a dynalically linked library

DICe standalone use

To use DICe in standalone mode (assuming the dice executable is built and in the system path), the user simply has to write an input file and invoke DICe with

The input file is an .xml formated set of parameters. To generate a template set of input files (which includes the input file and the correaltion paramters file, described below) add the -g option to the dice call.

The sample files will not be written if the folder permissions are read-only (as is the case if one executes dice inside the system install directory).

Commented notes on all of the paramters in the input files are given with the -g option in the template files. The input file specifies, the location of the images and results and also the images to use. The user can select a correlation of a list of specific images or a sequence of images based on which paramters are specified.

If MPI is installed and enabled (see below) DICe can be run in parallel with

Where num_procs specifies the number of processors.

Here is a helpful link for using MS-MPI on windows: Microsoft MPI. MPI must be installed to run in parallel.

Running DICe in parallel with MPI enabled

To run DICe in parallel with MPI enabled, Trilinos must be installed with MPI enabled by setting the approprate flag in the trilinos CMake script

When DICe is configured before building, CMake will default to using the same compilers that were used to build Trilinos. Nothing extra has to be done in the DICe CMake scripts to enable MPI parallelism.

There are three ways that the subsets can be decomposed over the set of processors. If the initialization method is USE_FIELD_VALUES (no neighbor solution is needed to initialize each frame’s solution) the subsets will be split up evenly across the number of processors. If the initialization method is USE_NEIGHBOR_VALUES the subsets will be split into groups that share a common seed. In this case, the maximum number of processors in use will be equal to the number of seeds that are specified in the input. If more processors are requested than seeds, the extra processors will remain idle with not subsets to evaluate. A third case involves the initialization method being USE_NEIGHBOR_VALUE_FIRST_STEP_ONLY . In this case, the decomposition of subsets will be the same as USE_NEIGHBOR_VALUES , but only for the first frame. If the analysis involves more than one frame, for subsequent frames the subsets will be split evenly over the number of processors available, regardless of how many seeds are specified.

The solution output from a parallel run will be concatenated into one file, written by process 0. Timing files, on the other hand, are written for each processor individually. The file naming convention for parallel runs is

The file naming convention for timing files is

User Manual

Input syntax

Capitalization is disregarded in the input files, all input text is converted to upper-case when read (with the exception of file names in which case the capitalization is preserved). Text files already in the DICe repository should work for Linux or Windows without modification. If this is not the case, you may have the wrong setting in your git repository regarding line endings, see the section on git below. Text files generated on Windows should follow the Windows line ending convention, the same for Linux (or Mac). The xml input files should follow standard xml format. The optional subset file uses the ‘#’ character to start a comment. Everything after the comment character is disregarded.

User specified correlation parameters

Correlation paramters (for example, the interpolation method or how the image gradients are computed) can be set in a separate .xml file. We refer to this file as the correlation parameters file. To include user defined correlation parameters, specify the following option in the input file (typically named input.xml )

If a correlation parameter is not specified in this file, the default value is used. The defaults can be found in the file DICe_ParameterUtilities.cpp . For the object tracking case a specific set of default parameters can be activated by using the

option in the parameters file. This option will set the default parameters to align with an object tracking analysis. Any other parameters that are set in the parameters file will overwrite these defaults.

User defined correlation point locations

In many cases, the locations of the correlation points are equally spaced on a grid in the image (when computing full-field displacements, not tracking objects). The user need not add any options to the input parameters for this default case. In other cases, the user may wish to specify the coordinates of the correlation points as well as use conformal subsets that trace objects in the image rather than use square subsets. Any combination of square and conformal subsets can be used in DICe for the same analysis. The user can specify custom coordinates and conformal subsets using a subset file with a certain syntax (Note: The -g option will not generate a template subset file). If a subset file is used, the coordinate list is mandatory, but the conformal subset definitions are optional. If a conformal subset definition is not provided in the subset file for a subset, it is assumed that the subset is square and will be sized according to the subset_size parameter in the input file. (If all of the subsets in the subset file are conformal, the user does not have to specify the subset_size in the input file.) The ids of the subsets are assigned according to the given coordinates, the first set of coordinates being subset id 0.

The name of the subset file is specified with the following option in the input file:

If a subset file is specified, the step_size parameter should not be used and will cause an error. The subset locations file should be a text file with the following syntax. Comments are denoted by # characters. Lines beginning with # or blank lines will not be parsed. Upper and lower case can be used, the parser will automatically convert all text to upper case during parsing (except when file names are specified, then the case is not changed).

The first section of the subset file is a mandatory listing of centroid coordinates. The coordinate system has its origin at the upper left corner of the image with x positive to the right and y positive downward. The coordinate listing should begin with the command BEGIN SUBSET_COORDINATES and end with END SUBSET_COORDINATES (or simply END , in which case the SUBSET_COORDINATES is inferred) and have the x and y coordinates of each subset listed in between. For example, if the user would like five subsets with centroids at (126,157), (125,250), (397,139), (177,314) and (395,405) this section of the subset file would look like:

If this is all the content in the subset file, five square subsets will be generated with the centroids as given above and a subset_size as specified in the input params file.

Regions of interest and seeding

In some instances, the user may wish to specify certain regions of an image to correlate, but without having to define the coordinates of each point. In this case the user can include a REGION_OF_INTEREST block in the subset file. This block uses common shapes to build up an active sub-area of the image. There are two parts to a REGION_OF_INTEREST definition, the boundary definition and an optional excluded area. Correlation points will be evenly spaced in the REGION_OF_INTEREST according to the step_size parameter. Multiple REGION_OF_INTEREST blocks can be included in the subset file. An example REGION_OF_INTEREST block is given below. Valid shapes include the same as those defined below for conformal subsets.

To seed the solution process, the following command block can be used

A seed gives the values that should be used for the initial guess for a subset. If the optional BEGIN_SEED command is used in a REGION_OF_INTEREST block, the correlation points will be computed in an order that begins with the seed location and branches out to the rest of the domain. The initializiation method for a seeded analysis should be USE_NEIGHBOR_VALUES or USE_NEIGHBOR_VALUES_FIRST_STEP_ONLY. Only one seed can be specified for each REGION_OF_INTEREST.

Shape functions

The user can select which shape functions are used to evaluate the correlation between subsets. By “shape functions” we are referring to the parameters used in the mapping of a subset from the reference to the deformed frame of reference. There are four sets of shape functions available in DICe: translation, rotation, normal strain and shear strain. To manually specify which shape functions should be used, the user can add the following options to the correlation parameters file

See DICe::Subset for information on how these parameters are used in constructing a deformed subset.

Conformal subsets

Conformal subsets are subsets with geometries that correspond to the outline of a particular part or region. For example the yellow and green lines in the image below outline conformal subsets. Both yellow outlines belong to one subset (they do not have to be contiguous).

Note: conformal subsets require that the coordinates of the subsets are specified with a SUBSET_COORDINATES block in the subset file as described above. Conformal subsets cannot be used in combination with regions of interest.

The user may wish to use conformal subsets for some or all of the subsets in an analysis. Conformal subsets can be helpful if the tracked object is of an odd shape. They enable more speckles to be included in the correlation. There are also a number of features for conformal subsets that are useful for trajectory tracking. For example, there are ways to enable tracking objects that cross each other’s path or become partially obscured by another object. Conformal subsets can also be evolved through an image sequence to build up the intensity profile if the object is not fully visible at the start of a sequence.

Conformal subsets are created by specifying the geometry using shapes. There are three attributes of a conformal subset that can be specified using sets of shapes. The first is the BOUNDARY of the subset. This represents the outer circumference of the subset. The second is the EXCLUDED area. This represents any area internal to the subset that the user wishes to ignore (see below regarding evolving subsets). Lastly, an OBSTRUCTED area can be defined. Obstructions are fixed regions in the image in which pixels should be deactivated if they fall in this region. For example if the user is tracking a vehicle through the frame, and it passed behind a light post, the light post should be defined using an obstructed area.

Note: For trajectory tracking and evolving subsets, the TRACKING_ROUTINE correlation_routine should be used, as well as the use_tracking_default_params option in the parameters file.

For each conformal subset, the following syntax should be used to define these three sets of shapes. Note, the BOUNDARY and SUBSET_ID are required, but the EXCLUDED and OBSTRUCTED sections are optional. Continuing with the example subset file above, after the coordinates are listed, we wish to denote that subset 2 (with centroid coordinates (397 139)) is a conformal subset made up of an odd shape made of two polygons and a circle. The subset file text for this subset would be

Available shapes include the following and their syntaxes

Note: for the RECTANGLE shape, if an even size is used for the width or height, the next largest odd number will be used for the size, to split the shape evenly on all sides of the centroid.

Sets of shapes used to define an attribute of a conformal subset can overlap. The pixels inside the overlap will only be included once. An example conformal subset definition that include all three attributes defined is as follows. This subset has a circular boundary, a triangular region to be excluded and an obstruction along the bottom edge.

The optional BLOCKING_SUBSETS section of the conformal subset definition above lists other subsets in the analysis that may cross paths with this subset. After each frame, pixels are deactivated from this subset if their location coincides with one of the listed blocking subsets. This is useful mostly for trajectory tracking.

An excluded area can be used to generate a hole in the subset and can also be used to denote an area that may be initially obstructed by an object in the image and therefore not visible. The reason a user may wish to treat this area as excluded rather than draw the subset around it is because if pixels in the excluded area become visible later in the image sequence the user can request that these pixels become activated. If the correlation parameter

is used, after each frame, the pixels in the excluded area are tested to see if they are now visible. If so, the pixel intensity value from the deformed image is used to evolve the reference pixel intensity that was initially not known. In this way, the subset intensity profile evolves as more regions become visible. There is some small error associated with evolving pixel intensities from the reference image using the deformed image, but in some cases, this is the only way to keep tracking an object, for example if all of the originally visible portions become blocked and only the newly exposed portions become visible.

When obstructions or blocking subsets are used, the user can specify the size of the skin that is constructed surrouding the obstructions, effectively enlarging them. To specify the skin size use the following option

The default value for the skin factor is 1.0. If a skin factor of 2.0 is used, the obstruction is effectively scaled to twice its size. Typical skin factors are in the range of 1.1 or 1.2. The scaling of blocking subsets is applied with the subset centroid as the origin for the scaling.

The solution values can be seeded for a conformal subset by adding a SEED command block to the CONFORMAL_SUBSET command block. For example,

Note that the location of the seed is automatically the subset centroid and cannot be specified (as it is in a REGION_OF_INTEREST). The displacement guess for a seed is required. All other initial values (shear strain, etc.) are optional.
The conformal subsets from the example above are shown in the image below. Note the subset shapes, exclusions etc. are simply a random example to illustrate the syntax, not a meaningful way to set up an analysis.

Path files

In the subset input file, for the TRACKING_ROUTINE , for conformal subsets, DICe provides the user the ability to define a path file. Path files are used for two purposes. The first is to define an expected trajectory for a part being tracked. The second is to provide a baseline to compare the trajectory with to test for anomalous behavior. For example if the object being tracked should follow a straight line, this can be specified in the path file. The initial guess for the subset associated with this object will be taken from a point on the specified path. Once the position solution is computed it will be tested against the closest point on the path.

To define a path file, create a text file with three columns (separated by spaces) u , v , and theta . The order of the points does not matter. There should be no header information in the file. Each point represents a valid solution configuration (a point on the object’s expected path). This file is read into DICe and filtered for unique sets of values and rounded to the nearest half pixel or tenth of a degree of rotation.

To use a path for a conformal subset, the syntax is given inside the CONFORMAL_SUBSET block as in the example below

To test the computed solution for a subset for each frame in terms of the distance from the path, the user can request the following option in the parameters file:

If this option is not specified, the solution will not be compared to the path.

Skip solve

In a conformal subset definition (inside the BEGIN CONFORMAL_SUBSET and END CONFORMAL_SUBSET keywords) the keyword SKIP_SOLVE can be used to take the initial guess for the solution as the final solution and skip the actual solve. If the user would like to skip the solve for all subsets, the following correlation parameter can be set in the parameters file.

Optical flow

If the user has requested the TRACKING_ROUTINE correlation_routine , optical flow can be used as an initializer for the solution or as a tracking method by itself. To use optical flow to determine the initial guess, the keyword USE_OPTICAL_FLOW can be added to the subset file as follows.

To use optical flow as the initializer for all subsets, use the following correlation parameter:

To use optical flow as the motion solution, not just as an initializer for SIMPLEX or GRADIENT DIC, use one of the methods above to turn on optical flow, but also use a SKIP_SOLVE keyword in the subset file (for a particular subset) or the skip_all_solves correlation parameter to skip the DIC solves and use the optical flow solution as the motion solution.

The optical flow method in DICe follows the implementation of Matlab for the Lucas-Kanade algorithm. To determine the points to use for optical flow (two are needed to compute the angle of rotation) all of the pixels in the subset are scanned to find the two with the highest gradients that are at least 10 pixels away from each other. If the subset is small, the pixels can be as close as 2 pixels away from each other which lowers the accuracy substantially. As long as these two pixels stay visible throughout the sequence, the positions about which optical flow is calculated do not change. If one of the pixels becomes obstructed during the sequence, new locations are selected from among the visible pixels.

If optical flow is being used as the initialization method and the initialization fails (due to the subset being too small for two optical flow points or the gradient information is not good enough for optical flow), the initialization defaults back to using the field values from the last frame.

Motion windows

For many tracking application (using the TRACKING_ROUTINE , for conformal subsets), only a small portion of the frame is occupied by the object in motion and the user may wish to use only the region around this portion for anlysis. To save time reading frames and computing image filters, a motion window can be defined for the subset. To define a motion window for a subset the syntax is as follows

The optional motion tolerance is the total number of intensity counts of the difference between the current and previous image that is used as a threshold for detecting motion. In most cases, this can be automatically computed by DICe without specifying this optional parameter. If motion detection is not used (see below), this tolerance is ignored.

Motion detection

The user can request that the correaltion only be performed if there is motion detected in the vicinity of the subset. This can be helpful in speeding up an analyis if the object being tracked sits idle for most of the video and only moves for a small portion of frames. To test for motion, the user simply adds the TEST_FOR_MOTION keyword to the subset definition. The subset must also have a MOTION_WINDOW defined for motion detection.

Since there can potentially be several subsets inside of one window, many subsets can share a particular motion test window. If multiple subsets share a window, the window needs only to be defined for one subset as in the example above, the rest of the subsets can simply identify the id of the subset with which to share the motion window. For example if subset 0 defines the window, subsets 1 and 2 can refere to the window of 0 using the following syntax

Skip solves for a particular conformal subset

If the user would like to turn tracking on or off for certain conformal subsets at different points in the analysis, the SKIP_SOLVE keyword can be added to the subset definition. The SKIP_SOLVE keyword is useful when a subset is in motion for only a portion of the video sequence. The syntax for this keyword is the keyword followed by a set of id numbers that represent frame ids. The first number turns tracking off and subsequent ids turn tracking on or off in an alternating fashion. In the following example, the user would like to only track the subset for frames 1000 to 2000 and then from 2500 to 3000 and stop tracking for the rest of the video.

The user can also specify this keyword followed by a file name that contains the ids in text form, with one frame id per line.

Filtering images

To Gauss filter the images add the following option to the parameters file. The default has no filtering.

When filtering is enabled, the size of the filtering mask is seven pixels by seven pixels. The coefficients of the mask are given in DICe_ImageSerial.cpp or DICe_ImageKokkos.cpp. It is possible to filter images with different window sizes, but so far this is only enabled for the DICe::Image class. It has not been enabled in the correlation parameters for the executable.

Detecting initialization or correlation failure

To force DICe to abort if the quality of the initialization step or the correlation does not provide a value for gamma that is above a certain threshold, set one or both of the following parameters in the correlation parameters file:

The initial_gamma_threshold will test the gamma value for the pre-solve, intial guess. The final_gamma_threshold will test the value of gamma post-solve.

Output files

The output produced are space delimited text files. The default output is one file per deformed image listing the solution variables for each subset. In the default case, the index at the end of the file name refers to the image id or frame number.

The user can alternatively request output as a separate file for each subset listing the solution variables for each deformed image or frame. To do so set the following option in the input file

If the separate file for each subset option is used, the index at the end of the filename refers to the subset id.

The file name prefix to use for the output files can be set with the following option

The user can also define an output specification to list the fields in a specific order and choose which fields to output. To define the output fields and the order, an output_spec can be added to the correlation parameters. The following is a syntax example that outputs only three fields in the order COORDINATE_X , COORDINATE_Y , DISPLACEMENT_X :

The order of the columns in the output file is determined by the ordering in the list of field names. If the boolean value is set to false , the field will not be output.

The delimiter used in the output file can be set in the parameters file with the following option:

The user can also request that the row id in the output files be omitted with

Fields

For almost any analysis, the following fields are available for output:

Some of the parameters require activation in the correlation parameters. To output BETA the following parameter must be set in the correlation parameters file

If all zeros are reported for a certain field, for example SHEAR_STRAIN_XY , it usually means that particular shape function is not activated in the correlation paramters. The CONDITION_NUMBER field is not used in the SIMPLEX optimization method so all zeros will be reported for SIMPLEX . Values of -1.0 typically imply failure of some kind. For example if the NEIGHBOR_ID field is -1.0 the subset corresponding to that field value does not have a neighbor. Another example would be if SIGMA is -1.0 it implies that the correlation failed for that particular step.

Coordinate system and positive rotation: Image coordinates are measured from the top left corner of the image with x positive to the right and y positive downward. Rotations are positive clockwise (opposite of the right-hand rule).

To see a complete list of available fields for output see the field_name_strings here: https://github.com/dicengine/dice/blob/master/src/mesh/DICe_MeshEnums.cpp. Note that fields are only available for output if they have been generated by the parameters of the analysis.

Determining the quality of the solution

The following output parameters are useful in estimating the quality of the displacement solution:

** SIGMA :** (Predicted displacement variation) This variable estimates the predicted variation in the displacement solution given variations in the data due to noise and interpolation bias. SIGMA can be used as an uncertainty metric if the following conditions are met: the displacements are smaller than one pixel in magnitude and the noise level in the images is less than roughly three or four percent. Use SIGMA to determine the confidence in the computed solution. For SIGMA lower values are better and imply lower uncertainty.

** GAMMA :** (Template matching quality) This variable measures how well the template or subset from the reference image matches the deformed image. GAMMA is the value of the correlation criteria (or objective functional or cost function magnitude). GAMMA provides the user with a way to tell if the subset is still registering on the correct location in the deformed image. For GAMMA lower values are better, 0.0 implies an identical or perfect match between the reference and deformed subset.

** BETA :** (Sensitivity of the cost function) This variable provides a measure of how sensitive the cost function (or correlation criteria) is to small perturbations in the displacement solution. If the cost function is highly sensitive, it will increase dramatically for even slight errors in the displacement solution. If the cost function is not sensitive, it implies that it cannot differentiate between many potential solutions. Again, for BETA lower values are better. Common sources of high BETA are poor contrast, lack of randomness in the speckle pattern, subset sizes too small, or high image noise levels.

Plotting results with python

Note: Some extra python modules must be installed to use these python scripts.

The following python script can be used to plot the results from a DICe output file. (Note: these are not intended for conformal subsets if they are defined). If the output files are in the default format of one output file per image with a listing of subset variables in columns with one subset per row, the following python script can be used to create a two-dimensional contour plot. This example also assumes that the output specification has the fields in the default column order (id,x,y,u,v. ) .

If separate output files were generated for each subset, a time history of the solution variables can be plotted for all subsets with the following script. A separate plot will be created for each subset.

The python scripts above are only meant to provide a simple example. Obviously, there are many ways python can be used to generate more sophisticated plots.

Building DICe

The easiest way to get started with DICe is to use the package installers on the releases page of the github site. Pre-built binaries of the command line executables and the GUI combined are available for Mac OS X and Windows. For a linux distribution (or to make custom mods to the executables or GUI) one has to build DICe from source. The instructions below provide a means to build the command line executables and libraries for DICe. To build the GUI from source a separate step is necessary and the details are provided below.

Requirements

DICe can be built and run on Mac OS X, Windows, and Linux. The prerequisite dependencies required for installing DICe include:

  • CMake Version 2.8 or greater (tested with 3.3.2)
  • Trilinos Version 12.0 or greater (tested with 12.2)
  • LAPACK or CLAPACK (for Windows, only CLAPACK is supported Version 3.2.1 or greater)
  • OpenCV Version 3.2.0 (tested with 3.2.0, newer versions give configure errors on linux)
  • Boost is no longer a dependency. It was removed Sept. 2018
  • LibTiff, LifJpeg, and LibPng are also not required anymore, they get built with OpenCV

For windows users, see the instructions below for using the DICe Developer Packs, which avoid having to build any of the dependencies listed above (Trilinos, OpenCV, or LAPACK).

CMake

DICe makes use of CMake for build configuration. Sample CMake scripts for building Trilinos and DICe are in the folder dice\scripts

Installing Trilinos

Trilinos contains a set of software packages within an object-oriented software framework used for the solution of large-scale, complex multi-physics engineering and scientific problems. DICe uses some of the packages avaiable in Trilinos and requires that Trilinos be installed.

Trilinos can be downloaded from http://trilinos.org and build instructions can be found on the getting started page. DICe requires that Trilinos be built with the following packages enabled:

  • Epetra
  • BLAS
  • LAPACK
  • TeuchosCore
  • TeuchosComm
  • TeuchosParameterList
  • TeuchosNumerics

Obtaining DICe source code

The central repository for the source code is on GitHub.

DICe can be forked from the following git repository

Feel free to send a pull request on GitHub if you add new features to a forked repo of DICe that you would like to see added to the main repo.

Setting up your git repository

If this is your first time using git, you will need to set up your git configurations. To set your user name and email use

Windows users

To enable the tests that diff text files to pass, it will be important for Windows git users to add another option to their git configuration

This checks out text files in Windows CRLF format and checks files into the git repo with unix LF format.

Note: this must be done before executing the pull command above, doing so after pulling the DICe repository will not work.

Building DICe

Mac OSX or Linux

To build DICe on Mac OSX or Linux, create a folder in the main directory called build

Change directory into the build directory and copy the CMake script from the scripts folder.

Edit the script to have the correct path locations. Then build DICe

It is important to execute the install target to put all of the dice libraries in one folder. This is helpful when incorporating dice in an external application since all the libraries will be in one place.

Ubuntu

Several users have had problems installing DICe on Ubuntu linux. For that reason, we include a detailed set of instructions to build the command line executable for DICe on Ubuntu here. We will add the GUI build instructions for linux later. Example configuration scripts for opencv, trilinos, and DICe on Ubuntu are located in the /scripts/ubuntu folder. The build configuration from these scripts will include the global methods, and opencv features. These instructions have been tested on Ubuntu 16.04.3 LTS. These instructions also work for RHEL if you change the apt-get commands to the corresponding rpm commands.

First, set up your package installer if not already updated.

Install git if not already installed. After installing git, follow the instructions above to set up your git configuration (username, ssh-keys, etc.).

Install an openmpi compiler.

Install cmake, if not already installed.

Install the BLAS and LAPACK libraries.

Install the image libraries.

Install NetCDF. After installing netcdf, change some of the defaults in /usr/include/netcdf.h.

Set the variables NC_MAX_DIMS to 65536, NC_MAX_ATTRS to 8192, NC_MAX_VARS to 524288, and NC_MAX_NAME to 256

Download Trilinos 12.4.2 from trilinos.org and create a build folder in the top of the directory tree of the source code. Place the do-trilinos-cmake file inside this build folder and change the path variables in the do-trilinos-cmake file.

Download OpenCV 3.2.0 from here. Create a build folder in the top level of the source code and copy the sample do-opencv-cmake script to this folder. Change the path variables and run the configuration and build.

Clone DICe and build.

Windows

  1. You currently have CMake installed
  2. git is installed and available through Windows terminal
  3. The DICe source code has been cloned
  1. Download and extract the DICeDevPack (there are VS2013 and VS2017 options) – https://github.com/dicengine/dice/releases/tag/wdp1.0
  2. (If you already have Visual Studio 13 or 17 installed with command line tools you can skip this step) Once extracted, open the dev pack and navigate to the vs2013 or vs2017 folder. Run the Visual Studio install executable.
  3. In the DICe source code folder (cloned above, not the dev pack directory), add a new directory named “build” such that the structure is: . \DICe\build
  4. Copy the following file from the DICeDevPack scripts folder into the build dir you just created: DICeDevPack/scripts/do-cmake.bat
  5. Open the “do-cmake.bat” that you just copied in a text editor and update the following information: a. “CMAKE_INSTALL_PREFIX:FILEPATH=” with the path to your dice install (for example: CMAKE_INSTALL_PREFIX:FILEPATH=C:\Users\user\software\DICe\build ) b. “DICE_DEVPACK_DIR:STRING=” with the path to your DICeDevPack (for example: DICE_DEVPACK_DIR:STRING=”C:\Users\user\software\DICeDevPack”)
  6. Launch Visual Studio’s native tools command prompt. For example, on a 64 bit system with VS2013 installed: click the start button, scroll through the list of applications to “Visual Studio 2013”, click the down arrow, and selected “Visual Studio Tools”. From this directory, I select “VS2013 x64 Native Tools Command Prompt”.
  7. From the VS Command prompt, navigate to the DICe\build folder with do-cmake.bat file from step 4 and run the .bat file.
  8. Once complete, run the command “nmake” from the terminal.
  9. Update your PATH env variable to include the path to the DICe build by entering the following at the command line:
  1. cd into test directory that should be in the same dir as the do-cmake.bat you ran.
  2. From this directory, run ctest

If everything has been set up properly, all tests should pass.

    If you received the error message: “CMake Error: The source directory “

” does not appear to contain CMakeLists.txt.” You didn’t put the do-cmake.bat file in the DICe dir in a build folder.
Every test fails with an error message saying “Could not find executable

” Check to make sure the excutable is in the specified path. If not, rerun do-cmake.bat and nmake.

  • If the first several tests fail you probably failed to set your PATH to include the bin. Verify you set it as outlined in step 8.
  • Building the GUI from source

    To build the DICe gui from source, four steps are required. First a separate DICe_gui repo needs to be cloned. Then nodejs has to be installed. After this, a pre-built binary for electron needs to be installed. The paths to the DICe executables needs to be set in a javascript file. At this point, the gui is ready to run. Detailed instructions are provided for Ubuntu linux, but the instructions for other operating systems can easily be replicated.

    First clone the DICe_gui repo

    Next install nodejs, either by downloading the package installer from nodejs.org or using apt-get

    Then download the pre-build electron binary from https://github.com/electron/electron/releases. For our Ubuntu example, we download electron-v1.7.10-linux-x64.zip. Unpack the zip file and inside is an electron executable file. Note the path to this executable, which we will call electron_exec below.

    Open the global.js file in the DICe_gui repo and modify the linux_path variable.

    The DICe_path in the variable above should point to the DICe command line executables and libraries that you built in the steps above (the DICe repo build folder, not the DICe_gui repo).

    Next, simply cd into the DICe_gui folder and run the electron executable

    DICe Configuration Options

    The following configuration options can be set in the CMake configuration script for DICe:

    This option, when activated, outputs large quantities of debugging information and is useful when setting up a new analysis.

    The default data type in DICe is float , double can be used by activating this option.

    Testing

    To test the installation, from the directory \dice\build\test execute this command (this assumes that DICe has been built in the build directory, otherwise specify the correct path)

    Compiling DICe with debug messages

    To enable debug messages in DICe, simply set the CMake flag -D DICE_DEBUG_MSG:BOOL=ON in the do-cmake script.

    Using DICe as a dynamically linked library

    DICe can also be used in library mode, whereby an application like LabView can call dice_correlate() as a function from libdice . There are two versions of the function call, one that takes an array of subset locations as input, and another that uses a subset file to define conformal subsets. If the array of points function is used, the subset size is constant.

    The code for the library is in file dice/src/api/DICe_api.cpp . This file can be used as a template for developing a more advanced interface. Another option is to write a custom DIC application that follows the example here. Using DICe as a library involves writing an interface that sets the parameters and orders the data from the correlation in the right order for the particular application (for the purposes of data exchange). Once this interface is compiled as a library, it can be linked and used in an external application simply by calling the correlation function exposed via the interface.

    Legal

    Sandia National Laboratories is a multi-mission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.

    License

    Copyright 2015 National Technology & Engineering Solutions of Sandia, LLC (NTESS)

    Under the terms of Contract DE-NA0003525 with NTESS, the U.S. Government retains certain rights in this software.

    Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:

    1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
    2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
    3. Neither the name of the Corporation nor the names of the contributors may be used to endorse or promote products derived from this software without specific prior written permission.

    THIS SOFTWARE IS PROVIDED BY NTESS “AS IS” AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL NTESS OR THE CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

    Introduction DICe is an open source digital image correlation (DIC) tool intended for use as a module in an external application or as a standalone analysis code. Its primary capability is

    The White

    The White is a potent hybrid marijuana strain known for having a distinct lack of odor or flavor. The White gives off a high that is usually described as providing equal body and head sensations. This mysterious strain — originally called “Triangle” and supposedly from somewhere in Florida — is aptly named. The White is covered with so many white trichomes that buds look like they were rolled in confectioner’s sugar. While it looks much like an OG in structure and certainly has the potency associated with the best OG Kush cuts, it has little of the smell or flavor found in those West Coast favorites. Those who do not enjoy the flavor and aroma of cannabis opt for this strain over more pungent varieties. Medical marijuana patients choose this strain to help relieve symptoms associated with insomnia.

    The White is a potent hybrid marijuana strain known for having a distinct lack of odor or flavor. The White gives off a high that is usually described as providing equal body and head sensations. This mysterious strain — originally called “Triangle” and supposedly from somewhere in Florida — is aptly named. The White is covered with so many white trichomes that buds look like they were rolled in confectioner’s sugar. While it looks much like an OG in structure and certainly has the potency associated with the best OG Kush cuts, it has little of the smell or flavor found in those West Coast favorites. Those who do not enjoy the flavor and aroma of cannabis opt for this strain over more pungent varieties. Medical marijuana patients choose this strain to help relieve symptoms associated with insomnia.

    The White effects

    • Feelings
    • Negatives
    • Helps with

    The White reviews 363

    Strain spotlight

    Most popular in

    • Portland, OR
    • Eugene, OR
    • Seattle, WA
    • Salem, OR
    • Bend, OR
    • Everett, WA
    • Denver, CO
    • Los Angeles, CA
    • Spokane, WA
    • Belfair, WA

    We’ll email you updates on local dispensary deals, new products, and industry news.

    By providing us with your email address, you agree to Leafly’s privacy policy and terms & conditions.

    • FAQ
    • Contact
    • About us
    • Careers
    • Investor relations
    • Marijuana strain database
    • List your store
    • List your CBD store
    • List your brand
    • List your practice
    • Certify your lab
    • Business log in
    • Colorado
    • Nevada
    • Illinois
    • Arizona
    • Massachusetts
    • Oregon
    • Terms of use
    • Commercial terms of use
    • Privacy policy
    • Do not sell my personal information

    We’re a long way from equity in cannabis. At Leafly, we’re on a mission to strengthen the industry. Want to join forces, get informed, get involved? Head over to Leafly | Social impact.

    The White is a hybrid cannabis strain.