PD-DIC User Manual

1Getting Started

PDDIC measures full-field displacement and strain from experimental images, and locates cracks automatically. You supply a reference image and one or more deformed images; PDDIC tracks the displacement between them, computes the strain field, and flags regions where the deformation is no longer physically compatible — which is where damage is.

Everything is done through the graphical interface. There is nothing to script and nothing to compile.

1.1System Requirements

1.2Installing PDDIC

1. Download PDDIC-Setup-1.0.0.exe from the link in your purchase email.
2. Double-click the installer. Windows will ask for administrator permission — accept it.
3. Read and accept the End User License Agreement.
4. Choose the installation folder, or accept the default (C:\Program Files\PDDIC).
5. Optionally tick Create a desktop shortcut.
6. Click Install, then Finish.

PDDIC appears in the Start menu under PDDIC. Launch it from there, or from the desktop shortcut if you created one.

1.3Activating Your License

The first time you launch PDDIC, the Activate PDDIC window opens. PDDIC will not run until it is activated.

Activate online (normal method)

1. Open the Activate Online tab.
2. Type or paste the license key from your purchase email. It has the form ABCDE-FGHJK-LMNPQ-RSTUV. Lower-case letters are converted to upper-case as you type, so you can paste it in whatever form you received it.
3. Click Activate.

PDDIC contacts the license server, claims one seat on your license for this machine, and opens the main window. This takes a few seconds.

Activate offline (air-gapped machines)

If the computer running PDDIC has no internet access:

1. Open the Offline Activation tab.
2. Enter your license key.
3. Click Save Request File… and save the .json request somewhere you can reach it.
4. Email that file to licensing@gertechnologies.com from any machine with internet access.
5. You will receive a lease file in reply. Copy it to the offline machine.
6. Return to the Offline Activation tab, click Import Lease File…, and select the file you were sent.

PDDIC starts as soon as the lease is imported.

1.4Managing Your License

Open File ▸ License… at any time to see the current status:

How the lease works

After activation PDDIC holds a 14-day lease. Once the lease is a week old, PDDIC renews it quietly in the background the next time it starts with an internet connection — you never see this happen. If the server cannot be reached, your existing lease keeps working until it expires, so a network outage never stops your work mid-session.

Moving PDDIC to a different computer

Each activation uses one seat. To free that seat, click Deactivate This Machine in the License window and confirm. PDDIC stops working on that computer, and you can activate the same key on the replacement machine.

1.5The Main Window

1.6Where Results Are Saved

PDDIC writes its results to a per-user data folder, not to the installation directory. Open File ▸ Open Output Folder to go straight there in Windows Explorer.

Keeping results out of C:\Program Files means PDDIC never needs administrator rights to run, and your results survive an uninstall or upgrade untouched.

1.7Getting Help

The Help menu contains:

2Workflow — Step by Step

PDDIC works through eight steps. Steps 1–3 define the geometry, Steps 4–7 run the solvers, and Step 8 detects cracks. Run them in order — each step consumes the previous one's output. Step 6 (FD Strain) is the one exception: it is optional and feeds nothing downstream.

2.1Step 1 — Select Working Directory

Choose the folder holding your reference and deformed images. TIF, PNG, JPG, and BMP are all recognised.

Procedure

1. Click Browse and select the folder.
2. Every recognised image appears in the list with a checkbox.
3. Tick each image you want to analyse. The first ticked image is the reference (undeformed) state; the rest are deformed states, analysed in list order.
4. Click Next.

Use Select All / Deselect All to work through long sequences quickly. The reference image appears in the Images tab as soon as it is selected.

2.2Step 2 — Domain Selection

Define the rectangular region of interest. Everything downstream — tracking, regression, differentiation, crack detection — is confined to this region.

Procedure

1. Interactively: drag a rectangle directly on the reference image. The coordinate fields update as you drag.
2. By typing: enter values in xmin, xmax, ymin, ymax, then click Apply Domain.

Keep the domain inside the specimen. Including background or grip hardware wastes computation and adds points that cannot be tracked.

2.3Step 3 — Cut-outs

Cut-outs are exclusion zones — holes, notches, pre-existing voids, grip regions. Points inside a cut-out are excluded from tracking, and also from the regression neighbourhood of nearby points, so displacement is never blended across a physical boundary.

Adding a cut-out

1. Choose Rectangle or Circle (an ellipse fitted to the bounding box).
2. Click Add Cut-out, then drag on the image — or type the bounding box coordinates and click Add Cut-out.
3. Drag the corner handles to fine-tune the shape.

Managing cut-outs

You can define as many cut-outs as you need, mixing rectangles and circles freely. Their geometry carries through to every later step automatically.

2.4Step 4 — Subset Tracking (DIC)

Tracking measures how far small image patches move between the reference image and each deformed image. The result is a sparse field of displacement vectors (U, V).

Procedure

1. Set Subset size and Min point distance — these are the two parameters that matter most.
2. Adjust anything under Advanced Settings only if you need to.
3. Click Run Tracking. A progress bar and status message track the run.
4. Results appear in the Tracking Results tab. Choose U or V from the drop-down to view horizontal or vertical displacement.

Advanced Settings

The remaining parameters are collapsed behind the Advanced Settings toggle, which starts closed. Click it to expand the section, click it again to fold it away. The defaults suit most experiments — open this section when the defaults are not working, not before.

2.5Step 5 — PD Regression

Regression spreads the sparse tracked displacements onto a dense, regular grid using the zeroth-order Peridynamic Differential Operator. The output is a smooth displacement map covering the whole domain.

Procedure

1. Set the parameters below.
2. Click Run PD Regression. The solver runs in the background.
3. The dense U and V fields appear in the PD Regression Results tab.

2.6Step 6 — FD Strain (Optional Preview)

A fast strain estimate computed by finite differences on the regression grid. Treat it as a preview — Step 7 produces the strain fields you should actually report.

Output fields

2.7Step 7 — PD Differentiation

The main strain analysis. The first-order PDDO differentiates the dense displacement field from Step 5, producing the full strain field plus the compatibility residual and Damage Confidence Factor that Step 8 needs.

Procedure

1. Set the parameters below.
2. Click Run PD Differentiation. The solver runs in the background; a status message confirms completion.
3. Results appear in the PD Differentiation Results tab.

Output fields

2.8Step 8 — Crack Detection

Step 8 finds cracks from the compatibility residual computed in Step 7. It needs no prior knowledge of where the crack is, which way it runs, or how long it is.

What happens when you click Detect Cracks

1. Boundary masking — points within 4× the point spacing of the domain edge or any cut-out are dropped, since boundaries produce compatibility artefacts that look like damage.
2. Damage confidence — the residuals of the remaining interior points are standardised into a two-tailed p-value. A point with a large residual gets a low CDF, meaning high damage probability.
3. Candidate selection — points with CDF at or below the threshold are flagged and displayed.
4. Crack path fitting — clustering isolates the dense crack region from scattered noise, the principal crack direction is identified, and a smooth curve is fitted along it to predict the continuous crack path.

Output views

The black lines on the last two views mark the masked region: the domain shrunk by 4× the point spacing, and each cut-out enlarged by the same margin.

3Post-Processing Tools

All of these live on the canvas toolbar and act on whichever result tab is currently open.

3.1Canvas Navigation

3.2Plot Settings

Click Settings on the toolbar. Changes apply immediately to every result canvas.

Color scheme

Number of colors

Color range

Unit

Displacement fields can be displayed in physical units instead of pixels. Choose the unit and enter the conversion factor.

Once you pick a unit, it propagates everywhere: colorbar ticks, the min/max overlay labels, Line Plot and Box Plot axes, and the coordinate input fields.

Coordinate axes

3.3Line Plot

Extracts field values along a straight line you draw — ideal for a profile across a crack tip or along a gauge length.

Procedure

1. Click the Line Plot toggle. A coordinate bar appears below the toolbar.
2. Click once on the canvas to set Point 1 (P1), then click again to set Point 2 (P2). White handles mark the endpoints.
3. Drag the handles to adjust, or type coordinates into the P1/P2 fields and click Move to.
4. Choose the Result set and Field.
5. Optionally tick Normalize Y to divide values by their maximum absolute value.
6. Click Plot.

Line Profile window controls

3.4Box Plot

Extracts every data point inside a region you draw and shows their spatial distribution as a colour-mapped scatter plot — useful for inspecting a notch tip or a specific gauge section statistically.

Procedure

1. Click the Box Plot toggle. A coordinate bar appears below the toolbar.
2. Select the region shape: Rectangle or Circle.
3. Drag on the canvas to draw the region, or type the bounds and click Move to.
4. Choose the Result set and Field.
5. Click Plot.

The scatter plot uses the colormap and unit from Plot Settings, and updates live as you drag the region's corner handles.

3.5Save PNG

Exports the current canvas view — colorbar overlay and axis helper included — as a PNG image file. A file-save dialog prompts for the output path.

3.6Save Results

Exports numerical data to a plain text file. A dialog lists the datasets you have computed; pick one.

The file opens with two comment lines describing its content, then the point count, then one row per point. Coordinates and displacements are in pixels; strains are dimensionless.

4Tips and Troubleshooting

4.1Preparing Your Images

• Use a high-contrast random speckle pattern. Speckles should span roughly 3–5 pixels so that each subset contains several of them.
• Minimise out-of-plane motion, rigid-body rotation, and lighting changes between frames.
• 8-bit or 16-bit greyscale is ideal. Colour images are converted to greyscale internally.
• Keep image dimensions identical across the sequence.

4.2Choosing a Subset Size

• Too small: not enough texture, correlation scores fall, points get rejected.
• Too large: displacements are averaged over too wide an area and sharp gradients are lost.
• Rule of thumb: the subset should contain 3–5 speckles. Start at 21–31 px and adjust.

4.3Choosing a Horizon Ratio

• The default of 3 suits most work.
• Raise it to 4–5 when the strain field is noisy or the points are irregularly spread.
• Lower it to 2 when sharp gradients near a crack tip are being smoothed away.

4.4Choosing a CDF Threshold

• Start at the default 0.75.
• If real cracks are missed, lower it (try 0.5).
• If false positives appear away from the crack, raise it (try 0.85–0.90).
• Always inspect the CDF map before interpreting the result.

4.5Common Problems

5Glossary

PDDIC is licensed by Global Engineering Research and Technologies, LLC. Support: admin@gertechnologies.com · Licensing: licensing@gertechnologies.com