Remove Surface Form

You are currently viewing the documentation for the latest version (2.2.2). To access a different version, click the "Switch version" button located in the upper-right corner of the page.

■ If you are not sure which version of the product you are currently using, please feel free to contact Mech-Mind Technical Support.

Description

This Step is used to remove the overall surface form, such as tilt and curvature, from surface data while preserving local height variations for subsequent feature extraction and measurement.

Usage Scenario

This Step is applicable when subsequent Steps need to focus on local height variations. It can reduce the impact of background forms, such as overall tilt and curvature, on feature extraction and measurement results.

Workflow

The process of configuring this Step is as follows:

remove-surface-form-workflow
  1. Configure the input. Connect the corresponding Step port in the graphical programming workspace, or select the corresponding input under Input in the parameter configuration panel.

  2. Determine whether to use a feature region. See Feature Region to learn about the related settings.

  3. Set other parameters under Parameters.

  4. Under Output, select the Surface Data output item (selected by default).

  5. Run the Step and view output.

Input Description

Input Item Description

Surface Data

Surface data from which the form is to be removed.

Alignment Parameter Group

Optional input. When this input is provided, the pose of the feature region is adjusted according to the pose transformation of the target object.

Parameter Description

Common Parameters

Parameter Description

Use Feature Region

Determines whether to use a feature region to limit the surface data involved in processing. When this parameter is selected, the Step processes only the data inside the feature region.

Feature Region

Visible only when Use Feature Region is selected.

See Set Feature Region to learn how to set and adjust a feature region.

Remove Outliers

When enabled, removes extremely low and high values based on the low and high percentiles. This is suitable when the input surface contains burrs, flying points, isolated noise points, or abnormally deep pits, and reduces the impact of outliers on subsequent form removal and filtering.

  • Low Percentile: Sets the low percentile threshold for outlier removal. Data points below this percentile are treated as extremely low values and removed.

  • High Percentile: Sets the high percentile threshold for outlier removal. Data points above this percentile are treated as extremely high values and removed.

Short-Wave Cutoff Wavelength

Sets the cutoff wavelength for short-wave filtering. Variations shorter than this wavelength are treated as short-wave variations and filtered out to remove high-frequency noise.

The recommended value is 3–5 times the input image resolution.

Long-Wave Cutoff Wavelength

Sets the cutoff wavelength for long-wave filtering. It is used to remove waviness, curvature, and other large-scale background variations.

Advanced Parameters

In most cases, retain the default values of the advanced parameters. If obvious distortion occurs at the boundaries, the background form is not sufficiently removed, or local features are excessively weakened, adjust the advanced parameters.

Parameter Description

Gap Filling Method

Selects how to fill gaps or invalid points in the surface.

  • Linear interpolation filling: Estimates the height of gap positions based on surrounding valid points, suitable for small gaps with continuous surrounding surfaces.

  • Min value filling: Fills gap positions with the minimum height in the surface data, suitable when gaps should be treated as low-lying or background areas.

Polynomial Fitting

When enabled, uses Polynomial Order and Sampling Step to fit the overall form.

  • Polynomial Order: Sets the fitting complexity of the overall form. A higher order can fit more complex background forms and is suitable for surfaces with greater curvature. An order that is too high may weaken local features such as glue beads, weld seams, or dents.

  • Sampling Step: Sets the point sampling interval used when fitting the overall form. A larger step uses fewer data points and provides faster computation.

Boundary Handling Method

Selects how to handle surface boundaries during filtering to reduce distortion caused by missing neighboring data.

  • Remove boundary data: Removes data near the surface boundary.

  • Zero padding (ZPA): Pads the area outside the boundary with zeros.

  • Linear extrapolation (LEX): Extends data outward based on the trend of the data near the boundary.

  • Line-symmetric reflection (LSR): Extends data from inside the boundary to the outside using the boundary line as the axis of symmetry.

  • Point-symmetric reflection (PSR): Extends data from inside the boundary to the outside using the boundary point as the center of symmetry.

Long-Wave Filter Model

Selects the model used for long-wave filtering. Long waves usually correspond to background variations with a large lateral span, such as waviness or curvature.

Two filtering models are supported: Gaussian and Isotropic spline.

Output Description

The output of this Step is surface data with the overall form removed, which can be used for subsequent feature extraction and measurement.

Troubleshooting

  • For common error codes of all Steps, see Error Code List.

  • If the error code is inconsistent with the error message, please contact Technical Support for help.

CV-W4601

Error: The value of "Low Percentile" is outside the valid range [0, 0.05].

Possible cause: Low Percentile is set too low or too high, causing the threshold for removing extremely low values to exceed the allowed range.

Solution: Reset Low Percentile and ensure that it is within the valid range.

CV-W4602

Error: The value of "High Percentile" is outside the valid range [0.95, 1].

Possible cause: High Percentile is set too low or too high, causing the threshold for removing extremely high values to exceed the allowed range.

Solution: Reset High Percentile and ensure that it is within the valid range.

CV-W4603

Error: The value of "Gap Filling Method" is invalid.

Possible cause: The selected filling method is not a Gap Filling Method supported by the current Step.

Solution: Select a valid Gap Filling Method as needed, such as Linear interpolation filling or Min value filling.

CV-W4604

Error: The value of "Polynomial Order" is outside the valid range [1, 9].

Possible cause: Polynomial Order is set too low or too high and cannot be used to fit the overall form.

Solution: Reset Polynomial Order and ensure that it is within the valid range.

CV-W4605

Error: The value of "Boundary Handling Method" is invalid.

Possible cause: The selected method is not a Boundary Handling Method supported by the current Step.

Solution: Select a valid Boundary Handling Method as needed, such as Remove boundary data, Zero padding (ZPA), Linear extrapolation (LEX), Line-symmetric reflection (LSR), or Point-symmetric reflection (PSR).

CV-W4606

Error: The value of "Short-Wave Cutoff Wavelength" is outside the valid range [0.001, 100].

Possible cause: Short-Wave Cutoff Wavelength is set too low or too high and cannot be used for short-wave filtering.

Solution: Reset Short-Wave Cutoff Wavelength and ensure that it is within the valid range.

CV-W4607

Error: The value of "Long-Wave Cutoff Wavelength" is outside the valid range [0.001, 100].

Possible cause: Long-Wave Cutoff Wavelength is set too low or too high and cannot be used for long-wave filtering.

Solution: Reset Long-Wave Cutoff Wavelength and ensure that it is within the valid range.

CV-W4608

Error: The feature region is invalid.

Possible cause: The feature region is empty or contains no valid surface data.

Solution: Reset the feature region and ensure that it is within the valid surface data range.

CV-W4609

Error: The value of "Sampling Step" is outside the valid range [1, 9].

Possible cause: Sampling Step is set too low or too high and cannot be used for sampling during polynomial fitting.

Solution: Reset Sampling Step and ensure that it is within the valid range.

CV-W4610

Error: The value of "Long-Wave Filter Model" is invalid.

Possible cause: The selected filtering model is not a Long-Wave Filter Model supported by the current Step.

Solution: Select a valid Long-Wave Filter Model as needed, such as Gaussian or Isotropic spline.

CV-W4611

Error: The feature region is too small, or the Short-Wave/Long-Wave Cutoff Wavelength is too large for the current data range.

Possible cause: The feature region covers too small a data range, resulting in insufficient valid data points for processing; or the Short-Wave/Long-Wave Cutoff Wavelength is set too large relative to the current data range, preventing a valid filtering result.

Solution: Adjust the feature region to ensure that it contains sufficient valid surface data, or decrease Short-Wave Cutoff Wavelength and Long-Wave Cutoff Wavelength.

CV-W4612

Error: The value of "Long-Wave Cutoff Wavelength" must be greater than the value of "Short-Wave Cutoff Wavelength".

Possible cause: The Long-Wave Cutoff Wavelength and Short-Wave Cutoff Wavelength settings are invalid, preventing the long-wave filtering range from covering the short-wave filtering range.

Solution: Increase Long-Wave Cutoff Wavelength or decrease Short-Wave Cutoff Wavelength to ensure that the long-wave cutoff wavelength is greater than the short-wave cutoff wavelength.

CV-W4613

Error: The value of "Low Percentile" must not be greater than the value of "High Percentile".

Possible cause: The Low Percentile and High Percentile settings are invalid, resulting in an invalid outlier removal range.

Solution: Decrease Low Percentile or increase High Percentile to ensure that the low percentile is lower than the high percentile.

Is this page helpful?

You can give a feedback in any of the following ways:

We Value Your Privacy

We use cookies to provide you with the best possible experience on our website. By continuing to use the site, you acknowledge that you agree to the use of cookies. If you decline, a single cookie will be used to ensure you're not tracked or remembered when you visit this website.