Image Morphological Transformation

Function Description

This step can process images using morphological operations.

MorphologicalTrans

Use Cases

This step is typically used during image preprocessing. Depending on the task, operations such as dilation, erosion, opening, and closing can be used for denoising, gap filling, and shape cleanup.

Inputs and Outputs

  • Input (Original Image): Image to undergo morphological processing. Supports depth images, color images, and mask images, specified by the parameter “Set Input Image Type”.

  • Output (Processed Image): Image after morphological processing.

Parameters

  • Morphological Operation Type

    • MORPH_ERODE

      • Name: Erode

      • Description: Shrinks bright regions, expands dark regions

      • Function: Eliminates small noise points, separates connected objects, smooths object boundaries, reduces target size

      • Application: Removes salt-and-pepper noise, image break fine connections, target size precision, preprocessing step

        Erosion effect image

    • MORPH_DILATE

      • Name: Dilate

      • Description: Expands bright regions, shrinks dark regions

      • Function: Fills internal holes, connects adjacent objects, expands target size, repairs broken parts

      • Application: Fills text stroke breaks, connects segmented cells, expands target area, post-processing step

        Dilation effect image

    • MORPH_OPEN

      • Name: Open Operation

      • Description: Erosion followed by dilation

      • Function: Removes small objects, smooths object contours, disconnects narrow connections, maintains large target shape

      • Application: Removes small noise points, separates slightly adhered objects, smooths edge burrs, text image denoising

        Open operation effect image

    • MORPH_CLOSE

      • Name: Close Operation

      • Description: Dilation followed by erosion

      • Function: Fills small holes, connects neighboring objects, smooths contour gaps, fuses narrow breaks

      • Application: Fills target internal holes, connects broken strokes, closes contour gaps, medical image repair

        Close operation effect image

    • MORPH_GRADIENT

      • Name: Gradient

      • Description: Dilation image minus erosion image

      • Function: Extracts object boundaries, edge detection, contour enhancement, thickness measurement

      • Application: Target contour extraction, alternative edge detection, industrial part dimension measurement, shape analysis preprocessing

        Gradient effect image

    • MORPH_TOPHAT

      • Name: Top Hat

      • Description: Original image minus opening operation

      • Function: Extracts bright small features, background correction, enhances small bright targets, uneven illumination processing

      • Application: Microscopic image small cell extraction, industrial inspection bright defect recognition, text image stroke enhancement, uneven illumination correction

        Top hat effect image

    • MORPH_BLACKHAT

      • Name: Black Hat

      • Description: Closing operation minus original image

      • Function: Extracts dark small features, detects holes and depressions, enhances small dark targets, shadow region analysis

      • Application: Industrial part surface defect detection, medical image blood vessel extraction, dark scratch and crack recognition, dark spot detection

        Black hat effect image

  • Set Input Image Type

    Select the type of input image, which determines which type of image the morphological operation acts on.

    • RGBImage: Performs morphological processing on color images, commonly used for removing small objects in color images, smoothing boundaries, etc.

    • DepthImage: Performs morphological processing on depth images, commonly used for removing noise in depth images, filling missing areas, etc.

    • MaskImage: Performs morphological processing on mask images, commonly used for correcting segmentation results, connecting broken areas, etc.

  • Operation Kernel Size

    The size of the operation kernel, in pixels (Pixel).

    Adjustment note: Set according to detection requirements. When the operation kernel is larger, the dilation-erosion effect is stronger.

  • Operation Kernel Shape

    Select the shape of the operation kernel.

    • MORPH_RECT: Rectangle.

    • MORPH_CROSS: Cross.

    • MORPH_ELLIPSE: Ellipse.

    Adjustment note: It is recommended to select the default value. If there are special circumstances, set according to actual requirements.