Crop

Overview

Crop node extracts a region of interest from the input frame using either a rectangular crop or a perspective-corrected crop.

Use this node to focus downstream processing on a specific area of the frame and reduce unnecessary computation, or to de-skew a region (for example a tilted document or angled label) into a flat rectangular output.

Input

Input Image

image required

The image frame to crop. Connect this to a camera or upstream processing output.

Crop Type

string

Determines which crop mode and input fields are used.

Values:

• cropBox (default) — rectangular crop defined by Crop Box.
• cropPerspective — perspective-corrected crop defined by Crop Perspective.

Crop Box

object

Used when Crop Type is cropBox. Defines the rectangular region to extract.

Fields (all in image pixel coordinates):

• x — left edge of the crop region.
• y — top edge of the crop region.
• width — width of the crop region.
• height — height of the crop region.

Crop Perspective

array

Used when Crop Type is cropPerspective. Defines the four corners of a quadrilateral in the source image that will be warped into a flat rectangle in the output.

Four points in image pixel coordinates, in order: top-left, top-right, bottom-right, bottom-left. Each point is [x, y].

Output

Frame Out

image

Cropped (and optionally de-skewed) image.