dilate

Usage

Dilates images. It is widely used for feature enhancement and shape analysis in image processing and computer vision. Dilatation is a morphological operation that can enhance the bright area in an image. It is usually used to remove small noise, fill holes in an object, or connect adjacent objects. This operation expands the border of the bright area using structure elements (convolution kernel).

Interface

cv2.dilate(src, kernel, anchor=None, iterations=1, borderType=cv2.BORDER_CONSTANT, borderValue=None)

Parameters

Parameter	Description	Value Range	Input/Output
src	Input image (usually a binary or grayscale image).	Not null	Input
kernel	Structure elements (usually a matrix) used for dilation.	Not null	Input
anchor	Anchor, the center of structure elements. The default value is none.	Tuple type in the (x, y) format x: horizontal offset. Value range: [0, kernel_width–1] y: vertical offset. Value range: [0, kernel_height–1] kernel_width and kernel_height are determined by the kernel parameter.	Input
iterations	Number of dilation iterations. The default value is 1. Increasing the value further dilates the image.	Non-negative integer (usually 1 or larger)	Input
borderType	Border type. The default value is cv2.BORDER_CONSTANT. KPCV supports only the BORDER_CONSTANT type.	cv2.BORDER_* constant, such as cv2.BORDER_CONSTANT, cv2.BORDER_REPLICATE, and cv2.BORDER_REFLECT	Input
borderValue	Constant value for border padding. The default value is 0. This parameter is valid only when borderType is set to BORDER_CONSTANT.	Float or int number	Input

Return Values

Success: KP_CV_SUCCESS
Failure: an error code

Error Codes

Error Code	Description
INVALID_PARAM_MSG	The src parameter has a null pointer.
NOT_SUPPORT_MSG	src is not a 3-channel uint8 image; the input image is incorrect; the type is not supported by borderType; or the input kernel is incorrect.

Example

import numpy as np 
import cv2
# Create a 5x5 image.
src = np.array([[0, 0, 0, 0, 0],
                [0, 1, 1, 1, 0],
                [0, 1, 0, 1, 0],
                [0, 1, 1, 1, 0],
                [0, 0, 0, 0, 0]], dtype=np.uint8)
# Kernel of the dilation operation
kernel = np.ones((3, 3), np.uint8) 
dst = cv2.dilate(src, kernel, iterations=1)

print(dst)

Output:

[[1 1 1 1 1]
 [1 1 1 1 1]
 [1 1 1 1 1]
 [1 1 1 1 1]
 [1 1 1 1 1]]

Parent topic: Interface Definition