Morphological transformation in opencv

To detect the shape of any object morphological transformation is performed. Morphological transformation is performed on binary images and their objects must be in white pixel and background in black pixel.
It needs two inputs 1. original image 2. Kernel.

To basic morphological transformations are:
1. Erosion:
Erosion works on an image object's edges or you can say boundary .Erosion will fix or sharpen the boundary of any object present in the image. Suppose you have a binary image where the background color is black and the object color is white. Now the boundary is not properly visible or you can say that some objects don't have the proper boundary. It means some black pixels are on the border. In this type of case, erosion will try to remove the black pixel and will try to create a proper white border of that object.
2. Dialation:
Dialation works inside of an object. It means erosion fixes/sharp the damage boundary of the objects and dialation fix the damage present inside of an object.
Suppose you have a binary image where the background color is black and the object color is white. Now inside the object, we have some black pixels but there must be white pixels. For this reason, we can say that object inside the area is not properly visible. In this type of case, dialation will try to remove the black pixel and will try to create a proper white area inside of that object.

Erosion

Dialation

Let's see some morphological transformation practical using OpenCV