Image segmentation can be defined as a process of dividing the image into non-overlapping but still related and meaningful regions. We can say that the main objective of this process is to split an image into sections reliable for further analysis and processing in order to get feedback or data outcome with necessary segments of desirable information.

Image segmentation technology is been widely used in video surveillance, face recognition technology, medical image processing, traffic control systems, object detections, etc. It became a skyrocketing popular after the development of a smart cities concept because the main pillars of that concept will need image segmentation technology on a daily basis, and maybe become dependable on it. Current image segmentation methods and techniques are segmentations based on supervised learning, based on clustering, edge detection segmentation, and segmentation based on clustering. The simplest and most common method is called Threshold image segmentation. This kind of segmentation algorithm directly divides the image grayscale information processing based on gray values of different targets. This method can be split into 2 methods:

Global threshold method– This method divides the image into 2 regions of background and target by a single threshold

Local threshold method– This method implies the need for selecting the multiple segmentation threshold and divides the image into multiple targeted backgrounds and regions by multiple thresholds.

Except for the Threshold image segmentation model, there is another, commonly used method for image segmentation. It’s called a Regional growth method. We can say that is a typical region segmentation algorithm with a basic idea of having similar properties of pixels in order to form a region. Requirements for using this method are the selection of seed pixels and then merging the similar pixels around the seed pixel in order to create a region around the seed pixel location.