Land use segmentation using image edge detection, marker-controlled watershed and k-means clustering techniques

Chong, Zi Qing (2021) Land use segmentation using image edge detection, marker-controlled watershed and k-means clustering techniques. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

	Text (24 Pages) Land use segmentation using image edge detection, marker-controlled watershed and k-means clustering techniques.pdf - Submitted Version Download (412kB)
	Text (Full Text) Land use segmentation using image edge detection, marker-controlled watershed and k-means clustering techniques.pdf - Submitted Version Restricted to Repository staff only Download (8MB)

Abstract

Land use is one of the basic data in geographic information system (GIS). It is important for the planning to meet the needs of people and safeguard the future resources. Satellite image is an important source of input but it is difficult to segment the land based on different uses due to high resolution satellite data. Therefore, image processing is necessary to extract this basic information from satellite image as it is a method that use any algorithm to perform operation which can solve this problem. In this project, image segmentation is focused because it is defined as the process that partition or divide an image into homogeneous pixel groups (Verdonck, L., et. al., 2019). The main purpose of the project is to segment the Malacca land and then identify the different types of land use. There are many techniques for image segmentation thus several of them have been identified. Image Edge Detection, Marker-Controlled Watershed as well as K-means clustering are the techniques used to determine the contours of objects within the satellite image in this project. Then, evaluate the three (3) techniques of segmentation by conducting comparison on the output of each technique. The best technique is identified after the analysis conducted.

Actions (login required)

View Item