Page 283 - Contributed Paper Session (CPS) - Volume 2

CPS1857 Nicholas J. et al.

Land-Use change detection using least median of
squares regression
Nicholas Jovianto, Janson Hendryli, Dyah Erny Herwindiati
Tarumanagara University, Jakarta, Indonesia

Abstract
This paper proposes a simple least median of squares regression algorithm to
perform a high performance computation for classifying land types and
mapping the land-use changes in a certain period of time. The main land-use
changes classification process is done with a conventional two phase
processes: training and mapping process. The outcome of the training process
is the regression model that will be used for classifying the land-use changes.
The regression model is built from the least median of squares estimator from
each subset of data. The case study of this research is the Jabodetabek which
is a metropolitan area consisting of megacities surrounding Jakarta, using the
Landsat-8 satellite image bands 1, 2, 3, 4, and 5 as the input data. The
classification results or the land types consist of impervious and non-
impervious land (green land, water, and soil land). The mapping evaluation
shows that the least median of squares regression is a simple and effective
method for the land-use changes mapping of the Jabodetabek area.

Keywords
Land-use Change; Impervious; Non-Impervious; Least Median of Squares
Regression; Outlier.

1. Introduction
Urban development is a process of transferring land functions to meet
community needs for infrastructure and facilities in the development of the
city [1]. In the process, urban development often sacrifices the ecological value
of land to fulfil the needs of human life. This will eventually cause the land to
experience a change in function. Changes in land functions or land-use
changes can be interpreted as changes in the function of all or part of the land
area from its original function to another function that can have a certain
impact on the environment and the potential of the land itself.
One of the famous megapolitan areas in Indonesia is the Jakarta, Bogor,
Depok, Tangerang and Bekasi (Jabodetabek) area (Fig. 1). In the Jabodetabek
area, there have been many changes in land function throughout the year. This
research is conducted with the aim to observe and determine the pattern of
land-use change of the Jabodetabek area, especially in the period of 2014 to
2017.

272 | I S I W S C 2 0 1 9