TITLE:
Application of Parametric and Non Parametric Classifiers for Assessing Land Use/Land Cover Categories in Cocoa Landscape of Juaboso and Bia West Districts of Ghana
AUTHORS:
Emmanuel Donkor, Edward Matthew Osei Jnr, Stephen Adu-Bredu, Samuel A. Andam-Akorful, Efiba Vidda Senkyire Kwarteng, Lily Lisa Yevugah
KEYWORDS:
Support Vector Machine, Random Forest, Artificial Neural Network, Maximum Likelihood, Image Classification, Cocoa Landscape
JOURNAL NAME:
Journal of Geoscience and Environment Protection,
Vol.10 No.11,
November
30,
2022
ABSTRACT: Satellite image classification has been used for long time in the field
of remote sensing since classification results are used in environmental
research, agriculture, climate change and natural resource management. The
cocoa landscape of Ghana is complex and diverse in nature, composing of mixture
of closed forest, open forest, settlements, croplands and cocoa farms which
make mapping the landscape difficult. The purpose of this research is to assess
and compare the classification performances of three machine learning
classifiers: Support Vector Machine (SVM), Random Forest (RF), Artificial
Neural Network (ANN) and a statistical classification algorithm: Maximum
Likelihood (ML) to know which classifier is best suited for mapping the cocoa
landscape of Ghana using Juaboso and Bia West districts of Ghana as study area.
A representative sampling approach was adopted to collect 1246 sample points
for the various Land Use/Land Cover (LULC) types. These sample points were
divided at random into 869 which form 70% for classification and 377 which
constitute 30% of the total sample points for validation. The Stacked sentinel-2
image, classification data and validation data storing the identities of the
LULC classes were imported in R to run supervised classification for each
classifier. The classification results show that the highest overall accuracy
and kappa statistics were produced by the support vector machine (86.47%,
0.7902); next is the artificial neural network (85.15%, 0.7700), followed by the
random forest (84.08%, 0.7559) and finally the maximum likelihood (78.51%,
0.6668). The final LULC map produced
under this study can be used to monitor cocoa driven deforestation especially
in the gazetted forest and game reserves. This map will also be very useful in
the national forest monitoring framework for the REDD + cocoa landscape project.