Environmental study is crucial in order to understand deeply about the flora and fauna living in the Earth. Mangrove forest is a unique and natural ecosystem that can be used to produce forestry product such as charcoal, timber, supply food to their surrounding marine life, and protect the inland from disturbance like erosion, flood, and tsunami. Due to the uncontrolled planning of human activity, many mangrove forests had been deforested for development of industry area, urban land and agriculture. In this study, Multi-layer Feed Forward/Multilayer Perceptrons (MLP) network system in Artificial Neural Network technique was used to map out the current state of mangrove trees. This network system require user to have a ground truth data such as in supervised classification in order to generate the training area for classification. Generally, this network comprise of a simple structure layer which consist of three layers namely input layer, hidden layer and output layer. Multi-layer Feed Forward algorithm has at least one hidden layer of neuron between the input and output layer. Each successive layer of neurons is fully interconnected with connection weight determine the strength of the connection. 2010 Thailand Earth Observing System (THEOS) satellite imagery was used as the source for the data processing with the aid of PCI Geomatica version 10.3.2 software packaging. The classification result of Multi-layer Feed Forward yield 5 category of classes. Post-classification analysis was further carried out to validate the classified data with reference data. High overall accuracy of 93.5% and kappa coefficient of 0.900 was obtained for the mangrove cover mapping. Final thematic map was produce to quantify and display the current distribution of mangrove land. The result indicates that neural network approach is suitable and reliable used for mangrove mapping.
[1]
D. Alongi.
Present state and future of the world's mangrove forests
,
2002,
Environmental Conservation.
[2]
Zhigang Liu,et al.
Object-based classification for mangrove with VHR remotely sensed image
,
2007,
Geoinformatics.
[3]
Martti Hallikainen,et al.
Artificial neural network-based techniques for the retrieval of SWE and snow depth from SSM/I data
,
2004
.
[4]
Le Wang,et al.
Photogrammetric Engineering & Remote Sensing Neural Network Classification of Mangrove Species from Multi-seasonal Ikonos Imagery
,
2022
.
[5]
Nasser M. Nasrabadi,et al.
Hopfield network for stereo vision correspondence
,
1992,
IEEE Trans. Neural Networks.
[6]
Tim Madge,et al.
Theory and Practices
,
1989
.
[7]
Virginie Bouchard,et al.
Ecological engineering applied to river and wetland restoration
,
2002
.
[8]
Hongbo Shao,et al.
Applying Neural Network Classification to Obtain Mangrove Landscape Characteristics for Monitoring the Travel Environment Quality on the Beihai Coast of Guangxi, P. R. China
,
2010
.
[9]
Ryan R. Jensen,et al.
Spectral analysis of coastal vegetation and land cover using AISA+ hyperspectral data
,
2007
.
[10]
I. Kanellopoulos,et al.
Land-cover discrimination in SPOT HRV imagery using an artificial neural network - a 20-class experiment
,
1992
.