Estimation of vegetation density using image analysis and its effect on hydraulic characteristics of an open channel
Vegetation density is one of the factors that affect the flow behaviour and resistance in vegetated channels and wetlands. Many approaches have been explored to quantify the vegetation density such as by counting the number of vegetation or determining the area covered by the vegetation per un...
Saved in:
Main Author: | |
---|---|
Format: | Thesis |
Language: | English |
Published: |
2015
|
Online Access: | http://psasir.upm.edu.my/id/eprint/67776/1/fk%202015%20113%20ir.pdf http://psasir.upm.edu.my/id/eprint/67776/ |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Summary: | Vegetation density is one of the factors that affect the flow behaviour and
resistance in vegetated channels and wetlands. Many approaches have been
explored to quantify the vegetation density such as by counting the number
of vegetation or determining the area covered by the vegetation per unit area.
However, in reality, the aquatic vegetation in the field is diverse in types and
has varying properties, and if the vegetation is submerged, the vegetation
density estimation becomes even more challenging. The use of remote
sensing imagery for vegetation mapping is now gaining popularity due to the
rapid development of remote sensing technology and readily available
remote sensing imageries from various sources. In this study, the use of
satellite image of a PLEIADES is explored to estimate the vegetation density
in the Putrajaya Wetlands. Arc Map 10.1 software is used for data entry,
image display and output. The vegetation type classification is derived using
the Supervised Maximum Likelihood Classification and Support Vector
Machine. The environment for Visualizing Images (ENVI) software is used to
derive the Normalised Difference Vegetation Index (NDVI) for the selected
study area and the area covered by the vegetation for different NDVI. NDVI
is the ratio between the maximum absorption of radiation in the red (R)
spectral band (0.66 μm) versus the maximum reflection of radiation in the
near infrared (NIR) spectral band (~0.83 μm). The percentage area covered by
vegetation obtained through ENVI software is then validated with ground
truth. In the field survey, areas with different densities are chosen and
divided into small cells of 1 m2. The percentage area covered by vegetation
for each cell is estimated by observation. Then, the relationship between the
percentage area covered by the vegetation and NDVI is established and it is
found that the NDVI has a polynomial relationship with the percentage area
covered by vegetation. Also, the relationship between NDVI and hydraulic
parameters in the wetland (velocity and resistance coefficient) is derived. The velocity at various locations of different densities in the wetland is measured
on site using the Acoustic Doppler Velocitimeter (ADV).
From the image analysis and ground truth validation, a map of vegetation
distribution based on types for the selected zone of Putrajaya Wetland has
been produced. It has been found that the land cover classes in the study area
fall into four classifications; that are the Hanguana Malayana, Phragmites
Karka, Scirpus Grossus and water bodies. The accuracy measured using
matrix confusion for the overall classification accuracy for Maximum
Likelihood is 78.81% and Kappa coefficient is 0.7187% and for Support
Vector Machine 81.36% for overall and 0.7512 for kappa coefficient. In
addition, the NDVI values representing the vegetation distribution and
density have also been generated and the NDVI values for the selected area
are found to be in the range of -0.1058 - 0.7825. The velocity at various
locations of different densities in the wetland showed that the density of
vegetation had a significant effect on the velocity. In addition, it was
observed that the flow resistance increases with the increasing in the density,
which is also showed in the increasing of NDVI values. |
---|