Extraction of Impervious Surface Areas from Remotely...
Transcript of Extraction of Impervious Surface Areas from Remotely...
Chuphan Chompuchan (蘇潘) 8101042007
Dept. of Soil and Water Conservation
National Chung Hsing University
Seminar IVFall Semester
(02 Oct. 2015)
Introduction
What is Impervious Surface Area (ISA)?
Why ISA is important?
Review methods for extracting ISA
Linear spectral mixture analysis (LSMA)
Artificial neural networks (ANN)
Object based image analysis (OBIA)
Spectral indices approaches
Case study of Puli township (preliminary analysis)
Discussion and conclusion 2
3
Impervious Surface Area (ISA)
Land surface which water cannot infiltrate
Artificial structures covered by impenetrable
materials such as asphalt, concrete, brick and stone
Mainly are building rooftops and pavement (roads,
highways, sidewalks and parking lots)
4
ISA indicator of the degree of urbanization
and environmental quality
Hydrological impact (urban runoff, GW recharge)
Water quality (non-point source pollution)
Urban heat island effect (sensible and latent heat fluxes)
5
Linear spectral mixture
analysis (LSMA)
Assumes that the
spectrum of each pixel is a
linear combination of the
“endmembers” spectra
Ridd (1995) proposed
Vegetation–Impervious
surface–Soil (V–I–S)
model for urban
ecosystem analysis
6
Each endmember represents
a pure land cover type
Endmembers selection
Maximum Noise fraction (MNF)
Tasseled Cap (TC)
LSMA can be expressed as
C. Deng & C. Wu (2012)D. Lu, et.al. (2008)
Rj = the reflectance for each band j
N = the number of endmembers
fi = the fraction of endmember i
Rij = the reflectance of endmember i in band jej = the unmodeled residual
7
M.C. Huang (2002)
Salt Lake City area, USA
LANDSAT image
LANDSAT ETM+ V-I-S model
Percentages of selected urban features in a V-I-S diagram
8
Artificial neural networks
(ANN)
Multilayer perceptron (MLP)
network is structured with 3
types of layers: input,
hidden, and output layers.
The back-propagation (BP)
learning algorithm is a key
to the success of an ANN
model require fewer
training samples & yield the
best result3-layer neural-network structure
9
Weng Q. and X. Hu (2008)
Indianapolis/Marion County, IN, USA
Terra’s ASTER and Landsat ETM+
Comparison between LSMA and ANN
Sample selection of
impervious surfaces
10
ASTER LSMA
ASTER ANN ETM+ ANN
ETM+ LSMA
R2 = 0.6623
R2 = 0.6036R2 = 0.7707
R2 = 0.7713
Weng Q. and X. Hu (2008)
11
Object based image analysis (OBIA)
From per-pixel image to object (segmentation) which
concerning color and shape homogeneity.
How to choose an optimal scale parameter and shape
factor is critical to the quality of classification
image object
12
OBIA vs. Maximum likelihood classification
Minnesota State University, Mankato campus
Quickbird imagery acquired on October 6, 2003
F. Yuan (2006)
Object orientedMaximum likelihood
Overall accuracy = 87.1% Kappa = 0.804
Overall accuracy = 92.5% Kappa = 0.887
13
Y. Deng et al. (2015)
0.42-0.52 0.53-0.61 0.63-0.69 0.78-0.90 1.55-1.75 2.09-2.35
Visible (BLUE)
Visible (GREEN)
Visible (RED)
Near-Infrared(NIR )
Short WaveInfrared (SWIR1)
Short WaveInfrared (SWIR2)
LANDSAT TM/ETM+ Spectral band specification
14
R. C. Estoque and Y. Murayama(2015)
LANDSAT Spectral band specification
*Thermal Infrared (TIR)
at-satellite brightness temperature (K)
15
NDBI: Normalized
Difference Built-up Index
NDBaI: Normalized
Difference Bareness Index
IBI: Index-based Built-up Index
EBBI: Enhanced Built-Up
and Bareness Index
Zha et al. (2003)
Zhao&Chen (2005)
Xu (2008)
As-Syakur et al. (2012)
NIR1SWIR
NIR1SWIRNDBI
TIR1SWIR
TIR1SWIRNDBaI
SWIR1)G/(G R) NIR/(NIR NIR) WIR12xSWIR1/(S
SWIR1)G/(G R) NIR/(NIR - NIR) WIR12xSWIR1/(SIBI
TIR1SWIR10
NIR1SWIREBBI
UI: Urban Index Kawamura et al. (1996) NIR2SWIR
NIR2SWIRUI
NBI: New Built-Up Index Jieli et al. (2010)NIR
R1SWIRNBI
16
As-Syakur et al. (2012)
Denpasar (Bali, Indonesia)
Landsat ETM+
17
As-Syakur et al. (2012)
18
As-Syakur et al. (2012)
Varshney & RajeshAs-Syakur (2014)
Delhi northern capital region of India
Landsat-5 TM
19
As-Syakur et al. (2012)
R.C. Estoque and Y. Murayama (2015)
The optimal threshold t* for image segmentation (the Otsu method)
Limitation of the index value for each type of index transformation
20
Universal Soil Loss Equation (USLE)
A = R x K x LS x C x Prainfall
erodibility factorsoil erodibility
factortopographic
factor cover factor
supporting practice factor
C-factor will range
between 1 and almost 0
C=1 means no cover
effect and a soil loss
comparable to that from
a tilled bare fallow
C=0 means a very
strong cover effect
resulting in no erosionPuli township
21
NDVI : Normalized Difference Vegetation Index
1
RNIR
RNIRNDVI
0
-1
Forest
Grassland/ Agriculture
Bare land/ ISA
Water/ Cloud
NDVI could not separate between
bare land and ISA
Bare land = high erosion (C=1)
ISA = no erosion (C=0)
= 2
= 1
Relationship between NDVI and C-Factor (Van der Knijff et al., 1999)
22
C-Factor derived from NDVI C-Factor with ISA* & water body** masking
1
0
* ISA extraction from IBI (Index-based
Built-up Index)
** Water body extraction from MNDWI
(Modified Normalized Difference Water Index)
Average C-Factor
No masking = 0.041
Masking = 0.024
23
Some coverages were classified as ISA
Jiu Fen Er Shan 九份二山
LSMA, ANN and OMIA required training samples
LSMA suitable for medium/low resolution imagery
OMIA suitable for high resolution imagery
24
D. Liu & F. Xia (2010)
Limitation of spectral indices approach
Detect shadow on high resolution imagery
Seasonal sensitivity for estimating impervious surfaces
Optimum threshold value needed
As-Syakur A.R., I. Adnyana, I.W. Arthana and I.W. Nuarsa (2012). Enhanced Built-up and Bareness Index (EBBI) for mapping built-up
and bare land in an urban area. Remote Sensing. 4: 2957–2970.
Deng Y., C. Wu, M. Li and R. Chen (2015). RNDSI: A ratio normalized difference soil index for remote sensing of urban/suburban
environments. International Journal of Applied Earth Observation and Geoinformation, 39: 40–48.
Estoque R.C. and Y. Murayama (2015). Classification and change detection of built-up lands from Landsat-7ETM+ and Landsat-8
OLI/TIRS imageries: A comparative assessment of various spectral indices. Ecological Indicators, 56: 205–217
Hung M. C. (2002) Urban land cover analysis from satellite images. In: Pecora15/Land Satellite Information IV/ISPRS Commission
I/FIEOS 2002 Conference Proceedings. Denver, CO, USA.
Jieli C., L. Manchun, L. Yongxue, S. Chenglei and H. Wei (2010). Extract residential areas automatically by new built-up index. In:
Proceeding of 18th International Conference on Geoinformatics (Geoinformatics 2010).
Kawamura M., S. Jayamana and Y. Tsujiko (1996). Relation between social and environmental conditions in Colombo Sri Lanka and the
urban index estimated by satellite remote sensing data. International Achieves of Photogrammetry Remote Sensing, 31 (Part B7):
321–326.
Liu D. and F. Xia (2010) Assessing object-based classification: advantages and limitations. Remote Sensing Letters, 1(4): 187-194.
Lu D., K. Song, L. Zeng, D. Liu, S. Khan, B. Zhang, Z. Wang and C. Jin (2008). Estimating Impervious Surface for the Urban Area
Expansion: Examples from Changchun, Northeast China. In: The International Archives of the Photogrammetry, Remote Sensing
and Spatial Information Sciences, Vol. XXXVII, Part B8, pp. 385-391.
Ridd M. (1995). Exploring a V–I–S (vegetation–impervious–surface-soil) model for urban ecosystem analysis through remote sensing:
Comparative anatomy for cities. International Journal of Remote Sensing, 16: 2165–2185.
Varshney A. and E. Rajesh (2014). A Comparative study of Built-up Index approaches for automated extraction of built-up regions from
remote sensing data. Journal of Indian Society of Remote Sensing, 42(3):659–663
Weng Q. and X. Hu (2008). Medium spatial resolution satellite imagery for estimating and mapping urban impervious surfaces using
LSMA and ANN. IEEE Transaction on Geosciences and Remote Sensing, 46(8): 2397–2406.
Xu H. (2008). A new index for delineating built-up land features in satellite imagery. International Journal of Remote Sensing, 29(14):
4269–4276.
Yuan F. and M. E. Bauer (2006). Mapping impervious surface area using high resolution imagery: A comparison of object-based and per
pixel classification. In: American Society for Photogrammetry and Remote Sensing Annual Conference Proceedings (ASPRS),
Reno, Nevada.
Zhao H.M. and X.L. Chen (2005) Use of Normalized Difference Bareness Index in quickly mapping bare areas from TM/ETM+. In:
Proceedings of 2005 IEEE International Geoscience and Remote Sensing Symposium, Seoul, Korea, Volume 3, pp. 1666−1668.25
26