[PPT] - Vinayaraj Poliyapram 1 , Venkatesh Raghavan 1 and Masumoto Shinji 2 1 PowerPoint Presentation

SLIDE 1

1

Geographical Weighted Regression Model for Improved Near-shore Water Depth Estimation from Multispectral Imagery

Vinayaraj Poliyapram1, Venkatesh Raghavan1 and Masumoto Shinji2

1Graduate School for Creative Cities, Osaka City University

3-3-138 Sugimoto, Osaka 558-8585, Japan Email: vinay223333@gmail.com

2Graduate School of Science, Osaka City University.

SLIDE 2

2

Presentation Outline:

INTRODUCTION

Optical Remote Sensing for near-shore depth Physical principles

MATERIALS AND METHOD

FOSS4G used (GRASS GIS, R and Qgis) Spectral radiance correction method Study area and data used Global regression model Motivation to geographically fitted models  Geographical Weighted Regression (GWR) model

RESULT AND DISCUSSION

Evaluation of estimated depth at different scenarios Evaluation by near-shore cross profiles

CONCLUSIONS

SLIDE 3

3

Source: Fugro Source: NOAA

Remote Sensing of Water Depth

LiDAR bathymetry Multibeam echo sounding

Available Bathymetry data 1.DBDB5 2.ETOPO5 3.ETOPO2 4.ETOPO1 5.GEBCO

SLIDE 4

4

Bathymetric Mapping using Multi-spectral Imagery

Advantages

Wide availability of data: IKONOS,

QuickBird, RapidEye, Landsat, etc

Relatively low cost
Large spatial coverage, high spatial

resolution

Derive water depth up to 20 m,

depending on water turbidity and atmospheric conditions

Disadvantages

Relatively low accuracy
Need water depth truth

SLIDE 5

5

Bathymetric Mapping using Multi-spectral Imagery

Physical Principle

Attenuation of light in water is a function of wavelength, at shorter wavelength region

in electromagnetic spectrum attenuates slowly and consistently increases the attenuation speed when wavelength increases.

Meanwhile, as depth increases light attenuates rapidly and separability of bottom type

spectra declines.

Schematic view of radiance components

bserved using the visible sensor over
ptically shallow water (Kanno et al., 2012)

Behavior of attenuation of light at near-shore water

SLIDE 6

6

Study Area

Fig.4. Puerto Rico, Northeastern Caribbean Sea

Puerto Rico

Study area

Puerto Rico

17 ° 49.2' 18 ° 5.4' 67° 18' 67 ° 00' 17 ° 49.2' 18 ° 5.4' 67° 18' 67 ° 00'

0 2 4 6 8 KM'

SLIDE 7

7

Data sets

Data Date of collection Resol ution Spectral bands used for estimation Band used for correction Landsat 8 25 November, 2013 30 M 0.43-0.45µm (coastal Aerosol) 0.45-0.51µm (blue) 0.53-0.59µm (green) 0.64-0.67µm (red) 0.85 - 0.88µm (NIR) 1.57-1.65µm

(SWI)

RapidEye 01 May, 2010 5 M 0.44-0.51µm (blue) 0.52-0.59µm(green) 0.63-0.69µm (red) 0.69-0.73µm (Red-edge) 0.76-0.85µm

(NIR)

LiDAR depth 12 November, 2006 4 M

SLIDE 8

8

Correction method to remove unwanted components

Lyzenga (1978; 1981; 1985) , Philpot (1989) and Kanno and Tanaka (2012)

Area of interest (Shallow water)

Landsat 8 SWI =1.57-1.65µm RapidEye NIR= 0.76-0.85µm

Where, L λi is the radiance value of selected region,α0 is the intercept and α1 is the slope of the regression.

Transformed Band (X(λ)i)

Transformed band (X(λ)i) = log ((L λi -α0 - α1 L(λSWI))/ L λi) Transformed band (X(λ)i)= log ((L λi -α0 - α1 L(λNIR))/ L λi)

SLIDE 9

9

Water Depth Retrieval Models

Global Regression Model

Only one regression equation is used for the entire image
The conventional regression models are a common practice by several

authors (Clark , (1987; 1988), Lyaenga, (2006) and Kanno and Tanaka, (2012).

Multi-band method

Depth=β0+ β1 X(λ)1+ β2 X(λ)2+…….+βnX(λ)i

β0, β1, β2… βn

+

Transformed bands (X(i)

LiDAR

Reality

heterogeneous bottom types and varying water quality

NIR band Red band Green band Blue band Coastal band

SLIDE 10

10

Motivation to the Geographically Fitted Models

Assessment of spatial propagation of heterogeneity

Residual map=Depth estimated by global regression model-LiDAR depth

Residual map of global regression model

Clear indication of different spatial clusters of residuals

Solution

R=0.88

LiDAR reference depth (m)

SLIDE 11

11

Motivation to the geographically fitted models

Assessment of spatial heterogeneity by classes

6 residual classes clustering

SLIDE 12

12

Motivation to the Geographically Fitted Models

R=0.94

Method R R2 RMSE (m) Global model 0.88 0.78 2.63 Class based model 0.94 0.90 1.70 Residual class based model

Depthclass=β0+ β1 X(λ)1+ β2 X(λ)2

+…….+βnX(λ)i

SLIDE 13

13

Geographically Weighted Regression (GWR Models)

Depth=β0(x,y)+β1(x,y) X1(x,y)+β2(x,y)+X2(x,y)+ +……..+βn(x,y)+ Xi(x,y)

Gaussian w i = exp(-0.5 * (d / bw)2)

d

i

x,y

(x, y): geographical coordinates of the centroid point for each local area

Kernel

Yellow points are LiDAR depth displayed

n Landsat 8

SLIDE 14

14

Results and comparison of models at various scenarios

SLIDE 15

15

Results and comparison of models at various scenarios

Scenario1

Randomly distributed LiDAR

depth points used for calibration

Landsat 8: 10,000 points
RapidEye: 60,000 point
Evaluation carried out with

separate depth points

Yellow points are LiDAR depth displayed on RGB (NIR,Red,Green) of Landsat 8 Satellite data Global model GWR model R R2 RMSE (m) R R2 RMSE (m) Landsat 8 0.88 0.78 2.63 0.97 0.93 1.41 RapidEye 0.88 0.78 2.48 0.97 0.93 1.34

SLIDE 16

16

Results and comparison of models at scenario1

Landsat 8 RapidEye Landsat 8 RapidEye

LiDAR depth m LiDAR depth m LiDAR depth m LiDAR depth m Estimated depth (Global model) m Estimated depth (Global model) m Estimated depth (GWR model) m Estimated depth (GWR model) m

LiDAR depth LiDAR depth

R=0.88 R=0.97 R=0.88 R=0.97

SLIDE 17

17

Results and comparison of models at various scenarios

Scenario2

LiDAR depth points are at

equal interval of 300m

Landsat 8: 1750 points
RapidEye: 1750 point
Evaluation carried out with

separate depth points

Yellow points are LiDAR depth displayed on RGB (NIR,Red,Green) of Landsat 8 Satellite data Global model GWR model R R2 RMSE (m) R R2 RMSE (m) Landsat 8 0.88 0.78 2.63 0.96 0.91 1.72 RapidEye 0.89 0.78 2.51 0.95 0.91 1.61

SLIDE 18

18

Results and comparison of models at scenario2

Landsat 8 RapidEye Landsat 8 RapidEye

LiDAR depth m LiDAR depth m LiDAR depth m LiDAR depth m Estimated depth (Global model) m Estimated depth (Global model) m Estimated depth (GWR model) m Estimated depth (GWR model) m

LiDAR depth LiDAR depth

R=0.96 R=0.88 R=0.89 R=0.95

SLIDE 19

19

Results and comparison of models at various scenarios

Scenario3

LiDAR depth points are at

equal interval of 600m

Landsat 8: 450 points
RapidEye: 450 points
Evaluation carried out with

separate depth points

Yellow points are LiDAR depth displayed on RGB (NIR,Red,Green) of Landsat 8 Satellite data Global model GWR model R R2 RMSE (m) R R2 RMSE (m) Landsat 8 0.88 0.78 2.62 0.94 0.88 1.91 RapidEye 0.88 0.78 2.48 0.93 0.87 1.87

SLIDE 20

20

Results and comparison of models at scenario3

Landsat 8 RapidEye Landsat 8 RapidEye

LiDAR depth m LiDAR depth m LiDAR depth m LiDAR depth m Estimated depth (Global model) m Estimated depth (Global model) m Estimated depth (GWR model) m Estimated depth (GWR model) m

LiDAR depth LiDAR depth

R=0.88 R=0.88 R=0.94 R=0.93

SLIDE 21

21

Results and comparison of models at various scenarios

Scenario4

Random LiDAR depth points

are at small portion of study area

Landsat 8: 3300 points
RapidEye: 20,000 points
Evaluation carried out with

separate depth points

Yellow points are LiDAR depth displayed on RGB (NIR,Red,Green) of Landsat 8 Satellite data Global model GWR model R R2 RMSE (m) R R2 RMSE (m) Landsat 8 0.83 0.68 3.49 0.87 0.75 2.86 RapidEye 0.88 0.78 2.60 0.90 0.81 2.26

SLIDE 22

22

Results and comparison of models at scenario4

Landsat 8 RapidEye Landsat 8 RapidEye

LiDAR depth m LiDAR depth m LiDAR depth m LiDAR depth m Estimated depth (Global model) m Estimated depth (Global model) m Estimated depth (GWR model) m Estimated depth (GWR model) m

LiDAR depth LiDAR depth

R=0.88 R=0.90 R=0.83 R=0.88

SLIDE 23

23

Results and comparison of models by using near- shore cross profiles spaced 2.5km

Global model LiDAR depth GWR model Global model LiDAR depth GWR model Global model LiDAR depth GWR model Global model LiDAR depth GWR model

SLIDE 24

24

Results and comparison of models by using near- shore cross profiles spaced 2.5km

Global model LiDAR depth GWR model

Global model LiDAR depth GWR model Global model LiDAR depth GWR model Global model LiDAR depth GWR model

SLIDE 25

25

12-bit dynamic range was provided by Landsat 8 and

RapidEye images were used to estimate good accuracy near- shore water depth.

Global regression model was used to estimate depth was not

fully able to address the spatial non-stationary introduced by various bottom types and water quality.

Meanwhile, GWR model was able to address the spatial non-

stationary introduced by various bottom types and water

quality. Therefore produces high accuracy depth estimation.
GWR models can give better and more consistent depth

estimates than the conventional global regression model.

Conclusions and future study

SLIDE 26

Thank you for your time

26