Have a personal or library account? Click to login
Topographic Correction of LAPAN-A3/LAPAN-IPB Multispectral Image: A Comparison of Five Different Algorithms Cover

Topographic Correction of LAPAN-A3/LAPAN-IPB Multispectral Image: A Comparison of Five Different Algorithms

Quaestiones Geographicae
Volume 39 (2020): Issue 3 (September 2020)
By: Zylshal Zylshal  
Open Access
|Aug 2020

Figures & Tables

Fig. 1

Study area.A – Overview of study area on South Sulawesi, Indonesia. B – LAPAN-A3 RGB composite of NIR-R-G, C – Landuse/landcover of the study area taken from visually interpreted SPOT-6 pansharpened iamge (1.5 meter) acquired at September 14th 2019. Red rectangle in left image indicates the area of interest. The right image is the LAPAN-A3 RGB composite of NIR-R-G.
Study area.A – Overview of study area on South Sulawesi, Indonesia. B – LAPAN-A3 RGB composite of NIR-R-G, C – Landuse/landcover of the study area taken from visually interpreted SPOT-6 pansharpened iamge (1.5 meter) acquired at September 14th 2019. Red rectangle in left image indicates the area of interest. The right image is the LAPAN-A3 RGB composite of NIR-R-G.

Fig. 2

The terrain elevation and its derivatives used in this study.A – ALOS World 3D 30 meter, B – Slope map, C – Aspect map, D – Histogram distribution of terrain elevation based on AW3D30, E – Histogram distribution of slopes, F – Histogram distribution of aspects.
The terrain elevation and its derivatives used in this study.A – ALOS World 3D 30 meter, B – Slope map, C – Aspect map, D – Histogram distribution of terrain elevation based on AW3D30, E – Histogram distribution of slopes, F – Histogram distribution of aspects.

Fig. 3

Geometric illustration shows all angles involved in calculating the incidence angle between normal to the ground and sunrays in slopped area. Image reproduced from Riaño et al. (2003).
Geometric illustration shows all angles involved in calculating the incidence angle between normal to the ground and sunrays in slopped area. Image reproduced from Riaño et al. (2003).

Fig. 4

A – NIR-R-B False color composite of LAPAN-A3 and after applying topographic correction algorithms: B – cosine correction, C – improved cosine correction, D – Minnaert correction, E – modifed Minnaert correction, and F – two-stage normalization.
A – NIR-R-B False color composite of LAPAN-A3 and after applying topographic correction algorithms: B – cosine correction, C – improved cosine correction, D – Minnaert correction, E – modifed Minnaert correction, and F – two-stage normalization.

Fig. 5

Scatter plots and regression lines between the pixel value and local illumination (cos γi) over Near-infrared band for 3942 taken from both Precorrected (A) and corrected image using: B – cosine correction (Teillet et al. 1982), C – modified-cosine correction (Civco 1989), D – Minnaert correction (Minnaert 1941), E – modified Minnaert correction (Riaño et al. 2003), and F – two-stages normalization (Civco 1989) modified by (Law & Nichol 2004).
Scatter plots and regression lines between the pixel value and local illumination (cos γi) over Near-infrared band for 3942 taken from both Precorrected (A) and corrected image using: B – cosine correction (Teillet et al. 1982), C – modified-cosine correction (Civco 1989), D – Minnaert correction (Minnaert 1941), E – modified Minnaert correction (Riaño et al. 2003), and F – two-stages normalization (Civco 1989) modified by (Law & Nichol 2004).

Statistics of pixel value before and after topographic correction_ A negative value indicates an increase in coefficient variation after correction_ Bold and highlighted value indicated the best performed algorithm (σ is the standard deviation, cv is the correlation coefficient, and R is the correlation coefficient)_

AlgorithmStatsBand
NIRRGB
PrecorrectedOrimean13340.198434.899639.64277.14
σ2070.843007.361579.32533.12
cv15.5235.6516.38192.37
Correctedcosine correctionmean14757.979210.3610686.45280.60
R0.690.290.650.15
σ3688.533444.892992.48540.96
cv24.9937.4028.00192.79
cvdifference−9.47−1.75−11.62−0.42
improved cosine correctionmean13061.568196.829467.50258.26
R0.650.140.600.14
σ2239.932773.341798.05500.44
cv17.1533.8318.99193.77
cvdifference−1.631.82−2.61−1.40
Minnaert correctionmean13929.208749.8610073.39278.58
R0.240.060.280.16
σ2019.772899.011615.29536.34
cv14.5033.1316.04192.53
cvdifference1.022.520.35−0.16
modified Minnaert correctionmean13542.258549.519819.09276.68
R0.200.080.210.16
σ1764.992921.021603.82534.63
cv13.0334.1716.33193.23
cvdifference2.491.490.05−0.86
two-stages normalizationmean13061.568196.829467.50258.26
R0.660.140.600.14
σ2239.932773.341798.05500.44
cv17.1533.8318.99193.77
cvdifference−1.631.82−2.61−1.40
DOI: https://doi.org/10.2478/quageo-2020-0021 | Journal eISSN: 2081-6383 | Journal ISSN: 2082-2103
Journal RSS Feed
Language: English
Page range: 33 - 45
Submitted on: Feb 25, 2020
Published on: Aug 5, 2020
Published by: Adam Mickiewicz University
In partnership with: Paradigm Publishing Services
Publication frequency: 4 issues per year
Keywords:
LAPAN-A3/LAPAN-IPB,
AW3D30,
Minnaert,
cosine correction,
two-stage normalization,
coefficient of variation
Related subjects:
Geosciences,
Geography

© 2020 Zylshal Zylshal, published by Adam Mickiewicz University
This work is licensed under the Creative Commons Attribution 3.0 License.

Previous articleVolume 39 (2020): Issue 3 (September 2020)Next article