Classification Sensing Image of Remote Using Landsat 8 through Unsupervised Classification Technique (Case Study of Bangkalan Regency)
0 likes56 views
This study analyzes Landsat 8 satellite imagery of Bangkalan Regency, Indonesia, to classify land use and vegetation coverage using an unsupervised classification technique. The results indicate that rice field vegetation predominates, with significant areas of vacant land and limited residential zones. The research highlights the potential for agricultural development in the region and suggests future studies to assess changes in land usage over time.
![International Journal of Advanced Engineering, Management and Science (IJAEMS) [Vol-3, Issue-8, Aug- 2017]
https://dx.doi.org/10.24001/ijaems.3.8.10 ISSN: 2454-1311
www.ijaems.com Page | 878
Classification Sensing Image of Remote Using
Landsat 8 through Unsupervised Classification
Technique (Case Study of Bangkalan Regency)
Rosida Vivin Nahari1
, Riza Alfita2
Department of Engineering, Trunojoyo University, Indonesia
Abstract— Bangkalan regency is classified as a new
regency which is located East Java, Indonesia. This
regency possesses several potential areas in agriculture,
plantation, and fishery. This research employs image
analysis process of remote sensing satellite Landsat 8 in
Bangkalan regency. This research uses Landsat 8 satellite
image processing method from image data collection
stage to classification stage by using unsupervised
classification technique. This method produces land
appearance, such as agriculture, ponds, and settlements
in Bangkalan regency. This research classification result
can be used as a reference of vegetation coverage in
Bangkalan regency. Based on the research result, rice
field vegetation is very dominant compared to other areas
in Bangkalan Regency. Rice field vegetation coverage is
much more dominant than other coverage such as
residential area. The main objective of this study is to
obtain the scale of comparison or area percentage in
Bangkalan.
Keywords— Landsat 8, remote sensing, unsupervised
classification
I. INTRODUCTION
One of the districts in East Java is Bangkalan regency
which is geographically located in the westernmost part
of Madura Island. Bangkalan Regency possess an area of
1,260.14 km2 is located between 112 ° 40 '06 "- 113 ° 08'
04" East Longitude and 6 ° 51 '39 "- 7 ° 11' 39" South
Latitude. The Regency is adjacent to Sampang Regency
in the east, Madura Strait / Gresik Regency in the west,
the Java Sea in the north and Madura Strait / Surabaya
City in the south. Bangkalan Regency administratively
consists of 18 districts, 273 villages and 8 hamlets.
Bangkalan is located at one end of Madura island. It is a
very profitable location because it is adjacent to Surabaya
city which is a trading center in East Java. Bangkalan
Regency is a development area of Kertasusila Gate and is
included in Surabaya City Development or better known
as Surabaya Urban Development Policy. With the
construction of the Suramadu bridge connecting the land
routes between Surabaya and Bangkalan as well as the
international seaports and container terminals, Bangkalan
possesses positive impact on economic development
especially investment in Bangkalan Regency.
The agricultural sector plays an important role in
Bangkalan regency development. Leading commodities
was acquired from food crops such as rice, corn, cassava,
peanuts, sweet potatoes, soybeans, and green beans.
Potential production of horticultural crops is also quite
prominent. These include fruits, vegetables, and
medicinal plants. This regency possesses the potential for
planting superior commodities or crops such as coconut,
cashew nut, kapok, and areca nut.
Bangkalan regency is one of rice producing districts in
East Java. BPS Bangkalan regency notes that rice
production in Bangkalan regency in 2013 is 132,901
Tonnes. In addition, BPS also noted that in 2015 the
population of Bangkalan Regency is 954.305 people with
growth rate 2014-2015 of 0.9% per year (BPS East Java,
2015).
Based on BPS data, Bangkalan District outline condition
was observed. To ensure the geographical monitoring of
the area, satellite images data processing were taken
directly from Landsat 8 using remote sensing method.
The method produced geographical vegetation
appearance. Landsat 8 satellite image data processing
utilized Unsupervised Classification technique, which
aims to classify division of element classes or t land cover
types such as; Urban, water bodies, wetlands, etc. This
method had obtained how land cover types in the
concerned area. Wiradisastra and Noviar stated that
satellite image data processing accuracy depends on
respective reviewed region’s field condition, area, and
characteristics (Wiradisastra & Noviar, 2005).
Another study stated image processing analysis accuracy
is supported by the amount of cloud covering a study area
(Hanindito, Tanaamah, & Papilaya, 2010).
Remote sensing is the science of observing and gathering
information regarding objects on earth’s surface, using
certain sensors, without direct contact with the observed
objects (Andree Ekadinata, Sonya Dewi, Danan Prasetyo
Hadi, Dudy Kurnia Nugroho, 2008).](https://image.slidesharecdn.com/10classificationsensingimageofremote-170906171759/85/Classification-Sensing-Image-of-Remote-Using-Landsat-8-through-Unsupervised-Classification-Technique-Case-Study-of-Bangkalan-Regency-1-320.jpg)
![International Journal of Advanced Engineering, Management and Science (IJAEMS) [Vol-3, Issue-8, Aug- 2017]
https://dx.doi.org/10.24001/ijaems.3.8.10 ISSN: 2454-1311
www.ijaems.com Page | 879
II. RESEARCH METHOD
The following Figure 1 describes research stages.
Composite Band
654
Sharpening
Image from
Landsat 8 (OLI)
Image of
Landsat 8
Band 8
Cropping
Shapefile
Boundary
Unsupervised
Classification
Reclassification
(4 class)
ValidationReal Data
Land Use Map of
Madura 2015
.
Fig. 1: Research Stage Chart
2.1 Image Conversion
Downloaded Landsat 8 satellite image is satellite images
consisting of multiple bands of satellite sensor recordings.
The image is are in* .tiff file extension and cannot be
analyzed. Therefore image conversion was conducted to
facilitate analysis process. The conversion was conducted
by combining between several bands of an image in an
appearance with * file extension.
The following is a list of R G B sequences of band
combinations in the available Landsat 8 Satellite Imagery:
Table.1: band combinations in the available Landsat 8
BAND
COMBINATI
ON
FUNCTION /
IMPLEMENTATION
Natural Color
– 4 3 2
Produce an image with true color
False Color Produce images with distinct
(urban) –
7 6 4
differences in urban areas
Color Infrared
(vegetation) –
5 4 3
Used to see the mass, density, and
dominance of vegetation. The contrast
between vegetation dominance will be
seen in infrared, making it effective for
large-scale forestry or agricultural
vegetation analyzes
Agriculture –
6 5 2
Produce an image with distinct
vegetation indicated by greenish color
Atmospheric
Penetration –
7 6 5
Clarify the image of cloud thickness,
clarify shoreline, and vegetation cover.
This combination can clarify the image
from weather disturbance
Healthy
Vegetation –
5 6 2
Produce an image that reveals healthy
vegetation
Land/Water –
5 6 4
Produces images with distinct
differences on water and land regions.
Natural With
Atmospheric
Removal –
7 5 3
Produce images with natural colors and
reduce the appearance of clouds
Shortwave
Infrared –
7 5 4
To get the biomass with clear contrast
and cleaner images of cloud cover
Vegetation
Analysis –
6 5 4
Used to analyze plants
This study utilized Landsat satellite images RGB 654
composites. Three bands included were near-infrared and
visible spectra. These have wavelengths corresponded to
4, 5 and 3 bands wavelengths on satellite images Landsat
7 ETM +.
Fig. 2:band 6, band 5 and band 4 combination
2.2 Image Contrast Improvement
This process is done in order to obtain images possessing
color quality similar to the original appearance on the
earth surface and support the following image
classification process. This process is aimed at providing
a sharper coloring in an endeavor in enabling easier image](https://image.slidesharecdn.com/10classificationsensingimageofremote-170906171759/85/Classification-Sensing-Image-of-Remote-Using-Landsat-8-through-Unsupervised-Classification-Technique-Case-Study-of-Bangkalan-Regency-2-320.jpg)
![International Journal of Advanced Engineering, Management and Science (IJAEMS) [Vol-3, Issue-8, Aug- 2017]
https://dx.doi.org/10.24001/ijaems.3.8.10 ISSN: 2454-1311
www.ijaems.com Page | 880
classification process. Figure 3 illustrates the process of
improving image contrast.
Fig. 3: raster data and vector data
2.3 Image Cropping
Process on cropping study area images is conducted on
this stage. This process aims to facilitate the analysis
process by focusing study area by eliminating regions
unused in research. This process was conducted by
combining raster data (satellite image data) with data
vector which is administrative data of Bangkalan Regency
borders.
2.4 Image Classification
This process is a review of image appearance based on
the visible phenomenon. The image is generated and
analyzed using true color composite terminology or image
appearance in accordance with its original appearance on
earth surface. Classification process was conducted by
distinguishing each color contained in the image. Figure 5
illustrations will illustrate the appearance of satellite
imagery and its classification.
III. RESULTS AND DISCUSSION
This research produces categorized images, and
determined vegetation classification contained in
Bangkalan Regency, East Java. Figure 6 exhibits Landsat
8 satellite image data processing classification.
Fig. 4: image cropping results
Fig. 5: analysis of unsupervised classification image
imagery
Fig. 6: map of reclassification validation results
Fig. 7: image classification legends
Forest vegetation is dominant in Bangkalan Regency.
Based on image classification, Bangkalan district is
dominated by rice field area, with small regions of
housings. Followed by several other imagery
appearances. The following table exhibit Bangkalan
Regency’s Landsat 8 satellite image classification
analysis and graphs on the landscape appearance.](https://image.slidesharecdn.com/10classificationsensingimageofremote-170906171759/85/Classification-Sensing-Image-of-Remote-Using-Landsat-8-through-Unsupervised-Classification-Technique-Case-Study-of-Bangkalan-Regency-3-320.jpg)
![International Journal of Advanced Engineering, Management and Science (IJAEMS) [Vol-3, Issue-8, Aug- 2017]
https://dx.doi.org/10.24001/ijaems.3.8.10 ISSN: 2454-1311
www.ijaems.com Page | 881
Table.2: Landscape Appearance of Area Percentage
Classification Result
No Image Total Area Percentage
1 Dyke & Lake 6.106448
2 Farming 51.301141
3 Vacant 28.38344
4 Housing 14.208971
Fig. 8: Bangkalan Regency Land Acquisition Percentage
Graph
IV. CONCLUSION
Based on Landsat 8 satellite image observation, almost
every Bangkalan district consists of vacant land. This
exhibits land untouched by human hands to be processed
into productive land such as plantation and agricultural
land clearing. Satellite images were taken on June 16,
2015. It is suggested for future research to compare the
results of previous year's image analysis, to find out the
extent agricultural land area extension by utilizing vacant
land empowerment.
.
REFERENCES
[1] Alonso, W. (1964). Location and land use. Toward a
general theory of land rent. Location and land use.
Toward a general theory of land rent.
[2] Cohen, W. B., Fiorella, M., Gray, J., Helmer, E., &
Anderson, K. (1998). An efficient and accurate
method for mapping forest clearcuts in the Pacific
Northwest using Landsat imagery. Photogrammetric
engineering and remote sensing, 64(4), 293-299.
[3] Hanindito, G. A., Tanaamah, A. R., & Papilaya, F. S.
(2010). Pengolahan Data Citra Satelit Landsat TM
Dalam Pemantauan Area Kebakaran Hutan Berbasis
GIS (Studi Area Kecamatan Arut Utara dan Seruyan
Tengah, Propinsi Kalimantan Tengah).Universitas
Kristen Satya Wacana.
[4] Turner, B. L., & Meyer, W. B. (1991). Land use and
land cover in global environmental change:
considerations for study. International Social Science
Journal, 43(130), 669-679.
[5] Statistik, B. P. (2015). Statistik kesejahteraan rakyat.
Biro Pusat Statistik.
[6] Wiradisastra, & Noviar, H.( 2005).Kemampuan
Interpretasi Kebun Semangka Dari Citra Satelit
Landsat-7 ETM+. Pertemuan Ilmiah Tahunan
MAPIN XIV, (September), 132–140.
[7] Andree Ekadinata, Sonya Dewi, Danan Prasetyo
Hadi, Dudy Kurnia Nugroho, F. J. (2008). Sistem
Informasi Geografis Untuk Pengelolaan Bentang
Lahan Berbasis Sumber Daya Alam. (1st ed., p. 70)
Bogor.
[8] Lindeijer, E. (2000). Review of land use impact
methodologies. Journal of Cleaner Production, 8(4),
273-281.
[9] Lillesand, T., Kiefer, R. W., & Chipman, J.
(2014). Remote sensing and image interpretation.
John Wiley & Sons.
[10]Woodcock, C. E., Allen, R., Anderson, M., Belward,
A., Bindschadler, R., Cohen, W., ... & Nemani, R.
(2008). Free access to Landsat imagery.
6%
51%
29%
14%
PERCENTAGE OF AREA
Tambak
Lahan
Kosong
Sawah](https://image.slidesharecdn.com/10classificationsensingimageofremote-170906171759/85/Classification-Sensing-Image-of-Remote-Using-Landsat-8-through-Unsupervised-Classification-Technique-Case-Study-of-Bangkalan-Regency-4-320.jpg)
Classification Sensing Image of Remote Using Landsat 8 through Unsupervised Classification Technique (Case Study of Bangkalan Regency)
- 1. International Journal of Advanced Engineering, Management and Science (IJAEMS) [Vol-3, Issue-8, Aug- 2017] https://dx.doi.org/10.24001/ijaems.3.8.10 ISSN: 2454-1311 www.ijaems.com Page | 878 Classification Sensing Image of Remote Using Landsat 8 through Unsupervised Classification Technique (Case Study of Bangkalan Regency) Rosida Vivin Nahari1 , Riza Alfita2 Department of Engineering, Trunojoyo University, Indonesia Abstract— Bangkalan regency is classified as a new regency which is located East Java, Indonesia. This regency possesses several potential areas in agriculture, plantation, and fishery. This research employs image analysis process of remote sensing satellite Landsat 8 in Bangkalan regency. This research uses Landsat 8 satellite image processing method from image data collection stage to classification stage by using unsupervised classification technique. This method produces land appearance, such as agriculture, ponds, and settlements in Bangkalan regency. This research classification result can be used as a reference of vegetation coverage in Bangkalan regency. Based on the research result, rice field vegetation is very dominant compared to other areas in Bangkalan Regency. Rice field vegetation coverage is much more dominant than other coverage such as residential area. The main objective of this study is to obtain the scale of comparison or area percentage in Bangkalan. Keywords— Landsat 8, remote sensing, unsupervised classification I. INTRODUCTION One of the districts in East Java is Bangkalan regency which is geographically located in the westernmost part of Madura Island. Bangkalan Regency possess an area of 1,260.14 km2 is located between 112 ° 40 '06 "- 113 ° 08' 04" East Longitude and 6 ° 51 '39 "- 7 ° 11' 39" South Latitude. The Regency is adjacent to Sampang Regency in the east, Madura Strait / Gresik Regency in the west, the Java Sea in the north and Madura Strait / Surabaya City in the south. Bangkalan Regency administratively consists of 18 districts, 273 villages and 8 hamlets. Bangkalan is located at one end of Madura island. It is a very profitable location because it is adjacent to Surabaya city which is a trading center in East Java. Bangkalan Regency is a development area of Kertasusila Gate and is included in Surabaya City Development or better known as Surabaya Urban Development Policy. With the construction of the Suramadu bridge connecting the land routes between Surabaya and Bangkalan as well as the international seaports and container terminals, Bangkalan possesses positive impact on economic development especially investment in Bangkalan Regency. The agricultural sector plays an important role in Bangkalan regency development. Leading commodities was acquired from food crops such as rice, corn, cassava, peanuts, sweet potatoes, soybeans, and green beans. Potential production of horticultural crops is also quite prominent. These include fruits, vegetables, and medicinal plants. This regency possesses the potential for planting superior commodities or crops such as coconut, cashew nut, kapok, and areca nut. Bangkalan regency is one of rice producing districts in East Java. BPS Bangkalan regency notes that rice production in Bangkalan regency in 2013 is 132,901 Tonnes. In addition, BPS also noted that in 2015 the population of Bangkalan Regency is 954.305 people with growth rate 2014-2015 of 0.9% per year (BPS East Java, 2015). Based on BPS data, Bangkalan District outline condition was observed. To ensure the geographical monitoring of the area, satellite images data processing were taken directly from Landsat 8 using remote sensing method. The method produced geographical vegetation appearance. Landsat 8 satellite image data processing utilized Unsupervised Classification technique, which aims to classify division of element classes or t land cover types such as; Urban, water bodies, wetlands, etc. This method had obtained how land cover types in the concerned area. Wiradisastra and Noviar stated that satellite image data processing accuracy depends on respective reviewed region’s field condition, area, and characteristics (Wiradisastra & Noviar, 2005). Another study stated image processing analysis accuracy is supported by the amount of cloud covering a study area (Hanindito, Tanaamah, & Papilaya, 2010). Remote sensing is the science of observing and gathering information regarding objects on earth’s surface, using certain sensors, without direct contact with the observed objects (Andree Ekadinata, Sonya Dewi, Danan Prasetyo Hadi, Dudy Kurnia Nugroho, 2008).
- 2. International Journal of Advanced Engineering, Management and Science (IJAEMS) [Vol-3, Issue-8, Aug- 2017] https://dx.doi.org/10.24001/ijaems.3.8.10 ISSN: 2454-1311 www.ijaems.com Page | 879 II. RESEARCH METHOD The following Figure 1 describes research stages. Composite Band 654 Sharpening Image from Landsat 8 (OLI) Image of Landsat 8 Band 8 Cropping Shapefile Boundary Unsupervised Classification Reclassification (4 class) ValidationReal Data Land Use Map of Madura 2015 . Fig. 1: Research Stage Chart 2.1 Image Conversion Downloaded Landsat 8 satellite image is satellite images consisting of multiple bands of satellite sensor recordings. The image is are in* .tiff file extension and cannot be analyzed. Therefore image conversion was conducted to facilitate analysis process. The conversion was conducted by combining between several bands of an image in an appearance with * file extension. The following is a list of R G B sequences of band combinations in the available Landsat 8 Satellite Imagery: Table.1: band combinations in the available Landsat 8 BAND COMBINATI ON FUNCTION / IMPLEMENTATION Natural Color – 4 3 2 Produce an image with true color False Color Produce images with distinct (urban) – 7 6 4 differences in urban areas Color Infrared (vegetation) – 5 4 3 Used to see the mass, density, and dominance of vegetation. The contrast between vegetation dominance will be seen in infrared, making it effective for large-scale forestry or agricultural vegetation analyzes Agriculture – 6 5 2 Produce an image with distinct vegetation indicated by greenish color Atmospheric Penetration – 7 6 5 Clarify the image of cloud thickness, clarify shoreline, and vegetation cover. This combination can clarify the image from weather disturbance Healthy Vegetation – 5 6 2 Produce an image that reveals healthy vegetation Land/Water – 5 6 4 Produces images with distinct differences on water and land regions. Natural With Atmospheric Removal – 7 5 3 Produce images with natural colors and reduce the appearance of clouds Shortwave Infrared – 7 5 4 To get the biomass with clear contrast and cleaner images of cloud cover Vegetation Analysis – 6 5 4 Used to analyze plants This study utilized Landsat satellite images RGB 654 composites. Three bands included were near-infrared and visible spectra. These have wavelengths corresponded to 4, 5 and 3 bands wavelengths on satellite images Landsat 7 ETM +. Fig. 2:band 6, band 5 and band 4 combination 2.2 Image Contrast Improvement This process is done in order to obtain images possessing color quality similar to the original appearance on the earth surface and support the following image classification process. This process is aimed at providing a sharper coloring in an endeavor in enabling easier image
- 3. International Journal of Advanced Engineering, Management and Science (IJAEMS) [Vol-3, Issue-8, Aug- 2017] https://dx.doi.org/10.24001/ijaems.3.8.10 ISSN: 2454-1311 www.ijaems.com Page | 880 classification process. Figure 3 illustrates the process of improving image contrast. Fig. 3: raster data and vector data 2.3 Image Cropping Process on cropping study area images is conducted on this stage. This process aims to facilitate the analysis process by focusing study area by eliminating regions unused in research. This process was conducted by combining raster data (satellite image data) with data vector which is administrative data of Bangkalan Regency borders. 2.4 Image Classification This process is a review of image appearance based on the visible phenomenon. The image is generated and analyzed using true color composite terminology or image appearance in accordance with its original appearance on earth surface. Classification process was conducted by distinguishing each color contained in the image. Figure 5 illustrations will illustrate the appearance of satellite imagery and its classification. III. RESULTS AND DISCUSSION This research produces categorized images, and determined vegetation classification contained in Bangkalan Regency, East Java. Figure 6 exhibits Landsat 8 satellite image data processing classification. Fig. 4: image cropping results Fig. 5: analysis of unsupervised classification image imagery Fig. 6: map of reclassification validation results Fig. 7: image classification legends Forest vegetation is dominant in Bangkalan Regency. Based on image classification, Bangkalan district is dominated by rice field area, with small regions of housings. Followed by several other imagery appearances. The following table exhibit Bangkalan Regency’s Landsat 8 satellite image classification analysis and graphs on the landscape appearance.
- 4. International Journal of Advanced Engineering, Management and Science (IJAEMS) [Vol-3, Issue-8, Aug- 2017] https://dx.doi.org/10.24001/ijaems.3.8.10 ISSN: 2454-1311 www.ijaems.com Page | 881 Table.2: Landscape Appearance of Area Percentage Classification Result No Image Total Area Percentage 1 Dyke & Lake 6.106448 2 Farming 51.301141 3 Vacant 28.38344 4 Housing 14.208971 Fig. 8: Bangkalan Regency Land Acquisition Percentage Graph IV. CONCLUSION Based on Landsat 8 satellite image observation, almost every Bangkalan district consists of vacant land. This exhibits land untouched by human hands to be processed into productive land such as plantation and agricultural land clearing. Satellite images were taken on June 16, 2015. It is suggested for future research to compare the results of previous year's image analysis, to find out the extent agricultural land area extension by utilizing vacant land empowerment. . REFERENCES [1] Alonso, W. (1964). Location and land use. Toward a general theory of land rent. Location and land use. Toward a general theory of land rent. [2] Cohen, W. B., Fiorella, M., Gray, J., Helmer, E., & Anderson, K. (1998). An efficient and accurate method for mapping forest clearcuts in the Pacific Northwest using Landsat imagery. Photogrammetric engineering and remote sensing, 64(4), 293-299. [3] Hanindito, G. A., Tanaamah, A. R., & Papilaya, F. S. (2010). Pengolahan Data Citra Satelit Landsat TM Dalam Pemantauan Area Kebakaran Hutan Berbasis GIS (Studi Area Kecamatan Arut Utara dan Seruyan Tengah, Propinsi Kalimantan Tengah).Universitas Kristen Satya Wacana. [4] Turner, B. L., & Meyer, W. B. (1991). Land use and land cover in global environmental change: considerations for study. International Social Science Journal, 43(130), 669-679. [5] Statistik, B. P. (2015). Statistik kesejahteraan rakyat. Biro Pusat Statistik. [6] Wiradisastra, & Noviar, H.( 2005).Kemampuan Interpretasi Kebun Semangka Dari Citra Satelit Landsat-7 ETM+. Pertemuan Ilmiah Tahunan MAPIN XIV, (September), 132–140. [7] Andree Ekadinata, Sonya Dewi, Danan Prasetyo Hadi, Dudy Kurnia Nugroho, F. J. (2008). Sistem Informasi Geografis Untuk Pengelolaan Bentang Lahan Berbasis Sumber Daya Alam. (1st ed., p. 70) Bogor. [8] Lindeijer, E. (2000). Review of land use impact methodologies. Journal of Cleaner Production, 8(4), 273-281. [9] Lillesand, T., Kiefer, R. W., & Chipman, J. (2014). Remote sensing and image interpretation. John Wiley & Sons. [10]Woodcock, C. E., Allen, R., Anderson, M., Belward, A., Bindschadler, R., Cohen, W., ... & Nemani, R. (2008). Free access to Landsat imagery. 6% 51% 29% 14% PERCENTAGE OF AREA Tambak Lahan Kosong Sawah