WATER BODIES EXTRACTION OF JHARKHAND USING RS AND GIS
http://www.gjms.co.in/index.php/gjms2016/article/view/1989/1493
GLOBAL JOURNAL OF MULTIDISCIPLINARY STUDIES ISSN: - 2348-0459
Volume-5, Issue-7, June- 2016 Impact Factor: 2.389
Abstract: Surveying
of water-bodies and delineating its features is vital for any planning,
especially for India, where the land-cover is dominated by
water-bodies. Extraction of water bodies using spatial inputs has been
the most important method in the investigation of water resources, flood
hazard prediction assessment and water planning with better
effectiveness. A methodology for water extraction mapping using an
integrated GIS and RS approach is presented, where in an endeavour has
been made to put forward an automatic approach to extract the water
bodies from satellite imagery. In this study, various methods were used
for water body extraction like as Supervised Classification,
Unsupervised Classification, Normalized Difference Water Index (NDWI),
and Water Bodies Extraction Index (WBEI). WBEI it is a newly developed
index used mainly for water body extraction. Landsat satellite data was
used as inputs to the methodology. In this research water bodies were
identified and corresponding thematic data layers were created. These
data layers represent the hydrological conditions of the study area. The
result obtained from different methodology was compared with each
other. Part of the Jharkhand state was selected as the study area to
evaluate and analyze which water extraction method is most effective and
quick for water bodies’ identification. On the basis of visual
interpretation of different methodology results, WBEI methodology showed
good result and exact area of water body as has been depicted in the
paper.
Keywords: Water bodies; Supervised; Unsupervised; NDWI; WBEI;
INTRODUCTION:
Because of high population pressure in Indian context, water resources are subject to over exploitation and monitoring at regular intervals becomes imperative for sustainable management. Water extraction is useful for to measure area under the water bodies, also to find out exact location of water bodies, water bodies mapping, accurate water planning and before-after flood mapping etc. Geospatial tools are advantageous for impact assessment for the implementation of water conservation measures. Visual interpretation of satellite data provides the better delineation of water bodies of varied sizes but if worked with high resolution data, it is time consuming. Supervised classification presents more accurate and reliable outputs than unsupervised method but may vary when used for high resolution data. Accurate information on the extent of water bodies is important for flood prediction, monitoring, and relief (Smith, 1997; Tholey et al., 1997; Paul Shane Frazier, et al.,). The first objective of this study was to extract water bodies from each sub-basin using different extraction methods and Index. The second objective was to generate new index for water extraction. The final objective was to compare outputs and find out accurate water extracting method.
Because of high population pressure in Indian context, water resources are subject to over exploitation and monitoring at regular intervals becomes imperative for sustainable management. Water extraction is useful for to measure area under the water bodies, also to find out exact location of water bodies, water bodies mapping, accurate water planning and before-after flood mapping etc. Geospatial tools are advantageous for impact assessment for the implementation of water conservation measures. Visual interpretation of satellite data provides the better delineation of water bodies of varied sizes but if worked with high resolution data, it is time consuming. Supervised classification presents more accurate and reliable outputs than unsupervised method but may vary when used for high resolution data. Accurate information on the extent of water bodies is important for flood prediction, monitoring, and relief (Smith, 1997; Tholey et al., 1997; Paul Shane Frazier, et al.,). The first objective of this study was to extract water bodies from each sub-basin using different extraction methods and Index. The second objective was to generate new index for water extraction. The final objective was to compare outputs and find out accurate water extracting method.
2.0 The Study Area:
Jharkhand is a state in eastern India. The industrial city of Ranchi is its capital. Jharkhand has an area of 79,710 km2. Jharkhand is a mountain region which is covered with a dense growth of forests and named Jharkhand, which literally means 'the territory of forests'. There are three seasons in this state. The cold-weather season, from November to February, is the most pleasant part of the year. The hot climate season stretches from March to mid of June. Maximum rainfall takes place during the months from July to September that accounts for more than 90% of total rainfall in the state.
Jharkhand is a state in eastern India. The industrial city of Ranchi is its capital. Jharkhand has an area of 79,710 km2. Jharkhand is a mountain region which is covered with a dense growth of forests and named Jharkhand, which literally means 'the territory of forests'. There are three seasons in this state. The cold-weather season, from November to February, is the most pleasant part of the year. The hot climate season stretches from March to mid of June. Maximum rainfall takes place during the months from July to September that accounts for more than 90% of total rainfall in the state.
3.0 Data and Methodology:
This study uses the OLI (Operational Land Imager) image of Landsat series to study the effect of water bodies’ extraction. Landsat 8 data is freely available on USGS website (http://earthexplorer.usgs.gov/). For this research, Nov/Dec 2014 time period of images were used. In this research, Supervised Classification, Unsupervised Classification, Normalized Difference Water Index (NDWI), and Water Bodies Extraction Index (WBEI) these methods were used for water extraction. Erdas Imagine 9.2 and ArcGIS 9.3 were used to complete this research.
This study uses the OLI (Operational Land Imager) image of Landsat series to study the effect of water bodies’ extraction. Landsat 8 data is freely available on USGS website (http://earthexplorer.usgs.gov/). For this research, Nov/Dec 2014 time period of images were used. In this research, Supervised Classification, Unsupervised Classification, Normalized Difference Water Index (NDWI), and Water Bodies Extraction Index (WBEI) these methods were used for water extraction. Erdas Imagine 9.2 and ArcGIS 9.3 were used to complete this research.
3.1 TOA reflectance (Top-of-atmosphere reflectance):
TOA reflectance is the reflectance measured by a space-based sensor flying higher than the earth's atmosphere. These reflectance values will include contributions from clouds and atmospheric aerosols and gases. OLI band data can also be converted to TOA planetary reflectance using reflectance rescaling coefficients provided in the product of metadata file. The following equation is used to convert DN values to TOA reflectance as follows: ρλ' = MρQcal + Aρ Where: ρλ' = TOA planetary reflectance, without correction for solar angle
Mρ = Band-specific multiplicative rescaling factor from the metadata (REFLECTANCE_MULT_BAND_x, where x is the band number) Aρ = Band-specific additive rescaling factor from the metadata (REFLECTANCE_ADD_BAND_x, where x is the band number) Qcal = Quantized and calibrated standard product pixel values (DN)
TOA reflectance is the reflectance measured by a space-based sensor flying higher than the earth's atmosphere. These reflectance values will include contributions from clouds and atmospheric aerosols and gases. OLI band data can also be converted to TOA planetary reflectance using reflectance rescaling coefficients provided in the product of metadata file. The following equation is used to convert DN values to TOA reflectance as follows: ρλ' = MρQcal + Aρ Where: ρλ' = TOA planetary reflectance, without correction for solar angle
Mρ = Band-specific multiplicative rescaling factor from the metadata (REFLECTANCE_MULT_BAND_x, where x is the band number) Aρ = Band-specific additive rescaling factor from the metadata (REFLECTANCE_ADD_BAND_x, where x is the band number) Qcal = Quantized and calibrated standard product pixel values (DN)
3.2 Supervised Classification:
Classification is an image processing function which creates thematic maps from remotely sensed images. In this research supervised classification done on the basis of Landsat Imagery reflectance. The spatial distribution of LULC can be obtained via classification of satellite images which can be defined as the process of assigning each pixels or group of pixels of the image to thematic classes (Richards, 1999). Maximum likelihood algorithm and sample water signatures were used for supervised classification in Erdas Imagine.
Classification is an image processing function which creates thematic maps from remotely sensed images. In this research supervised classification done on the basis of Landsat Imagery reflectance. The spatial distribution of LULC can be obtained via classification of satellite images which can be defined as the process of assigning each pixels or group of pixels of the image to thematic classes (Richards, 1999). Maximum likelihood algorithm and sample water signatures were used for supervised classification in Erdas Imagine.
3.3 Unsupervised Classification:
The
most famous types of classification techniques are the unsupervised
classification which doesn’t need a prior knowledge of the area and the
supervised classification which needs prior knowledge of the area
(Lillesand and Kiefer, 2000). The number of classes in the
classification processes was set to 15 classes, the maximum number of
iteration was set to 100, and the convergence threshold was set to 0.95.
After the process, assigned classes were grouped into two categories
(water and other) according to their spectral appearance on screen.
3.4 Normalized Difference Water Index (NDWI):
The
NDWI is expressed as follows (Mc Feeters, 1996). NDWI=
(Green-NIR)/(Green+NIR) Where Green is a Green band such as OLI band
3(Wavelength 0.53 - 0.59 μm), and NIR is a Near Infrared band such as
OLI band 5 (Wavelength 0.88 - 0.85 μm). Ratio-based index commonly is
also called normalized, one of which is Normalized Difference Water
Index proposed by K. McFeeters. The Normalized Difference Water Index
(NDWI) is a new method that has been developed to delineate open water
features and enhance their presence in remotely-sensed digital imagery
(K. McFEETERS, 1996).
3.5 Water Bodies Extraction Index (WBEI):
The Water Bodies Extraction Index (WBEI) is a new method that has been developed to delineate water features in satellite imagery. For WBEI Index used reference of modified Difference Water Index (MNDWI) which is proposed by Hanqiu Xu in 2005. MNDWI = (Green – MIR)/(Green + MIR). He worked on Landsat 5 TM and used Green band as well as MIR (Mid-Infrared) band. Landsat 8 doesn’t have MIR band but wavelength is quite similar of SWIR 1 band so as an experiment SWIR 1 band, green band were used for new index which is Water Bodies Extraction Index. WBEI = (Green-(SWIR) 1) / (Green+(SWIR) 1)Where Green is a Green band such as OLI band 3(Wavelength 0.53 - 0.59 μm), (SWIR-1) is a Short-wave Infrared such as OLI band 6 (Wavelength 1.57 - 1.65 μm).
The Water Bodies Extraction Index (WBEI) is a new method that has been developed to delineate water features in satellite imagery. For WBEI Index used reference of modified Difference Water Index (MNDWI) which is proposed by Hanqiu Xu in 2005. MNDWI = (Green – MIR)/(Green + MIR). He worked on Landsat 5 TM and used Green band as well as MIR (Mid-Infrared) band. Landsat 8 doesn’t have MIR band but wavelength is quite similar of SWIR 1 band so as an experiment SWIR 1 band, green band were used for new index which is Water Bodies Extraction Index. WBEI = (Green-(SWIR) 1) / (Green+(SWIR) 1)Where Green is a Green band such as OLI band 3(Wavelength 0.53 - 0.59 μm), (SWIR-1) is a Short-wave Infrared such as OLI band 6 (Wavelength 1.57 - 1.65 μm).
3.6 Determining threshold value:
After calculation of this both index it was converted to a binary images. For NDWI all the values above 0.05 are changed to 1 and below 0.05 to zero and for WBEI all the values above 0.3 are changed to 1 and below 0.3 to zero. These values were used on the basis of NDWI and WBEI histogram. Then its result comes out with water body and non-water body feature. The threshold value determines whether a pixel belongs to a water body or not. There were two raster objects and one function used in threshold model. Raster object used for input, output image and in function used for condition as shown in figure 3 and figure 4.
After calculation of this both index it was converted to a binary images. For NDWI all the values above 0.05 are changed to 1 and below 0.05 to zero and for WBEI all the values above 0.3 are changed to 1 and below 0.3 to zero. These values were used on the basis of NDWI and WBEI histogram. Then its result comes out with water body and non-water body feature. The threshold value determines whether a pixel belongs to a water body or not. There were two raster objects and one function used in threshold model. Raster object used for input, output image and in function used for condition as shown in figure 3 and figure 4.
4.0 Result and Discussion:
In
comparison of Supervised and Unsupervised Classification, Unsupervised
Classification is quite good effective, and fast process. In this paper
experiment with TOA reflectance band2 and band4 of Landsat 8 OLI; NDWI
is good for water extraction. TOA reflectance band 2 and band 6 of
Landsat 8 OLI are successfully used for Water Body Extraction Index.
WBEI is a new index which has better accuracy in extraction of minor
water body. WBEI image has accurate and more water information than NDWI
image. Water Bodies Extraction Index removes the interference of the
shadow as well.
Table 1, is shows area under water bodies, these
statistical values represent outputs which were derived using different
water extraction method. This particular statistics is only valid for
Nov / Dec 2014. Figure 11, is shows comparison between four extraction
methods. Images are clearly showing that major water body are capturing
in all methods but small stream and minor water bodies are more accurate
in WBEI. Figure 12, is shows that a stream is taken for comparison with
all methodology to find which methods is accurate. On the basis of
visual interpretation Figure 11 and Figure 12, shows that WBEI is more
accurate and clearly shows water bodies.
5.0 CONCLUSION:
Quick and accurate water extraction method is useful for after and before flood assessment, monitoring, water resource survey. There are several methods available for water information extraction. In this research there were four methods has been studied and applied for Jharkhand state. Jharkhand study area was classified and water index model has been done using satellite imagery, GIS and RS techniques. NDWI and WBEI both index are capable for extracting water information. NDWI and WBEI getting accurate water extraction results using appropriate threshold. On the basis of visual interpretation of different methodology results, WBEI methodology shows good result and exact area of water body. This research comes on final conclusion that WBEI is quick and more accurate index for water Extraction. The proposed model has been tested using Landsat 8 data, are preferred by researchers for better accuracy of water feature mapping. Extraction of water bodies from Landsat data is challenging, if suggested model works out well with water bodies extraction.
Quick and accurate water extraction method is useful for after and before flood assessment, monitoring, water resource survey. There are several methods available for water information extraction. In this research there were four methods has been studied and applied for Jharkhand state. Jharkhand study area was classified and water index model has been done using satellite imagery, GIS and RS techniques. NDWI and WBEI both index are capable for extracting water information. NDWI and WBEI getting accurate water extraction results using appropriate threshold. On the basis of visual interpretation of different methodology results, WBEI methodology shows good result and exact area of water body. This research comes on final conclusion that WBEI is quick and more accurate index for water Extraction. The proposed model has been tested using Landsat 8 data, are preferred by researchers for better accuracy of water feature mapping. Extraction of water bodies from Landsat data is challenging, if suggested model works out well with water bodies extraction.
Acknowledgments:
The authors wish to thank to all who helped directly or indirectly to complete this research. We are also thankful to U.S. Geological Survey server (http://earthexplorer.usgs.gov/) for free access and to download Landsat 8 OLI data.
The authors wish to thank to all who helped directly or indirectly to complete this research. We are also thankful to U.S. Geological Survey server (http://earthexplorer.usgs.gov/) for free access and to download Landsat 8 OLI data.
REFERENCES:
[1]
H. Xu., 2006. “Modification of normalised difference water index (NDWI)
to enhance open water features in remotely sensed imagery,”
International Journal of Remote Sensing, vol. 27, no. 14, pp. 3025–3033.
[2] K. McFEETERS., 1996. The Use of Normalized Difference Water Index (NDWI) in the Delineation of Open Water Features. International Journal of Remote Sensing, 17 (7): ISSN: 1425-1432.
[3] Lillesand, T.M., and Kiefer, R.W., 2000. Remote Sensing and Digital Image Interpretation, Wiley, New York, 724 p.
[4] Paul Shane Frazier., and Kenneth John Page., 2000. Water Body Detection and Delineation with Landsat TM Data. Photogrammetric Engineering& Remote Sensing.
[5] Richards, J. A., 1999. Remote Sensing and Digital Image Analysis: An Introduction, 2nd ed. Springer, Berlin, Heidelberg, 363 p.
[6] Smith, L.C., 1997. Satellite remote sensing of river inundation area, stage, and discharge: A review, Hydrological Processes, 11:1427-1439.
[7] Tholey, N., S. Clandillon., and P. De Fraipont., 1997. The contribution of spaceborne SAR and optical data in monitoring flood events: Examples in northern and southern France, Hydrological Processes, 11:1409-1413.
[2] K. McFEETERS., 1996. The Use of Normalized Difference Water Index (NDWI) in the Delineation of Open Water Features. International Journal of Remote Sensing, 17 (7): ISSN: 1425-1432.
[3] Lillesand, T.M., and Kiefer, R.W., 2000. Remote Sensing and Digital Image Interpretation, Wiley, New York, 724 p.
[4] Paul Shane Frazier., and Kenneth John Page., 2000. Water Body Detection and Delineation with Landsat TM Data. Photogrammetric Engineering& Remote Sensing.
[5] Richards, J. A., 1999. Remote Sensing and Digital Image Analysis: An Introduction, 2nd ed. Springer, Berlin, Heidelberg, 363 p.
[6] Smith, L.C., 1997. Satellite remote sensing of river inundation area, stage, and discharge: A review, Hydrological Processes, 11:1427-1439.
[7] Tholey, N., S. Clandillon., and P. De Fraipont., 1997. The contribution of spaceborne SAR and optical data in monitoring flood events: Examples in northern and southern France, Hydrological Processes, 11:1409-1413.
WEBSITES:
[1] http://landsat.usgs.gov/Landsat8_Using_Product.php
[2] http://www.exelisvis.com/Home/NewsUpdates/TabId/170/ArtMID/735/ArticleID/13592/Digital-Number-Radiance-and-Reflectance.asp
[1] http://landsat.usgs.gov/Landsat8_Using_Product.php
[2] http://www.exelisvis.com/Home/NewsUpdates/TabId/170/ArtMID/735/ArticleID/13592/Digital-Number-Radiance-and-Reflectance.asp
No comments:
Post a Comment