See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/312704684

Analysis and modeling of agricultural land use using remote sensing and
geographic information system: a review

Article · January 2013

CITATIONS READS

63 5,552

2 authors, including:

Amol Vibhute
Symbiosis International University
69 PUBLICATIONS 568 CITATIONS

SEE PROFILE

All content following this page was uploaded by Amol Vibhute on 09 November 2017.

The user has requested enhancement of the downloaded file.

 Amol D. Vibhute, Dr. Bharti W. Gawali / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 3, May-Jun 2013, pp.081-091
Analysis and Modeling of Agricultural Land use using Remote
Sensing and Geographic Information System: a Review
Amol D. Vibhute*, Dr. Bharti W. Gawali**
1
Research Scholar, 2Associate Professor
*, ** (Department of Computer Science & Information Technology,
Dr. Babasaheb Ambedkar Marathwada University, Aurangabad, (M.S), INDIA-431004)

ABSTRACT
GIS, remote sensing and Global built-up areas, trees, etc. Land use means those areas
positioning System are the most widely useful which are used by humans for their need. Peoples
tools for land use planning and decision support agreed to produce, maintain the change on land
system. Remotely sensed imagery is beneficial for cover through the adjustments, movements and
agricultural production. It gives the accurate inputs [3].
information of agricultural activities such as Both developed and undeveloped countries
different crop identification and classification, are dependent on agriculture for the economic as
crop condition monitoring, crop growth, crop well as food security. Agriculture is the primary
area and yield estimation, mapping of soil source of many countries to maintain the food for
characteristics and precision farming. everyone and it plays a dominant role in almost
Information from remotely sensed imagery, every country to develop the economical condition.
geographic information system and global Food production in a cost-effective manner is the
positioning system allows farmers to carry only goal of every farmer as well as large scale farmer
affected areas of a field. Problems within the field and agricultural agencies, so there is need to be
may be identified before they create a big informed a farmer in an efficient way to get the
problem in the agricultural production using knowledge and inform them about food production.
remotely sensed images. This paper attempts to Remote sensing, GIS and GPS these technologies
review different techniques for various will useful for farmer to monitor their crops before
applications of GIS and Remote sensing for land they got damage, yield estimation, soil condition as
use/land cover change detection, crop well as precision farming. Commodity brokers are
identification and classification, crop condition also very interested in how well farms are
monitoring, crop growth, crop area and yield producing, as yield (both quantity and quality)
estimation, mapping of soil characteristics and estimates for all products control price and
precision farming. Thus implementating GIS and worldwide trading [4], [5]. So there is need to secure
RS for better production of the crops as well as the food in an effective way. Because good crop
land use/land cover change detection can be health is a part of every person’s diet. In every
achieved. country food production through sufficient quantity
and quality is essential for the well-being of the
Keywords – Classification, Crop acreage and yield people. To growth of crop in the field of agricultural,
estimation, Precision farming, RS and GIS, Soil as living organisms, it requires water and nutrients
mapping and is sensitive to extreme weather phenomena,
diseases and pests. So to overcome these problems
1. INTRODUCTION and identify as well as monitor crops, remote
Due to rapid growth of population, sensing data is very much useful. Geographical
urbanization, industrization, the agricultural land is Information System along with Remote Sensing and
decreasing day by day. In twenty first century, land other types of data will helpful in decision support
use and land cover plays an important role in global system about crops and agricultural strategies [6].
environmental change. The use of information Extracting land use/land cover (LU/LC) information
technology to support decision making in detecting is crucial exercise for agricultural land which is most
land use and cover is essential and recent. For local, useful for decision support system, planning and
regional and macro level planning and management development in agriculture [7]. In the study of Land
Land cover/land use play an important role [1] by use/ Land cover pattern analysis different sources
using RS & GIS it is possible to make planning and are important for effective planning and
decision process more realistic, effective [2]. Land management like soil survey manuals, topographic
cover is nothing but the physical material on the maps, aerial photographs, vegetation surveys, flood
Earth surface which is covered by different parts of maps, hydrology maps, and property surveys etc.,
the nature or also the man made parts on the Earth [8]. Due to the population and urbanization land use
surface i.e. soil, rocks, water bodies, vegetation, and land cover rapidly change, to acquire the exact

81 | P a g e
 Amol D. Vibhute, Dr. Bharti W. Gawali / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 3, May-Jun 2013, pp.081-091
land pattern remote sensing & GIS technique is most remote sensing imagery. The overall fuzzy highest
essential [9], [10]. classification accuracy 93.12% was achieved from
Remote sensing is the technology to acquire the SAVI temporal data base of data set 2 [26].
exact information from the earth surface patterns Maximum Likelihood and Artificial Neural Network
without making any physical contact to it. In today’s algorithms are used with SPOT HRV and Land sat
date, this technology uses the satellites to detect and TM data for classifying agricultural crops combining
classify objects on the earth surface and in multi-temporal, multi-spectral and multi-source
the atmosphere and oceans by using the remotely-sensed data to identifying general
signals of electromagnetic radiation emitted agricultural crop classes over an area in East Anglia
from aircraft or satellites [11]. (UK) [27]. Sugarcane, cassava and maize crops are
identified through Radarsat-2 multipolarization data
Geographic Information System (GIS) is a in Thailand, because high cloud cover availability in
computer system designed for capturing, storing, nature and these sensors cannot give the whole
integrating, analyzing and displaying data from a information about the objects. Therefore, the radar
geographic angle. The measurement of natural and sensor is the alternative option to obtain satellite data
human made phenomena and processes from a especially RADARSAT-2 which provides full
spatial angle. These measurements emphasize three polarizations (HH, VV, HV, VH) [28]. Unsupervised
types of properties commonly associated with these ISODATA and NDVI (normalized difference
types of systems: elements, attributes, and vegetation index), these two methods was used to
relationships. The storage of measurements in digital classify crop. Method first based on unsupervised
form in a computer database. These measurements classification of Landsat visible and Near-infrared
are often linked to features on a digital map. The satellite images obtained at multiple dates in the
features can be of three types: points, lines, or areas growing season, followed by traditional, manual
(polygons). The analysis of collected measurements class identification. And second method based on
to produce more data and to discover new unsupervised classification as well as NDVI maps
relationships by numerically manipulating and performed MODIS images. Tree analysis was
modeling different pieces of data. The depiction of applied for grouping similar classes into clusters
the measured or analyzed data in some type of using NDVI mask data and for crop type of each
display - maps, graphs, lists, or summary statistics cluster was identified from ground truth. They got
[12]. 3% more overall classification accuracy using
unsupervised classification and NDVI masks which
2. WORK DONE SO FAR is the second method than first method [29].
There is lot of techniques used to identify Multispectral images obtained from the SPOT 5
land use/land cover change detection, and land satellite acquired two different date images has been
use/land cover monitoring from satellite imagery. used to identify crop area using pixel based i.e.
Fuzzy C-mean supervised classification algorithm unsupervised and maximum supervised
has been used in land use/land cover change classification and object based classification
detection through Landsat TM5 satellite images with techniques [30]. Hyperspectral remote sensing has
overall 90% and 89% accuracy [13]. To classify the also helped enhance more detailed analysis of crop
land use/land cover from remotely sensed classification [31]. Maximum Likelihood classifier
multispectral imagery and to detect the change on was performed on hyperspectral data of the Hymap
LULC from multitemporal imagery researchers has sensor for different crop classification. This method
been used supervised maximum Likelihood was used to extract a subset of individual bands from
Classification [1], [7], [14], [15], [16], [17], [18], hyperspectral data and the classification results are
[19], [20], [21], minimum distance [16], [21], compared to those achieved by using principal
[22],[23], parallepiped [16], [21], fishers components transformation as a well established
classification [16], unsupervised classification for feature extraction technique to optimize the
ISODATA [14],[23],[24], Hybrid classification [14], performance of the maximum likelihood classifier
Box classifier[25] mostly. [32]. Rice Crop Identification and Classification
using Hyper-Spectral Image Processing System, was
Many researchers identified and classified reviewed through the comparison of multispectral
crops by using multispectral and multitemporal data and hyper spectral image processing techniques [33].
and supervised and unsupervised classification Support vector machine method is better than neural
techniques. networks classifiers as multilayer perceptrons
Fuzzy classifier, SR (Simple Ratio), NDVI (MLP), Radial Basis Functions (RBF) and Co-
(Normalized Difference Vegetation index), TNDVI Active Neural Fuzzy Inference Systems (CANFIS)
(Transformed Normalized Difference Vegetation for crop classification [34]. A number of studies
Index), SAVI (Soil-Adjusted Vegetation Index) and have been carried out on different crop acreage
TVI (Triangular Vegetation Index) are used to estimation using remote sensing and GIS
identify cotton crop using temporal multi-spectral technology.

82 | P a g e
 Amol D. Vibhute, Dr. Bharti W. Gawali / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 3, May-Jun 2013, pp.081-091
Supervised classification i.e. maximum likelihood MODIS (TERRA) time series (multitemporal) data
classification algorithm was used to estimate the of three years using NDVI values and supervised,
wheat crop acreage of Indore district, M.P. using unsupervised as well as Hybrid classification
single date, and cloud free Resourcesat-1 (IRS P6) techniques has been used for estimation of soya bin
LISS-III data. This work has been carried out using acreage and crop yield estimation in districts of M.P.
the ground truth through GPS(Global Positioning They performed NDVI on all time series images and
System) and georeferenced by image-to-image compared the scene average NDVI values for
registration existing master images then the individual district over the years, to get the predicted
important steps were used, like, extraction of district yield of soyabin. Crop yield was calculated by
mask of the image, generation of training signatures, comparison of the NDVI values of soyabin for these
test of statistical class separability and maximum years [36]. Optical and microwave remote sensing
likelihood supervised classification [5]. Maximum data has been used for corn monitoring and Crop
Likelihood Classification and unsupervised yield estimation and production efficiency model
classification techniques were used for Crop Area (PEM) has been used to estimate crop yield on
Estimation on IRS-P6 LISS-III(Resourcesat-1) data QuickBird imagery and getting higher accuracy of
of different seasons kharif, Rabbi And Zaid in year predicted results than Landsat TM imagery [37].
in Bundi Tahsil of Rajsthan. The visual Precision farming is a best technique to establish the
interpretation also performed for this particular study agricultural management and variable rate
and classified different land use and land cover application (VRA) is also plays a significant role in
objects for extracting exactly crop area from precision farming [38].
agricultural land Finally conclude that 688.92 sq.km
area was occupied in agriculture while 73.12sq.km. Due to the lower absorption of light
Area is fallow land [7]. NDVI algorithm and comparing other objects on earth surface agricultural
unsupervised classification technique were used to crops have higher spectral reflectance of signature.
identify the sugarcane area as well as its condition Vegetation frequently decreases absorption and
assessment by using toposheet and IRS LISS II increase reflectance and transmittance in the visible
remote sensing data were used. NDVI image is very wavelength due to the dietetic scarcity in the
much useful in sugarcane crop area identification vegetation. Hence, agricultural crops have higher
because of its canopy cover, crop biomass and vigor. reflectance of the spectral signature. In this
For this particular study two bands were used i.e. Situation, hyperspectral remote sensing plays an
NIR and RED of LISS-II data then georeferenced important role as tool for monitoring and estimating
through ground control points, NDVI image has agricultural land use and cover because it can
been generated and sugarcane area were extracted provide earlier information than conventional land
from NDVI image. For crop area estimation NDVI use mapping methods. So, using hyperspectral
value of Band 4(NIR) and Band 3(RED) of IRS remote sensing imagery development should be
LISS-II data have been used and for classification increased in the field of agricultural land use and
the NDVI image cluster using ground truth have land cover for future planning and management [39].
been used. And for assessing the sugarcane crop
condition, sugarcane area map has been used as a For managing the natural resources on the
‘mask’ over the NDVI value map to generate earth surface mapping of vegetation and land cover
respective NDVI mask for sugarcane. NDVI value is is essential. For planning, management and
very much sensitive to crop canopy, biomass and development the information on the land use and
vigour i.e. health condition of the crop. So, NDVI land cover in the form of maps and data is very
mask map was then classified and given the class important [40].
names as ‘very good’, ‘good’, ‘moderate’ and ‘poor’
for sugarcane health condition according to verified 3. AGRICULTURAL APPLICATIONS
crop condition by field observation and farmers 3.1. Crop Inventory
knowledge. In referring healthy crops, reflectance in The importance of spectral reflectance of
the blue and red parts of the spectrum is low since data to identify and classify different crops is very
chlorophyll absorbs this energy. In contrast, much useful in crop area estimation, crop condition
reflectance in the green and near-infrared spectral monitoring, and crop yield estimation. Remote
regions is high. Stressed or damage crop experience sensing data have unique spectral signature which
a decrease in chlorophyll content and changes in the useful in crop identification and crop classification.
internal leaf structure. So examining the ratio of When two crops have same spectral signature occur
reflected infrared to red wavelengths is an excellent in a given date, then multi date data is required to
measure of vegetation health. Healthy plants have a identify them [41]. Remotely sensed images are used
high NDVI value because of their high reflectance of as mapping tools to classify crops, monitor the crop
infrared light, and relatively low reflectance of red health and capability and monitor farming practices.
light. [35]. These are the some main agricultural applications in
remote sensing: 1) Crop type classification and Crop

83 | P a g e
 Amol D. Vibhute, Dr. Bharti W. Gawali / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 3, May-Jun 2013, pp.081-091
identification, 2) crop condition (monitoring) such places in the farm where the growth is
assessment, 3) crop area and yield estimation, 4) decreased or slow; this will allow farmer to provide
mapping of soil characteristics, 5) Precision farming adequate amount of fertilizers, pesticides and
practices, etc. herbicides to the crop in such areas. It will not only
increase the productivity of the farm land but it will
3.1.1. Crop type classification and Crop also reduce the farm input cost and the
identification environmental impact will be minimum. The remote
There are number of reasons to identify and sensing image provides the required spatial overview
classify the crops. National/multinational of the farm land. It allows a farmer to observe image
agricultural agencies and other agriculture related of his fields and make timely decisions about
agencies are responsible to produce different crop managing the crops. Remote sensing can be helpful
type maps to prepare an inventory of what was in identifying crops affected by conditions that are
grown in certain areas and when [4]. Properly too dry or wet or that can be affected by insect, weed
identification of crops is very much beneficial for or fungal infestations or weather related damage.
crop acreage and crop production estimation [42]. Images obtained throughout the growing seasons
Remote sensing technique plays a vital role in will help to detect to problems but also to monitor
identifying and classifies different crops in acreage. the success ratio of the treatment [4].
It shows the vegetation health and growth based on To monitor the crop condition we need a cloud free
the spectral reflectance of the vegetation and other data, but crop signatures are not unique or the
patterns. For crop identification and classification variance in the signature of a single crop is too large,
they use the multispectral and multitemporal data so it’s difficult to monitor the crop. In this condition,
and mainly supervised and unsupervised radar remote sensing plays a significant role to
classification techniques. In Supervised technique vegetation biomass and structure as well as these
machine require to give training signature which sensors are very much useful in crop monitoring
class is in the pixel, then identify examples of the [37]. What is the crop condition means, there is no
Information classes (i.e., crop type) of interest in the clear definition for crop condition and researchers in
image. These are called "training signatures". But the field of this area don’t have any idea about this
unsupervised classification is a method which definition. If we don’t have any exact information
examines a large number of unknown pixels and about crop condition, then how could we monitor the
divides into a number of classed based on natural crop condition? Therefore it is urgent requirement to
groupings present in the image values. Unsupervised know the crop condition means [44]. The use of
classification does not require analyst-specified remote sensing methods in crop monitoring has been
training data [7], [43]. reported by many researchers. These are the some
main useful methods for crop condition monitoring
3.1.2. Crop condition (monitoring) assessment as following:
Crop condition monitoring is one of the
main important advantages using remote sensing. To 3.1.2.1. Direct monitoring method
preserve the countries food it is necessary to monitor Vegetation indices which are very much
the crops timely and accurately [42]. Crop condition useful in crop condition monitoring, because if
monitoring is very much useful in crop acreage and vegetation indices higher, then the crop condition is
crop yield estimation. Availability of water and better. This direct monitoring method is depends on
nutrients, pest attack, diseases outbreak and weather vegetation indices which are easy to use and
conditions these are the some factors that could be promised but if large area or complex area will be
affect to crop condition [41]. Crop condition is big then it should be failed [44]. Following table 1
primarily focused on individual physical parameters shows the main important indices useful in crop
as well as different indices of the crop [31]. condition monitoring.
It is difficult to ensure good agricultural
productivity due to crop health assessment and early
detection of crop infestations. Some diseases should
be detected early enough to provide an opportunity
for the farmer to mitigate due to for example,
moisture deficiencies, insects, fungal and weed
infestations. To early detection of diseases remote
sensing imagery provides frequently images at
minimum within 2 days. The growth of the crops in
the field is not even, it can vary in the field from one
place to another place of the farm. The differences in
growth are result of various factors such as soil
nutrition or other forms of stress. The remote
sensing technology can help the farmer to identify

84 | P a g e
 Amol D. Vibhute, Dr. Bharti W. Gawali / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 3, May-Jun 2013, pp.081-091
TABLE 1. Useful Vegetation Indices for Crop Condition Monitoring

Vegetation Index Formula Usable

Bands
RVI(Ratio Vegetation Index) NIR NIR, RED
RVI  [26]
RED
NDVI(Normalized NIR  RED NIR, RED
Difference Vegetation NDVI  [26]
Index) NIR  RED
NDWI(Normalized NIR  SWIR NIR,
Difference Water Index) NDWI  SWIR
NIR  SWIR
SAVI(Soil Adjusted (1  L)(NIR  RED) NIR, RED
Vegetation Index) SAVI  [26]
NIR  RED  L
TVI (Triangular Vegetation TVI  0.5(120( NIR  GREEN))  200( RED  GREEN) NIR, RED,
Index) GREEN
[26]
TNDVI(Transformed  NIR  RED   NIR, RED
Normalized Difference TNDVI     0.5 1/2 [26]
Vegetation Index)  NIR  RED  
3.1.2.2. Image classification method
To monitor the crop growth status 3.1.2.6. Diagnosis model
generally having two methods which are very The Diagnosis model assesses the crop
important i.e. supervised and unsupervised condition using the characteristics of condition and
classification methods using multitemporal satellite environment that influence crop growth [44].
imagery [44].
3.1.3. Crop area and yield estimation
3.1.2.3. Same-period comparing method Crop area estimation is very much
While we are comparing previous year important for number of reasons; it is especially
remote sensing data (NDVI, LAI for instance) with helpful for yield estimation. Reliable and timely
new one then we come to know the crop growing crop acreage estimation could be helpful in
state. Some vegetation indices such as differences, agricultural planning and decision support system
ratios are mostly useful in this method [44]. to planners and policy makers for purchasing,
storing, import export of food. Crop area estimation
3.1.2.4. Crop growth profile monitoring method is the backbone of agricultural activities [5]. This
The crop growth profiling method is the crop acreage estimation procedure is basically
different between the year and year for the crop divided into following main steps are 1) Single date
growth profile can it can be seen by the crop data selection with maximum vegetative crop
growing in the prolong time duration of the growth, 2) Identification of crop through image
growing season. The crop growth profiles are using ground truth, 3) Signature generation for the
produced by getting the statistic of NDVI at the training site, 4) Image classification through
level of province. The time series data of NDVI training statistics, 5) Crop area estimation using
during the crop season is used for crop growth administrative boundary like district masks [41].
profile. Different crops have different There are three main methods of remote sensing to
characteristics in the crop growth profile using estimate the crop areas are followings [45].
NDVI; even the same crop grown in different
environment is having the different crop growth 3.1.3.1. Pixel counting and sub-pixel analysis
profile. The NDVI profiles are useful in the crop Using remote sensing imagery pixel
growing conditioning [44]. counting or sub-pixel analysis can be done to
estimate area. But there are some limitations of this
3.1.2.5. Crop growing models method method for image classification due to same order
In this method the crop growing model is of the commission or the omission errors [45].
used to simulate the various stages in the crop life
cycle. The growing status and the estimated crop 3.1.3.2. Regression, calibration and small area
condition are the result of the simulation [44]. estimates
These methods are the combining catholic
of regression, calibration and small area estimator’s

85 | P a g e
 Amol D. Vibhute, Dr. Bharti W. Gawali / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 3, May-Jun 2013, pp.081-091
but inaccurate information from remote sensing using microwave remote sensing data for soil
imagery with accurate information on a sample quality assessment are explained below.
through ground surveys. Combination of fine or
medium resolution images provides findings of 3.1.4.1. Soil Salinity
area estimation when ground surveys are not Crop growth and productivity that could
possible due high cost [45]. be affected by soil degradation; soil salinity is a
main problem of soil degradation. For melioration
3.1.3.3. Supporting area frame surveys of the salt affected soil it is essential to identify of
The different ways by which remote soil type and intensity of soil with their exact
sensing imagery can support area frame surveys location and areal extent. Due to unwanted spectral
may be stated as: to define sampling units, for mixing with sand and little spectral reflectance in
layer; as graphic documents for the ground survey black soils areas it is difficult to the illustration of
or for quality control [45]. Farmer have salt affected soil in littoral areas, infertile areas as
information on crop yield at an early stage then it well as black clay rich soils regions, hence using
will be profitable for farmer as well as countries optical remote sensing imagery it is feasible to
development which is dependent on agricultural illustration of salt affected soils [48].
production, to propitiate the national requirements
for the crops and also for income through exports 3.1.4.2. Ravine Erosion Inventory
[46]. The vegetation indices (VIs) generated from For utilization of depiction and mapping
remotely sensed images are usable for crop yield of ravine affected areas we commonly use optical
estimation after analyzing the vegetation indices multispectral as well as panchromatic fine
and these indices correspond to the maximum resolution imagery and its classification is mainly
vegetation stage of the crop. The average area based on manual interpretation of the image
weighted vegetation indices calculated for each features as well as its depth categories, because
district is regressed with final yield at district. these methods are qualitative in nature. Terrain
Direct weather variables such as rainfall, ruggedness and vegetation penetration capability
temperature and derived variables such as ET are are highly sensitive to microwave SAR, by taking
used either independently or in combination in advantage of microwave remote sensing such as
multiple regressions. Sometimes, in the· absence of SAR and InSAR (Interferrometric SAR) for
weather data, purely statistical models such as Auto illustration and characterization of ravines
Regressive Integrated Moving Average (ARIMA) quantitative terms such as ravine density, ravine
is fitted with time series data using standard depth, and ravine surface cover [48].
techniques. Crop wise production estimates are
thus made based on the remote sensing based 3.1.4.3. Sand Dune Characterization
acreage and yield estimates using a combination of Huge infertile areas are covered by sand
models [42]. dune and these fixed, semi fixed sand dunes are
distributed around the desert periphery with mobile
3.1.4. Mapping of soil characteristics dunes in the inner part of the deserts. Normally
An elevated understanding of the soil used desert areas are difficult to access; hence to
to increasingly better scale. It is beneficial in understand the environmental changes in dry areas
agricultural management and development. information about the dune properties is very much
Traditional soil sampling and laboratory analyses helpful. So, remote sensing imagery are helpful for
methods are very slow, valuable, and they could monitoring desert environments [48].
not reacquire all spatial as well as temporal
changeability of the soil quality, so this methods 3.1.4.4. Soil texture and Hydraulic Properties
cannot efficiently provides the required information An accurate estimation of spatially
[47], [48]. In this condition Microwave (active and variable soil physical properties such as texture and
passive) remote sensing plays a significant role in hydraulic properties is necessary to develop
mapping of soil quality parameters. Mapping of reliable models of water flow and the efficient
soil characteristics is useful for different purposes management of soil resources for improving crop
such as soil and crop management for improving productivity with an environmental quality. To
crop yield estimation, sustainable land use Measure soil physical properties are time-
planning, soil erosion and runoff modeling in consuming, expensive but a large number of
watershed management, land – atmosphere gaseous measurements are necessary to quantify their
exchange study for climate change modeling, space-time variability [48].
biogeochemical cycles study and precision
agriculture etc [48]. 3.1.4.5. Soil Drainage
Research has been carried out using microwave Although the various characteristics, soil
(active and passive) remote sensing for mapping of drainage which directly affect of crop growth,
soil characteristics. Some soil property parameters water flow and solute transport in soils. Drainage is

86 | P a g e
 Amol D. Vibhute, Dr. Bharti W. Gawali / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 3, May-Jun 2013, pp.081-091
major issue at every where so soil information will 4. USEFUL TECHNIQUES FOR LULC
help to took decision at various levels. Therefore, ANALYSIS
radar remote sensing has the effective way to map To analyze any particular area we need a
and analysis soil properties, such as soil drainage GIS or remote sensing data. Those data are in the
[48]. form of toposheet, maps and satellite or airborne
images. For remote sensing data there is lot of
3.1.4.6. Soil Surface Roughness techniques used by researchers to analysis the
Soil surface roughness (SSR) has LULC in any area. The main techniques are
influenced soil thermal properties, infiltration rate, supervised and unsupervised classification, and
surface run-off and susceptibility of soil to erosion others are hybrid classification, box classifier,
[48]. fuzzy classifier, Support vector machine,
Normalized data vegetation index (NDVI), etc.
3.1.5. Precision farming practices These techniques are used for identification and
Precision farming or agriculture is nothing change detection in land cover and land use,
but the management of farm which is depends on classification of land use/land cover, agricultural
noticing and giving response on various changes in areas, patches, water bodies, as well as
the intra-field with the goal of increasing returns on classification of different crops and crop
inputs without changing the resources. Precision monitoring etc.
agriculture is helpful to locate the exact position of
the farmer in the field which is totally based on 4.1. Supervised classification
remote sensing, GIS and GPS or GNSS technology With supervised classification, we identify
[49]. To increase yield, quality, and profit and examples of the Information classes (i.e., land
reduce waste it is essential to control the crop cover type) of interest in the image. These are
production using water, seed, fertilizer, etc. In this called "training sites or different signatures" for
condition, precision farming will be helpful for each class. For this method we require a priori
taking right decision, in the right situation, in the knowledge of the interested region [30]. There are
right position, in the right way, at the right time to number of supervised classifiers based on the
become eco-friendly [50]. Hence precision farming statistics, but the maximum likelihood classifier,
opposes to conventional farming practices which minimum distance classifier, parellopiped
include crop treatments like irrigation, use of classifier, mahalanobis classifier are most useful
fertilizers, pesticides and herbicides were classifiers [7].
equivalently applied to the whole area without
noticing variability within the area. Remote sensing 4.1.1. Maximum likelihood classifier
technologies are becomes advanced and the cost of The maximum likelihood classifier is one
sensors are reduced, so we can easily use of these of the most popular methods of classification in
facility in the farming to identify the particular area remote sensing, in which a pixel with the maximum
in which the specific treatment is required for that likelihood is classified into the corresponding class.
crop as well as use of chemicals in required area, to It is a statistical decision criterion to assist in the
reduce the amount chemicals used. So these classification of overlapping signatures; pixels are
technologies will be helpful for protecting assigned to the class of highest probability.
environment and saving the cost of the application Gaussian maximum likelihood classifier classifies
[51]. Precision farming is an emerging an unknown pixel based on the variance and
methodology which is designed to link covariance of the category spectral response
management actions to site-specific soil and crop patterns. This classification is based on probability
conditions, and place inputs of fertilizers, density function associated with a particular
herbicides, and pesticides when and where they are signature (training site). Pixels are assigned to most
needed most to maximize the farm efficiency and likely class based on a comparison of the posterior
minimize the environmental contamination. The probability that it belongs to each of the signatures
core technologies which play an important role in being considered. Under this assumption, the
precision agriculture are GIS, GPS, and remote distribution of a category response pattern can be
sensing. The importance of these technologies in completely described by the mean vector and the
agriculture was underscored when NASA (Stennis covariance matrix [7], [52], [53].
Space Center), in the early portion of the current
century, embarked upon the Ag 2020 program in an 4.1.2. Minimum distance classifier
effort to commercialize the geospatial technologies, The minimum distance classifier is used to
develop practical tools for producers, and classify unknown image data to classes which
undertake projects with various types of crops to minimize the distance between the image data and
illustrate the utility of the technologies [31]. the class in multi-feature space. The distance is
defined as an index of similarity so that the
minimum distance is identical to the maximum

87 | P a g e
 Amol D. Vibhute, Dr. Bharti W. Gawali / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 3, May-Jun 2013, pp.081-091
similarity. It uses mean vector in each class
signature, while standard deviation and covariance 4.5. Normalized data vegetation index (NDVI)
matrix are ignored [53]. NDVI is based on the principle of spectral
difference based on strong vegetation absorbance
4.1.3. Parellopiped classifier in the red and strong reflectance in the near-
A parallelepiped is a geometrical shape infrared part of the spectrum. NDVI (Normalized
whose opposite sides are straight and parallel. The Difference Vegetation Index) is mostly useful
parallelepiped classifier uses the class limits and method to classify the land cover vegetations [56],
stored in each class signature to determine if a using this formula NDVI= NIR-RED/NIR+RED
given pixel falls within the class or not. The class [57], [58], [59] and calculated using the RED (0.6
limits specify the dimensions (in standard deviation to 0.7 µm) NIR (0.7 to 1.1 µm) wavelengths [29],
units) of each side of a parallelepiped surrounding [35], [59]. NDVI is a single band image having the
the mean of the class in feature space. The values between -1 to +1, the value closer to +1 will
parallelepiped classifier is typically used when tells about presence of healthy green vegetation
speed is required. The drawback is (in many cases) [57], [35], [59].
poor accuracy and a large number of pixels
classified as ties (or overlap, class 255) [53]. 5. CONCLUSION
There is lot of remote sensing techniques
4.1.4. Mahalanobis classifier used by many researchers for land use/land cover
It is based on correlations between classification, identification, classification and
variables by which different patterns can be analysis of any area or any region but Maximum
identified and analyzed. It gauges similarity of an likelihood classification technique is most useful
unknown sample set to a known one. It differs technique in supervised classification, because it
from Euclidean distance. It takes into account the gives greater accuracy than other techniques.
correlations of the data set and is scale-invariant Unsupervised technique is the second important
[54]. method for classification as well as identification.
Fuzzy classification (also known as fuzzy
4.2. Unsupervised classification clustering) algorithm is the best useful classifier for
Unsupervised classification is a method same elements. And Hybrid classifier is the lowest
which examines a large number of unknown pixels useful technique for classification. Normalized data
and divides into a number of classed based on vegetation technique (NDVI) technique is the most
natural groupings present in the image values. useful technique for land cover analysis as well as
Unsupervised classification does not require crop condition monitoring. Soil characteristics are
analyst-specified training data [7], [53]. The two very much useful for crop growth and crop yield.
common types of these classifiers are K-means Precision farming is useful to enhance the
clustering and ISODATA (Iterative Self- agricultural production, to reduce the chemical use
Organizing Data Analysis) [7], [30]. in crop production, to use water resources, to
improve quality, quantity & reduced cost of
4.3. Hybrid classifier production in agricultural crops with the help of
The combination of supervised and RS, GIS and GPS technology.
unsupervised classifier is a hybrid classifier. This
classification produced more accurate REFERENCES
classifications than the supervised classification; [1] Meliadis Ioannis, Miltiadis Meliadis,
however, it did not improve the accuracy Multi-temporal Landsat image
significantly in com-parison to the unsupervised classification and change analysis of land
classification [14]. cover/use in the Prefecture of
Thessaloiniki, Greece, Proceedings of the
4.4. Fuzzy classifier International Academy of Ecology and
Fuzzy logic is a soft computing technique. Environmental Sciences, 1(1): 15-25,
These soft computing techniques are better than 2011.
hard computing techniques, because these soft [2] D.K. Tripathi and Manish Kumar, Remote
computing techniques like as fuzzy logic, neural Sensing Based Analysis of Land Use /
network, machine learning, can gives better facility Land Cover Dynamics in Takula Block,
to handle mixed pixels [26]. Fuzzy classification is Almora District (Uttarakhand), J Hum
the process of grouping elements into a fuzzy Ecol, 38(3): 207-212, 2012.
set whose membership function is defined by the [3] http://en.wikipedia.org/wiki/Land_cover/L
truth value of a fuzzy propositional function and and_use.
uses degree of membership functions in between [4] http://www.nrcan.gc.ca/sites/www.nrcan.g
1.0 and 0.0 for full membership and nonmember c.ca.earthsciences/files/pdf/resource/tutor/
ship respectively [55]. fundam/pdf/fundamentals_e.pdf.

88 | P a g e
Amol D. Vibhute, Dr. Bharti W. Gawali / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 3, May-Jun 2013, pp.081-091
http://geography.huji.ac.il/personal/Noam Sensing And GIS Inputs, Earsel
%20Levin/1999-fundamentals-of-remote Eproceedings 10, 1/2011.
sensing.pdf.,p.p-3,4,5. [15] BHAGAWAT RIMAL, Application of
[5] Goswami S. B., Dr. Aruna Saxena and Dr. Remote Sensing And GIS, Land Use/Land
G. D. Bairagi, Remote Sensing and GIS Cover Change In Kathmandu
based wheat crop acreage estimation of Metropolitan City, Nepal, Journal of
Indore district, M.P., International Theoretical and Applied Information
Journal of Emerging Technology and Technology, JATIT & LLS, 2005 – 2011.
Advanced Engineering, Volume 2, Issue 3, [16] Dr. Lazarus M. Ojigi, Analysis of Spatial
March 2012. Variations Of Abuja Land Use And Land
[6] http://www.seosproject.eu/modules/agricu Cover From Image Classification
lture/agriculture-c00-p01-a.html. Algorithms, ISPRS Commission Vii Mid-
[7] Sandeep Kr. Soni, Crop Area Estimation Term Symposium, Enschede, The
for Bundi Tahsil of Rajasthan using Netherlands, 8 – 11th May 2006.
Remote Sensing and GIS Technique, [17] Sudeesh S and Sudhakar Reddy C,
Geospatial World Forum, Hyderabad, Vegetation and Land Cover Mapping of
India, 18-21-janeuary 2011. Nagarjunasagar-Srisailam Tiger Reserve,
[8] Amber L. (Williams) Coleman and John Andhra Pradesh, India using Remote
M. Galbraith, Using GIS as an Sensing and GIS, International Journal of
Agricultural Land-Use Planning Tool, Geomatics And Geosciences Volume 2, No
Virginia Agricultural Experiment Station, 4, 2012.
Bulletin 00-2, Department of Crop and [18] Atiqur Rahman, Sunil Kumar, Shahab
Soil Environmental Science College of Fazal and Masood A Siddiqui,
Agriculture and Life Sciences Virginia Assessment of Land use/land cover
Tech, December-2000. Change in the North-West District of
http://www.vaes.vt.edu/research/publicati Delhi Using Remote Sensing and GIS
ons/index.html. Techniques, Indian Society of Remote
[9] Ujoh Fanan, Kwabe Isa Dlama and Sensing, DOI 10.1007/s12524-011-0165-
Ifatimehin Olarewaju Oluseyi, Urban 4, 05 January 2012.
expansion and vegetal cover loss in and [19] T. V. Ramachandra, Uttam Kumar,
around Nigeria’s Federal Capital City, Geographic Resources Decision Support
Journal of Ecology and the Natural System for land use, land cover dynamics
Environment Vol. 3(1), pp. 1-10, January analysis, Proceedings of the FOSS/GRASS
2011. Users Conference - Bangkok, Thailand,
http://www.academicjournals.org/jene. 12-14 September 2004.
[10] Lakshumanan.C, Pradeep Kishore.V, [20] Daniel Ayalew Mengistu and Ayobami T.
Viveganandan.S, Krishnakumar.P, Salami, Application of remote sensing and
Muthusankar.G, Landuse / Land cover GIS inland use/land cover mapping and
dynamics study in Nilgiris district part of change detection in a part of south western
Western Ghats, Tamilnadu, International Nigeria, African Journal of Environmental
Journal of Geomatics And Geosciences, Science and Technology Vol. 1 (5), pp.
Volume 2, No 3, 2012. 099-109, December, 2007.
[11] Emilio Chuvieco and Alfredo Huete, [21] Seyed Zeynalabedin Hossein, Sayed
Fundamentals of Satellite Remote Sensing Jamaleddin Khajeddin, Hossein Azarnivad
(CRC Press, Taylor & Francis Group, and Seyed Ali Khalilpour, Land Use
2010). Mapping Using ETM+ Data (Case
[12] Ian Heywood, Sarah Cornelius and Steve Study:Chamestan Area, Iran), College Of
Carver, An Introduction to Geographic Natural Resources & Desert Studies,
Information Technology (Third edition, University Of Yazd, P.O.Box 89195-741
Pearson, 2010). Yazd, Iran.
[13] Arun Mondal, Subhanil Guha, Prabhash http://www.isprs.org/proceedings/XXXV/c
Kumar Mishra, Sananda Kundu, Land ongress/comm7/papers/75.pdf.
Use/Land Cover Changes In Hugli [22] Haruhisa Shimoda And Kiyonari Fukue,
Estuary Using Fuzzy C-Mean Algorithm, Eastern Asia Land Cover Classification
International Journal of Geomatics And Using Modis Surface Reflectance
Geosciences, Volume 2, No 2, 2011. Products, ISPRS Archives Xxxviii-8/W3
[14] Offer Rozenstein, And Arnon Karnieli, Workshop Proceedings: Impact Of
Comparison Of Methods For Land-Use Climate Change On Agriculture, Tokyo
Classification Incorporating Remote 151, Japan.
http://www.isprs.org/proceedings/XXXVII

89 | P a g e
Amol D. Vibhute, Dr. Bharti W. Gawali / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 3, May-Jun 2013, pp.081-091
I/8-W3/B5/1-6-125----------------- Resources. Paper 217, The SAGE
12%20Dec%20formatted.pdf. Handbook of Remote Sensing, (SAGE
[23] Edwin Martínez Martínez, Remote Publications. 14 Apr. 2010), pp.4, 5, 13,
Sensing Techniques For Land Use 14., 1-1-2009.
Classification of Rio Jauca Watershed [32] Sebastian Mader, Michael Vohland,
Using IKONOS Images, Agricultural and Thomas Jarmer and Markus Casper, Crop
Biosystems Engineering Department, Classification With Hyperspectral Data of
University of Puerto Rico-Mayagüez. The Hymap Sensor Using Different
http://gers.uprm.edu/geol6225/pdfs/e_mar Feature Extraction Techniques,
tinez.pdf. Proceedings of the 2nd Workshop of the
[24] Anil Z Chitade, Impact Analysis Of Open EARSeL SIG on Land Use and Land
Cast Coal Mines On Land Use/ Land Cover, University of Trier, Germany, 28-
Cover Using Remote Sensing And GIS 30 September 2006.
Technique: A Case Study, International [33] Shwetank, Jain Kamal and Bhatia K. J.,
Journal of Engineering Science And Review of Rice Crop Identification and
Technology, Vol. 2 (12), 7171-7176, 2010. Classification using Hyper-Spectral Image
[25] Vemu Sreenivasulu and Pinnamaneni Processing System, International Journal
Udaya Bhaskar, Change Detection in of Computer Science & Communication,
Landuse and landcover using Remote Vol. 1, No.1, January-June 2010, pp. 253-
Sensing and GIS Techniques, 258.
International Journal of Engineering [34] G. Camps-Valls, L. G´omez-Chova, J.
Science and Technology, Vol. 2(12), 7758- Calpe-Maravilla, E. Soria-Olivas, J. D.
7762, 2010. Mart´ın-Guerrero and J. Moreno, Support
[26] Vijaya Musande, Anil Kumar and Vector Machines for Crop Classification
Karbhari Kale, Cotton Crop Using Hyperspectral Data, Information
Discrimination Using Fuzzy Classification Society Technologies (IST) programme of
Approach, Indian Society of Remote the European Community, Spain.
Sensing, DOI 10.1007/s12524-012-0201- [35] Md. Rejaur Rahman, A.H.M. Hedayutul
z, 6 January 2012. Islam and Md. Ataur Rahman, NDVI
[27] Carlos Antonio Oliveira Vieira, Paul Derived Sugarcane Area Identification and
Mather and Paul Aplin, Agricultural Crop Crop Condition Assessment, Dept. of
Classification Using The Spectral- Geography & Environmental Studies,
Temporal Response Surface, Anais XI University of Rajshahi,Rajshahi-
SBSR, Belo Horizonte, Brazil, INPE, 05- 6205,Bangladesh.
10 April 2003. http://www.ida.liu.se/~746A27/Literature/
[28] Tippathoo TEEKAPAKVISIT, Crop NDVI%20derived%20sugar%20cane%20
Identification Using Radarsat-2 Multi- area%20identification.pdf.
Polarization Data : A Case Study Of [36] Abhishek Kumar Maurya, Estimation of
Lopburi Province, Thailand, Geo- Acreage & Crop Production Through
Informatics Scientist, Institute of Space Remote Sensing & GIS Technique,
Knowledge Development, Geo-Informatics Geospatial World Forum, Hyderabad,
and Space Technology Development India, 18-21-Janeuary 2011.
Agency (Public organization), 120 [37] Jesus Soria-Ruiz, Yolanda Fernandez-
(Building B) Chaeng Wattana Road, Laksi Ordonez and Heather McNairn, Corn
District, Bangkok 10210, Thailand. Monitoring and Crop Yield Using Optical
[29] G. S. Jang, K. A. Sudduth, E. J. Sadler and and Microwave Remote Sensing,
R. N. Lerch, Watershed-Scale Crop Type Geoscience and Remote Sensing, Pie-Gee
Classification Using Seasonal Trends In Peter Ho(Ed.), ISBN:978-953-307-003-2,
Remote Sensing‐Derived Vegetation In Tech, Mexico, Canada,
Indices, American Society of Agricultural 01,October,2009.
and Biological Engineers, Vol. 52(5): [38] Çağatay TANRIVERDİ, A Review of
1535-1544, 2009. Remote Sensing and Vegetation Indices in
[30] Cankut ormeci, ugur alganci and elif Precision Farming, KSU. Journal of
sertel, Identification of Crop Areas Using Science and Engineering 9(1), KSÜ. Fen
SPOT – 5 Data, TS 4H – Remote Sensing ve Mühendislik Dergisi, 9(1), 2006.
and Imagery I, Sydney, Australia, 11-16 [39] Kamaruzaman Jusoff, Land Use and
April 2010. Cover Mapping with Airborne
[31] M. Duane Nellis, Kevin P. Price and Hyperspectral Imager in Setiu, Malaysia,
Donald Rundquist, Remote Sensing of journal of agricultural science,
Cropland Agriculture, Papers in Natural, vol.1,No.1,December-2009.

90 | P a g e
 Amol D. Vibhute, Dr. Bharti W. Gawali / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 3, May-Jun 2013, pp.081-091
[40] Sudeesh S And Sudhakar Reddy C, [53] http://www.sc.chula.ac.th/courseware/230
Vegetation And Land Cover Mapping Of 9507/Lecture/remote18.htm.
Nagarjunasagar-Srisailam Tiger Reserve, [54] http://en.wikipedia.org/wiki/Mahalanobis_
Andhra Pradesh, India Using Remote distance.
Sensing And GIS, International Journal [55] Md. Rejaur Rahman and S.K. Saha,
of Geomatics And Geosciences, Volume 2, Multi-resolution Segmentation for Object-
No 4, 2012 . based Classifi cation and Accuracy
[41] Shesha Sai MVR, Ramana KV and Assessment of Land Use/Land Cover
Hebbar R, Remote Sensing Applications in Classifi cation using Remotely Sensed
Agriculture, (National Remote Sensing Data, J. Indian Soc. Remote Sens. (June
Centre, ISRO, Department of Space, 2008) 36:189-201, 30 November 2007.
Government of India, Balanagar, [56] Hojat Ahmadi And Kaveh Mollazade,
Hyderabad-500 625, A.P., India), 1-21. Application of Remote Sensing In The
[42] R. Kalpana, S. Natarajan, S. Mythili, D.E. Agricultural Land-Use Assessment,
Shekinah and J. Krishnarajan, Remote Journal of Theoretical and Applied
Sensing For Crop Monitoring - A Review, Information Technology, Jatit, 2005 -
Agric. Rev, 24(1):31-39, 2003. 2009. Www.Jatit.Org.
[43] http://www.sc.chula.ac.th/courseware/230 [57] Ramchandra T.V., Regional land cover
9507/Lecture/remote18.htm. mapping using remote sensing data,
[44] Meng Ji-Hua and Wu Bing-Fang, Study Journal of Agricultural, food and
On The Crop Condition Monitoring Environmental sciences, Volume 2, Issue
Methods With Remote Sensing, The 1, 2008.
International Archives of the [58] J.T. Aranha, H.F. Viana And R.
Photogrammetry, Remote Sensing And Rodrigues, Vegetation Classification and
Spatial Information Sciences. Vol. Quantification By Satellite Image
XXXVII., Part B8., 2008. Processing. A Case Study In North
[45] F.J. Gallego, Review of The Main Remote Portugal, International Conference And
Sensing Methods For Crop Area Exhibition On Bioenergy, April 6th – 9th
Estimates, ISPRS Archives Xxxvi-8/W48 2008.
Workshop Proceedings: Remote Sensing [59] Guifeng Han and Jianhua Xu, Vegetation
Support To Crop Yield Forecast And Area classification in eastern china using time
Estimates, Agriculture Unit, IPSC, JRC, series NDVI Images, Proc. SPIE 6790,
21020 Ispra (Va), Italy, 2006. MIPPR 2007: Remote Sensing and GIS
[46] http://www.seos- Data Processing and Applications; and
project.eu/modules/agriculture/agriculture Innovative Multispectral Technology and
-c02-p01.html. Applications, 67901N,
[47] Mehrez Zribi, Nicolas Baghdadi and doi:10.1117/12.749124, 14 November
Michel Nolin, Editorial Remote Sensing 2007.
of Soil, Applied and Environmental Soil
Science, Volume 2011, Article ID 904561,
2 pages, France, 21 November 2011.
[48] S K Saha, Microwave Remote Sensing In
Soil Quality Assessment, ISPRS Archives
Xxxviii- 8/W20; Workshop Proceedings:
Earth Observation For Terrestrial
Ecosystems, Agriculture And Soils
Division, Indian Institute Of Remote
Sensing (Iirs), Indian Space Research
Organization(Isro), 4, Kalidas Road,
Dehradun – 248001, India.
[49] http://en.wikipedia.org/wiki/Precision_agr
iculture.
[50] http://www.ncpahindia.com/v2/precision-
farming.
[51] http://www.seos-
project.eu/modules/agriculture/agriculture
-c04-p01.html.
[52] http://wgbis.ces.iisc.ernet.in/energy/paper/
MODIS/intro.htm#gmlc.

91 | P a g e

View publication stats