LAND USE / LAND COVER CHANGE
DETECTION USING REMOTE SENSING
TECHNIQUES
(As a Case Study of Dehradun City Area from Year 2000 to 2006)
Supervised By
Mrs. MINAKSHI KUMAR
Scientist / Engineer SE
Photogrammetry and Remote Sensing Division
INDIAN INSTITUTE OF REMOTE SENSING
A. HARINDA LAKMAL SRI LANKA
A.M.K.B. ABEYSINGHE SRI LANKA
FAUSTO LAZARO ACOSTA AGUILA CUBA
JITENDRA KUMAR FIJI ISLANDS
MANOJ PRAKASH SINGH NEPAL
INTRODUCTION
LAND USE
Changes on the earths surface features brought about by the
intense human activities like agriculture, urbanization,
industrialization, hydroelectric and thermal power generation,
transportation network activity etc.
LAND COVER
Earths surface features which are not affected by
human activities. However, the land cover does change
by natural force like those associated with geological
activities, weather and climate.
Objective and Scopes
To determine the land use / land cover
changes during the specific period, in a
particular case study area by using
remote sensing techniques.
Scopes
Identify the case study area and satellite data
availability.
Classified each satellite data using Hybrid
Classification.
Identify land use/ land cover changes with respect to
the Level I Classification scheme.
Quantify the changes with in the individual classes.
OVERALL METHODOLOGY
IDENTIFICATION OF CASE STUDY
AREA & DATA AVAILABILITY
IMAGE CO-REGISTRATION &
RESAMPLING
HYBRID CLASSIFICATION,
ACCURACY ASSESSMENT &
CHANGE DETECTION
EXPERT CLASSIFICATION, CHANGE
DETECTION & QUANTIFCATION
STUDY AREA
Our study area in DehraDun lies between
33361281.41 N to 3550992.91 N and
211069.82 E to 222726.32 E coordinates
with respect to projection of zone no. 44N
UTM on WGS 84 datum.
It covers approximately 120 Sq. Km. of
DehraDun city area.
DehraDun is the capital of Uttarakhand
State, which is northern west part of India.
DATA AVAILABILITY
Satellite Data
Criteria
Landsat-7
Dataset
IRS 1C/1D
Dataset
Sensor
MSS
LISS
Year of captured
2000
2006
Spectral Resolution
Spatial Resolution
28.5
23.5
Radiometric
Resolution
Projection System
UTM,44N,WGS8
4
Not available
Capacity
1.01MB
0.6MB
DATA AVAILABILITY
Topographic sheets
53 F/15 and 53 J/3
Scale: 1:50,000
Software:
ERDAS EMAGINE Version 9.1
Satellite Data Preparation
Land sat-7 2000 was consisted with isolated 7
bands in geo TIFF format. Select Band No. 1-5
and 7. All 6 bands were staked into 1 image
and subset was used in this study.
IRS 1C/1D 2006 data consists of 4 band
image, which was co-registered with respect to
Landsat-7 image and relevant subset was used
in this study.
Study Area in LandSat 7 image
Study Area in IRS 1C, 1D image
Image Co-Registration and Resample
Firstly we selected 1st order polynomial
as a co-registration function according
to the same coordinate as Landsat 7
image, zone UTM 44N and WGS 84
Ellipsoid.
Then we select prominent control points
in LISS III 2006 IRS 1C/1D data with
respect to the reference points in
LandSat 7 - 2000 data.
Image Co-Registration and Resample
Likewise we gave around 9 control points
to the LISS 3 - 2006 data and obtained
RMS error of 0.4678 which was less than
0.5 as per the threshold value.
Image Co-Registration and Resample
Then we resampled the image using
nearest neighborhood method. However
this method uses the data file value of
the pixel closest to the target pixel to
assign to the output pixel size of 28.5.
Finally this is how we managed to coregister and resampled our image for
classification.
Image Co-Registration and Resample
HYBRID CLASSIFICATION
Hybrid Classification Flow diagram
LandSat 7 image
Co-registered IRS 1C, 1D
image
Selection of Training Data /
Signatures
Selection of Training Data /
Signatures
Evaluation of Training Data
/ Signatures
Evaluation of Training Data
/ Signatures
Supervised Classification
output of LandSat 7 image
Supervised Classification
output of IRS 1C, 1D image
at 7 image
Hybrid component, digitize,
Recode and overlay
Hybrid Classification
output of LandSat 7
image
94.3%
0.9309
Following signatures were used for
their respective classes,
Signature Name
Urban (UR)
Degraded (DE)
Agriculture (AG)
Stream (ST)
Fallow (FA)
Grassland (GR)
Dense Vegetation
(DVEG)
Va
lue
1
2
4
5
6
7
9
Or
de
r
25
15
23
28
24
33
22
No. of Pixels
2000
419
421
216
127
210
86
384
2006
378
364
133
127
110
132
287
Accuracy assessments were computed by
using contingency matrix and overall Kappa
statistic of each satellite image.
Hybrid Classification
output of IRS 1C, 1D
image
96.1%
OVERALL ACCURACY
0.9530
OVERALL Kappa Statistic
Final 2000
Final 2006
Class Name
URBAN AREA
DEGRADED
DENSE
VEGETATION
STREAM
FALLOW
GRASSLAND
AGRICULTURE
Landsat Data (2000)
LISS 3 Data(2006)
Difference
No. of
Area
No. of
Area
No.of
Area
Precentage
Pixel
(Sq. m)
Pixel
(Sq. m)
Pixel
(Sq. m)
(%)
33738
961533
49634
1414569
15896
453036
47.12
6600
188100
7555
215317.5
955
27217.5
14.47
25334
5974
37698
6948
31717
722019
170259
1074393
198018
903934.5
35171
5937
27496
3802
18481
1002373.5
169204.5
783636
108357
526708.5
9837
-37
-10202
-3146
-13236
280354.5
-1054.5
-290757
-89661
-377226
38.83
-0.62
-27.06
-45.28
-41.73
EXPERT CLASSIFICATION
Objective
Identify the pixel based qualitative changes of
two images.
Target
Identify the pixel based land use changes
from year 2000 to year 2006.
e.g.
Year 2000
Year 2006
URBAN
UR.
DE.
AG.
ST.
FA.
GR.
DV.
EXPERT CLASSIFICATION cont.
VARIBALES
YEAR
2000
CLASS
VALUE
RULE
YEAR
2006
CLASS
VALUE
1
2
4
5
6
7
9
Classified Year 2000
Image
Urban - 1
Degraded - 2
Agriculture - 4
Stream - 5
Fallow - 6
Grassland - 7
Dense Vegetation - 9
Classified Year 2006
Image
Urban - 1
Degraded - 2
Agriculture - 4
Stream - 5
Fallow - 6
Grassland - 7
Dense Vegetation - 9
EXPERT CLASSIFICATION cont.
HYPOTHESIS
YEAR
2000
CLASS
VALUE
YEAR
2006
CLASS
VALUE
1
2
4
5
6
7
9
URBAN
URBAN TO DEGRADE
URBAN TO AGRICULTURE
URBAN TO STRAM
URBAN TO FALLOW
URBAN TO GRASS LAND
URBAN TO DENSE VEG.
EXPERT CLASSIFICATION cont.
(KNOWLEDGE BASE MODEL)
QUANTITATIVE EVALUATION
OF RESULTS
Class Name
Class
No.
No. of Pixels
Area (sq. m)
Urban
29953
24329324.25
Urban to Degrade
85
69041.25
Urban to Agriculture
909
738335.25
Urban to Stream
34
27616.50
Urban to Fallow
1521
1235432.25
Urban to Grassland
35
28428.75
Urban to Dense Vegetation
1175
954393.75
Degrade
4271
3469119.75
Degrade to Urban
862
700159.50
Degrade to Agriculture
10
171
138894.75
Degrade to Stream
11
14
11371.50
Degrade to Fallow
12
759
616497.75
Degrade to Grassland
13
39
31677.75
OUTPUT MAP OF
EXPERT
CLASSIFICATION
CONCLUSIONS
RESULTS FROM THE HYBRID CLASSIFICATION, AREA OF
URBAN, DEGRADED AND DENSE VEGETATION WERE
INCREASED AND FALLOW, AGRICULTURE; GRASSLAND AND
STREAM WERE DECREASED DURING THE PERIOD OF 2000
2006.
AGRICULTURE AND FALLOW AREAS IN THE STUDY AREA WAS
HIGHLY DECREASED DURING 2000-2006.
EXPERT
CLASSIFICATION
CLEARLY
INDICATES
THAT
INCREASE OF URBAN AREA IS MAINLY DUE TO CHANGE FROM
AGRICULTURAL AND FALLOW LANDS TO URBAN AREAS.
THANK YOU.
