Home Articles Monitoring of changes in land use/ land cover using multi- sensor satellite...

Monitoring of changes in land use/ land cover using multi- sensor satellite data

 
Srivastava, S. K.
Allahabad Development Authority, Indira Bhawan, VII/VIII Floor, Civil Lines, Allahabad-211004, India,
E-mail: [email protected]

Gupta, R. D. b
Senior Lecturer, Department of Civil Engineering, Motilal Nehru National Institute of Technology, Allahabad-211 004, India
E-mail: [email protected]

Introduction
Human communities cannot exist in isolation, economically, socially or physically. They have to be inter-dependent for the mutual satisfaction of their requirement and also for meeting their basic needs. For meeting their basic needs, they are also dependent on the nature specially the immediate environment. Flux of population in urban territories due to migration from rural areas coupled with rapid growth in population has disturbed the ecological balance. This process hampers the socio economic sustainable development of any region. Expansion of suburban territory with encroachment in prime land is a matter of concern for all and in particular for the authorities associated with the urban planning and development.

Spatial distribution of land use/ land cover information and its changes is desirable for any planning, management and monitoring programmes at local, regional and national levels. This information not only provides a better understanding of land utilization aspects but also play a vital role in the formulation of policies and program required for developmental planning. Planning means the assessment of future and making provisions for it. For ensuring sustainable development, it is necessary to monitor the ongoing changes in land use/ land cover pattern over a period of time.

Urban planning and development is a continuous process and involves planners, administrators, developers, investors and of course, the residents. In order to achieve sustainable urban planning and to check haphazard development, it is necessary that authorities associated with the urban development generate such planning models so that every bit of the available land is used in most rational and optimal way. This requires the present and past land use/ land cover information of the area and pattern of changes with respect to urban settlements and other local resources (Chaurasia, et. al, 1996).

Remote Sensing Based Change Analysis
There are several urban applications where satellite based remotely sensed data are being applied, namely, urban sprawl/ urban growth trends, mapping and monitoring land use/ land cover, urban change detection and updation, urban utility and infrastructure planning, urban land use zoning, urban environment and impact assessment, urban hydrology, urban management and modeling (Raghavswamy, 1994). Remote sensing techniques offer benefits in the field of land use/ land cover mapping and their change analysis. One of the major advantages of remote sensing systems is their capability for repetitive coverage, which is necessary for change detection studies at global and regional scales.

Detection of changes in the land use/ land cover involves use of at least two period data sets (Jenson, 1986). The changes in land use/ land cover due to natural and human activities can be observed using current and archived remotely sensed data (Luong, 1993). Land use/ land cover change is critically linked to natural and human influences on environment. With the availability of multi-sensor satellite data at very high spatial, spectral and temporal resolutions, it is now possible to prepare up-to-date and accurate land use/ land cover map in less time, at lower cost and with better accuracy.

Keeping the above in view, the present work has been undertaken to prepare the multi-date land use/ land cover maps of Allahabad city from multi-sensor satellite data and to monitor the changes in various land use/ land cover classes using digital remote sensing techniques.

Study Area and Data Used
Allahabad district, one of the important parts of Uttar Pradesh State, is situated at the confluence fork of river Ganga, Yamuna and Saraswati, which are held in highest streams by religious minded people in India. For the present work, Allahabad city consisting of urban settlements has been selected. The study area is bounded between the latitudes 25°18’N to 25°35’N and longitudes 81°42’E to 82°00’E.

For monitoring the changes in land use/ land cover, LISS-III digital Data of 2nd December 2000 of IRS -1D satellite along with IRS-1B LISS-II digital data of 20th December 1994 has been used. The path/ row number of IRS-1D imagery is 101/ 54 while the same is 24/50 for IRS-1B. The Survey of India (SOI) topographical map 63-G at the scale 1:250,000 has been used along with other collateral data such as Modified Master Plan 2001 of Allahabad district.

Methodology
The overall methodology adopted for the preparation of land use/ land cover map and change analysis is shown with the help of a flow chart in Figure 1. Digital image processing techniques have been used for preparation of land use/ land cover maps from the multi-date, multi-sensor satellite data. The Anderson’s multilevel classification system has been adopted (Anderson, et. al., 1976). ERDAS IMAGINE 8.5 image processing software and its GIS analysis capabilities (VECTOR module) have been used for the preparation of multi-date land use/ land cover maps and to monitor the change pattern.

1 Study Area Image Extraction
The SOI topographical map 63-G was first scanned at 400 dpi resolution and then registered using ERDAS IMAGINE software. For registration, 9 well-distributed Ground Control Points (GCPs) at the intersection of latitude-longitude lines have been selected. The second order polynomial transformation with nearest neighborhood resampling technique has been adopted and a registration error of 0.42 pixels was obtained. Process of image-to-image registration is then adopted for the registration of satellite imagery and modified master plan of Allahabad district on already registered SOI topographical map. The registration error (r.m.s.) obtained was 0.49 pixels for IRS-1D LISS-III data and 0.62 pixel for modified master plan. After this, LISS-II imagery is registered over LISS-III imagery by using 16 GCPs with registration error as 0.38 pixels.

 
VECTOR module of ERDAS IMAGINE software has been used for on-screen digitization and to obtain the boundary of study area from the modified master plan 2001 of Allahabad. By using this as raster mask, LISS-III and LISS-II imageries of the study area have been extracted.


Figure 1: Methodology Adopted

2 Preparation of Multi-date Land Use/ Land Cover Maps
For monitoring the changes in the pattern of land utilization, preparation of multi-date land use/ land cover maps are necessary. These maps are also required for ensuring planned development of any region. After the registration of digital images of the Allahabad city, the training samples for each land use and land cover class are collected for both LISS-II and LISS-III imageries. Limited field visits have been undertaken for collecting the training samples. The various land use/ land cover classes selected include built-up, agriculture, sand, water, scrub and open/ barren land.

The multispectral classification was carried out using supervised classification techniques with maximum likelihood classifier. The overall accuracy of the classification is finally obtained through the computation of confusion matrix to assess the reliability of the prepared maps.

Analysis of Results
.The land use/ land cover maps prepared using the methodology described above have been shown in Figure 2 (LISS-II data of 1994) and Figure 3 (LISS-III data of 2000). The various land use and land cover classes delineated include built-up, agriculture, sand, water, scrub and open/ barren land. The spatial coverage of each class may be visualized on both maps. The area of each class for LISS-II (year 1994) and LISS-III (year 2000) data has been compiled in Table 1. An unbiased set of testing samples has been selected from the reference data for assessing the classification accuracy, which comes out to be 79.5% for LISS-II and 78.2 for LISS-III data respectively. Table 1 also shows the change pattern in the land utilization from year 1994 to 2000 for the Allahabad city.

From LISS-III land use/ land cover map of year 2000, it may be seen that built up area constitutes 7.5% of total area, while the agriculture and scrub covers 47.0% and 13.9% area respectively. In year 1994, the built up area obtained from digital image processing techniques has been 6.3% of the total area. The agriculture and scrub covers 50.0% and 14.3% area respectively. The open/ barren land constitutes 22.3% in year 2000 while it was 19.7% in 1994.


Figure 2: Land Use/Land Cover Map

 

VECTOR module of ERDAS IMAGINE software has been used for on-screen digitization and to obtain the boundary of study area from the modified master plan 2001 of Allahabad. By using this as raster mask, LISS-III and LISS-II imageries of the study area have been extracted.


Figure 1: Methodology Adopted

2 Preparation of Multi-date Land Use/ Land Cover Maps
For monitoring the changes in the pattern of land utilization, preparation of multi-date land use/ land cover maps are necessary. These maps are also required for ensuring planned development of any region. After the registration of digital images of the Allahabad city, the training samples for each land use and land cover class are collected for both LISS-II and LISS-III imageries. Limited field visits have been undertaken for collecting the training samples. The various land use/ land cover classes selected include built-up, agriculture, sand, water, scrub and open/ barren land.

The multispectral classification was carried out using supervised classification techniques with maximum likelihood classifier. The overall accuracy of the classification is finally obtained through the computation of confusion matrix to assess the reliability of the prepared maps.

Analysis of Results
.The land use/ land cover maps prepared using the methodology described above have been shown in Figure 2 (LISS-II data of 1994) and Figure 3 (LISS-III data of 2000). The various land use and land cover classes delineated include built-up, agriculture, sand, water, scrub and open/ barren land. The spatial coverage of each class may be visualized on both maps. The area of each class for LISS-II (year 1994) and LISS-III (year 2000) data has been compiled in Table 1. An unbiased set of testing samples has been selected from the reference data for assessing the classification accuracy, which comes out to be 79.5% for LISS-II and 78.2 for LISS-III data respectively. Table 1 also shows the change pattern in the land utilization from year 1994 to 2000 for the Allahabad city.

From LISS-III land use/ land cover map of year 2000, it may be seen that built up area constitutes 7.5% of total area, while the agriculture and scrub covers 47.0% and 13.9% area respectively. In year 1994, the built up area obtained from digital image processing techniques has been 6.3% of the total area. The agriculture and scrub covers 50.0% and 14.3% area respectively. The open/ barren land constitutes 22.3% in year 2000 while it was 19.7% in 1994.


Figure 2: Land Use/Land Cover Map