Abstract: Accuracy assessment is a tool of validation of remotely sensed data and other spatial
information and analyses the ‘goodness factor’ of the maps derived from satellite images. If we
talk about the remotely sensed data of land use/land cover then accuracy assessment helps in
analyzing the spatial details of particular landscape which in lieu serve the various purposes of
policy making and administrative planning. The data related to land use/land cover is of utmost
importance and helps in environmental modelling specifically models dealing with climate
change at all levels viz. micro level, meso level and macro level. The present research study
aims to analyze the accuracy assessment of remotely sensed data of land use/land cover of
national capital territory of Delhi, India. Three different sets of Landsat images for the years 1990,
2000 and 2015 have been evaluated by level of overall accuracy, producer’s accuracy,
consumer’s accuracy and kappa coefficient.
Key words: Accuracy Assessment, Validation, Land Use/Land Cover, Kappa Coefficient.
Introduction
Accuracy assessment or validation of spatial data facilitates the decision making process. The
estimates based on accuracy assessment are more accurate, efficient and effective. Mapping
related to different issues viz. forestation/deforestation, wetlands and estimation of land use/land
cover will be inappropriate and incomplete without accuracy assessment or validation. This
investigates validity of map with all possible error. The present study of land use/land cover of
NCT of Delhi has been discussed in the light of quantitative accuracy assessment.
Study Area
The study looks at the city of Delhi, national capital of India (Figure 1) due to its high degree of
urbanization, dominant grey infrastructure and high density of population. It is located between
the latitude 28°24´17´´ N and 28°53´00´´ N and longitude 76°50´2´´ E and 77°20´37´´ E. It has an
area of about 1483 km2 with maximum length of 51.90 km and maximum width of 48.48 km. The
study employs Landsat-4, 5, TM for 7 November 1990; Landsat-7 ETM+ for 16 November 2000
and Landsat-8 OLI for 9 November 2015 (see Table 1). All images were obtained from the
archives of United States Geological Survey (USGS). The Three scenes fell within the path 147
and row 40 of the WRS-2 (Worldwide Reference System) from which the data for the location
under the study could be extracted. All bands 1-5 and 7 have spatial resolution of 30 m and the
thermal infrared band (band 6) has a spatial resolution of 60m for Landsat 7 and 120m for
Landsat 5. For Landsat OLI, thermal infrared band (10, 11) has a spatial resolution of 100m.
Objective of the Study
The objectives of the present study are as follows:
1. To identify the land use/land cover classes (1990-2015) through the Anderson’s satellite
image classification in the National Capital Territory of Delhi, India.
2. To analyze the accuracy assessment of land use/land cover classification (1990-2015) of
National Capital Territory of Delhi, India.
By:
Sandesh Yadav and Shams Perwaiz2
1Freelance Researcher, New Delhi, India
2Teacher, Sr. High School, Jokihat, Araria (Bihar) India
Corresponding author’s Email: sandesh_official@yahoo.in