Y. Baby Kalpana and S.M. Nandhagopal
LULC Image Classifications using K-Means Clustering and KNN Algorithm
Dynamic Systems and Applications 30 (2021) No. 10, 1640 – 1652

https://doi.org/10.46719/dsa202130.10.07

ABSTRACT.
The human consumption of ecological aspects includes soil, water resources and greenery. Normally, on the earth, the lands may be used and covered by the human being and there are numerous changes in spatial distribution during a period of time. The physical characteristics of a particular land area may be detected and monitored using special cameras or sensors and such technique is called remote sensing. The fields like geography, ecology, land surveys, oceanography and all other earth science disciplines may use remote sensing to obtain the required information’s accordingly. It is also used for military, intelligence, commercial and economic human applications. The Geographic Information System (GIS) is used for encapsulating, accumulating, verifying, and displaying data related to positions on Earth’s surface. GIS helps the researchers to find special kinds of images on a single satellite pictures like lanes, constructions, and plants. Supervised Classification technique is used for the quantitative analysis with the theory of spectral domain segmentation into regions which is associated with ground cover classes of a defined application. A pixel based classification is employed to specify the number of spectral regions based on the number of information’s of the sensed data. This technique is called unsupervised classification in which the information’s are created with the help of pixel values for each spectral bands. In this work, K means clustering algorithm and KNN algorithms are used to tabulate the Land used for Coimbatore district at a specific range of 11.01° N and 76.9°E. The two algorithms are employed for the two different classification techniques which are compared with the Tamilnadu gazette data.

Keywords: Image Differencing, Classification Approaches, Multi-Scale Amalgamation, Land Use, Remote Sensing Images.