Classification of Land and Crops Based on Satellite Images Landsat 8: Case Study SD Timisoara

  • Mihai Hrebei University of Agricultural Sciences and Veterinary Medicine, 3-5 Manastur St., 400372 Cluj-Napoc
  • Florin Sala
Keywords: algorithms, classification, crops, GIS, precision agriculture, satellite images

Abstract

Agriculture benefits increasingly from the several services of related areas, including satellite technology and informatics, which have a substantial contribution lately. The present study aimed to analyze and classify the habitat, and crops of the SD Timisoara based on spectral information from satellite images. Analysis of the habitat and agricultural crops was done with ArcGIS 10 software based on satellite images Landsat8. Of the tested algorithms, Maximum Likelihood was used and as method, the supervised classification was used in order to analyze the spectral information contained by satellite images. The high and stable level of the correlation between spectral bands and satellite index NDVI recorded in May, has determined that satellite images from the certain month should be used for the classification the territory, evaluation of crop structure and areas occupied by them. There was an interdependent relationship between NDVI index and Band 543 having high statistical accuracy (p<< 0.001, R2 = 0.820).  Analysis done by supervised method based on Maximum likelihood algorithm has facilitated land and crop classification in four classes in the year 2013 and six classes in the year 2014 and the determination of crop areas with high level of accuracy.

Published
2016-05-18