Search published articles


Showing 2 results for Maximum Likelihood

Dr Vahid Riahi, Dr Parviz Zeaiean Firouzabadi, Dr Farhad Azizpour, Ms Parastoo Darouei,
Volume 19, Issue 52 (3-2019)
Abstract

The cognition of cropping pattern is important for planning and resource management .Remote sensing as a science and technology of spatial information and geographic information system due to having the analytical facilities can play a key role in determining the distribution of crops and their lands under cultivation. In this research, in order to identify and separate the lands under cultivation of the dominant crops in Lenjanat of Isfahan province, the multi-temporal images of Landsat 8 satellite, OLI sensor were used in the dates of April 17, July 6, and August 23 in 2016. Using maximum likelihood classification and normalized difference vegetation index (NDVI) of the agriculture crops in different periods of growth and according to their cropping calendar, the map of the cropping pattern of the area was determined. To evaluate the accuracy of the results, the produced maps were examined with reference data. Kappa coefficient and overall accuracy were 0.88 and 90%, respectively, in maximum likelihood classification, and 0.90 and 93%, respectively, in NDVI. Furthermore, statistics presented by Agricultural Jihad Organization of Isfahan province in the 2015-2016 crop year was used for evaluation. The results showed that there were differences equal to 10.2%, 18.6% and 1.8%, in the area under cultivation of wheat and barley, rice, and potato and forage, respectively, in maximum likelihood classification, comparing with the statistics of Agriculture Jihad while the results of NDVI comparing with Jihad statistics showed the errors equal to 6.6 %, 6.5 % and 3.2%, respectively, that indicated the better performance of temporal vegetation indices in estimation of area under cultivation according to its phenology. Investigation of land use and cropping pattern of this area indicate a high centralization of agricultural lands with high water requirements and industries on the proximity of Zayanderud River which necessitates the spatial analysis of land use in this area.


Khadijeh Mikaeli Hajikandi, Behrooz Sobhani, Saeid Varamesh,
Volume 23, Issue 68 (4-2023)
Abstract

Study of land use/cover changes is widely used in environmental planning. During the last decade, growing increase of aridity in Uromiyah Basin has become a major regional and even national problem. The purpose of this study is to reveal the changes in land use/cover in the southern and southeastern parts of the basin with using 2 images for month of July of 2000 to 2017. Landsat TM and OLI data and NDVI were used for classification this study. Land use/cover maps in the two studied years were provided using Maximum Likelihood Classifier (MLC) algorithm applied on two series data including spectral bands (data series 1) also spectral bands and filter texture layer (data series 2) and six categories of land use/cover containing Irrigated Farmland, Dry Farmland, garden, rangeland, bare land and water bodies were distinguished.. The accuracy of the produced maps were assessed and compared with the training samples derived from Google Earth images and Kappa Index, overral accuracy, producer accuracy and user accuracy. The results demonstrated that the maps produced using the data series 1 have higher accuracy and the overall accuracy of the maps of 2000 and 2017 using the data series 2 are 98.93 and 98.29 and these values for data series 1 were gained 99.28 and 91.45, respectively. In additional, texture filtering decreased amount of mixing between classes of rangeland, Irrigated Farmland and garden. The results of change detection showed considerable increase in the area of Irrigated Farmland (13.44) and garden 1.85 (27.24) an also at the studied period, the area of the water bodies and rangeland were decreased to 1.58 and 22.94%.
 

Page 1 from 1     

© 2024 CC BY-NC 4.0 | Journal of Applied researches in Geographical Sciences

Designed & Developed by : Yektaweb