Spatiotemporal Land Cover Transformation Assessment of Kirkuk City Using Remote Sensing Data
Abbas Mohammed Noori
Keywords: Change Detection, LULC, ML, NN, Landsat, Remote Sensing.
Abstract
Both Environmental assessment and land management benefit from a thorough thoughtful of land use and land cover change (LULC). Human actions, such as population growth and the demand for new capacity are the main cause for this occurrence. Changes in land cover are detected and classified using Landsat images collected between 2014 and 2022. Two types of supervised classification were applied to categorize the Landsat images of the study area which are Maximum Likelihood (ML) algorithm and Neural Net (NN) classification. Specifically, water, bare land, soil, farmland, urban areas, and vegetation were chosen as six of the signature classifications to be categorized. However, the overall accuracy assessments of ML classifier were 96.19%, 95.18% and 96.93% for period 2014, 2018 and 2022 respectively between 2014 and 2022. In contrast, the overall accuracy assessments of NN classifier were 95.01%, 94.23% and 94.52% for period 2014, 2018 and 2022 respectively. It has found that ML algorithm is accurate than NN classification. The results demonstrate an important rise in the urban area, vegetation and farmland, while, a commensurate decrease in the bare land and soil areas. Land use shifts significantly in response to both population increase and economic development. In order to ensure the long-term viability of city growth, LULC research is extremely helpful for urban planners for the next generation.
