Diagnosis of landscapes of the province of Al-Diwaniyah (Iraq) by using of Landsat-8 multispectral images

Remote sensing monitoring and prediction of crop yields are widely used throughout the world. However, such techniques are not universal and require testing in each landscape region. This paper deals with approaches to using multispectral satellite images to classify landscapes of Al-Diwaniyah, one of the main centers of irrigated agriculture in Iraq. As a data source for the classification, we choose of Landsat-8 images received in March 2014, 2015 and 2016 with similar radiation characteristics. Snapshot dates coincided with the period of crops maximum vegetation, allowing a certain allocation of agricultural land based on Normalized Difference Vegetation Index (NDVI). The analysis of NDVI distributions demonstrate their sensitivity to changes of cereal crops areas and seasonal precipitation sum, that allow us to establish four types of regional landscapes based on NDVI thresholds: landscapes with health vegetation with moderate-to-high projective cover (NDVI > 0,34); landscapes with stressed or low projective cover vegetation (0,19 < NDVI $\le$ 0,34), landscapes without vegetation (0,1 < NDVI $\le$ 0,19) and water landscapes (NDVI $\le$ –0,1). Comparison of classification with the traditional man-made interpretation of landscapes showed that the overall accuracy and F-measure of the proposed automatic classification is about 90 %.