Vegetation Loss Detection of Nineveh Province Using Remote Sensing Images Based on Ant Colony Algorithm

Abstract

Nineveh province in Iraq has experienced a process of land cover conversion and vegetation loss especially in last decades. It is important to get accurate information on vegetation loss and changes in areas that are used for agriculture. Among the most effective methods to study and get information about this phenomenon is remote sensing technology.Since classical approaches lack of accuracy, artificial intelligence has been introduced to strengthen feature detection which leads to better classification. This paper uses ant colony algorithm to study and classify part of Nineveh province land into six classes. These are Agriculture land/flood plane, Water, Outcrop, Origin of early sand sheet, Desertable area, and Sand dunes. The variation in these six classes from 1987 to 2009 is shown. Results show that agriculture region and flood plain decreased from around 31% in 1987 to 11.2% of total area in 2009 while origin of early sand sheet and desertable area increased from 42.7% to around 49%. Beside that sand dune appears in 2009 to form about 26.47% of total area under study.

Multi-spectral images classification based on intelligent water drops algorithm

“…The difference in the spectral beam gives better results. The information from each package separately from the different sensors is a fraction of the earth's factual information [1].…”

“…The use of multimode images in classification is an important step in producing thematic spatial information from satellite imagery data. Over the past two decades, artificial intelligence improvement algorithms used to solve various problems in many disciplines [1].…”

Multispectral Image Classification Based on the Bat Algorithm

“…In [16], Some researchers revealed the loss of vegetation cover for Nineveh Governorate by using remote sensing images based on the algorithm of an ant colony According to the results, the agricultural area and floodplain decreased from about 31% in 1987 to 11.2% in 2009, while the origin of the early sandy plate and desert region increased from 42.7% to 49%. Sand dunes also appeared in 2009 in about 26.47% of the overall study area.…”