Modeling of the movement of quicksands using Landsat satellite imagery and artificial neural network (Study area: Lut plain)

Alireza Rezaei ¹ (Ph.D candidate in urban planning Islamic Azad University Larestan branch)
marzieh mogholi ² (Faculty Member of Islamic Azad University, Larestan Branch)

Submited date : 2021-09-03 accepted date : 2022-07-16

Keywords: Modeling, Artificial Neural Networks, Wind Erosion, remote sensing, lute desert,

Abstract :

To provide digital elevation model, the method of digitalising the alignment lines in topographic maps and then their interpolation is used. First, topographic maps of 1: 50000 area were converted to TIN by GIS software, then TIN map was prepared and then DEM of the region was prepared. Corrections and processing were done on the images. For artificial neural network models, the Artificial Neural Network Toolkit has been used with spss16 software. The displacement of sand dunes was modeled using artificial neural networks and regression models. In a modeling process, either statistical or artificial neural networks, data are usually divided into two parts. One part for training or calibration or calibration of the model and another part for testing or verifying the model. The results showed that remote sensing, artificial neural networks and regression models have good ability to model the displacement of sand dunes. In this research, remote sensing technology was used in the study and monitoring of the morphology of active sand gravel and its source in the region. In order to provide Quaternary geomorphological maps, improve geological map, separation of sand types in the Lut plain and, finally, Harvesting and sedimentation have been used.