Open AccessApplying Support Vector Machine Algorithm on Multispectral Remotely sensed satellite image for Geospatial Analysis
Author(s) -
Fatima Hashim,
Hayder Dibs,
Hussein Sabah Jaber
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1963/1/012110
Subject(s) - support vector machine , land cover , multispectral image , confusion matrix , remote sensing , computer science , geospatial analysis , kernel (algebra) , satellite , artificial intelligence , cohen's kappa , pattern recognition (psychology) , land use , geography , mathematics , machine learning , civil engineering , combinatorics , aerospace engineering , engineering
In this research support vector machine (SVM) method apply to classify the satellite image and produce land use and land cover (LULC) map. The used data is the multispectral Landsat-8 OLI satellite image with a spatial resolution of (30 × 30) m 2 . However, the Karbala city was the study area. The SVM Applied with the default parameters of Kernel type, gamma in kernel function, penalty parameter and classification probability threshold. The SVM method achieved high accuracy in separating the categories of the study area based on the test samples collected from the study area in the Karbala province, Iraq. The classification training sites were selected based on visual interpretation and Google Earth Program. The image classification carried for six classes of the study area (Urban Area, Vegetation Area, Soil-1, Soil-2, Water Bodies and Roads). The results show a good accuracy of using SVM method based on relying on the capabilities and the precision of each pixel within the categories. The result evaluation was performed using the confusion matrix, the Kappa coefficient and the overall were 0.89 and 90.61% respectively. The SVM method is able to classify the land use and land cover of the study area with good and accurate results.