Integrated technique of segmentation and classification methods with connected components analysis for road extraction from orthophoto images

Road networks are one of the main urban features. Therefore, road parts extraction from high-resolution remotely sensed imagery and updated database beneficial for many GIS applications. However, owing to presence various types obstacles in images, such as shadows, cars, trees, with similar transparency spectral values class, achieving accurate using different classification segmentation methods is still difficult. This paper proposes an integrated method combining connected components analysis extract class orthophoto images. The proposed technique threefold. First, multiresolution was applied segment Then, methods, namely, decision trees (DT), k-nearest neighbors (KNN), support vector machines (SVM), were implemented based on spectral, geometric, textural information classify obtained results into two classes: non-road. Three accuracy evaluation measures, recall, precision, F1-score, evaluated determine performance method, respective average 87.62%, 89.71%, 88.61%, respectively, DT; 86.61%, 88.17%, 87.30%, KNN; 89.83%, 89.52%, 89.67%, SVM. Finally, labelling used component parts, morphological operation employed delete non-road noises improve performance. These also compared other prior works, which confirmed that effective technique.