Parallelized remote sensing classifier based on rough set theory algorithm

Abstract

Supervised classification in remote sensing imagery is receiving increasing attention in current research. In order to improve the classification accuracy, a lot of spatial-features (e.g., texture information generated by GLCM) are often utilized. Unfortunately, too many spatial-features usually reduce the computation speed of remote sensing classification, that is, the time complexity may be increased due to the high dimensionality of the data. It is thus necessary to improve the computational performance of traditional classification algorithms which are single process-based, by making use of multiple CPU resources. This study presents a novel parallelized remote sensing classifier based on rough set (PRSCBRS). Feature set is firstly split sub-feature sets into in PRSCBRS; a sub-classifier is then trained with a sub-feature set; and multiple sub-classifier's decisions ensemble are finally utilized to avoid the instable performance a single classifier. The experimental results show that both the classification accuracy and computation speed are all improved in remote sensing classification, compared with the traditional ANN and SVM method.