Accuracy Improvement For Remote Sensing Based Lithological Mapping By Using Randomisation And Categorical Coincidence

Rio Priandri Nugroho (Universitas Pertamina, Jakarta Indonesia)
Mohamad Roviansah (Universitas Pertamina, Jakarta Indonesia)

TEKNOLOGIA
Vol 3 No 1 (2020): Jurnal Teknologia

Remote sensing based lithological mapping is commonly applied in the field of earth science as it requires less resource in contrast with real field work. Limitation regarding low accuracy is a challenge that should be tackled in applying remote-sensing based classification. In this paper, an attempt to improve overall accuracy of image classification using randomisation and categorical coincidence analysis was performed. It yielded final majority classification map which has higher overall accuracy compared to the overall accuracy of the population average and the overall accuracy of the map created by including all training data.