Detection and classification of land cover change using ISOCLUST classifier and Land Change Modeler

Abstract

The agricultural sector is a mainstay sector in the Indonesian economy. The government can detect and classify land cover change phenomena using remote sensing technology to present agricultural improvement solutions. Remote sensing is an effective tool for monitoring change that occurs continuously and in a large area. This technology can record the earth’s surface from a height of hundreds of kilometers in outer space. This paper uses the ISOCLUST classifier (by producing 6 classes) and Land Change Modeler on Landsat image data in the study area located in Malang Regency, East Java Province. ISOCLUST is an iterative supervised/unsupervised hybrid classifier. Land Change Modeler is an innovative land planning and decision support software that simplifies change analysis, resource management, and habitat assessment. The results of the study using ISOCLUST showed that the overall Kappa was in medium category, 0.4018 and 0.5711. The agriculture area increased to 14.94%, the open forest area increased to 6.46% between 2015 and 2020, while sparse vegetation decreased by 19.36%. The process of detection and classification of land cover using the ISOCLUST classifier and Land Change Modeler shows that the use of remote sensing satellite data is still a consideration in government policy recommendations.