CROP TYPE MAPPING USING MACHINE LEARNING-BASED APPROACH AND SENTINEL-2: STUDY IN LUMAJANG, EAST JAVA, INDONESIA
PEMETAAN JENIS TANAMAN MENGGUNAKAN PENDEKATAN MACHINE LEARNING DAN SENTINEL-2: STUDI DI LUMAJANG, JAWA TIMUR, INDONESIA
DOI : https://doi.org/10.35633/inmateh-72-12
Authors
(*) Corresponding authors:
Abstract
In general, sentinel-2 imagery can be used for crop mapping. Crop types mapping aims to develop future strategies for sustainable agricultural systems. This study used Sentinel-2 from June 25 to July 6, 2023, with 10% cloud cover. The research was conducted in Pasrujambe and Candipuro sub-districts (± 242.23 km2). The image is processed using a random forest on the GEE platform. Accuracy was generated using a confusion matrix with an overall accuracy of 85.82% and a kappa of 71.19%. Five main types of land use/cover were produced, namely: paddy (17.31%), sugarcane (0.93%), vegetation (69.74%), sand (7.4%) and built-up land (4.59%).
Abstract in Indonesian