Green open space detection and mapping using planetscope-3A image with vegetation index approach and supervised classification in Banda Aceh, Indonesia
Abstract: Role of Green Open Space (GOS) is essential in creating a comfortable environment in cities. It is detected using a high-resolution satellite image, Planetscope-3A. This study aimed to classify landcover in Banda Aceh using a multispectral classification of Planetscope-3A image and to assess the applicability of Planetscope-3A image to identify GOS in Banda Aceh. The multispectral classification was used with a supervised classification-maximum likelihood algorithm that refers to normalized difference vegetation index (NDVI) transformation to obtain eight landcover classes. Additionally, field observation was used to retrieve sample points determined by Stratified Random Sampling. The classification detected eight landcover classes comprising non-GOS objects (water, developed, barren) and GOS objects (trees, shrubland, herbaceous, wetland, and cultivated). The result was combined with 128 samples data of field, producing an accuracy of 76.036 % and a kappa value of 0.726. Landcover was dominated by developed class with 29.739 km2 or 53.6 % of study area total with an accuracy of 94.094 %. Furthermore, GOS in Banda Aceh included 19.589 km2 or 35.291 % of the study area, consisting of trees (6.863 km2, accuracy 79.396 %), Shrubland (8.216 km2, accuracy 59.413 %), Herbaceous (4.132 km2, accuracy 73.564 %), Cultivated (0.291 km2, accuracy 73.475 %), and Wetlands (0.088 km2, accuracy 70.185 %). This concludes that Banda Aceh has a sufficient area of GOS. The result of GOS detection using Planetscope-3A image with supervised classification-maximum likelihood algorithm could be reference data and recommendation in managing sustainable development in Banda Aceh.