Assessment of the Land Cover Suitability in Elderly-Friendly Tourist Areas, Using A Convolutional Neural Network

Abstract

Purpose - This study aimed to identify land cover in tourist areas of Indonesia popular among elderly individuals. However, many of these locations do not meet the requirements to be considered elderly- friendly tourist sites. By detecting the land cover, we can determine which tourist sites suit elderly visitors., it is important to ensure that tourist destinations meet the criteria for being elderly-friendly. Design/methodology/approach - This study utilized a dataset of 420 UAV imagery taken from www.openaerialmap.org. The dataset was processed using a Convolutional Neural Network (CNN), specifically the YOLOv6 model. YOLOv6 is a deep-learning model renowned for its speed and accuracy in object detection. It is a CNN architecture that can generate real-time object predictions in a single feedforward, making it advantageous over other models. YOLOv6 can detect objects in various scales and aspects without the need for resampling. In this study, the YOLOv6 architecture was used to detect seven classes of land cover. The data was split into training and testing sets with a ratio of 80:20. The model produced an accuracy rate of 85,10%. Findings - Indonesia's tourist sites are not entirely friendly towards the elderly, as many locations lack accessibility for this age group. Elderly tourism offers engaging activities for those with reduced activity levels. Originality/value - At present, there are no established standards for identifying tourist locations that are elderly-friendly. This model leverages the power of artificial intelligence to determine the suitability of tourist sites and represents a revolutionary development in tourism science.