Drone-based Cultural Landscape Heritage Analysis using Computer Vision and Cultural Landscape Genes

Abstract

Huizhou traditional villages hold a lot of cultural and historical significance as they represent the rich architectural traditions and heritage of the region. These settlements pose unique challenges to management and conservation because of their intricate cultural landscape history. The main goal of this work is to use computer vision techniques in order to determine the genetic information inherent in traditional villages located in Huizhou for practical planning for their preservation. We propose Multi-Criteria Decision Making (MCDM) approaches that are intelligent have been employed, thus addressing decision-making complexities within cultural asset protection environments. Utilization of computer vision technology allows the research to automatically analyse and evaluate what constitutes genes of Traditional Villages’ nationalities in Huizhou’s cultural landscapes. This study presents the architectural features, spatial patterns, cultural artifacts as well as issue on historical context among others. Four approaches are compared: Convolutional Neural network (CNN), Recurrent neural network (RNN), CNN and RNN, and MCDM. The most accurate and consistent approach was shown by MCDM with accuracy being 0.88% while its CR equals 0.08%.