In order to analyze the characteristics of the spatial pattern of traditional village landscapes and to protect the integrity and stability of village landscapes， based on remote sensing technology and the ArcGIS platform， the spatial pattern of Jiangtou Ancient Village in Guilin was analyzed using a region-based feature fusion extraction method. The improved Cellular Automata （CA） model was used to simulate the landscape evolution process of Jiangtou Ancient Village. The application results showed that the rapid feature extraction method based on regional fusion effectively extracted landscape distribution features and obtained various landscape distribution result maps. The Producer’s Accuracy （PA）， Overall Accuracy （OA）， Figure of Merit （Fom）， and Kappa index values of the improved CA model reached the highest values of 0.88， 0.86， 0.65， and 0.58， respectively. The results indicate that this model， which achieved higher overall accuracy than other models， can provide technical guidance for the scientific planning of rural spatial patterns in China.
ZHANG et al. (Sun,) studied this question.