Ensemble Deep Learning Using faster RCNN Model and Fuzzy rule System for Health Monitoring of Heritage Castles

Heritage castles detection involves the use of various technologies and methodologies to identify, document, and preserve historic castles and other structures of cultural significance. This process is critical for preserving these structures for future generations and ensuring that their historical and cultural significance is not lost. The heritage castles are recognized through experienced evaluators or computer vision techniques. The process of experienced evaluator is highly laborious and too costly. The computer vision techniques help to recognize the heritage and non-heritage castles automatically and resolve the experienced evaluator issues. In this study, a hybrid approach is based on faster rcnn and resnet-50 models along with fuzzy rules to recognize the health monitoring of heritage castles. Firstly, the heritage castles in form of four objects ("thick walls", "moat", "towers", and "drawbridge") have been detected and their features are classified through Resnet-50 pretrained model. Based on classified features, fuzzy rules are produced which is helpful to determine the health conditions of castle features. The proposed approach has been trained and tested on 4300 preprocessed images. Several performance parameters have been used to employ to measure the performance of the proposed hybrid approach. During the testing process to recognize the different castles recognition, a total of 94.16% accuracy has been achieved in tower feature images. Overall, heritage castle detection is an important process for preserving the rich cultural heritage of the past and ensuring that future generations can continue to appreciate and learn from these historic structures.