Monuments are a vital part of the heritage of any culture that witnesses the history of our time. These monuments are inherited from previous generations, and it is essential to preserve them through monitoring. The analysis of monument degradation employs a nondestructive method. One of the most effective Non-Destructive Techniques is Deep Learning Technique. The collected images from various monuments are preprocessed and fused with local binary pattern and Luminance. The integrated K-Means clustering used in the proposed approach automatically segments the moss and cracks in monuments, and multilayer neural networks are used to classify the decay. Performance parameters are assessed in terms of precision, accuracy, recall, and F1-Score after the decay parameters are identified using MLNN and validated. The novelty of the proposed work classifies the decay as moss or crack with an accuracy of approximately 97%.