Main Article Content
The appearance of cracks and other distortions caused by building movement can be visually unattractive and disconcerting for occupants and if left untreated they can affect the integrity, safety and stability of the structure. Accuracy is crucial for disaster response and recovery efforts. Inspection of surface cracks humanly is laboursive and danger. So, here we compare whether the proposed model can assess the damage accurately like human or even better than human using python and deep learning. In recent years, deep learning techniques have shown great potential in automating damage assessment tasks. In this, we propose convolutional neural networks and recurrent neural network. Firstly, we compile the dataset of crack images with varying degrees of damage. The dataset is pre-processed to normalize the images and augment the dataset to improve model generalization. Experimental results on the compiled dataset demonstrate the effectiveness of the proposed approach for damage assessment in buildings. This shows >90% of accurate prediction. The developed Python-based deep learning model shows promising accuracy and efficiency in identifying different types of damage in buildings.