Analysis of Deep Learning Methods for Fracture Segmentation of Road surface under on Encoder and Decoder

Road imperfections have a significant impact on driving in terms of both safety and convenience. It's for this reason that road degradation identification and restoration assessments are carried out on a regular basis. Road accidents, especially in developing countries like India, are often caused by poorly designed roads and highways that have been damaged by cracks or potholes. Road accident and fatality rates may be reduced if interested parties get timely and accurate information about dangerous road conditions. A new topology for neural networks, as well as training and prediction approaches is discussed in this research. Road surface deterioration conditions may be detected using a unique deep neural network approach proposed here. 1300 training photos and 400 testing photographs with various sorts of road distress are used to train the neural network. Nine deep learning models from diverse disciplines are compared with the suggested approach. The suggested method exceeds all others with a pixel accuracy of 97.61%, according to the comparison findings. The findings of this research might play a significant role in the future in assuring safe driving by quickly identifying dangerous highway conditions.