This dissertation investigates the use of weigh-in-motion (WIM) and continuous count station (CCS) data, remote sensing, and machine learning techniques to maintain resilient infrastructures and employs a deep learning methodology to detect the surface distresses of pavements. WIM and CCS technology obtain much useful traffic information as vehicles move along the roads. Remote sensing methods offer tools to replace or complement existing traditional methods of pavement management systems and can serve many needs of transportation agencies. Machine learning techniques, also facilitate pavement management routines by creating models and predicting the roads' performance and their deterioration status. This study investigates the predictions of resilient infrastructure distress situations by analyzing the historical spectra of flexible pavement road sections derived from Sentinel-2 and PlanetScope satellite images. The study also attempts to utilize a deep learning-based algorithm for predicting the pavement distresses using the images taken from interstate roads in Georgia. The recognition accuracy of the prediction model determines whether the remote sensing platform accompanied by machine learning approaches and Deep Neural Network (DNN) algorithms can be utilized as smart monitoring platforms for distress evaluation of the asphalt infrastructures and further maintenance activities.