Accidents and fatalities from motor vehicle accidents are major concerns despite substantial advancements in safety technology. Because of this, the industry has made significant investments in creating new safety features, like cutting-edge driver assistance systems, and raising public awareness of safe driving habits. In general, car accidents can result in severe damage to the vehicles involved, and assessing and repairing that damage can be time-consuming and expensive. Manual inspection of vehicles is prone to errors and often requires trained professionals to identify the extent and location of the damage. Therefore, there is a need to develop an automated system that can detect and assess the damage caused to vehicles using AI and deep learning techniques. An image-based processing technique, YOLOv3, is proposed in this work to automate damage detection on automobiles. In the work, we used CNN to create a Mask R-convolutional neural Networks model to identify the location of damage on a car. The damaged area is precisely marked in the images. The base weights from the Mask R-CNN COCO dataset are used to train the model. 21 epochs are used to process the images. The surface of the damage is highlighted in the final image using a color splash technique after processing. Auto insurance firms, vehicle rental companies, and repair shops would all benefit from this automated method of determining the degree of exterior vehicle damage and then calculating the severity of that damage. The value of fraudulent auto insurance claims can also be reduced.