Road Damage Detection in Low-Light Environments Using Deep Vision Models and Spatial Mapping

Abstract

Road infrastructure is highly crucial in maintenance. safe transportation, and economy stable. It's important to find damage of road, such as potholes and cracks, early to prevent accidents. and keep repair costs down. Nevertheless, a lot of systems today that don't use deep vision models, which are mostly used to detect road damage. operate effectively even in low-light conditions e.g. at night or in the dark. city areas. The approach described in this paper is a new way of combining. high vision models having tools to enhance low-light images and a system to map locations. It is based on a deep learning method. model to clarify images and make them more visible, and then employs a top- carrying out object detection system to detect various types of. road damage. It also washes spatial mapping to relate the identified. damage using real-world locations, which can be properly mapped. of highway issues on the whole highway system. Testing this method on normal and customcrafted low-light data. demonstrates that it is far more effective than conventional techniques trained. only on daylight images. It is possible to use this system in self-driving, vehicles, urban design, and intelligent city road maintenance.