Measurement and Management of the Lane Markings' Stripping Ratio from In-vehicle Camera Image

Natural weathering and time-related deterioration by running of vehicles can be seen in road markings. By deterioration of compartment lines, not only the drivers' visibility and the function of visual guidance are reduced but also the line recognition becomes non-functional in spite of important role for the autonomous driving. In this study, we aim to measure and manage the state of deterioration of lane markings. The stripping ratio is measured by image processing technique such as comparison between the pixels of white lines of the road image and the theoretical image that we generate. The road images are continuously synthesized with projection transform from road images taken by an in-vehicle camera and CAN (Controller Area Network) data logged when the vehicle is being driven. The theoretical images are automatically generated by seeking pixels where lane markings should be placed in the synthesis image. The Stripping ratio and the synthesized road images are managed by GIS (Geographic Information System). It is useful to survey the stripping ratio distribution for repainting. This paper, also confirmed the effectiveness of our proposed method through an experiment where the stripping ratio of lane markings was calculated from data obtained by a running vehicle.