HD-Map assisted lane marking worn out detection from MMS data

This study focus on developing an efficient and quantifiable method for lane marking worn out detection. In this study, the proposed method for lane marking worn out detection is as follows: 1. Extract ground points from the data collected by MMS by using Clothes Simulation Filter(CSF). 2. Generate DEM from the ground points. 3. Based on the lane markings position provided by the High-definition map, extract image segments that contain lane markings. 4. Project image pixels of lane markings onto the DEM and calculate the coverage area. 5. Calculate lane markings worn-out ratio and divide into 10 levels. Worn-out ratio computation is based on the area information of lane markings provided by the High-definition map.