LiDAR Intensity Calibration for Road Marking Extraction

Abstract

The rising demand for precise three-dimensional (3D) maps is a result of the growth of autonomous vehicles into a major focus. These precise 3D maps are often created using Mobile Mapping System (MMS) instruments equipped with various sensors. Among these sensors, Light Detection and Ranging (LiDAR) is the most widely adopted sensor for the generation of 3D maps. This paper describes the intensity calibration process for 3D LiDAR sensors installed on MMS. A typical LiDAR sensor measures the distance to an object in addition to intensity, which is the magnitude of the reflection from the surface of an object. Physically, the value of intensity depends on three factors: reflectance, the distance from the surface, and the angle of incidence. Therefore, the intensity values of road markings with the same reflectance may result in different intensity values depending on distance and angle of incidence. The method proposed in this paper aims to correct the measured intensity value of objects to ensure the measurements of the same object reports the same intensity value. To solve this issue, the intensity characteristic of each sensor with regards to a reference surface was exploited and the reflectance of the surface was estimated. The calibration result was validated via road marking extraction with Otsu threshold method.