Lane width evaluation is one of the crucial aspects in road safety inspection. Moreover, lane width is one of the most important attributes of a road digital map for many applications, such as driver assistance systems and risk assessment. Using mobile mapping systems (MMSs) equipped with laser scanners is a safe and cost-effective method for rapidly collecting detailed information about various characteristics along any transportation corridor. This report presents an approach to derive lane width estimates using point clouds acquired from a calibrated mobile mapping system. To derive point clouds with high positional accuracy, estimation of the mounting parameters relating the different laser scanners to the onboard GNSS/INS unit is the foremost and necessary step. This report proposes a multi-unit LiDAR system calibration procedure where the mounting parameters can be estimated through minimizing the normal distance between conjugate planar/linear features in overlapping point clouds derived from different drive-runs. After generating the LiDAR point cloud using the estimated mounting parameters, the road surface can be extracted with the assistance of navigation data, which in turn is used to identify lane markings. Lane markings have a high retro-reflective property that will be exhibited as high-intensity points when scanned by a laser scanner. This property can be used to distinguish lane markings from the extracted road surface in the LiDAR point cloud. Then, non-lane marking points among the extracted high-intensity points are identified and removed. Next, the lane marking centerline is derived for lane width estimation.