Extraction of linear features from vehicle-borne laser data
Abstract
In this paper, we focus our discussion on auto- extraction of linear features like guard-rails (a fence line at
the edge of the road or middle of the road) from vehicle- borne laser data. The vehicle-borne laser data is
quite heterogeneous in nature as we scan the objects while the vehicle is moving. In order to extract,
linear features, the laser data are projected on the horizontal plane and then rasterized. The raster data
contains grid density image and maximum height image, which are used for assisting in decision-making
process for linear features. The raster data is further converted to binary image using threshold values
for linear features. Radon transformation is applied on the binary image to identify the seed position and
orientation of the most probable linear features. Arbitrary seed lines are drawn from these seed points.
These seed points (and lines) coordinate information are then converted back to the vector data (original
laser points). A circle growing technique is applied on the seed points to correct the seed position of the
linear feature points at certain horizontal spacing. Once all the seed points are corrected on the original
data, straight lines are fitted (locally) to represent the linear features. The height of the linear feature is
computed by fitting the maximum height values of the points that fall inside the circle (during the circle
growing process). This gives us 3-D modeling of linear features.
It is possible to identify linear features from vehicle-borne laser data. The
algorithm is successful in extracting the linear features automatically for
continuous linear features. If the linear features are non- continuous (or
smaller spans of a few meters) or data are occluded, auto-extraction will
be quite complex and might even fail to identify. In this case, a semi-automated
extraction is recommended.
Introduction
Laser point data scanned from vehicle-borne platform can be used for 3- D
modeling of various urban features. Apart from building faces, roads and
trees, there are many other features that can be modeled from laser data.
Some of these are cables, poles, fence or guardrails, tunnels, vehicles and
pedestrians. Refer Manandhar & Shibasaki, 2001 for details on extraction
of some of these features. In this paper we are focusing on the possibility
of automated extraction of linear features (especially guard rails) from laser
data. The range data used no other information except the range distance
itself. The data are bare 3-D real world coordinates. Figure 1 shows the mapping vehicle equipped with the laser scanning system.
Figure 1: Vehicle- borne
Laser Mapping System
Linear Feature Extraction
- Definition
We define linear features that exhibit laser points with linear geometry when viewed along the vehicle
trajectory (along track). For example, laser points reflected by cables, guardrails etc are defined as linear
features. However, laser points reflected by poles are not classified as linear features since they exhibit
points linearly along the scanning direction but not along the vehicle trajectory (across track).
- Linear Feature Extraction
There are different approaches to segment range data. These approaches basically depend on the type
of range data and the features we would like to extract. Refer Hoover et al for comparative study of
various range image segmentation algorithms. These algorithms are developed for fixed platform. Range
data may be either in grid format (2.5D) or point cloud format (3D). The range data we use are point
cloud data that have only 3- D coordinates. The data have already been filtered for the road and non -road
data. We use only the non-road data to identify linear features.
The feature extraction is basically done in three major steps, (a) conversion to raster image and image
analysis (b) Identify seed points by performing radon transformation and ( c) correct seed points / lines by
fitting the identified points / lines.
