Data Fusion

The integration of Light Detection and Ranging (LiDAR) captured data from aerial, mobile and/or terrestrial platforms, plus traditional survey techniques, has exposed new vistas in data processing and product generation for the 2D, 3D, and 4D environment. In particular, the fusing of aerial LiDAR and mobile terrestrial LiDAR datasets has afforded an opportunity to leverage similar technologies captured from differing vantage points, to create a single comprehensive dataset that provides widespread coverage over a large area, and high resolution detail where its needed.

Data fusion is an effective way to optimize utilization of large volumes of data from multiple sensors/sources/technologies. Multi-sensor data fusion seeks to combine information from multiple sensors and sources to achieve inferences that are not feasible from a single sensor or source. Benefitting from the synergy between the two technologies has provided data fusion opportunities between the differing capture methodologies. When used in tandem, the resulting dataset(s) provide a comprehensive insight that is greater than the sum of the parts. Aerial LiDAR data provides a foundation for the topographic information, while the mobile LiDAR data provides detailed information about the infrastructure.

Performing data fusion operations between Aerial and Mobile LiDAR, or any disparate data sources, requires a sound understanding of the technologies, capabilities, and limitations of each source to facilitate the fusion of data into a single common model. Though aerial and mobile LiDAR collections generate similar deliverables (spatially located point features in a common file format), there are differences in parameters such as collection speed, point density, age of data, and nadir or oblique views, etc., as well as positional accuracy that present challenges.

Mobile Terrestrial LiDAR and Aerial LiDAR

Earlier this year, Baker was awarded a Mobile LiDAR project encompassing approximately 115 miles of an existing high speed rail corridor. Along twenty-two (22) separate interlockings on the corridor, ground control targets were conventionally surveyed to improve the overall spatial accuracy of the collections, which would be further used to develop planimetric features in support of engineering/design activities. However, large stretches would have gone unconstrained, relying solely on the accuracy achieved from the GPS, IMU and DMI, if it were not for existing Aerial LiDAR data. The process of fusing the two datasets first required an assessment of general spatial accuracies. Baker was keenly aware that the Mobile LiDAR data would be more accurate than the provided Aerial LiDAR data at the specific work locations along the corridor for which Baker established ground control. For the remaining sections of the corridor, Baker assessed diagnostic outputs from the GPS observerations captured throughout the scanning activities, which provided insight into the potential spatial accuracies of the Mobile LiDAR data.

Similar in process to utilizing ground control points to ortho-rectify aerial photography, LiDAR-identifiable features visible in both datasets were utilized as tie-points to anchor one dataset to the other. Additionally, because the disparate datasets comprised varying spatial accuracies, data densities, and were captured several years apart, it was determined that to produce the most sanitary dataset, some of the resulting overlap between the point-clouds (i.e. real-world objects captured by both the Aerial LiDAR and Mobile LiDAR) that existed within the railway right-of-way, would be trimmed in an effort to reduce downstream confusion for novice users. The process leverages the enhanced assets of both technologies, whereas the high-density, high-resolution Mobile LiDAR data was retained witin the bulk of the existing right-of-way, the Aerial LiDAR data was employed to cover areas extending well past the confines of the corridor.

The image above presents the fusion of Aerial LiDAR (red) and Mobile LiDAR (white). The tops of structures, including poles and push braces, are completely saturated with LiDAR points while the rail corridor (immediately adjacent to the buildings) contains dense information.

Traditional Surveying What LiDAR Cant Do

Although the fusion of Mobile and Aerial LiDAR presents the possibility to perform highly detailed feature extractions, additional surveying services outside of the right-of-way were later requested to support engineering design activities. Features such as property boundaries, underground utilities and right-of-way monuments were required to compile a complete as-built survey of an existing station platform undergoing a major reconstruction. Therefore, the fused LiDAR dataset was further supplemented with traditional surveying.

Field surveying staff utilized local ground control, including monuments observed in the Mobile LiDAR ground control survey, to perform detailed boundary retracement and field location of features. The use of common monuments and coordinate systems, the CAD drawings developed seamlessly integrated with the LiDAR point cloud, derived planimetrics, digital terrain model and contours.

The image above depicts the fusion of Mobile LiDAR and Traditional Surveying (Aerial LiDAR points removed for clarity). The LiDAR data was supplemented outside of the Mobile LiDAR corridor to locate utilities, property boundaries and other features not captured by Aerial means within the larger corridor.

Points, Lines and Polygons (and Text)

More often than not, the survey drawings we compile are made up of features represented as points, lines and polygons annotated or described with text. How the individual objects are captured can be incredibly varied. The common thread between these objects is not the technology used, but rather the coordinate(s) to which they are referenced (represented by monuments in the ground). By employing the various technologies, we effectively address a number of questions with respect to accuracy and completeness; and, each of these technologies are just another tool in our toolbox.