Automated LIDAR Processing System

The Automated LIDAR Processing System (ALPS) system is a combination of Information Technology capabilities and hard scientific algorithmic computation to achieve the goal of maximum automation of processing for LIDAR data.  The processing of LIDAR data is necessarily complex, as there are numerous factors involved.  LIDAR data is in the form of random points, which are the specific locations at which a laser beam was sensed by the hardware and used to calculate height information.   In this form, the data is only partially usable.  Most systems require some other form of the data, including gridded and Triangular Irregular Network (TIN) form.  While it is not inherently difficult to transform the data into these formats, the high resolution nature of the data makes a simple transformation less than useful.  The largest problem encountered is the inclusion of non-ground items in the data.  As an example, an automobile driving along a highway at an inopportune time will cause a reading of 6ft higher than the actual ground at this location.  Similarly for houses, buildings, bushes and trees.

 

Automated LIDAR Processing System

In order to make LIDAR data useful, non-ground objects must be removed from the data, and replaced with interpolated values that nearly approximate where the actual ground exists.  This is a very complex problem with many variables.  Algorithms exist to perform object detection and removal, however these algorithms are generally only valuable in specific instances.  For example, some algorithms are best at detecting and removing building features (with hard edges) while other algorithms are best suited for more complex forested areas where more of a smoothing vs. an object removal approach is warranted.

An additional complexity involves the intended use for the data.  In some instances the removal of objects is intended, where other applications do not wish for this to happen.  In some instances other processes may need to be applied to the data, including processing which has yet to be developed or has been developed in a proprietary way.

LIDAR Visual Data Editor

The key problem is that there does not, and likely will never exist, the perfect algorithm for processing all types of LIDAR data.  An automation system must provide a mechanism to allow the user to interactively select, apply and interact with various algorithms and tools to apply to the data.  This is where the IT components of the ALPS system come in.  The ALPS project has developed a software system, as well as a set of standard algorithms, for allowing the user to customize the tasks associated with processing LIDAR data.The ALPS processing engine is a sophisticated graphical user interface which allows the user to develop a processing flow for LIDAR data.  The system uses a graphical modeling mechanism to make this complex task simple and intuitive.  The user simply loads data, connects data to processing components, connects those processing components to other components, etc. until a processing flow is created.  The system allows process flows to be “nested” such that a complex set of processing tasks can be embedded within another process.  This allows for any level of complexity in defining process flows.

LIDAR Processing Algorithm Design Tool

 

The final result is a complete system that allows for the development of custom processing of LIDAR data, allows for the visualization and editing of data, and provides capabilities for ingesting and extracting LIDAR data from various file formats.

 

Documents:

 

 

Posted in Our Projects