Feature Type Interpreter brings speed and accuracy to automated feature extraction 

Object Raku Technology's Feature Type Interpreter (FTI) is the culmination of eight years of research & development in the field of feature extraction. In 2003, Object Raku first delivered its LiDAR Tool suite as a component of the Sextant Geospatial software suite. FTI builds on that foundation and takes automated extraction accuracy to a whole new level.

Using traditional digitizing tools, operators typically spend several hours per square mile of LiDAR data analyzed. Identifying buildings and vegetation in dense data sets requires meticulous attention to detail and is a crucial step to realize the full value of the LiDAR collected. In creating FTI, the development team focused on solving the problem of feature extraction with the highest level of speed and accuracy possible.

Raw sensor output and pre-processed data simply aren't end-products. Regardless of the industry, analysts need LiDAR to become information rather than merely (point) data. Using FTI to extract feature detail including size, location, and direction, is the first step to actionable insight and intelligence from a particular LiDAR collection.

FTI produces reports on each feature type extracted. The report generated from a building analysis includes total building count, average building height, and total building area coverage. Vegetation reports include total crown count and area of coverage as well as both count and area for the three pre-defined sub classes (high, medium, and low). These reports are available from the heads-up display with FTI or from a text file.
Feature counts are provided at each analysis.

Object Raku partners with Simon Fraser University for LiDAR research effort  

Object Raku Technology is pleased to announce the creation of a research and development partnership with Dr. Richard Zhang of Simon Fraser University's School of Computing Science. Dr. Zhang's team will leverage their extensive knowledge in the field of geometric analysis and processing to assist Object Raku's on-going innovation in automatic feature extraction from LiDAR data.  

"Object Raku uses a rapid-prototyping approach in its software development and we intend to insert Dr. Zhang’s research results as early and as often as possible", says Object Raku Managing Director Mike Parlow. "As fast as technology moves, rarely is meaningful research & innovation achieved without a long-term focus. To leap forward and lead the charge on the industry's requirements, we searched for a research partner with the necessary experience, capability, and a shared long-term vision. We found that the team at SFU has key experience in geometry processing and Dr. Zhang's goals fit so well with our own that we could enable this collaboration."

"The five-year research project hopes to tap into innovative solutions for feature extraction. Object Raku's emphasis in the Feature Type Interpreter software has been on leveraging the Graphics Processing Unit (GPU)", says CTO Trevor Hooper. "Richard's team will take us in some pretty exciting directions and our challenge will be to translate those gains into innovations that yield a competitive edge in the industry."

"Dr. Zhang's work has been published in the most prestigious venues in his fields, including SIGGRAPH, SIGGRAPH Asia, Eurographics, and Symposium on Geometry Processing. His team’s expertise in geometric processing and analysis is a perfect fit for us. Their expertise is supported by the equipment and resources already in place at the Graphics Usability and Visualization (GrUVi) lab at SFU."

Test Patches and Optimizing Parameters  

FTI uses a test patch approach to parameter optimization for feature extraction. Test patches are very quick sub-set analyses done on small areas within the data set. They can be compared to the various source layers as well as to each other. This approach allows the operator to rapidly progress from the raw data to a finished feature layer.

Each type of analysis has its own set of extraction parameters. These parameters have been ordered by impact - the parameters at the top of the list generally have greater impact than those at the bottom. Further, FTI saves the parameter settings for every test patch and full analysis undertaken so the user can easily use successful settings from previous projects in new projects.