Dr. Bingcai Zhang, GXP engineering fellow, was invited to speak at the University of Southern California’s Integrated Media Systems Center (IMSC) annual advisory board retreat on February 17, 2011.
His talk focused on automated 3-D object recognition from LIDAR point clouds, which is the research behind a new building extraction tool for SOCET GXP v4.0 that increases the productivity of geospatial data. The 3-D object recognition process generates city maps and site models and extends to the navigation of unmanned ground vehicles and aircraft.
According to Dr. Zhang, the new Automatic Feature Extraction tool (AFE), together with BAE Systems’ Next-Generation Automatic Terrain Extraction (NGATE) functionality, completes the next-generation photogrammetric automation system that the GXP R&D team began six years ago. NGATE uses a hybrid matching process to create precise elevation data for 3-D terrain and surface models and decreases manual editing time, resulting in shorter workflows. AFE automatically extracts 3-D buildings and houses using a digital surface model which can be from either LIDAR or NGATE. Automatic feature extraction is considered by many to be the Holy Grail in photogrammetry.
During the IMSC retreat, Bingcai fielded questions about his research from technology innovators, Facebook, Qualcomm, HP and IBM, who expressed an interest in geospatial technology. The fusion between geospatial information and social information could become the next new computing paradigm.
IMSC’s mission is to inspire and encourage students, faculty and industry to apply sophisticated IT techniques to designated geographical areas.