Italy’s Leaning Tower of Pisa is already in a pretty precarious position – hence the name – but imagine the tragedy if an earthquake or other disaster knocked it over. Preservationists would need precise information to restore it to its historic glory. And now, courtesy of a spring-mounted mobile laser mapping system, a comprehensive 3-D map could help with such an effort. “We knew that the Pisa Tower had been 3-D mapped from the outside a number of years ago, but that nobody had managed to map the interior,” said Dr. Jonathan Roberts, research program leader in the Computational Informatics Division at the Commonwealth Scientific and Indus-trial Research Organization (CSIRO) in Brisbane, Australia. Conventional mapping systems have been unable to capture the bell tower’s interior, thwarted by the narrow staircase and complicated architecture. The map includes the tower’s interior features, from the structure and composition to tiny details in the stonework. It was created using a handheld 3-D mapping system called Zebedee, which incorporates a laser scanner that sways on a spring. The mounting is lightweight and ensures a wide scanning field of view, Roberts said: “The spring converts the natural motions of the operator into a suitable sweeping motion of the scanner.” An inertial sensor provides rough measurements of the spring’s rotation. The final Zebedee 3-D interior map of the Leaning Tower of Pisa could be very useful to historians, preservationists and anyone interested in the historic building. Courtesy of CSIRO. The Zebedee system is based on lidar sensing technology; software converts the raw range and inertial measurements into a 3-D map or point cloud. It requires no external position systems such as GPS, enabling it to scan nearly any environment accessible to the operator, including indoors and underground. Zebedee can map an interior as quickly as an operator can walk through it, its creators say. It was developed at CSIRO in collaboration with a team from the Sant’Anna School of Advanced Studies in Pisa. “The main breakthrough,” Roberts said, “was developing software that determines where Zebedee has been as it moves through the environment – with this information, it is possible to build a map.” The Zebedee system solves what is known as the SLAM (simultaneous localization and mapping) problem by using re-observations of surfaces in the environment. For example, Roberts explained, if a single wall is observed 5 m in front of the sensor, and then the same wall is rescanned from 3 m away, the system can determine that the sensor has moved 2 m in that direction. By accumulating many thousands of such re-observations, the scanner’s six-degrees-of-freedom motion can be accurately estimated. “Given the trajectory estimate, the raw range measurements can also be projected into a common frame to produce a globally consistent 3-D point cloud,” Roberts said. CSIRO has worked with England-based 3D Laser Mapping to commercialize the work into the ZEB1 product. The team is investigating Zebedee-like systems for robots, and 3-D printing has enabled them to produce models of some sites they have scanned.