Recursive 3d model optimization

Abstract

Systems and methods for generating and optimizingD models of objects. A laser scanner is controlled by a data processing subsystem that controls the scanning and processes the data generated. From an initial model, either a raster scan of the object or an initial scan, the system scans the object multiple times, each time adjusting the laser scanner to maximize a correlation between the data generated and the model. The model is also updated using the scan data and taking into account the adjustments which were applied to the laser scanner. Other adjustments, such as those which would remove the effect of relative motion on the scan data, are also accounted for when incorporating scan data into the model. The model is recursively adjusted and optimized using previous scan data (both pre-recorded and fresh scan data) and previous adjustments. The model may be modified and adjusted using previous scan data and an Iterative Closest Point method. The method and systems may be used in uncontrolled environments where unpredictable motions and vibrations can occur. The invention may be used with scanners mounted on a robotic arm, scaffolding, unstable tripods, at the tips of booms or lifts, or the scanners may be simply held by hand.