Registration of terrestrial laser scanning data in forest areas using smartphone positioning and orientation data

Abstract

Terrestrial laser scanning (TLS) is an effective tool in forest inventory, but multiple scans are commonly required to cover the whole plot. Thus, it is necessary to register different scans into uniform coordinate system. Including auxiliary information to simplify registration is an effective idea, e.g., levelling scanner and measuring initial scanning direction, but requires additional survey equipment. In this letter, a registration method for forest TLS data is proposed, using smartphone to obtain auxiliary information. Scanner position and initial scanning direction are measured by smartphone for each scan and then used to calculate pair-wise coarse transformation parameters. In fine registration, we construct a series of transformation parameters based on coarse registration result and evaluate them by two-level objective functions, using stem positions as input. In this stage, stems far away from origin are excluded with filtering threshold set as the horizontal distance between scanner positions. The proposed method is validated on six scan pairs, with distance between scan positions ranging from 19 m to 40 m. The mean rotation, horizontal and vertical translation errors are 4.3′, 8.3 mm and 35 mm. Comparison with previous study shows auxiliary information from smartphone is beneficial for registration and filtering distant stems can improve registration accuracy.