A Low-Cost MMS Approach to the Simultaneous Localization and Mapping Problem

The increasing computational capacity and the reduction of production costs of electronic devices allow the adoption of new approaches to the architectural survey. The purpose of this paper is to analyze and test different low-cost sensors able to acquire odometry, depth data and RGB images processed through SLAM (Simultaneous Localization and Mapping) algorithms to obtain metrically correct point clouds applicable in the normal workflow of the architectural survey. The advantage of these new systems is the time needed to obtain the result. The post-processing time is virtually cleared, and it is not necessary to use targets and control points during acquisition. The capabilities of these sensors are tested both qualitatively and quantitatively by comparing them with the data obtained with traditional surveying systems (laser scanning, photogrammetry, etc.) set as “ground truth”. First, the acquisition efficiency of the low-cost sensors under different conditions of use are verified. Subsequently, the different point clouds obtained using different acquisition software are compared. Lastly, pros and cons are evaluated, also from an economic point of view, in relation to traditional surveying systems. The comparisons highlight that the performance of these devices varies considerably depending on the size of the acquired environment. Therefore, they are excellent for the acquisition of small environments as they allow to obtain an accuracy of about ±3 cm in an area of 100 m2