Multi-step approach for automated scaling of photogrammetric micro-measurements

Photogrammetry can be used for the measurement of small objects with micro-features, with good results, low costs, and thepossible addition of texture information to the 3D models. The performance of this technique is strongly affected by the scalingmethod, since it retrieves a model that must be scaled after its elaboration. In this paper, a fully automated multi-step scalingsystem is presented, which is based on machine vision algorithms for retrieving blurred areas. This method allows researchers tofind the correct scale factor for a photogrammetric micro model and is experimentally compared to the existing manual methodbasing on the German guideline VDI/VDE 2634, Part 3. The experimental tests are performed on millimeter-sized certifiedworkpieces, finding micrometric errors, when referred to reference measurements. As a consequence, the method is candidate tobe used for measurements of micro-features. The proposed tool improves the performance of the manual method by eliminatingoperator-dependent procedures. The software tool is available online as supplementary material and represents a powerful tool toface scaling issues of micro-photogrammetric activities.