Abstract:

Object of study:
The contribution describes the didactic experience of teaching the basis of automated structure from motion survey to the first year students of architecture in the Italian context of Politecnico di Torino. This experience is part of the wider educational proposal of the Architectural Drawing and Survey Laboratory (held by Authors), an annual course that guides freshmen in acquiring the foundations of architectural language (drawing) and the basis of the useful tools for analysing the built (survey). The course structure includes descriptive geometry, hand drawing, 2D/3D CAD drawing, direct survey procedures (trilateration, progressive and partial measuring) and indirect survey techniques (photography, level, picture rectification and automated structure from motion) (Comparetto et al. 2020, Zich et al. 2019, Zich 2017).

State of the art:
In the architectural field, professionals conduct survey operations for many purposes. One of them is to reach a certain metric/geometric knowledge of the surveyed object/building (Ippolito 2016). Professionals can use many tools to achieve this knowledge. Among them, the structure from motion survey is a range imaging methodology based on an automated photogrammetry approach. By allowing an elaboration of several frames together, it leads to the generation of a 3D model of the object of study (De Luca 2011). Nowadays, this process can be automated, thus providing tools that do not require in-depth knowledge of their processes but their workflow and limitations.

Methodology:
In order to introduce students to these automated survey techniques, we teach them the basis of photogrammetry, its properties and basic rules to define a good project for acquiring pictures (i.e. CIPA 3x3 rule). We also introduce the use of a specific automated software (Autodesk Recap Photo Educational) and we discuss its strengths and weaknesses. Since this software does not let the user set any kind of parameters nor recognize any homological points, moreover it does not allow him to verify its photogrammetric workflow, we show what kind of objects could be effectively elaborated by the software and what physical properties they must have, in order to provide the software both with workable object and sets of pictures. We also discuss geometric patterns/surface typologies the software can handle, so that students can be aware of errors and issues of the automated process without human control. We then ask students to elaborate an exercise dealing with this workflow, by analyzing and elaborating a small object on their own. We guide them during collegial revisions and we evaluate their works with a scale of 0-8 points. This contribution provides an analysis of such results, because we want to assess the quality of students’ reactions to this exercise by comparing their chosen objects with the rules provided and with the automated results obtained.

Objectives pursued and achieved:
We have been doing this experience with our students in the past two academic years and now we can critically analyze their first results. For the 2019/2020 a.y. we have 41 evaluated exercises, while for the 2020/2021 a.y. we count 34 evaluated homework. The main issue with the process can be found in choosing an object that could be easily elaborated, mainly because of shapes or too flat surfaces that do not support a correct finding of homologues points in different pictures. We then discuss possible improvements to this activity.

Keywords:

Architecture, Architectural Survey, Critical Thinking, Graphic Analysis, Exercise.