In Situ (May 2019)
Relevé photogrammétrique d’objets patrimoniaux : un protocole expérimental pour l’optimisation du placement de caméras pour le relevé aérien
Abstract
Professional drones have become indispensable tools for photogrammetric surveys and light detection and ranging. The trade practices of drone management, applied in agricultural or industrial uses, are now largely established, implementing well-rehearsed protocols for the optimisation of flight patterns for the measurement or surveillance of a given territory. In this context, the practices do not necessarily involve the best possible positioning of cameras since, more often than not, it is a question of repeating identical flight patterns over subjects of a broadly similar nature: quarries, agricultural surfaces, overhead power lines, road or railway infrastructures… Where the heritage is concerned, however, the heterogeneous nature of the objects observed means that drone operators have to make empirical operational choices according to the nature, the scale and the topology of the object being recorded or measured. Today, there are no preliminary procedures that allow for planning the best camera positions around a three-dimensional heritage object, particularly if these cameras are attached to drones. The experience of the users, whether they are professionals or enlightened amateurs, is usually the best judge here. The MAP is collaborating today with a team from the LIRIS laboratory (Laboratoire d’informatique en image et systèmes d’information) to work on an interdisciplinary IMAG’IN project, co-financed by the CNRS. The project is based on the experimental and empirical approaches developed over the last fifteen years by the laboratory in the field of aerial photogrammetry, combined with scientific advances. The aim is to facilitate the geometrical and topological understanding of the objects studied in order to offer a ‘field’ tool of assistance, guiding operators towards the most pertinent operational choices, on the ground or in the air, for surveying campaigns. This article presents a method tried out in the framework of an ongoing collaboration between the M2DisCo team (Modèles multirésolution, discrets et combinatoires) and Liris.
Keywords
