Automatic Processing of Many Images for 2D/3D Modelling

Abstract

The era of big data requires increasing automation for the analysis of huge information in a short time and this need becomes critical when dealing with geoinformation. This chapter describes the automatic geocoding of digital images based on high-end Photogrammetric and Remote Sensing methods. In particular, the so-called Structure-from-Motion (SfM) technique is developed to handle image data sets in close-range applications, and here, it is generalized to deal with multiscale applications. Some examples are proposed with panoramic images for the measurement of indoor narrow spaces, with smartphone cameras and UAV for the 3D reconstruction of complex monuments, as well as with airborne and satellite images for the survey at the territorial scale.

Techniques for Comparing Multi-Temporal Archive Aerial Imagery for Glacier Monitoring With Poor Ground Control

“…Structure-from-Motion Photogrammetry (SfM -see Granshaw, 2018) is commonly referred to as the entire automatic procedure for computing the EO of a block of images and to output a dense point cloud derived from the application of Multi-View-Stereo (MVS) matching techniques (see Eltner et al, 2016;James et al, 2019). While at the beginning SfM could be only applied to close-range blocks, successive development in processing and computing techniques extended its application to any types of images (Barazzetti et al, 2020). Today SfM can be used to process a photogrammetric block by using low-cost and open-source software packages running on a standard laptop.…”

Vertical geomediation: The automation and platformization of photogrammetry

“…SfM has also benefitted from the arrival of unmanned aerial systems (UASs) or drones, both large and small (Colomina and Molina, 2014), which increased access to affordable forms of vertical mediation. Novel algorithms have also been developed that have streamlined processes and have ‘significantly changed the traditional workflow of several surveying applications’ (Barazetti et al, 2020: 363). Among these are image-based terrain extraction algorithms that have been developed to reconstruct in 3D space the location of individual points from digital photographs (Chesley et al, 2017).…”