Abstract
Abstract. Images from archival aerial photogrammetric surveys are a unique and relatively unexplored means to chronicle 3D land-cover changes occurred since the mid 20th century. They provide a relatively dense temporal sampling of the territories with a very high spatial resolution. Thus, they offer time series data which can answer a large variety of long-term environmental monitoring studies. Besides, they are generally stereoscopic surveys, making it possible to derive 3D information (Digital Surface Models). In recent years, they have often been digitized, making them more suitable to be considered in automatic analyses processes. Some photogrammetric softwares make it possible to retrieve their geometry (pose and camera calibration) and to generate corresponding DSM and orthophotomosaic. Thus, archival aerial photogrammetric surveys appear as being a powerful remote sensing data source to study land use/cover evolution over the last century. However, several difficulties have to be faced to be able to use them in automatic analysis processes. Indeed, surveys available on a study area can exhibit very different characteristics: survey pattern, focal, spatial resolution, modality (panchromatic, colour, infrared…). Planimetric and altimetric accuracies of derived products strongly depend on these characteristics. Thus, analysis processes have to cope with these uncertainties. Another important gap states in the lack of training data. Deep learning methods and especially Convolutional Neural Networks (CNN) are at present the most efficient semantic segmentation methods as long as a sufficient training dataset is available. However, temporal gaps can be very important between existing available databases and archival data. In this study, two custom variants of simple yet effective U-net - Deconv-Net inspired DL architectures are developed to process ortho-image and DSM based information. They are then trained out of a groundtruth derived out of a recent database to process archival datasets.
Highlights
1.1 Archival photogrammetric campaigns: opportunuitiesArchival aerial photogrammetric surveys were initially acquired by mapping, cadastral or military agencies for topographic map generation
For each epoch, different configurations were tested. They are denoted as Method:im+mod, where Method is the used Convolutional Neural Networks (CNN) architecture (”UNet” for the mono input one, ”FuseNet” for the two inputs one), im is the ortho-image and mod is the used height information modality (”dsm” for raw DSM, ”sh” for the shading map). ”UNet:im” means that only the ortho-image is used by the CNN
When a configuration is trained out of the cleaned versions of the database, it is named Method:im+mod:clean. 4.2 Evaluation method Obtained results are visually assessed for all classes
Summary
1.1 Archival photogrammetric campaigns: opportunuitiesArchival aerial photogrammetric surveys were initially acquired by mapping, cadastral or military agencies for topographic map generation. Stereoscopic configurations were adopted so as to offer 3D plotting capacities These surveys have been a common practice in many countries over the last century. Several countries have digitised their film-based photos (e.g., >3 millions images in France), and facilitated their access through spatial data infrastructures and web services with basic metadata and visualisation capacities. These images are a unique yet unexplored means for long-term environmental monitoring and change analysis, as they chronicle Earth surface evolution in a comprehensive way. Some photogrammetric softwares make it possible to retrieve their geometry (pose and camera calibration) and to generate corresponding Digital Surface Models and orthophotomosaics (e.g. (Agisoft, 2016) or (Rupnik et al, 2017))
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call