Enabling reproducible obia with open-source software in docker containers

Abstract

While most approaches in Object-Based Image Analysis (OBIA) currently rely on proprietary software, the interest in free and opensource software (FOSS) for OBIA is growing. This interest stems not only from cost savings, but also from benefits concerning reproducibility and collaboration. However, technical challenges hamper practical reproducibility, especially when multiple software packages are involved. In this study, we use the Docker technology to containerise an OBIA workflow in a well-defined FOSS environment. Running the analysis inside a container eliminates the need to recreate the original software environment on the executing computer. We explore the approach using two software stacks (InterIMAGE, and QGIS in combination with Orfeo ToolBox, SAGA and Python libraries) to perform an exemplary analysis detecting destruction of buildings in bi-temporal images of a conflict area. The analysis combines feature extraction techniques with segmentation and object-based analysis to detect changes and to distinguish disappeared buildings from similarly changed non-target structures. The resulting workflow is published as FOSS comprising both the model and a ready to use Docker image including all required software and data. The presented solution advances OBIA in the following aspects: higher transparency of methodology; easier reuse and adaption of workflows; better transferability between operating systems; complete description of software environment; and easy adoption of OBIA workflows by image analysis experts and non-experts.