OBIA4RTM – towards an operational open-source solution for coupling object-based image analysis with radiative transfer modelling

Abstract

Radiative transfer models (RTM) provide universally applicable, highly accurate prospects for plant parameter retrieval. Due to the ill-posed nature of radiative transfer theory, however, the retrieval of plant parameters requires sophisticated strategies for model inversion. We argue that object-based image analysis (OBIA) works as an effective regularization measure to cope with this ill-posedness. Despite similar findings reported in the literature, OBIA and RTM are rarely used in a combined manner. Additionally, there is a clear lack of software solutions ready for operational usage. Therefore, we propose OBIA4RTM as an approach to combine OBIA and RTM using Python and PostgreSQL/PostGIS spatial databases in a fully Open Geospatial Consortium (OGC) compliant way. First results obtained in agricultural regions in southern Germany and Austria using Sentinel-2 data during the 2017 and 2018 growing season show root mean squared errors (RMSE) in the leaf area index (LAI) of 1.47 m²/m² in the case of silage maize and 1.31 m²/m² in the case of winter cereals. Issues of integrating space and time as well as defining appropriate validation strategies, however, require further research.

Multi-Feature Sample Database for Enhancing Deep Learning Tasks in Operational Humanitarian Applications

“…In particular, for monitoring purposes, measurements over time need to assure to actually measure the status of the system or object and not the divergence in observation. For vegetation and crop type monitoring, radiative transfer modelling (RTF) is being used as an example (Graf, Papp, & Lang, 2020;Verhoef and Bach, 2003). In general, when interpreting images and overcoming the semantic gap, rigorous classifiers based on solid spectral models, acting across sensors, are available.…”