Can we sense road cargo using Sentinel-2 data? Normally, targeting small objects at 10 m resolution does not seem promising. Still, Henrik Fisser, Master's student at the University of Würzburg, presented a method that enables the detection of moving trucks with Sentinel-2 data and won the first prize of the Euro Data Cube COVID-19 Custom Script Contest with this contribution. The method exploits an effect of the Sentinel-2 Multispectral Instrument (MSI) that sees moving objects once per spectral band at slightly different times. Due to this effect objects of sufficient size appear spectrally disassembled in the data, which makes them detectable based on the unusual spectral signature.
Fig. 1: Truck detection example in southern France
Having proven the concept, Henrik Fisser scaled up the truck detection to the area of the whole EU for integration into the Rapid Action on Coronavirus and EO (RACE) dashboard. The derived truck information may indicate how truck traffic has developed during the COVID-19 pandemic compared to normal years. All data processing was implemented on Euro Data Cube resources. In the Euro Data Cube session at ESA EO Phi-Week Henrik is going to present his work on Sentinel-2 truck detection and the upscaling to the whole EU.
Join the ESA EO Phi-Week side event 'Euro Data Cube - Furnishing the “EO Workplace” with Data and Tools for Effectiveness and Production' on Thursday, 1st of October, from 11.30 AM.