Join Us at ESA EO Phi-Week 2019
We are just a few days from one of the biggest European EO events of its kind — the European Space Agency (ESA) EO Phi-Week, which will take place at ESRIN, Frascati, Italy on September 9–13, 2019. The event will offer a review of the latest developments in Open Science trends with a focus on EO Open Science and FutureEO. We are pleased to extend ESA’s invitation and recommend you to join our activities.
Sinergise’s Activities at Phi-Week
We will hold a talk on Spatio-Temporal Deep Learning: Application to Land Cover Classification on Wednesday, September 11, and organize a side event Getting the Most Out of EO Data with the Help of Data Cube Service and Machine Learning (ML) Tools on Thursday, September 12, which will be full of impressive speakers. Don’t miss the poster session the Next Generation DataCube Service Based on Cloud-native Architecture after the side event.
Getting the Most Out of EO Data with the Help of Data Cube Service and ML Tools
Side Event: We will host a dedicated session on Data Cubes on Thursday, September 12, 2019, 9:30–13:00
This side event will demonstrate best practices of EO data processing using various data cube options — from on-the-fly virtual data cubes to xarrays embedded in the Python environment, from basic algorithms to advanced ML techniques — focusing on how these tools are exploited by various groups of users, from researchers, data scientists in the commercial sector to application developers.
Make sure you listen to speakers from GAF, Picterra, TomTom, Development Seed, CLAAS, SpaceTec, Brockmann Consult and us.
Spatio-Temporal Deep Learning: Application to Land Cover Classification
Oral Presentation: Wednesday, September 11, 2019 at 11:30
The increased availability of a large number of spatio-temporal images is driving the need for adequate tools to process, analyse and extract actionable information from the data. ML and, in particular, deep learning methods have become the state-of-the-art in many vision, language, and signal processing tasks, due to their ability to extract patterns from complex high-dimensional input data.
Classical ML methods, such as random forest and support vector machines have been used in many EO applications to analyse temporal series of remote sensing images. On the other hand, convolutional neural networks have been employed to analyse the spatial correlations between neighboring observations, however mainly in single temporal scene applications. This presentation will introduce our investigation of a deep learning architecture capable of simultaneously analysing the spatio-temporal relationships of satellite image series. To learn more see the abstract of our paper.
Next Generation DataCube Service Based on Cloud-native Architecture
Poster session: Thursday, September 12, 2019 at 18:15.
We will present how Sentinel Hub, running on various platforms (AWS, Mundi, CreoDIAS, ONDA), is able to produce instantenuous analysis ready data (ARD), long time-series analysis, how annual cloudless mosaic can be created ad-hoc and how it is possible to use data fusion to combine various data sources such as performing on-the-fly orthorectification of Sentinel-1 GRD data.
This session will also provide an opportunity for people interested in data cubes to interact with us directly.
Team members of our EO Research and Sentinel Hub team Grega Milčinski, Matic Lubej, Anja Vrečko, Matej Aleksandrov and Jovan Višnjić will be happy to discuss your ideas. You will be able to find them at the event, and if you would like to book a meeting just send us an email at info[at]sentinel-hub.com or contact us via Twitter.
We look forward to meeting you in ESRIN.