ERCNN-DRS Urban Change Monitoring

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F61989100%3A27740%2F21%3A10248384" target="_blank" >RIV/61989100:27740/21:10248384 - isvavai.cz</a>

Výsledek na webu

<a href="https://github.com/it4innovations/ERCNN-DRS_urban_change_monitoring" target="_blank" >https://github.com/it4innovations/ERCNN-DRS_urban_change_monitoring</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

ERCNN-DRS Urban Change Monitoring

Popis výsledku v původním jazyce

This project contains the Ensemble of Recurrent Convolutional Neural Networks for Deep Remote Sensing (ERCNN-DRS) used for urban change monitoring with ERS-1/2 &amp; Landsat 5 TM, and Sentinel 1 &amp; 2 remote sensing mission pairs. It was developed for demonstration purposes (study case) in the ESA Blockchain ENabled DEep Learning for Space Data (BLENDED) project. Two neural network models were trained for the two eras (ERS-1/2 &amp; Landsat 5 TM: 1991-2011, and Sentinel 1 &amp; 2: 2017-2021). The enclosed data was used for the MDPI Remote Sending publication Neural Network-Based Urban Change Monitoring with Deep-Temporal Multispectral and SAR Remote Sensing Data [2].

Název v anglickém jazyce

ERCNN-DRS Urban Change Monitoring

Popis výsledku anglicky

This project contains the Ensemble of Recurrent Convolutional Neural Networks for Deep Remote Sensing (ERCNN-DRS) used for urban change monitoring with ERS-1/2 &amp; Landsat 5 TM, and Sentinel 1 &amp; 2 remote sensing mission pairs. It was developed for demonstration purposes (study case) in the ESA Blockchain ENabled DEep Learning for Space Data (BLENDED) project. Two neural network models were trained for the two eras (ERS-1/2 &amp; Landsat 5 TM: 1991-2011, and Sentinel 1 &amp; 2: 2017-2021). The enclosed data was used for the MDPI Remote Sending publication Neural Network-Based Urban Change Monitoring with Deep-Temporal Multispectral and SAR Remote Sensing Data [2].

Druh

R - Software

CEP obor

—

OECD FORD obor

10200 - Computer and information sciences

Projekt

<a href="/cs/project/LQ1602" target="_blank" >LQ1602: IT4Innovations excellence in science</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2021

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Interní identifikační kód produktu

017/15-12-2021-SW

Technické parametry

Directories ERS12_LS5 and Sentinel1_2 contain the respective: - AoI shape files used for the project (three sites: Rotterdam, Liege and Limassol) - Pre-processing script (two steps) to create the training/validation TFRecord files used for training the neural network and for validation purposes - The description of how the synthetic and noisy labels are created for supervised training - The neural network architecture model - The training script for training the neural network with the pre-processed training data; next to the training scripts are also the snapshots of the pre-trained model used in our work Directory “external” contains a (modified) 3rd party component (Omnibus change detection) under MIT license.

Ekonomické parametry

roční zvýšení objemu výroby, zisku, exportu

IČO vlastníka výsledku

61989100

Název vlastníka

VŠB-TUO