CALCHAS · Computational Intelligence for Multi-Source Remote Sensing Data Analytics
„Хоризонт 2020“ — Действия „Мария Склодовска-Кюри“
- Период
- 2019-11-01 → 2022-04-30
- Финансиране от ЕС
- 215 492 €
- Участници
- 2
- Схема
- MSCA-IF-GF
Линиите свързват координатора с партньорите.
Накратко на български
Нови методи за машинно обучение анализират данни от сателити и наземни сензори за променливи като влажността на почвата и температурата. Това помага за по-точното проследяване на климатичните промени и тяхното влияние върху околната среда и здравето на хората.
Кратко обяснение, генерирано от езиков модел по текста на CORDIS. Оригиналът е по-долу.
Резултати накратко
Computational Intelligence for Multi-Source Remote Sensing Data Analytics
Climate change is having a profound impact in terms of human health and welfare, contributing to ecological collapse and the destruction of habitats. Understanding these phenomena necessitates the persistent monitoring of the Earth through the Essential Climate Variables (ECV) which include soil moisture, land surface temperature, and land cover. The field of Earth Observation (EO) is undergoing an unprecedented revolution in terms of the amount of data collected, facilitating the large-scale monitoring of the environment. Unfortunately, the increase in volume, variety, and complexity of measurements has led to a situation where data analysis is causing a bottleneck in the observation-to-knowledge pipeline. To address this challenge, innovative Machine Learning (ML) paradigms like deep neural network models, have established themselves as fundamental tools for EO data analytics, successfully tackling problems such as scene classification and object detection. Despite the success of these approaches, there is a clear need for establishing new paradigms which will address major challenges in ML-enabled EO data analytics. These challenges include integrating observations from multiple sources and modalities, addressing the diversity between spaceborne (tens of kilometers) and in-situ sampling scales, as well as analyzing time-series of dynamic observations. Within the CALCHAS project, we developed an innovative signal processing and machine learning framework that offered the ability to jointly analyze long sequences of EO measurements from satellite and on-ground (in-situ) sensors, achieving high resolution and accurate estimation of critical geophysical parameters. Our experimental analysis demonstrated that the proposed approach achieved an x2 reduction in retrieval error at x9 finer spatial scales compared to gold-standard NASA products. Furthermore, our models also demonstrated the ability to forecast ECVs, surpassing the performance of state-of-the-art methods by more than 10% in prediction error reduction. The framework was also considered for detecting flooded regions by modeling changes between land/permanent water and flooded water and achieved 90% accuracy in automatically detecting flooding at 10-meter spatial resolution.
Текст от CORDIS, на английски · Данни: CORDIS, © Европейски съюз
Цел на проекта
Earth Observation (EO) is undergoing a radical transformation due to the massive volume of observations acquired by remote sensing and in-situ sensor networks. While satellites provide coarse-resolution, yet global-scale monitoring of environmental processes, in-situ sensor networks acquire high-accuracy localized measurements. Extracting information from spaceborne and ground based instruments requires innovative solutions which will allow the autonomous integration of diverse in nature and scale observations in order to provide high-quality geophysical parameter estimation. CALCHAS will demonstrate cutting edge technologies targeting three major factors towards the vision of fully automated multi-source EO data understanding, namely (i) the fusion of observations from different sources and modalities, (ii) the efficient aggregation of the sampling scales associated with spaceborne and in-situ measurements, and (iii) the analysis of time-series of dynamic observations. To that end, the paradigm-shifting signal processing and learning framework of Deep Learning will be utilized and extended through powerful mathematical tools and appropriate methodologies like supervised and generative learning, dramatically extending the current scope of single source data analysis. The developed framework will be employed for analyzing time-series of measurements from active and passive microwave and multispectral spaceborne imaging instruments (SMAP, SMOS and Sentinels), and in-situ sensor measurements, targeting the high-accuracy spatial and temporal resolution enhancement for observations and soil moisture estimation. The merits of the developed technology will be demonstrated in two intelligent water management case studies, namely optimized irrigation management and water pipeline leakage detection.
Оригинален текст от CORDIS (на английски).
Участници
- IDRYMA TECHNOLOGIAS KAI EREVNAS · IRAKLEIOКоординаторГърция
- UNIVERSITY OF SOUTHERN CALIFORNIA · Los Angeles CaСъединени щати
Връзки
Данни: CORDIS, © Европейски съюз