Developing countries usually lack the proper governance means to generate and regularly update a national rooftop map. Using traditional photogrammetry and surveying methods to produce a building map at the federal level is costly and time consuming. Using earth observation and deep learning methods, we can bridge this gap and propose an automated pipeline to fetch such national urban maps. This paper aims to exploit the power of fully convolutional neural networks for multi-class buildings’ instance segmentation to leverage high object-wise accuracy results. Buildings’ instance segmentation from sub-meter high-resolution satellite images can be achieved with relatively high pixel-wise metric scores. We detail all engineering steps to replicate this work and ensure highly accurate results in dense and slum areas witnessed in regions that lack proper urban planning in the Global South. We applied a case study of the proposed pipeline to Lebanon and successfully produced the first comprehensive national building footprint map with approximately 1 Million units with an 84% accuracy. The proposed architecture relies on advanced augmentation techniques to overcome dataset scarcity, which is often the case in developing countries.

We propose a flexible solution to the problem of pixel regression in the context of heterogeneous image input and sparse or incomplete ground truth. This is a very broad and generic problem space in image processing, and we discuss its application to different Use Cases in Earth Observation.

As AI permeates industries and reaches the general public, the significance of responsible AI (RAI) principles becomes increasingly vital. This study offers valuable insights from 27 Earth Observation (EO) experts in 11 countries, unveiling diverse attitudes towards RAI principles and nuanced perspectives across genders and age groups. It highlights the variation in definitions and interpretations of core principles such as fairness, reliability, privacy, transparency, accountability, and sustainability. Moreover, the study identifies RAI concerns specific to the EO domain, emphasizing the need to integrate domain knowledge and effectively communicate issues of inequality and failure use cases. These findings contribute to a preliminary understanding of RAI attitudes in the AI for EO context. Future research should involve a larger pool of experts and investigate the attitudes of EO users and the general public to complement these initial findings.

As extreme weather events become more frequent, understanding their impact on human health becomes increasingly crucial. However, the utilization of Earth Observation to effectively analyze the environmental context in relation to health remains limited. This limitation is primarily due to the lack of fine-grained spatial and temporal data in public and population health studies, hindering a comprehensive understanding of health outcomes. Additionally, obtaining appropriate environmental indices across different geographical levels and timeframes poses a challenge. For the years 2019 (pre-COVID) and 2020 (COVID), we collected spatio-temporal indicators for all Lower Layer Super Output Areas in England. These indicators included: i) 111 sociodemographic features linked to health in existing literature, ii) 43 environmental point features (e.g., greenery and air pollution levels), iii) 4 seasonal composite satellite images each with 11 bands, and iv) prescription prevalence associated with five medical conditions (depression, anxiety, diabetes, hypertension, and asthma), opioids and total prescriptions. We combined these indicators into a single MedSat dataset, the availability of which presents an opportunity for the machine learning community to develop new techniques specific to public health. These techniques would address challenges such as handling large and complex data volumes, performing effective feature engineering on environmental and sociodemographic factors, capturing spatial and temporal dependencies in the models, addressing imbalanced data distributions, developing novel computer vision methods for health modeling based on satellite imagery, ensuring model explainability, and achieving generalization beyond the specific geographical region.

This research addresses the use of machine learning for downscaling pollution information from multiple satellite sources. The proposed solution uses a Convolutional Neural Network (CNN)-based segmentation architecture (U-Net [1]) to perform a resolution enhancement task. Readings obtained by spatially sparse ground stations are used as the ground truth to guide the task. A custom implementation of this solution won the AI4EO Air Quality & Health challenge [2] initiated in 2021 by the European Space Agency (ESA) [3], on a need expressed by the European Centre for Medium-Range Weather Forecasts (ECMWF) [4] and the Copernicus Atmosphere Monitoring Service (CAMS) [5]. In that successful case, the algorithm was trained and tested on volatile particulate matter (PM2.5, from Copernicus Atmosphere Monitoring Service) and nitrogen dioxide (NO2, from Sentinel-5P TROPOMI) data over three Areas of Interest: Northern Italy, California and South Africa.

This research addresses the use of machine learning for detecting building footprints from dual-polarization multitemporal SAR data. We assume that in SAR images, built-up areas exhibit very high backscattering values, making them brighter than other land cover classes in both the co- and cross-polarization channels. In addition the building class is considered to be stable in time. In this context, we propose to employ a convolutional neural network that integrates a set of SAR images of different dates and polarizations for delineating the building footprint. The algorithm was tested on Sentinel-1 datasets from eight different test sites located in Iraq. Four of the test sites along with their corresponding annotations (i.e. ground truth) constitute the training dataset, while the other four test sites are used for evaluation. All the datasets were provided in the framework of a humaritan AI4EO UNOSAT challenge launched in 2019 by Phi-Unet. The results evaluated by the F1 score with values of about 0.75 indicate that the proposed method is able to accurately detect the building footprint.

This paper describes automated procedures for the design and deployment of artificial skin for humanoid robots. This problem is challenging under different perspectives: on the one hand, different robots are characterized by different shapes, thereby requiring a high degree of skin customization; on the other hand, it is necessary to define optimal criteria specifying how the skin must be placed on robot parts. This paper addresses the problem of optimally covering robot parts with tactile sensors, discussing possible solutions with reference to a specific artificial skin technology for robots, which has been developed in the past few years. Results show that it is possible to automate the majority of the required steps, with promising results in view of a future complete automation of the process.