Segmentation of Earth observation (EO) satellite data is critical for natural hazard analysis and disaster response. However, processing EO data at ground stations introduces delays due to data transmission bottlenecks and communication windows. Using segmentation models capable of near-real-time data analysis onboard satellites can therefore improve response times. This study presents a proof-of-concept using MobileSAM, a lightweight, pre-trained segmentation model, onboard Unibap iX10-100 satellite hardware. We demonstrate the segmentation of water bodies from Sentinel-2 satellite imagery and integrate MobileSAM with PASEOS, an open-source Python module that simulates satellite operations. This integration allows us to evaluate MobileSAM’s performance under simulated conditions of a satellite constellation. Our research investigates the potential of fine-tuning MobileSAM in a decentralised way onboard multiple satellites in rapid response to a disaster. Our findings show that MobileSAM can be rapidly fine-tuned and benefits from decentralised learning, considering the constraints imposed by the simulated orbital environment. We observe improvements in segmentation performance with minimal training data and fast fine-tuning when satellites frequently communicate model updates. This study contributes to the field of onboard AI by emphasising the benefits of decentralized learning and fine-tuning pre-trained models for rapid response scenarios. Our work builds on recent related research at a critical time; as extreme weather events increase in frequency and magnitude, rapid response with onboard data analysis is essential.

Abstract—Coastal water quality monitoring is crucial for environmental sustainability and public health. This work introduces a very cutting-edge methodology, using ΦSat-2 multispectral data and quanvolutional neural networks to explore quantum-enhanced machine learning for water contaminant assessment. By integrating quantum preprocessing into a classical regression model, it is possible to achieve a significant reduction in model parameters while maintaining high predictive accuracy. Additionally, this work introduces an innovative dataset that integrates simulated ΦSat-2 spectral data with Copernicus Marine Service bio-geochemical products, ensuring a strong alignment between satellite observations and reference turbidity measurements. Our results show that quantum models use up to 98% fewer parameters than their classical counterparts, while achieving a 6.9% improvement in the Pearson correlation coefficient between the ΦSat-2 pre-processed bands and the ground-truth turbidity values, compared to the case without quantum pre-processing. Additionally, the Root Mean Square Error (RMSE) improves by 7.3% over the classical baseline. These findings highlight the potential of quantum-assisted remote sensing to enable more efficient and scalable analysis of large-scale water contaminant data, paving the way for advanced big data approaches in water quality monitoring.

Artificial Intelligence (AI) applications on Earth Observation (EO) satellite data, such as those for vessel detection, are gaining attention for their potential to meet strict bandwidth and latency requirements. While traditional on-ground computing pipelines often rely on heavy post-processing, implementing these techniques onboard satellites is challenging due to limited computing resources. To support the development of efficient onboard data processing strategies, this study compares the performance of object detection on raw data from Sentinel-2 and VENμS missions. The study demonstrates that the proposed two-stage approach with a focus on efficiency is capable of identifying vessels in raw data with minimal pre-processing. Specifically, our method achieved a remarkable Average Precision (AP) of 0.841 on the VENμS dataset.

Monitoring water contaminants is of paramount importance, ensuring public health and environmental well-being. Turbidity, a key parameter, poses a significant problem, affecting water quality. Its accurate assessment is crucial for safeguarding ecosystems and human consumption, demanding meticulous attention and action. For this, our study pioneers a novel approach to monitor the Turbidity contaminant, integrating CatBoost Machine Learning (ML) with high-resolution data from Sentinel-2 Level-2A. Traditional methods are labor-intensive while CatBoost offers an efficient solution, excelling in predictive accuracy. Leveraging atmospherically corrected Sentinel-2 data through the Google Earth Engine (GEE), our study contributes to scalable and precise Turbidity monitoring. A specific tabular dataset derived from Hong Kong contaminants monitoring stations enriches our study, providing region-specific insights. Results showcase the viability of this integrated approach, laying the foundation for adopting advanced techniques in global water quality management.

As space debris poses substantial risks to space-based assets, the need for efficient, high-resolution monitoring and prediction methods is pressing. This paper presents the findings from the project NEU4SST, exploring Neuromorphic Engineering, specifically event-based visual sensing coupled with Spiking Neural Networks (SNNs), as a solution for enhanced Space Domain Awareness (SDA). Our research concentrates on event-based visual sensors and SNNs, offering low power consumption and precise high-resolution data capture and processing. These technologies bolster the ability to detect and track objects in space, addressing key challenges in the Space domain. Our method exceeded previous models by 15% on the informedness metric, demonstrating its potential in improving SDA, and aiding safer, more efficient space operations. Continued research and development in this area are crucial for realising the full potential of Neuromorphic engineering for future space missions.

Hyperspectral sensors are increasing in popularity for Earth Observation applications due to their ability to gather data over multiple spectral bands. However, the processing of such amount of information is difficult to handle for the current computing capabilities of small satellites. Several Band Selection methodologies have been developed in the last years; although, some of them demand very low computational resources, they use, at least, Level 1 data products. Therefore, the Level 0 data needs to be processed and the spectral bands coregistered. Artificial Intelligence has shown its potential to reduce the computational burden while achieving high accuracies in EO applications. In this study, a Neural Network-based methodology is proposed to select a spectral band set directly using non coregistered data captured by hyperspectral sensors.

This paper introduces DTACSNet, a Convolutional Neural Network (CNN) model specifically developed for efficient onboard atmospheric correction and cloud detection in optical Earth observation satellites. The model is developed with Sentinel-2 data. Through a comparative analysis with the operational Sen2Cor processor, DTACSNet demonstrates a significantly better performance in cloud scene classification (F2 score of 0.89 for DTACSNet compared to 0.51 for Sen2Cor v2.8) and a surface reflectance estimation with average absolute error below 2% in reflectance units. Moreover, we tested DTACSNet on hardware-constrained systems similar to recent deployed missions and show that DTACSNet is 11 times faster than Sen2Cor with a significantly lower memory consumption footprint. These preliminary results highlight the potential of DTACSNet to provide enhanced efficiency, autonomy, and responsiveness in onboard data processing for Earth observation satellite missions.

Nowadays, the use of Artificial Intelligence on board Earth Observation satellites is under investigation for applications having strict bandwidth and latency requirements, such as vessel detection. However, many of the on-ground current computing pipelines rely on data post-processing techniques whose applications onboard satellites are tricky because of their limited computing power. To enable the analysis and the research of lightweight onboard data processing techniques, we provide VDS2Raw, the first Sentinel-2 Raw dataset for vessel detection applications. Finally, we also compared different object detection Deep Learning techniques in terms of detection performance and inference time to perform a feasibility analysis of performing onboard vessel detection on raw multi-spectral data.

Artificial intelligence (AI) is paving the way for a new era of algorithms focusing directly on the information contained in the data, autonomously extracting relevant features for a given application. While the initial paradigm was to have these applications run by a server hosted processor, recent advances in microelectronics provide hardware accelerators with an efficient ratio between computation and energy consumption, enabling the implementation of AI algorithms “at the edge.” In this way only the meaningful and useful data are transmitted to the end-user, minimizing the required data bandwidth, and reducing the latency with respect to the cloud computing model. In recent years, European Space Agency (ESA) is promoting the development of disruptive innovative technologies on-board earth observation (EO) missions. In this field, the most advanced experiment to date is the$\Phi $-sat-1, which has demonstrated the potential of artificial intelligence (AI) as a reliable and accurate tool for cloud detection on-board a hyperspectral imaging mission. The activities involved included demonstrating the robustness of the Intel Movidius Myriad 2 hardware accelerator against ionizing radiation, developing a Cloudscout segmentation neural network (NN), run on Myriad 2, to identify, classify, and eventually discard on-board the cloudy images, and assessing the innovative Hyperscout-2 hyperspectral sensor. This mission represents the first official attempt to successfully run an AI deep convolutional NN (CNN) directly inferencing on a dedicated accelerator on-board a satellite, opening the way for a new era of discovery and commercial applications driven by the deployment of on-board AI.

Convolution Neural Networks are a class of deep neural networks commonly used in audio and video elaborations. Their implementation on the edge represents a complex task due to the limited computational power and low power consumption requirement that characterize these applications. In this paper, a fully on-chip Convolutional Neural Network Field Programmable Gate Array-based hardware accelerator is presented. This approach allows to reduce power consumption due to off-chip memory accesses and aims to reduce design time. Advantages and limitations of the proposed architecture are discussed and a trade-off analysis is provided to give intuitions about the feasibility of this method.

Nowadays Voice User Interfaces (VUIs) have become popular thanks to their easiness of use that makes them accessible to the elderly and people with disability. Nevertheless, their use in embedded systems for the realization of portable devices is limited by the computation complexity, the memory requirements and power consumption of the keyword spotting (KWS) algorithms, usually based on deep neural networks. In this paper we propose a new algorithm based on convolutional neural networks for the keyword spotting task, that offers a good trade-off among accuracy, power consumption and memory footprint. To select our proposed solution, we compared different neural network architectures to select the best trade-off of these metrics. For further improvements of these performances we implemented our solution on a dedicated hardware platform as Myriad 2 by Movidius. The use of this chip has reduced inference time and energy per inference by 50%.

U-PHOS (Upgraded PHP Only for Space) is a project developed by a team of students from the University of Pisa with the goal to analyze and characterize the behavior of a Pulsating Heat Pipe (PHP), one of the most attractive two phases passive systems for thermal management in space applications. The PHP consists of a sealed serpentine capillary tube filled with a working fluid. The heat is efficiently transported by means of the combined action of phase change and capillary forces, so no extra equipment is required. The project aims at investigating the thermal response of such a device under a milli-gravity condition, in order to assess its effectiveness in space conditions. U-PHOS is one of the selected experiment of the REXUS/BEXUS programme, which allows European university students to carry out scientific and technical experiments on research rockets and balloons, thanks to a bilateral agency agreement between the German Aerospace Centre (DLR) and the Swedish National Space Board (SNSB) in collaboration with ESA. 19 students from the University of Pisa, with different backgrounds, compose the U-PHOS team. Students had the chance to completely design, build and test the experiment, which will flight up to space in March 2017. This paper intends to describe the work done by the students, their organization and how this experience empowered their careers, from both an academic and professional point of view.