One of the challenges in the fight against poverty is the precise localization and assessment of vulnerable communities’ sprawl. The characterization of vulnerability is traditionally accomplished using nationwide census exercises, a burdensome process that requires field visits by trained personnel. Unfortunately, most countrywide censuses exercises are conducted only sporadically, making it difficult to track the short-term effect of policies to reduce poverty. This paper introduces a definition of vulnerability following UN-Habitat criteria, assesses different CNN machine learning architectures, and establishes a mapping between satellite images and survey data. Starting with the information corresponding to the 2,178,508 residential blocks recorded in the 2010 Mexican census and multispectral Landsat-7 images, multiple CNN architectures are explored. The best performance is obtained with EfficientNet-B3 achieving an area under the ROC and Precision-Recall curves of 0.9421 and 0.9457, respectively. This article shows that publicly available information, in the form of census data and satellite images, along with standard CNN architectures, may be employed as a stepping stone for the countrywide characterization of vulnerability at the residential block level.

This article presents the development of an interactive, augmented reality, and collaborative Pong game. In our approach, an overhead projector displays the game on the floor, showing the game's current state as it develops. In the meantime, a computer vision system sensing throughout an overhead depth camera detects the players and their movements. We tested our system in a roofed but wall-less environment. Our assessment of the gaming experience includes natural language processing on interviews with children and qualitative analysis of their verbal responses.

It is commonly accepted that there is no significant gain in computing precisely the center of radial distortion (CoD). Indeed, practitioners take the center of the image (CI) and the principal point (PP) as proxies for camera lens distortion correction. This article assesses the relevance of accurately measuring the CoD to obtain reliable 3-D reconstructions from multiple views. We found the reconstruction error arising when displacing the CoD from its correct location with a simulation. Then, we study the cases when the PP and the CoD are considered the same or different parameters. Using real images of a chessboard-like pattern, we observe that, for a high distortion lens, the reconstruction error significantly increases when using the CI as a replacement to the CoD, especially at the corners of an image. Although the error lessens for a low distortion lens, it is still significant.

Leveraging satellite monitoring and machine learning (ML) techniques for water clarity assessment addresses the critical need for sustainable water management. This study aims to assess water clarity by predicting the Secchi disk depth (SDD) using satellite images and ML techniques. The primary methods involve data preparation and SSD inference. During data preparation, AquaSat samples, originally from the L1TP collection, were updated with the Landsat 8 satellite’s latest postprocessing, L2SP, which includes atmospheric corrections, resulting in 33,261 multispectral observations and corresponding SSD measurements. For inferring the SSD, regressors such as SVR, NN, and XGB, along with an ensemble of them, were trained. The ensemble demonstrated performance with an average determination coefficient of $R^2$ of around 0.76 and a standard deviation of around 0.03. Field data validation achieved an $R^2$ of 0.80. Furthermore, we show that the regressors trained with L1TP imagery for predicting SSD result in a favorable performance with respect to their counterparts trained on the L2SP collection. This document contributes to the transition from semi-analytical to data-driven methods in water clarity research, using an ML ensemble to assess the clarity of water bodies through satellite imagery.

Global extreme poverty has fallen dramatically over the past two centuries, yet hundreds of millions remain impoverished, underscoring the need for scalable monitoring tools. In Mexico, poverty metrics are available only sporadically in terms of time and space (e.g., every 5 years at the municipal level), making it difficult for decision-makers to access reliable, up-to-date, and sufficiently detailed information, highlighting the need for higher-resolution, timely methods. To address this problem, we propose a two-stage approach that combines socioeconomic and Earth Observations-based data. Initially, a machine learning model maps census variables to official poverty indicators belonging to a multidimensional model, yielding fine-scale poverty estimates. A census-based model trained with eXtreme Gradient Boosting (XGBoost) achieved a determination coefficient ($R^2$) of approximately $0.842$, indicating strong agreement with official poverty figures and providing high-resolution proxies. Afterward, we use features based on remote observations to predict these poverty estimates at a 469 m grid scale. In this case, advanced foundation models outperformed other machine learning (ML) approaches, achieving an $R^2$ of $0.683$. While foundation models enable more accurate, fine-scale poverty mapping and could accelerate poverty assessments, their use comes at a heavy price in terms of carbon emissions.