Recent progress in computer vision and embedded systems has facilitated real-time monitoring of bioprocesses; however, lightweight and scalable solutions for resource-constrained settings remain limited. This work presents a modular framework for monitoring Chlorella vulgaris growth by integrating RGB image processing with multimodal sensor fusion. The system incorporates a Logitech C920 camera and low-cost pH and temperature sensors within a compact photobioreactor. It extracts RGB channel statistics, luminance, and environmental data to generate a 10-dimensional feature vector. A feedforward artificial neural network (ANN) with ReLU activations, dropout layers, and SMOTE-based data balancing was trained to classify growth phases: lag, exponential, and stationary. The optimized model, quantized to 8 bits, was deployed on an ESP32 microcontroller, achieving 98.62% accuracy with 4.8 ms inference time and a 13.48 kB memory footprint. Robustness analysis confirmed tolerance to geometric transformations, though variable lighting reduced performance. Principal component analysis (PCA) retained 95% variance, supporting the discriminative power of the features. The proposed system outperformed previous vision-only methods, demonstrating the advantages of multimodal fusion for early detection. Limitations include sensitivity to lighting and validation limited to a single species. Future directions include incorporating active lighting control and extending the model to multi-species classification for broader applicability.

El cultivo de microalgas como Chlorella Vulgaris es una estrategia eficaz para la captura de dióxido de carbono atmosférico, contribuyendo a mitigar el cambio climático. Este estudio propone una metodología basada en redes neuronales artificiales para optimizar el crecimiento de Chlorella Vulgaris en fotobiorreactores. Utilizando imágenes en RGB, temperatura y pH del cultivo, se entrena un modelo de MLPRegressor para predecir la densidad celular con alta precisión (R²=0.985). La implementación del sistema incluye la captura de imágenes en tiempo real y el análisis de regiones de interés, mejorando la precisión y eficiencia del cultivo. Los resultados subrayan el potencial de la inteligencia artificial para la biotecnología de microalgas, ofreciendo una alternativa no invasiva al conteo celular manual y optimizando las condiciones del cultivo para maximizar la productividad.

Urban expansion and industrial activities in peri-urban environments modify land cover and influence surface thermal patterns. This study examines the thermal behavior of artisanal brick kilns in León, Guanajuato, Mexico, using a paired spatial approach designed to distinguish the contribution of kiln activity from that of the surrounding landscape. Landsat 8 Collection 2 Level-2 imagery acquired during the dry seasons (March-May) of 2024-2026 was processed in Google Earth Engine. Eight active brick kiln polygons were paired with eight bare-soil control sites and evaluated within concentric buffers of 100, 500, and 1000 m. Eight spectral and thermal indices, including Land Surface Temperature (LST), NDVI, and NDMI, were calculated. Wilcoxon signed-rank and Mann-Whitney U tests indicate that bare-soil control sites consistently exhibited higher LST values than active brick kilns, with a mean difference of approximately $\Delta LST\approx-3.7^{\circ}C$. Spearman correlation showed a moderate inverse relationship between LST and NDMI $(\rho\approx-0.49)$, indicating that higher surface temperatures are associated with reduced vegetation moisture. The results indicate that, under the analyzed daytime Landsat observations, exposed bare soil exhibited higher surface temperatures than the active brick kiln polygons. This three-year assessment provides a reference for monitoring peri-urban thermal conditions and supports the use of paired multi-buffer analyses to evaluate the effects of land cover on surface temperature.

Este estudio presenta un sistema automatizado para clasificar las etapas de crecimiento en cultivos de Chlorella vulgaris mediante análisis de imágenes RGB y redes neuronales artificiales (ANN) en un fotobiorreactor inteligente. La metodología integra la captura de imágenes en tiempo real con sensores ambientales (pH, temperatura e intensidad lumínica) para extraer características discriminativas (valores RGBI, densidad celular). Un modelo ANN multicapa, optimizado con activación ReLU, regularización dropout y el optimizador Adam, alcanzó una precisión promedio de 0.995 ± 0.001 en 40 experimentos, con métricas consistentes de precisión, recall y F1-score. Técnicas como SMOTE para balanceo de clases y validación cruzada k-fold aseguraron robustez. Un análisis de componentes principales (PCA) confirmó la separabilidad lineal entre las fases de crecimiento (retardo, exponencial y estacionaria). Los resultados muestran su potencial para el monitoreo automatizado y escalable de microalgas. Futuras investigaciones podrían adaptar este enfoque a otras especies o condiciones ambientales dinámicas.