CLCesar Ruben Sandoval Lozada

Cesar Ruben Sandoval Lozada

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Cesar Ruben Sandoval Lozada has a total of 3 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Spatial modeling of soil salinity in hyperarid agroecosystems of northern Peru using a hybrid machine learning ensemble

    by Cesar Ruben Sandoval-Lozada

    Soil salinization poses a critical threat to agricultural sustainability in arid agroecosystems, limiting food security and economic stability. On the northern coast of Peru, the synergy between inefficient water management and extreme climate variability has accelerated soil degradation, demanding precise spatial monitoring tools. This study proposes a Hybrid Smart Ensemble model to map soil electrical conductivity (EC), integrating optical data (Sentinel-2), dual-polarization SAR texture (Sentinel-1), and thermal imagery (Landsat-9 LST). Three modeling scenarios were evaluated: (1) Random Forest (RF), (2) Deep Neural Network (DNN), and (3) a Hybrid Ensemble. To overcome the typical scarcity of field data in developing regions, a Spatial Data Augmentation strategy based on Tobler's First Law was implemented. The results demonstrated that the hybrid architecture (Scenario 3) achieved superior performance (R² = 0.86; RMSE = 1.70 dS/m), exceeding the baselines of individual classifiers. While the Random Forest provided stability in moderate ranges, the inclusion of Deep Learning proved essential for capturing the nonlinear spectral response of extreme salt crusts (>12 dS/m). Variable importance analysis identified Ground Surface Temperature (LST) and Elevation as dominant predictors, validating the hypothesis that thermal inertia and topographic depressions are the main physical drivers of salinity in hyperarid environments. The resulting map, with a 10 m resolution, reveals "hot spots" in the Chulucanas and Lower Piura sectors, providing valuable information for implementing targeted detection and analysis zones to optimize drainage rehabilitation and soil recovery strategies.While themodel effectively delineates non-saline baselines and severe salt crusts, boundary zones (moderate salinity)remain challenging due to historical sampling limitations. Therefore, the applicability of the model is currentlylimited to hyperarid agroecosystems with edaphic and topographic characteristics similar to those of the LowerPiura Valley; results should be validated with additional field campaigns before implementation in other contexts.

  • Open Access

    Sistema de diagnóstico automático de enfermedades foliares en mango (Mangifera indica L.) basado en MobileNetV2

    by Cesar Ruben Sandoval Lozada, Ernesto Altamirano Flores, Elvira Cáceres Cayllahua, Luis José Llaque Ramos

    El cultivo del mango (Mangifera indica L.) es vulnerable a múltiples enfermedades foliares que impactan severamente su productividad, por lo cual la identificación temprana es fundamental para mitigar pérdidas económicas. El objetivo de esta investigación fue entrenar y evaluar un modelo de clasificación de enfermedades foliares en mango basado en la arquitectura MobileNetV2 mediante aprendizaje por transferencia. Se utilizó un conjunto de datos de 4000 imágenes con una partición estratificada del 70 % para entrenamiento, 15 % para validación y 15 % para prueba independiente (n= 600). Se implementó una estrategia de entrenamiento en dos fases en Python (TensorFlow/Keras) y el rendimiento se evaluó mediante métricas multidimensionales como el coeficiente Kappa de Cohen, MCC, curvas ROC y un análisis de confianza de las predicciones. El modelo demostró una alta capacidad de generalización en el conjunto de prueba independiente, alcanzando una exactitud del 87.67 %, un Kappa de 0.859 y un MCC de 0.862. El análisis ROC reveló un AUC micro-promedio de 0.9926. Además, se identificó que la principal dificultad diagnóstica reside en la distinción de síntomas incipientes de Gall Midge frente a hojas sanas (error del 22.67 %). Los resultados confirman que el análisis de confianza (media= 0.855 en aciertos vs media= 0.580 en errores) permite establecer umbrales de seguridad diagnóstica. Se concluye que MobileNetV2 ofrece un equilibrio óptimo entre eficiencia computacional y fiabilidad, lo que representa una solución viable para el monitoreo fitosanitario en agricultura de precisión.

  • Open Access

    Comparative Evaluation of CNN Architectures for Automatic Detection of Diseases in Mango Leaves

    by César Rubén Sandoval Lozada

    Early and accurate identification of leaf diseases in mango crops is essential for plant protection, improved agricultural yield, and reduced economic losses. This paper presents a comparative evaluation of five convolutional neural network (CNN) architectures in the automatic classification of eight categories of diseases and healthy leaves. The models evaluated include DenseNet201, ResNet50, InceptionV3, MobileNetV2, and EfficientNetB0. Using transfer learning and fine-tuning techniques, a maximum accuracy of 99.12% was achieved with the DenseNet201 and ResNet50 models. The comparative analysis considered not only accuracy metrics but also computational efficiency, training time, and number of parameters. The results demonstrate that ResNet50 offers the best balance between accuracy (99.12%) and computational efficiency (106.9 minutes of training), positioning itself as the most suitable architecture for practical implementations in agricultural monitoring systems in the field.