Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Salvador Castro-Tapia has a total of 44 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Sequence-Aware Dataset Auditing for Leakage-Free Benchmarking of YOLO Detectors for Bottle Detection

    by Rafael Reveles-Martínez, Sebastián Burciaga-Sosa, José M. Celaya-Padilla, Salvador Castro-Tapia, Huizilopoztli Luna-García, Humberto Morales-Magallanes, Mayra N. Regalado-Pérez, César Landeros-Soriano, Umanel A. Hernández-González, Flabio D. Mirelez-Delgado, et al.

    This work presents a reproducible YOLO-based pipeline for bottle detection in sandy environments, emphasizing dataset integrity, leakage-free evaluation, and deployment-oriented model selection. A one-class dataset of 1585 images and 3167 annotated bottles was audited to identify annotation-format defects and near-duplicate contamination between training and validation partitions. Sequence membership was reconstructed through perceptual-image similarity and used to assign complete image components to a sequence-aware train/validation split, eliminating the near-duplicate pairs found in the initial random partition. A controlled ablation holding model, seed, and corrected labels fixed showed that the random split reports 0.040 higher mAP@0.5:0.95 than the sequence-aware split (0.787 vs. 0.747), quantifying the leakage risk directly rather than only asserting it. Five YOLO configurations were then benchmarked under three independent seeds each; the observed mAP@0.5:0.95 differences among models (0.004–0.008) were small in absolute magnitude and, given only three seeds per model, are interpreted descriptively rather than as evidence of statistical equivalence or significance, so yolo11n_bottle was selected through a joint accuracy-parity, compactness, and exportability criterion (precision 0.982, recall 0.985, mAP@0.5 0.992, mAP@0.5:0.95 0.748), using approximately ten times fewer parameters than the largest configuration and producing a 5.2 MB checkpoint. ONNX export preserved detection geometry closely (100% count agreement, mean matched IoU ≥0.9998), without meeting strict metric-parity tolerances. A stratified sample of 108 frames from operational RealSense BAG footage was manually annotated by an independent reviewer and evaluated quantitatively: mAP@0.5 remained close to the internal validation figure (0.927 vs. 0.992), while mAP@0.5:0.95 fell substantially (0.483 vs. 0.747), revealing a localization gap between the curated benchmark and operational conditions that this manuscript reports transparently. Together, these results show that dataset auditing, sequence-aware partitioning, multiseed benchmarking, and manually annotated operational evidence are each necessary to interpret a detection benchmark built from continuous video acquisition, providing a traceable, reproducible workflow for selecting and evaluating compact visual-perception models for resource-constrained environmental applications.

  • Open Access

    Enhancing Embedded Systems Education Through Remote Laboratories with Real-Time Firmware Observability

    by Fabián García-Vázquez, Héctor A. Guerrero-Osuna, Luis F. Luque-Vega, Jesús Antonio Nava-Pintor, Jorge A. Lizarraga, Salvador Castro-Tapia, Mario Molina-Almaraz, Ana G. Castañeda-Miranda, Carlos A. Olvera-Olvera

    Microcontroller programming is a key element in engineering education, enabling students to acquire practical skills in embedded systems. Despite its importance, limited access to physical laboratories at many institutions restricts hands-on learning experiences. Remote laboratories have been introduced as a solution; however, most rely solely on external visual feedback, which constrains observability and limits analysis of internal program behavior. This paper introduces a remote laboratory for microcontroller education utilizing an STM32 platform within a Cloud–IoT framework. The proposed system allows real-time monitoring of internal firmware variables through compilation artifacts, without requiring direct interaction with traditional debugging tools. A pilot study involving 40 students in a microcontrollers course, who engaged with the system remotely using standard development workflows, was conducted. Usability was measured via the System Usability Scale (SUS), yielding an average score of 75.19 and reflecting high satisfaction. The platform enables real-time observation of internal system states during program execution, facilitating program analysis while maintaining an experience similar to traditional on-site laboratories. These findings indicate that the system offers a robust and scalable alternative for embedded systems education, with improved observability over existing remote laboratory solutions.


  • Open Access

    Deep Learning-Based Multistage Peach Ripeness Detection with Data Leakage Mitigation and Real-World Validation

    by Salvador Castro-Tapia, Germán Díaz-Florez, Rafael Reveles-Martínez, Héctor A. Guerrero-Osuna, Luis F. Luque-Vega, Humberto Morales-Magallanes, Jorge Pablo Vega-Borrego, Gilberto Vázquez-García, Carlos A. Olvera-Olvera

    Accurate fruit ripeness assessment is essential for optimizing harvest timing and ensuring postharvest quality, particularly in climacteric fruits such as peaches, which exhibit rapid and heterogeneous ripening. This study proposes a deep learning-based approach for multistage peach ripeness classification across five maturity levels (green, green-blushed, blushed, yellow-blushed, and fully yellow). Four datasets were constructed using controlled image acquisition, segmentation, data augmentation, and perceptual hashing to mitigate data leakage. The performance of AlexNet, EfficientNet-B0, and three YOLO (You Only Look Once) architectures (YOLOv8, YOLOv11, and YOLOv12) was evaluated using standard metrics, including accuracy, precision, recall, F1 score, mAP, and inference speed. Results show that YOLO-based models significantly outperform classical networks, achieving accuracies between 95.25% and 98.3% and mAP@0.5 above 98.25%, while also reducing inference time to 8.1–12.7 ms compared with 722.23 ms for AlexNet and 171.87 ms for EfficientNet-B0. In a practical sorting experiment with 214 peaches, YOLOv12 achieved 92.06% accuracy, demonstrating robust real-world performance. Misclassifications were primarily observed between adjacent ripeness stages. These findings indicate that YOLO-based models provide an effective and scalable solution for real-time fruit sorting, while the use of perceptual hashing enhances dataset reliability and model generalization for deployment in agricultural quality control systems.


  • Open Access

    A Feasibility Study of IoT-Based Classification of Residential Water-Use Activities in Storage Tank Systems: A Comparative Analysis of Decision Trees, Random Forest, SVM, KNN, and Neural Networks

    by Iván Neftalí Chávez-Flores, Héctor A. Guerrero-Osuna, Jesuś Antonio Nava-Pintor, Fabián García-Vázquez, Luis F. Luque-Vega, Rocío Carrasco-Navarro, Marcela E. Mata-Romero, Jorge A. Lizarraga, Salvador Castro-Tapia

    The increasing scarcity of urban water resources, particularly in regions with intermittent supply and household water storage tanks, demands monitoring approaches capable of identifying end-use consumption patterns beyond aggregated volume measurements. Framed primarily as a feasibility study, this research presents an IoT-based framework for the automated classification of residential water consumption activities using water-level dynamics and supervised machine learning. A non-intrusive sensing architecture based on hydrostatic pressure measurements was deployed in a domestic water tank and integrated with a cloud-based data acquisition and processing platform. Five representative household states and activities were considered: tank refilling, stable state, toilet flushing, washing clothes, and taking a bath. A labeled dataset comprising 4396 consumption events was used to train and evaluate Decision Tree, Random Forest, Support Vector Machine (SVM), k-Nearest Neighbors, and Recurrent Neural Network (LSTM) models using features derived from water-level variations. All models achieved high performance, with accuracies above 0.92 and weighted F1-scores up to 0.93. The evaluated models showed highly comparable results, with the SVM (RBF) achieving a slightly higher accuracy (0.9307) in this evaluation setting, while ROC analysis showed AUC values between 0.97 and 1.00 across all classes, indicating strong discriminative capability. Additionally, specific activities such as washing clothes and tank refilling achieved precision and recall values above 0.95. These findings confirm that hydrostatic pressure-based sensing, combined with machine learning, enables reliable identification of domestic water-use events under intermittent supply conditions. The proposed approach provides actionable insights for demand management, leak detection, and user awareness, supporting more efficient and sustainable residential water consumption strategies.


  • Open Access

    Relación del burnout y engagement en estudiantes universitarios de un instituto tecnológico en Zacatecas, México

    by José Iván Padilla Lugo, José Roberto González Hernández, Salvador Castro Tapia, María Esther Sánchez Cabral

    El burnout o cansancio emocional escolar es uno de los problemas principales en estudiantes universitarios. El engagement académico ha comprobado ser un constructo fundamental para impulsar el aprendizaje, el rendimiento y el bienestar emocional de estudiantes universitarios. El objetivo es identificar la relación entre burnout y engagement en estudiantes universitarios de un Instituto Tecnológico en el estado de Zacatecas, México. Se aplicaron 212 formularios a estudiantes de cinco carreras; aplicando el coeficiente de correlación de Pearson (r) y el análisis por conglomerados no jerárquico de K medias. Se comprueba la existencia de relación negativa entre burnout y engagement, se verifica el aumento de estrés prolongado de los estudiantes y la disminución de un estado motivacional satisfactorio persistente que experimentan en relación a su actividad académica. Se concluye que la prevención y gestión del burnout, junto con estrategias para promover el engagement, pueden ser importantes para el bienestar de los estudiantes.

  • Open Access

    Design and Validation of a Solar-Powered LoRa Weather Station for Environmental Monitoring and Agricultural Decision Support

    by Uriel E. Alcalá-Rodríguez, Héctor A. Guerrero-Osuna, Fabián García-Vázquez, Jesús A. Nava-Pintor, Luis F. Luque-Vega, Emmanuel Lopez-Neri, Salvador Castro-Tapia, Luis O. Solís-Sánchez, Ma. del Rosario Martínez-Blanco

    Due to changing weather conditions, productivity needs to be enhanced and resources must be used more efficiently in agriculture. Precision agriculture relies on systems that can gather real-time environmental data to address these issues. However, the high cost of commercial weather stations often limits their adoption in rural areas. This study introduces a low-cost weather station designed for precision agriculture applications. The system consists of three main modules. The first module is the weather station, which gathers data on temperature, relative humidity, barometric pressure, solar radiation, wind speed and direction, and precipitation. It then transmits this data via LoRa communication to the local console module. This console receives the data, displays it on a screen, and sends it through Wi-Fi to the cloud server module. The cloud server presents the information via an interactive interface and is responsible for storing, processing, and analyzing the data records collected. The system was installed in the municipality of Ojocaliente, Zacatecas, Mexico, where performance and validation tests were conducted over a one-month period using sensors and reference measurements to evaluate the accuracy and stability of the data. The results showed high operational reliability and a strong correlation between the recorded values and the reference data. This confirms that the proposed solution provides a scalable, low-cost, and reliable alternative for environmental monitoring in precision agriculture.


  • Modular Platform Architecture for a Product Family of Functional Growth Chambers for Sustainable Controlled Agriculture

    by M. F. Jara-Villagrana, C. A. Olvera-Olvera, S. Villgrana-Barraza, S. Castro-Tapia, L. O. Solís-Sánchez, G. Díaz-Flórez


  • Open Access

    Conceptual Design Based on Modular Platforms for a Prototype of a Functional Growth Chamber for Cuttings in Controlled Agriculture

    by María Fernanda Jara-Villagrana, Carlos Alberto Olvera-Olvera, Santiago Villagrana-Barraza, Salvador Castro-Tapia, Salvador Ibarra-Delgado, José Ricardo Gómez-Rodríguez, Remberto Sandoval-Aréchiga, Víktor I. Rodríguez-Abdalá, Germán Díaz-Flórez

    Agricultural research and propagation systems often suffer due to a lack of access to affordable, adaptable, and well-structured technological solutions. Traditional plant growth devices typically rely on ad hoc construction, which limits their scalability, reuse, and adaptability. This study employs a user-centered conceptual design methodology based on product platform development and modular architecture to design a growth chamber for plant cuttings. The approach followed three main phases: (i) identification and classification of user needs, (ii) functional modeling of the base system and its variants, and (iii) architectural modularization through heuristic principles. Interviews with researchers yielded 55 functional requirements, of which 26 were defined as essential. Functional models were developed for both a base system and two variant systems incorporating alternative irrigation and sensing technologies. Heuristic analysis identified independent modules, such as irrigation, lighting, environmental monitoring, and control. Subsequently, block diagrams were used to translate functional logic into spatially coherent conceptual designs. The resulting architecture supports modular integration, reconfiguration, and scalability for diverse experimental needs. This work demonstrates that structured design methodologies, which are commonly used in industrial contexts, can be effectively applied in agricultural research settings to produce solutions that are versatile, low-cost, and have enduring value, offering a pathway for innovation, reproducibility, and technology transfer in resource-limited environments.


  • Open Access

    Defining Feasible Joint and Geometric Workspaces Through Boundary Functions

    by Jorge A. Lizarraga, Dulce M. Navarro, Marcela E. Mata-Romero, Luis F. Luque-Vega, Luis Enrique González-Jiménez, Rocío Carrasco-Navarro, Salvador Castro-Tapia, Héctor A. Guerrero-Osuna, Emmanuel Lopez-Neri

    This work presents an alternative method for defining feasible joint-space boundaries and their corresponding geometric workspace in a planar robotic system. Instead of relying on traditional numerical approaches that require extensive sampling and collision detection, the proposed method constructs a continuous boundary by identifying the key intersection points of boundary functions. The feasibility region is further refined through centroid-based scaling, addressing singularity issues and ensuring a well-defined trajectory. Comparative analyses demonstrate that the final robot pose and reachability depend on the selected traversal path, highlighting the nonlinear nature of the workspace. Additionally, an evaluation of traditional numerical methods reveals their limitations in generating continuous boundary trajectories. The proposed approach provides a structured method for defining feasible workspaces, improving trajectory planning in robotic systems.


  • Open Access

    Defining Feasible Joint and Geometric Workspaces Through Boundary Functions

    by Jorge A. Lizarraga, Dulce M. Navarro, Marcela E. Mata-Romero, Luis F. Luque-Vega, Luis Enrique González-Jiménez, Rocío Carrasco-Navarro, Salvador Castro-Tapia, Héctor A. Guerrero-Osuna, Emmanuel López-Neri

    This work presents an alternative method for defining feasible joint-space boundaries and their corresponding geometric workspace in a planar robotic system. Instead of relying on traditional numerical approaches that require extensive sampling and collision detection, the proposed method constructs a continuous boundary by identifying key intersection points of boundary functions. The feasibility region is further refined through centroid-based scaling, addressing singularity issues and ensuring a structured trajectory. Comparative analyses demonstrate that the final robot pose and reachability depend on the selected traversal path, highlighting the nonlinear nature of the workspace. Additionally, an evaluation of traditional numerical methods reveals their limitations in generating continuous boundary trajectories. The proposed approach provides a structured method for defining feasible workspaces, improving trajectory planning in robotic systems.