848.207

Visitas

2.609.552

Descargas

182.228

Búsquedas

49.330

Documentos

Envíos recientes

Apoyo Técnico y Administrativo en la Supervisión de la Infraestructura Vial, Espacio Público y Gestión Documental de la Secretaría de Infraestructura Territorial de Tunja
(Universidad Santo Tomás, 2026-09-23) Ortiz Castro, Vanessa Carolina; Rojas Gamba, Néstor Iván; Universidad Santo Tomás; https://scienti.minciencias.gov.co/cvlac/visualizador/generarCurriculoCv.do?cod_rh=0001412931; https://scholar.google.com/citations?user=ffQAroUAAAAJ&hl=en
Road infrastructure and public spaces are fundamental components of urban development and quality of life in the municipality of Tunja. In this context, the main objective of this internship is to provide technical and administrative support to the Secretariat of Territorial Infrastructure in the supervision of road interventions, the management of technical information, and the collection of preliminary inputs for public infrastructure projects. The relevance of this project lies in the need to strengthen institutional supervision and monitoring processes, ensuring that construction works comply with quality standards and current technical regulations, while also guaranteeing the proper restoration of the initial conditions following interventions carried out by contractors or public service companies. To achieve this, the methodology employed combines office-based administrative work with fieldwork in urban and rural areas of the municipality. Technical inspections, photographic records, material verification, database consolidation, and the preparation of preliminary technical concepts will be carried out. The scope of the project includes the completion of 600 hours of regulatory internship activities, covering everything from on-site technical evaluations to support in data collection for market studies and contracting processes. As potential outcomes, the project is expected to optimize the technical monitoring of construction sites, improve the department's documentary traceability, and provide technical and rigorous tools that facilitate administrative decision-making by the Secretariat.
Monitoreo y Análisis Georreferenciado de Prodiplosis Longifila Mediante Visión Artificial en Tomate Bajo Invernadero
(Universidad Santo Tomás, 2026-09-23) González Torres, David Santiago; Díaz Rubiano, Kevin Andrés; Pardo Beainy, Camilo Ernesto; Gutiérrez Cáceres, Edgar Andrés; Universidad Santo Tomás; https://scienti.colciencias.gov.co/cvlac/visualizador/generarCurriculoCv.do?cod_rh=0001257986; https://scholar.google.com/citations?user=iLrVanwAAAAJ&hl=en; https://orcid.org/0000-0003-0012-3679
This thesis developed and evaluated a georeferenced acquisition, detection, and analysis system to support the monitoring of Prodiplosis longifila in greenhouse tomato crops using computer vision, deep learning, and precision agriculture tools. The problem was addressed as an integrated workflow involving macro image acquisition, small-object detection, metadata traceability, structured storage, and spatial visualization of larval foci. Data acquisition was performed using an iPhone 13 coupled with a 200x Apexel macro lens with integrated lighting, allowing high-resolution images (3024x4032 pixels) to be captured on young leaves, apical shoots, flowers, and tender stems. Images were preserved in HEIC format to retain EXIF metadata, including date, time, and GPS coordinates. These coordinates were used only as an approximate spatial reference for georeferencing the images. The initial dataset consisted of 410 original field images and, after preprocessing and data augmentation in Roboflow, reached 984 images distributed into training, validation, and test subsets. The labeling process followed an object detection approach using bounding boxes, distinguishing live and dead larvae according to a classification protocol applied during field review. Subsequently, YOLOv8x, YOLOv9e, YOLOv12x, and YOLO26x were evaluated for 100 epochs with a batch size of 3 and an effective input size of 704x704 pixels; this value resulted from automatically aligning imgsz=695 to the multiple of 32 required by the networks. YOLOv8x was selected as the final model because of its stability and stronger overall behavior, achieving, on the validation set, a global [email protected] of 0.942, with 0.995 for dead larvae and 0.890 for live larvae. The reported metrics correspond exclusively to the validation set; the test subset of 42 images and 72 instances was reserved but not evaluated because the required inference outputs were not preserved. The implemented system integrates a user interface, EXIF metadata extraction, YOLOv8x inference, MySQL storage, GPS point visualization, Kernel density map generation, before/after comparison of phytosanitary applications, and KML export. During the three monitoring sessions conducted in the same crop area between June 3 and June 17, 2026, with pre- and post-application records separated by approximately 48 hours, the georeferenced maps made it possible to observe changes in the quantity and spatial distribution of live- and dead-larva detections between the records obtained before and after the phytosanitary applications, together with the persistence of residual foci. These monitoring sessions corresponded to successive temporal observations rather than independent experiments. The changes are described in spatial terms and do not allow causality to be established or the efficacy of the applied product to be determined. The maps must be interpreted as monitoring support tools based on approximate spatial references rather than as high-precision topographic representations. Overall, the results demonstrate the feasibility of integrating computer vision and georeferencing as a support tool for phytosanitary monitoring, without replacing technical field inspection or complementary agronomic analyses.
Análisis de la tecnología DECCO (Naturcover+Fullcover) en aguacate Hass tipo exportación hacia Europa y Asia
(Universidad Santo Tomás, 2026-09-24) Holguín Ruíz, Edith Patricia; Buitrago Núñez, Diana Astrid; Universidad Santo Tomas
Colombian Hass avocado faces significant postharvest challenges due to strict phytosanitary and quality standards required by European and Asian markets, combined with transit times exceeding 40 days. This study aimed to implement DECCO technology (Naturcover + Fullcover) in export-grade Hass avocado, evaluating its efficiency across three dimensions: the application process, the physical characteristics of the fruit, and postharvest costs. An experimental design with mixed data collection was used, applied to 10 metric tons of fruit processed in one day at a packing facility, with a sample of 5 metric tons. Three treatments were evaluated: Naturcover coating applied with the Fullcover electrostatic system, thiabendazol fungicide (TBZ), and an untreated control group, measured at 0, 7, 15, 25, and 40 days under refrigerated conditions simulating container transit. Results showed that Naturcover significantly reduced weight loss and preserved color uniformity in the fruit, extending its postharvest shelf life without affecting organoleptic properties. Thiabendazol outperformed other treatments in controlling fungal diseases such as anthracnose and stem-end rot. The Fullcover electrostatic system improved coverage and application efficiency while minimizing water consumption. The Activity-Based Costing (ABC) analysis identified three main postharvest cost components: labor, chemical inputs, and refrigerated transport logistics, with the latter having the greatest economic impact. It is concluded that an integrated strategy combining Naturcover with a broad-spectrum fungicide represents the most competitive alternative for positioning Colombian Hass avocado in international markets.
Desconexión Entre el Diseño y la Implementación de Políticas Públicas para el Desarrollo de la Bioeconomía en Colombia: Análisis de Efectividad, Coherencia Institucional y Capacidades Territoriales (2010-2024)
(Universidad Santo Tomás, 2026-09-24) Hoyos Rojas, Carlos Andrés; Lobo Garrido, Gustavo Andrés; Universidad Santo Tomás; https://scienti.minciencias.gov.co/cvlac/visualizador/generarCurriculoCv.do?cod_rh=0001658185
The study analyzed the disconnect between regulatory design and the implementation of public policies for the bioeconomy in Colombia. Indexed literature from international databases was examined using bibliometric techniques, along with CONPES policy documents. A review of regulatory documents and qualitative research in the field of bioeconomy was conducted, considering relevant studies and public policy strategies. The findings revealed an evolutionary process in which public policies shifted from addressing specific sectoral problems to adopting a comprehensive sustainable development approach. The structural gap between policy design and measurable, sustainable outcomes at the territorial level is attributed to the inadequate implementation of SNC instruments. A temporal and thematic analysis showed a high concentration of research in traditional areas, such as agricultural biotechnology and bioprospecting. Strategic action for the bioeconomy focuses on industrial bioproducts, health-related value creation, the circular economy, and high-value-added bioindustries. However, other stated priorities contrast with this focus. The findings also reveal insufficient technical and leadership capabilities for intersectoral coordination, indicating significant systemic institutional shortcomings. The governance crisis within Colombia’s administrative system hinders the implementation of changes aligned with the constitutional principles of sustainable development. It also reflects a failure to comply with constitutional mandates concerning the rational use of natural resources and the protection of the nation’s environmental heritage. An adaptive governance model is proposed across four areas: strengthening science and technology capabilities, developing the domestic market for bio-based products, promoting flagship bioeconomy projects, and improving institutional coordination. Regulatory frameworks and operational capabilities must be aligned across Colombia’s territories to position the country as a global leader in sustainable bioeconomy, given its exceptional biodiversity.
Soluciones basadas en inteligencia artificial para reducir tiempos y costos en la interventoría y supervisión de contratos de obra pública en Colombia
(Universidad Santo Tomás, 2026-09-24) Morales Godoy, Jose Sebastian; Rodriguez Cujia, Alexander; Lizcano Prada, Fabio Andres; Universidad Santo Tomas
This research aimed to propose a conceptual model based on artificial intelligence (AI), particularly machine learning, natural language processing (NLP), and generative AI models to optimize public works supervision processes in Colombia. The study identifies limitations of the traditional supervision model, including delays, deficiencies in document management, and difficulties in contractual control. Likewise, the study proposes a conceptual model based on digital platforms such as Procore to strengthen operational efficiency, information traceability, and early risk detection through automation and real-time monitoring. The research used a qualitative, descriptive approach supported by a documentary review of academic, legal, and technological sources on digital transformation in the construction sector. The results show that integrating artificial intelligence and collaborative tools can improve decision-making, optimize administrative processes, and strengthen transparency in public infrastructure contracts. Finally, the study concludes that artificial intelligence does not replace the supervisor's professional judgment but instead supports preventive control and operational efficiency in public works supervision processes in Colombia.