Published:
2026-05-18Issue:
Vol. 22 No. 2 (2025)Section:
Technological presentAnotación YOLO de objetos pequeños en Ortofotos.
YOLO annotation of small objects in Orthophotos
Keywords:
Ortofotos, herramienta de anotación, Etiquetas YOLO, YOLO, UAV (es).Keywords:
Orthophotos, annotation tool, YOLO labels, YOLO, UAV (en).Downloads
Abstract (es)
Este trabajo presenta una herramienta manual para la anotación en formato de anotación YOLO de objetos pequeños en ortofotos de alta resolución. La solución se implementó en Python con arquitectura modular, teselado dinámico para navegación fluida y persistencia inmediata de etiquetas (.txt homónimo por imagen). Su desempeño se validó en equipos con Windows 10/11 de gama media, y mediante una prueba práctica con 10 estudiantes que anotaron tapas de registro y marcas viales en campañas de dos jornadas. Los resultados evidencian un tiempo de inicio de la aplicación inferior a 3 segundos desde que el usuario ejecuta el archivo hasta que la interfaz gráfica está lista para trabajar, junto con estabilidad sin cierres inesperados y una experiencia de uso positiva (zoom fluido, líneas guía, deshacer), favoreciendo la consistencia del conjunto de entrenamiento. La herramienta reduce la fricción operativa respecto a anotadores genéricos al centrarse en el flujo geoespacial con ortofotos grandes y producir salidas compatibles con YOLO, lo cual agiliza la construcción de datasets para tareas de detección en infraestructura urbana.
Abstract (en)
This work presents a manual tool for YOLO-format annotation of small objects in high-resolution orthophotos. The solution was implemented in Python with a modular architecture, dynamic tiling for smooth navigation, and immediate label persistence (per-image homonymous .txt). Its performance was validated on mid-range Windows 10/11 workstations and through a practical test with 10 students who annotated manhole covers and road markings over two field-day sessions. Results show an application startup time of less than 3 seconds from execution to a fully operational graphical interface, as well as stable operation without unexpected crashes and a positive user experience (fluid zoom, crosshair guides, undo), contributing to consistent training datasets. The tool reduces operational friction compared to generic annotators by focusing on geospatial workflows with large orthophotos and producing YOLO-compatible outputs, thereby streamlining the creation of datasets for object detection in urban infrastructure applications.
References
[1] M. Aljabri, M. AlAmir, M. AlGhamdi, M. Abdel Mottaleb, and F. Collado Mesa, “Towards a better understanding of annotation tools for medical imaging: A survey,” Multimedia Tools and Applications, vol. 81, no. 18, pp. 25877–25911, 2022, doi: 10.1007/s11042-022-12100-1.
[2] J. Weng, Z. Tang, and Y. Liu, “Application of improved YOLOv8 image model in urban manhole cover defect management and detection: Case study,” Sensors, vol. 25, no. 13, Art. 4144, 2025, doi: 10.3390/s25134144.
[3] Y. Zhang, F. Li, X. Wang, et al., “Real time detection of road manhole covers with a deep learning approach,” Scientific Reports, vol. 13, Art. 18791, 2023, doi: 10.1038/s41598-023-43173-z.
[4] H. Li, Q. Zhao, S. Chen, et al., “Manhole cover classification based on super resolution and UAV aerial imagery with YOLOv8,” Applied Sciences, vol. 14, no. 7, Art. 2769, 2024, doi: 10.3390/app14072769.
[5] V7 Labs, “A Friendly Guide to LabelImg,” 2022. [Online]. Available: https://www.v7labs.com/blog/labelimg-guide
[6] HumanSignal, “labelImg,” GitHub repository, 2025. [Online]. Available: https://github.com/HumanSignal/labelImg (original release 2015).
[7] VGG (Visual Geometry Group, Oxford), “VGG Image Annotator (VIA),” 2019–2020. [Online]. Available: https://www.robots.ox.ac.uk/~vgg/software/via/via_demo.html
[8] CVAT, “Installation Guide,” [Online]. Available: https://docs.cvat.ai/docs/administration/basics/installation/. Accessed: Jul. 28, 2025.
[9] Labelbox, “Plans & Pricing,” [Online]. Available: https://labelbox.com/pricing/. Accessed: Jul. 14, 2025.
[10] SuperAnnotate, “Pricing,” [Online]. Available: https://www.superannotate.com/pricing. Accessed: Jul. 14, 2025.
[11] SkalskiP, “Issue #271: It is simply very difficult to save the progress while annotating a large …,” GitHub: make sense, 2022. [Online]. Available: https://github.com/SkalskiP/make-sense/issues/271. Accessed: Jul. 14, 2025.
[12] J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, “You Only Look Once: Unified, Real Time Object Detection,” in Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR), 2016, pp. 779–788. doi: 10.1109/CVPR.2016.91.
[13] G. Jocher, A. Chaurasia, and J. Qiu, “YOLO by Ultralytics,” GitHub/Docs, 2023–2025. [Online]. Available: https://docs.ultralytics.com/ (acceso: Jul. 29, 2025).
How to Cite
IEEE
ACM
ACS
APA
ABNT
Chicago
Harvard
MLA
Turabian
Vancouver
Download Citation
Visitas
Downloads
License
Copyright (c) 2026 Revista Vínculos

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
3.png)
Creative Commons Attribution 4.0 Colombia
