EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

QHAWAY: An Instance Segmentation and Monocular Distance Estimation ADAS for Vulnerable Road Users in Informal Andean Urban Corridors

Abel De la Cruz-Moran, Hemerson Lizarbe-Alarcon iD, Wilmer Moncada iD, Victor Bellido-Aedo iD, Carlos Carrasco-Badajoz iD, Carolina Rayme-Chalco iD, Cristhian Aldana iD, Yesenia Saavedra, Edwin Saavedra iD, Alex Pereda

Sensors · 21.04.2026 · 26 (8) · 2569

DOI: 10.3390/s26082569 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 21.04.2026. Verfasst von Abel De la Cruz-Moran und 9 weiteren Autor:innen. Beteiligte Einrichtung: Research Group GECOTRAN (Studies in Management, Construction and Transport), Faculty of Mining, Geological and Civil Engineering, Universidad Naciona…. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Vulnerable road users in informal urban environments confront a distinct set of hazards that standard computer vision datasets are ill-equipped to represent: artisanal speed bumps constructed without regulatory compliance, deteriorated road markings, and the mototaxi—a three-wheeled motorized vehicle that constitutes the primary informal transport mode in intermediate Andean cities yet is absent from all major international repositories. This paper presents QHAWAY—from Quechua qhaway, a transitive verb meaning “to look; to observe”—an Advanced Driver Assistance System (ADAS) predicated on instance segmentation, monocular distance estimation via the pinhole camera model, and Time-to-Collision (TTC) computation, developed for the road environment of Ayacucho, Peru (2761 m a.s.l.), a city recognised by UNESCO as a Creative City of Crafts and Folk Art since 2019. A hybrid dataset comprising 25,602 images with 127,525 annotated instances across 12 classes was assembled by combining an original local collection of 4598 images (10,701 instances) captured through four complementary acquisition methods across the five urban districts of the Huamanga province with three established international datasets (BDD100K, BSTLD, RLMD; 21,004 images, 116,824 instances). A three-phase progressive training strategy with monotonically increasing resolution (640, 800, and 1024 pixels) was evaluated as an ablation study. A multi-architecture comparison spanning YOLOv8L-seg and the YOLO26 family (nano, small, large) identified YOLO26L-seg as the best-performing model, attaining mAP50 Box of 0.829 and mAP50 Mask of 0.788 at epoch 179. The integration of ByteTrack multi-object tracking with the pinhole equation D=(Hreal×f)/hpx delineates operational risk zones aligned with the NHTSA forward collision warning standard (danger:

Schlagwörter

Stichwörter der Autor:innen

Andean urban environmentGrad-CAMYOLO26YOLOv8advanced driver assistance systeminformal road infrastructureinstance segmentationmonocular distance estimationmototaxiroad safetytime-to-collisionvulnerable road users

Autor:innen und Institutionen

  • Research Group GECOTRAN (Studies in Management, Construction and Transport), Faculty of Mining, Geological and Civil Engineering, Universidad Nacional de San Cristóbal de Huamanga, Ayacucho 05000, Peru.
  • Remote Sensing and Renewable Energy Laboratory (LABTELER), Faculty of Mining, Geological and Civil Engineering, Universidad Nacional de San Cristóbal de Huamanga, Ayacucho 05000, Peru.
  • Biodiversity and Geographic Information Systems Laboratory (BioSIG), Faculty of Mining, Geological and Civil Engineering, Universidad Nacional de San Cristóbal de Huamanga, Ayacucho 05000, Peru.
  • Research Group on Productive Efficiency, Environment and Society (EPAS UNF), Universidad Nacional de Frontera, Sullana 20100, Peru.
  • Universidad Nacional de Frontera, Sullana 20100, Peru.
  • Statistics Programme, Faculty of Mining, Geological and Civil Engineering, Universidad Nacional de San Cristóbal de Huamanga, Ayacucho 05000, Peru.

ORCID-Profile: Hemerson Lizarbe-Alarcon, Wilmer Moncada, Victor Bellido-Aedo, Carlos Carrasco-Badajoz, Carolina Rayme-Chalco, Cristhian Aldana, Edwin Saavedra

Zitieren

APA

Cruz-Moran, A., Lizarbe-Alarcon, H., Moncada, W., Bellido-Aedo, V., Carrasco-Badajoz, C., Rayme-Chalco, C., Aldana, C., Saavedra, Y., Saavedra, E., & Pereda, A. (2026). QHAWAY: An Instance Segmentation and Monocular Distance Estimation ADAS for Vulnerable Road Users in Informal Andean Urban Corridors. Sensors, 26(8), 2569. https://doi.org/10.3390/s26082569

Vancouver

Cruz-Moran A, Lizarbe-Alarcon H, Moncada W, Bellido-Aedo V, Carrasco-Badajoz C, Rayme-Chalco C, et al. QHAWAY: An Instance Segmentation and Monocular Distance Estimation ADAS for Vulnerable Road Users in Informal Andean Urban Corridors. Sensors. 2026;26(8):2569. doi:10.3390/s26082569

BibTeX
@article{CruzMoran2026,
  author = {Cruz-Moran, Abel De la and Lizarbe-Alarcon, Hemerson and Moncada, Wilmer and Bellido-Aedo, Victor and Carrasco-Badajoz, Carlos and Rayme-Chalco, Carolina and Aldana, Cristhian and Saavedra, Yesenia and Saavedra, Edwin and Pereda, Alex},
  title = {QHAWAY: An Instance Segmentation and Monocular Distance Estimation ADAS for Vulnerable Road Users in Informal Andean Urban Corridors},
  journal = {Sensors},
  year = {2026},
  volume = {26},
  number = {8},
  pages = {2569},
  doi = {10.3390/s26082569}
}
Kostenlos starten →