Published June 5, 2026 | Version v1
Journal article Open

PINHOLE KAMERA GEOMETRIYASI VA BAYES FUSIONGA ASOSLANGAN REAL VAQTDA HAYDOVCHI HOLATINI MONITORING QILISHNING GIBRID MULTIMODAL MATEMATIK MODELI

Authors/Creators

  • 1. Muhammad Al-Xorazmiy nomidagi Toshkent Axborot Texnologiyalari Univeriteti

Description

Haydovchilarning charchoqlari va diqqatni chalg'itishi natijasida yuzaga keladigan yo'l-transport hodisalari global transport xavfsizligining asosiy muammolaridan biri bo'lib qolmoqda. Mavjud haydovchilarni kuzatish tizimlari ko'pincha faqat vizual signallarga, transport vositalarining dinamikasiga yoki fiziologik signallarga tayanadigan unimodal arxitekturalar bilan cheklangan bo'lib, bu yorug'lik o'zgarishi, okklyuziyalar va haydovchilarning murakkab xatti-harakatlari kabi real sharoitlarda beqarorlikka olib keladi. Ushbu tadqiqot pinhole kamera geometriyasi, moslashuvchan vaqtinchalik modellashtirish va dinamik Bayes sinteziga asoslangan real vaqt rejimida haydovchi holatini kuzatish uchun gibrid multimodal matematik modelni taklif qiladi. Taklif etilayotgan tizim bosh holatini baholash, moslashuvchan ikki tomonlama eksponensial tekislash (ADES) asosidagi chuqurlikni barqarorlashtirish, ko'z aspekt nisbati (EAR) va og'iz aspekt nisbati (MAR) asosidagi charchoqni aniqlash va chalg'itishni aniqlash uchun moslashtirilgan YOLOv11 neyron tarmog'ini birlashtiradi. Barcha hisoblash modullari o'rtasida fazoviy muvofiqlikni ta'minlash uchun pinhole kamera modeliga asoslangan yagona geometrik kalibrlash tizimi qo'llaniladi. Geometrik, ehtimollik va neyron modullarining chiqishlari uzluksiz haydovchi holati vektorini yaratish uchun dinamik ravishda og'irlashtirilgan Bayes xulosa mexanizmi yordamida birlashtiriladi. Eksperimental tahlil shuni ko'rsatadiki, taklif qilingan arxitektura real vaqt rejimida hisoblash qobiliyatini saqlab qolgan holda turli xil yorug'lik sharoitlarida, qisman yuz okklyuziyalari va dinamik haydash senariylarida mustahkam ishlashga erishadi.

Files

V4I38.pdf

Files (787.1 kB)

Name Size Download all
md5:012f4761584574bdfc54d188749bdbef
787.1 kB Preview Download

Additional details

References

  • Brown D. C. Close-Range Camera Calibration. Photogrammetric Engineering, Vol. 37, No. 8, 1971, pp. 855–866.
  • Hartley R., Zisserman A. Multiple View Geometry in Computer Vision. 2nd edition. Cambridge: Cambridge University Press, 2004, 672 p.
  • Civik E., Yuzkat M. Real-Time Driver Fatigue Detection System with Deep Learning on Embedded Platforms. Displays, Vol. 78, 2023, Article 102441.
  • Safarov F., Kuriyama N., Soliev A., et al. Real-Time Deep Learning-Based Drowsiness Detection. Sensors, Vol. 23, No. 13, 2023, Article 5983.
  • Redmon J., Divvala S., Girshick R., Farhadi A. You Only Look Once: Unified, Real-Time Object Detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016, pp. 779–788.
  • Viola P., Jones M. Rapid Object Detection using a Boosted Cascade of Simple Features. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2001, pp. 511–518.
  • Lepetit V., Moreno-Noguer F., Fua P. EPnP: An Accurate O(n) Solution to the PnP Problem. International Journal of Computer Vision, Vol. 81, 2009, pp. 155–166.
  • Zhang Z. A Flexible New Technique for Camera Calibration. IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 22, No. 11, 2000, pp. 1330–1334.
  • Soukupová T., Čech J. Real-Time Eye Blink Detection using Facial Landmarks. Proceedings of the 21st Computer Vision Winter Workshop, Rimske Toplice, Slovenia, 2016, pp. 1–8.
  • Ultralytics. YOLO11: Fast and Accurate Vision AI Model Documentation. Ultralytics Docs, 2024.