Maswali ya Visual Odometry

Majibu na maelezo huonekana unapojibu maswali.

1. Je, odometry ya kuona inakadiria nini kimsingi?
   • Mwendo wa nyongeza wa kamera (mzunguko na tafsiri) baada ya muda
   • Joto la rangi ya eneo
   • Idadi ya vitu vinavyoonekana
   • Mtengenezaji wa kamera
2. 'Drift' ni nini katika odometry ya kuona?
   • Kamera inateleza kwenye uso
   • Upotezaji wa ghafla wa data yote ya picha
   • Mkusanyiko wa makosa madogo kwa kila fremu ambayo hukua juu ya trajectory
   • Mzunguko wa makusudi wa lensi
3. Why is absolute scale ambiguous in ordinary monocular visual odometry without a metric reference?
   • A color sensor cannot detect image features
   • The image geometry can support differently scaled scenes and translations that project alike
   • Every monocular camera records absolute depth directly
   • A single-camera sequence has no rotation information at all
4. Je, odometry ya kuona-inertial inaboreshaje uimara?
   • Kwa kuongeza kamera zaidi za aina moja pekee
   • Kwa kupunguza kasi ya fremu
   • Kwa kubadilisha picha kuwa maandishi
   • Kwa kuunganisha data ya kamera na vipimo vya IMU (accelerometer/gyroscope).
5. Je, odometry ya kuona wazi inatofautianaje na SLAM kamili ya kuona?
   • VO inaunda ramani ya kimataifa na kufungwa kwa kitanzi; SLAM hufuatilia mwendo wa ndani pekee
   • Wao ni kitu kimoja kabisa
   • VO inaangazia mwendo wa ndani wa fremu hadi fremu, huku SLAM pia huunda na kudumisha ramani ya kimataifa kwa kufungwa kwa kitanzi.
   • SLAM haitumii kamera hata kidogo
6. Njia za moja kwa moja za odometry ya kuona hupunguza nini, badala ya kulinganisha vipengele vya wazi?
   • Hitilafu ya Photometric (pixel intensity) kati ya fremu
   • Idadi ya pikseli kwenye picha
   • Matumizi ya betri ya kamera
   • Maandishi yanaweka lebo kwenye eneo la tukio
7. What physical quantity gives a calibrated stereo visual-odometry system a metric reference?
   • The number of letters in the camera brand
   • The chosen image file extension
   • The scene's average pixel brightness alone
   • The known separation, or baseline, between the two cameras
8. Why can a moving vehicle in view interfere with a static-scene visual-odometry estimate?
   • Every tracked pixel must belong to a fixed background point
   • Its image motion includes object movement that is not explained by camera movement alone
   • Moving objects reveal the camera's manufacturer automatically
   • Object motion changes only the file's compressed byte count