GitHub/quartz
1
0
Fork 2
mirror of https://github.com/jackyzha0/quartz.git synced 2026-03-24 15:05:42 -05:00
Code Issues Actions Packages Projects Releases Wiki Activity

vault backup: 2023-04-06 10:27:01

Browse Source
This commit is contained in:
Jet Hughes 2023-04-06 10:27:01 +12:00
parent ddb3d48edf
commit 6e7f89e30e
1 changed files with 9 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
content/notes/12-advanced-sensing-vision-sensors.md
View File

@ -18,6 +18,8 @@ Sensor Fusion - Fusion with vision sensors
> [!INFO] can you camera help you know where you are without GPS.
> could build snensor fusion system using GPS and the camera to improve accuracy. can use 3d maps to compare you camera view with generate views of 3d maps
> [!INFO] combine compasss reading with magnetometer, to stabilise magnetometer
Vision-based Localisation and Tracking
- Reprise from our earlier lecture on localisation and tracking using hardware sensors:
- GPS: not working indoors, not very accurate, but (nearly) everywhere available
@ -39,4 +41,10 @@ Computer Vision
- Human body including face, fingers, hands and feet, or the whole body
- Activities such as typing, standing up, walking, entering a doorway or shaking hands
• Reprise: • Degrees of Freedom (DoF): Number of parameters of the system that may vary independently. • E.g. a Compass has 1DoF, Accelerometers 3DoF, • 6DoF: posX, posY, posZ, rotX, rotY, rotZ • Localisation: Computing ones pose assuming no prior knowledge (usually in 6DoF, and often not in real-time) • Tracking: Computing ones pose using prior knowledge (usually in 6DoF, and often not in real-time) V
Vision-based Localisation and Tracking
- Reprise:
- Degrees of Freedom (DoF): Number of parameters of the system that may vary independently.
- E.g. a Compass has 1DoF, Accelerometers 3DoF,
- 6DoF: posX, posY, posZ, rotX, rotY, rotZ
- Localisation: Computing ones pose assuming no prior knowledge (usually in 6DoF, and often not in real-time)
- Tracking: Computing ones pose using prior knowledge (usually in 6DoF, and often not in real-time)