IMU camera control / stabilisation

A simple, example application of the 9DOF MARG sensor array and orientation filter shown in a previous video.

When the angles of the camera pan/tilt/roll unit are controlled to be equal to the angles of the sensor, the camera points in the same direction as the sensor. An example use of this would be to have the sensor fixed to a persons head and the camera look in the same direction as the wearer.

When the sensor is fixed to base of the the pan/tilt/roll unit and the angles of the unit controlled to be equal to the negative of angles measured by the sensor, the camera direction is stabilised to always be level and pointing in the same direction. This may be useful if the camera is to be mounted on a moving platform; e.g. a helicopter.

Hardware used in video: Sparkfun 6DOF IMU Razor (ADXL335, LPR530 and LPY530) with gyroscope RC HP filters removed, Sparkfun HMC5843 breakout board (low ESR cap replacement), 3 R/C servos, x-io Board with .NET library (http://www.x-io.co.uk).

Category:

Science & Technology

Tags:

• IMU
• MARG
• AHRS
• camera
• stabilisation
• gimbal
• accelerometer
• gyroscope
• magnetometer
• Sparkfun
• x-io Board

[yngwievanhendrix]

This is amazing. I did something similar using 3d camera. One question for you. Using AHRS can i also estimate the position in 3D ???

[SebMadgwickResearch]

@yngwievanhendrix It is theoretically possible for an AHRS to track position (dead-reckoning) but often extremely difficult to implement practically. Check out some of my other videos for examples.

[revoracer523]

how can i do this with it reading just 1 axis.  what would i need? i have the arduino nano already but i dont know what to use for a gyro and accelerometer. its going to control camera tilt in 1 direction (left to right) on a motorcycle. any ideas.?

[SebMadgwickResearch]

@revoracer523 A 1 DOF version of this system would be very easy to achieve. You need to fuse the accelerometer and gyro data into single angular measurement of the horizon relative to the sensor. If the servo angle is constantly set to be equal to this angle then the camera will remain level. If you need more guidance then I suggest you ask this question in a forum.

[peter140988]

Excellent work! How did this fare with sensor drift? Did the alignment of the camera drift away from the initial position if the sensors were left alone for a while? (particularly in yaw unless you coupled the yaw gyro with a magnetometer?)

Excellent work regardless. Very applicable in the real world.

[SebMadgwickResearch]

The system uses an AHRS so there is no drift.

[Nomoreidsleft]

Nice, but you need faster servos. You can see the servo lag. The question I have, does it help in eliminating camera vibrations?

[vincentjanse]

that is cool

[tecnoshack]

good work

[obooth]

Wow, I'm keen on something similar for my Arducopter aerial photography project. Could almost be used in conjunction with target acquisition for tracking a subject. Great video man, exciting.