Laser Vibrometer - Is a scientific instrument that is used to make non-contact vibrations measurements of a surface using the Doppler effect.

Mirror Galvanometer - Is an electromechanical instrument that reacts to electrical current by deflecting a light beam with a mirror.

Material Used

Laser Vibrometer - Information Not Available at the moment.

Kinect - Camera produced by Microsoft, it will be used to measure depth.

Galvanometer-Information Not Available at the moment.

Arduino - Information Not Available at the moment.


Job 1 - Detect part and create a mesh on a 2D surface. 

  • First it would be necessary to detect the surface's contours, for that, it's suggested the use of the tool treshhold on a gray image using matlab. This would differentiate the intended surface from the background. Then the command regionprops would detect the properties of all detected objects. By selecting the biggest area detected, the noise background would be removed, and just the object with the biggest area would be available.
  •  After eliminating the noise, it would be necessary to create a mesh on the part's surface.
  • The time expected to do this job is one week and 3 days.

Job 2 - Treatment of the 3D data provided by the Kinect.

  • This task is in my opinion the hardest to perform. The 3D data form the infrared camera of the Kinect would need be treated.
  • After knowing the 3D coordinates, the mesh created on Job 1 could be refined so the distance between points would be the same along the surface.
  • Since this task is new for the me it is expected to last one month.

Job 3 - Control the mirrors by using an arduino.

  • The control of an arduino is also new for me. It would be necessary to translate the intended coordinates into analogue current.
  • The current generated would be sent to an galvanometer that would deflect the laser from the vibrometer, in order to perform the an visual effect of the mesh from Job 2.

Job 4 - Calibrate the mesh

  • The final task would be to calibrate the mesh. It is expected that, the first time the mesh would the sent to the part's surface, that it wouldn't be on the desired coordinates.
  • An idea to try to fix this is that by knowing were the mesh should be and knowing were it is, that the distance between the points could be known, so it would be necessary to get a photo of the part after the mesh is "printed".
  • This task is expected to require 3 days.