Continued Work Towards a "Virtual Gimbal."

One of the many projects I am currently working on is a, "virtual gimbal." The idea is that gyroscopic sensor data can be used in conjunction with post processing software to steady recorded video after the fact. This is different from existing software which attempts to determine motion based on reference points in the video, instead of direct data recorded during the flight. The reason for using a virtual gimbal as opposed to adding an actual gimbal to a UAS is to allow smaller systems, in the mini to micro range, to give relatively decent quality video without the weight expense of a gimbal system.

To date, my work has only considered the option of video being shot pointing in one direction, generally forwards, at a time. However, after seeing, and purchasing, the above shown camera at a toy store during Christmas shopping, I have decided I may have an additional feature to add. The Panosphere 360 camera uses a fisheye lens to record a single panoramic video, and then uses software to allow users to, "explore," this video afterwards. The results of this exploration can also be recorded for later playback as one continuous shot. As an example, I did a fly-by of a table in my house, with the small hexacopter I was using pointed in the same direction the entire time. As I played the video back (and recorded at the same time) I focused on candles on the table, and panned the shot around to follow them. The result looked as if a gimbal mounted camera had been turned during the flyby to keep focus on the candles. Granted, as this is just a toy, the quality was not great, and the low-light conditions did not help the situation. If anyone is interested in that video, I can post it up early next week after we return home from visiting relatives for Thanksgiving.

In my review above, I show my first basic attempt at shooting outdoors in better lighting. The camera was attached to a small hexacopter I recently bought, and briefly flown about twenty feet above my front yard. While this camera is just a toy, it was worth the $40 for the inspiration it has given me. I would like to see what might be possible with a better sensor, higher quality lens, and less buggy software.

The first iteration of my virtual gimbal software is being developed on the Android, as I am more familiar with coding on that platform, and the camera, sensors, and processing capability are all contained in one package. I will likely purchase a fisheye lens for my phone soon, and attempt to integrate this type of virtual panning feature. At this time, my software is in extreme alpha stage, but when I get something which works somewhat reasonable, and I develop a more intuitive interface, I will likely post a link for others to download and test.

E-mail me when people leave their comments –

You need to be a member of diydrones to add comments!

Join diydrones

Comments

  • Do you have any newer finfints on This subjekt? Thanks!
  • That sounds promising.  Give it a shot.

  • Ok, a little more research and I may have come up with a different approach, using a panoramic shot. It seems a panoramic shot, or video can be imported to Blender, and then a virtual camera can be added to view that scene. So, I am now thinking gimbal data which would normally run a real gimbal could be recorded and imported into Blender. Then the view from the virtual cameral/vitrual gimbal in Blender could be recorded. This would mean I would not have to re-invent the
    wheel. Just modify it slightly.
  • I am aware a virtual gimbal would never be as good as a gimbaled system, but my hope is for something that is good enough for many applications. I feel, from a sheer safety standpoint, small is going to be big.
  • I like the premise of weight saving.  That's a significant concern with multirotors and anything that "adds lightness" boosts flight time.  I hope you keep at it.

This reply was deleted.