SOLVED: How to Stabilize Time Lapse Images in THETA Z1 Interval Shooting

I am using THETA Z1 for timelapse at the construction site.
Currently, I am making a timelapse by taking single shooting regularly,
but image matching between before and after is not accurate, so it shakes very much.

So I am trying to use the interval shooting mode in another way, but after booting the camera back to the normal mode, I cannot use it.

The camera is only taken in the morning and afternoon, and the camera is turned off in the evening.
If you have a good idea, please.

API and plug-in development are possible.

For timelapse shooting, what did you set _topBottomCorrection to? Is it disabled?

https://api.ricoh/docs/theta-web-api-v2.1/options/_top_bottom_correction/

Hi there!
Timelapse shooting gives a video or a bunch of images? I need to record a timelapse video using SDK or API (controlling from a Raspberry Pi). Haven’t found yet correct command.

The original poster is asking about taking individual pictures and making a timelapse experience with the images.

What are your requirements for timelapse video?

What is the fps you need in the video? For example, are you trying to take 1 frame a minute?

Hi!

We need 2 fps in our video. It’ll be triggered from a raspberry in pre-defined areas to cover.

Can you confirm:

  • you are unable to use the standard 4K video and extract the required frames (either due to complexity in workflow or the resolution is not high enough)
  • your deployment is a low-volume use of cameras

If you are prototyping a high-volume commercial deployment that could involve many Z1s, please send a DM or email to @jcasman and me.

Yes, it’s a low-volume.

Our last option is code a script to extract the required frames, but i’d like to avoid that if camera can
set timelapse mode. it’ll save space in memory for walks in long areas.

  • if you take a stitched image, the Z1 can achieve 4 second intervals
  • if you use a plug-in, you can take unstitched images and reduce this to 2 seconds (1 frame every 2 seconds). You may be able to get slightly lower

However, if you walking around or have the camera on a robot, then the person or robot would still have to stop.

the two reasonable options you have to work with are:

  1. have the person or robot stop and take shots every 4 seconds (with the webapi)
    or
  2. take a normal 4K video and extract the frames prior to processing or sending to the cloud

The third option involves stitching dual-fisheye images for use by a human. If you need the frames for an AI to process, you could use a plug-in to output unstitched JPEG images and get under 4 second intervals. You may be able to achieve 1 second delays. Back in June 2019, I posted related information.

We know that the Google Streetview app used to take video with parameters that are not in the public API. Video Mode publishing Street View improved to 5.4K & 5FPS

The information on how to achieve that is not available. However, I imagine that with a high-volume sale potential a door for discussion might open.

1 Like

Thank you for your answer, I will try.
The current video status can be viewed at the address below.

Thank you. craig
We solved problem.
In the video, you can see that it has stabilized after 16:30.

Set THETA USB API 0xD825 TopBottomCorrection 0x04(Does not perform top/bottom correction).

1 Like

Thank you for posting your solution. It is great news. I am going to change the title of this topic to make it easier for people to find.

Good to know that you’re using the USB API.

Top_bottom_correction

If you want to mount the camera sideways or upside down at the construction site, you could also try value 0x01, which applies the same Top/bottom correction for all pictures.

Top/bottom correction is performed. However, for Interval shooting, the parameters used for top/bottom correction for the first image are saved and used for the 2nd and subsequent images.

Fantastic news that you were able to solve the problem.

Thank you. I understand exactly. ^^

1 Like