Originally published at: 360 Camera For Construction: 80 Images in 10 Minutes - RICOH THETA Developer Community
The RICOH THETA X is an excellent 360 camera for construction. Using RICOH THETA X 8K 2fps video frame extraction, I created an 80 image simulation of a construction job site capture in a few minutes.
The images of the job site capture were created by extracted frames from a 8K 2fps video. You can see the results at https://theta360developers.github.io/oppkey-office-indoor/
Using 360 cameras for construction involves inspection of either finished or unfinished buildings. The image quality is much lower than with the 11K still image from the THETA X. As the frames are extracted from video, the image quality is also generally lower than the 5.5K images. The advantage of video frame extraction is workflow speed.
As I used 100% free software and free hosting, my workflow takes longer and is more convoluted than if you used one of the many commercial systems to create virtual tours.
To simulate construction job site capture software, I am using Marzipano, which is designed for virtual tours. I converted the 80 THETA X video frames into 15,309 files. That’s a lot of files and I’m not sure why there are so many. Marzipano does divide a single frame into many smaller image tiles. That may account for the majority of the files. As I just drag and drop the files onto the Marzipano window than download a single zipped file, the number of individual files in my tour doesn’t impact my workflow.
Download Video and Frames from the 360 Camera for Construction Demo
The following media is available for download after registration.
- 8K video, 2fps, GPS enabled, 32Mbps bitrate (set to Fine), 188.3MB file
- 80 8K frames extracted with ffmpeg
Prior to uploading the video file to Maripano.net, I converted the video file into individual frames with ffmpeg. There are 2 frames per second. A 47 second video clip walking around my office produced about 94 frames, but I cut out the beginning and ending frames where I was adjusting the camera. For this test, I used the middle 80 frames, roughly 40 seconds of video.
The video below shows the workflow for video frame extraction in detail.
ffmpeg -i R0010598.MP4 -r 2/1 frames/R0010598-%05d.jpg