ton of mp4 files to analyse

[askr]

Hello,

I have lot of (more than 1tb) mp4 footage from two cameras and I would like to use zone minder as video motion analyser only. I would like to ask you is that possible to make automated analysis :

First camera has small, 2MB files, there is huge qty of this files in one dir

Second camera has done folders:
/date/hour/file.mp4 and that files is about 2GB in size.

I can make two instances of vm with ZM to analyze this two cameras separately.

I would like to analyze all files from camera and mark only this videos where is change or/and make image snapshots if that is possible.

There is no live video streaming from camera.

Is that possible in newer zoneminder?

I know that this is only small functionality usage but I'm thinking to move this two old iptv cameras into small system on ZM based. But firstly I need to purge lot of similar recordings to restore disk space <facepalm>

Any informations would be great help

Regards

[asker]

I suppose one approach could be:

a) write a script to use ffpmeg and concat the files into larger files (so they become easier to manage)
b) Create a ZM monitor using source as "File" and point it to one of the MP4s (thats why I suggested concating the files into just a few files, or one file)
c) Set it in modect mode
d) Wait till ZM does its magic and spits out image files for motion as events

I've never tried it, just a thought.

[askr]

hello, thanks for your help,

I have tried lot of times and I cant make an "file" input to recognize mp4 I suppose:

viewtopic.php?f=36&t=26239

regards

[iconnor]

The ffmpeg input will probably read a file.

Our file input should be updated to take a list or directory of files...

In short, it's doable, but will take some work.

[askr]

Hmm, So, what can I do ?

I have to concatenate the files, so I can make one huge file with about 30* 24hrs ~720hr/file...

[askr]

Now I see whats going on... but ... great


I have that instance for only one cam (I know, nonsense) and I would like to learn as many from that and later runs.

So, I would see two possible approaches to that case:

1. Batch/API/whatever create (duplicate- I would choose specfic parameters for that monitor) multiple monitors one per day (24hr material) - is that possible?
2. Pushing the zmind to catch list of files to analyse as an input on one monitor only.

What would be simpler and more efficient?

In that case I only need to extract to .jpg files frames with motion,

regards