Hi,
I'm using VitaminD with two cameras. One of them is working fine, and is a simple setup (email, save to disk and upload to an FTP site when motion is detected anywhere in the frame).
The second camera is more problematic - I have two rules... rule 1 is to save video to the local hard drive when any motion is detected. Rule 2 is to email, save to disk and upload to FTP when motion is detected in a certain section of the frame). The problem is that whenever there is motion detected in that section of the frame, vitamin d almost always uploads 2 video files to the FTP server. These files have slightly different filenames (the filename timestamps are usually about a second apart), but both contain the same motion that was detected, so it isn't a matter of the detected motion being split into two portions.
For example, this morning my camera saw a dog walking across the frame, partly across the section that is watched. It uploaded two copies of the video, one 0:14 long and 0:15 long. Both contained the entire sequence of the dog entering the frame through exiting, as well as a few seconds afterwards each.
Any suggestions about how I can get it to only upload one copy of a video per motion detection?
Thanks,
Dave
Duplicate uploads to FTP server
Moderator: Staff
2 posts
• Page 1 of 1
-
doug - Posts: 63
- Joined: Sun Sep 12, 2010 10:24 pm
Re: Duplicate uploads to FTP server
Dave,
I've seen the problem you're describing before. When I've seen it, it has to do with the fact that there are two different types of rules in Vitamin D Video: those that are looking for an exact instant and those that are looking for something that happens over a period of time.
Examples of instant rules are ones that involve entering/exiting a region or crossing a line. If you think about it, the act of entering / crossing happens in an instant. At one moment you are on the left side of a line. The next instant you're on the right side.
Anything that's not an instant rule happens over a period of time. If you want to know if an object is inside a region, that object can be inside a region for quite a while. You can recognize a non-instant rule by the fact that you're allowed to set a duration on it.
For instant rules, we'll fire off the rule's action right at the instant that the event happens (we might wait a few seconds to put some trailing video on for context, but that's it). This can be good--there is very little delay for instant rules. However, if there is some "jitter" in the object being tracked, it may look to Vitamin D Video that the object activated the rule twice rather than once. This can usually be seen most clearly by turning on "Show Region Zones" in the Vitamin D Video application, then searching with the rule in question (the one you have setup to FTP). Turning on "Show Region Zones" will show you your rule when playing back video. It will also mark a "+" on the center of objects so you can see exactly when they cross boundaries. Hopefully, you will see the "+" cross your boundary or line twice.
If you want less duplication (though maybe a bit more delay in sending the clip), try changing away from instant rules. For non-instant rules, Vitamin D Video will try harder to make larger clips, even if that might introduce a bit of delay before sending via FTP.
I hope that helps!
-Doug
I've seen the problem you're describing before. When I've seen it, it has to do with the fact that there are two different types of rules in Vitamin D Video: those that are looking for an exact instant and those that are looking for something that happens over a period of time.
Examples of instant rules are ones that involve entering/exiting a region or crossing a line. If you think about it, the act of entering / crossing happens in an instant. At one moment you are on the left side of a line. The next instant you're on the right side.
Anything that's not an instant rule happens over a period of time. If you want to know if an object is inside a region, that object can be inside a region for quite a while. You can recognize a non-instant rule by the fact that you're allowed to set a duration on it.
For instant rules, we'll fire off the rule's action right at the instant that the event happens (we might wait a few seconds to put some trailing video on for context, but that's it). This can be good--there is very little delay for instant rules. However, if there is some "jitter" in the object being tracked, it may look to Vitamin D Video that the object activated the rule twice rather than once. This can usually be seen most clearly by turning on "Show Region Zones" in the Vitamin D Video application, then searching with the rule in question (the one you have setup to FTP). Turning on "Show Region Zones" will show you your rule when playing back video. It will also mark a "+" on the center of objects so you can see exactly when they cross boundaries. Hopefully, you will see the "+" cross your boundary or line twice.
If you want less duplication (though maybe a bit more delay in sending the clip), try changing away from instant rules. For non-instant rules, Vitamin D Video will try harder to make larger clips, even if that might introduce a bit of delay before sending via FTP.
I hope that helps!
-Doug