Re: Free Multi-Camera Surveillance Software

Oops, looks like a real, real old post. Has anyone seen how the tracker software works with moving objects? With all due respect to the OP, I'm skeptical.

Dave spewed:

I haven't seen the product you are talking about, but I bet it will work better than you'd expect. The program that I use ( from

) has motion sensing capabilities - and the registered version actually lets you digitally zoom in on an object that is moving across the frame. Giving you a poor-mans motion tracking ability. It works pretty good, but it's incredibly annoying when used in situations not suited for it, such as the time-lapse video on my webcam page. (temp down)

But yeah, what I'm saying is... if this software package can do a digital zoom tracker with decent results, I'd bet that a dedicated tracking hardware/software system would work pretty good. :)

I'm working on a tracking recorder myself and I can track a moving object but if the camera moves all bets are off. If it moves then everything in the frame looks like it's moving, (to the motion detector and tracker)...because it is. So I'm curious if this mechanical tracking algorithm can really perform on a 1.6ghz. But maybe they have a special pattern search method that works differently. It would take a fair amount of CPU to recognize items in order to match the background image in both frames and it has to distinguish between the background and the moving object in order to understand what is moving and what isn't....all while the camera moves about. I wish I could test the hardware along with the software to see how well it works. If the company has a video I'd like to see it.

x-no-archive

BTW: Here's a couple of my test videos.

I've used open-source software called motion

records when it detects motion. With some pan/tilt cameras, it can even follow whatever is moving. I haven't tried that feature, as I only have a fixed camera.

Thanks for the link. I'm going to take a look and see how it all works.

Regards, Dave

I took a (brief) look at the Motion guide and it looks like the motion tracking-panning feature probably disables the motion detector momentarily while the camera is panning. I'm just theorizing based on reading (or skimming) the guide, but I suspect that minimum_motion_frames parameter would be used after a camera pan has reached it's new position in order to prevent a false trigger.

I saw a few features in the detector that look similar to the type of detector I'm using. I don't have a de speckle nor fully automatic smart-masking but I have something similar to the smart mask...only manually controlled. I call mine auto-masking.

Here's a screen shot of what mine currently looks like today. I'll be putting mine up for the purpose of trying to get other people to help test it eventually when I get more things working. I hope I can find people to help test it when the time comes....although I realize that the webcam software industry already has several others that are probably better than what I have.

Thanks again, Dave

x-no-archive

Oops, here's the screenshot of my unfinished webcam SW as it looks today.

Here is a video of how the motion tracker works. It tracks based on color. This was recorded using a Logitech QuickCam Pro 3000 with a TrackerPod.

Dave wrote:

Thanks for the video. I finally got back to this thread...

I think I understand how this tracking sw works now. It temporarily ignores the motion detector until the camera reaches it's new position, other wise the entire background scene would probably confuse the sw into thinking that everything is moving all over the entire frame.

I noticed that how blue blouse is the item being tracked. Is that because the blue has a higher difference in color than the yellowish background and yields a better s/n ratio in the motion detector? If the color sensitivity were more broad, I suppose that the tracker would show a box around both people, but in this case it's only around her blouse. Interesting.