motion detection and alarm handling

Roland · Post by **Roland** » Wed Feb 25, 2004 5:33 pm

Hi, I like this software, but I missing some functionality.

a.) Motion detection with a area mask
b.) Alarm handling i.e. Email and the picture file
c.) Multiple webcams

Is there a possibilty to get the source code to modify.

Thanks
Roland

Post by **malun** » Thu Feb 26, 2004 1:05 pm

I like your ideas! Motion detection with an area mask is something I have to add to the "to do list". Alarm handling sounds nice too. But I think multiple cameras will have to wait for some time...

In the beginning I had thoughts about releasing Yawcam as open source software, but I decided to release it as freeware. I hope you'll respect that decision. So to answer your question: You may not have the source code.

I hope you still will be using Yawcam and hopefully after some new releases you have the functionality you miss today.

/malun

Markku · Post by **Markku** » Sat Jan 15, 2005 5:51 pm

Any news on Motion Detection?

Post by **malun** » Sat Jan 15, 2005 10:45 pm

No, nothing new on the motion detection front...
...and right now it has not the highest priority. Sorry.

/malun

Jii · Post by **Jii** » Sun Jan 16, 2005 1:43 am

a.) Motion detection with a area mask

I have seen this feature in a program bundled with Creative Webcams and I have used it for testing of the features of it.

This is how I experienced it:

The masked area that the program was monitoring for movement had a sensitivity adjustment. That adjustment was made with a standard vertical slider with a green/red/yellow bar (with the green part being the slider-adjusted sensitivity and yellow being the observed movement).

This motion detection is based on a comparison of the masked area of two or three frames taken by the webcam. If there was a certain amount of difference between those images (a percentual amount, to be exact), it was compared to the amount set by the slider and if the difference amount exceeded the slider-set percentual amount, it was interpreted as movement and an alarm was triggered.

This whole thing was a complete blunder on my webcam. My webcam is a Creative Webcam NX Pro and it's a CMOS-sensor cell structure Webcam. CMOS-sensors are cheap to manufacture and are mostly found on cheapest of webcams, but they have a real nasty tendency of generating excess noise on the picture if the webcam is subjected to a low-light condition.

I was never able to balance this feature in a way that it would have reached even at least a decent level of operation. I set my webcam software to email me a picture everytime it "saw something move". On one sunny day, I received 23 email messages from that software and all the pictures were extremely dark - with no movement.

My webcam is situated in a place where there are not too much of light. The webcam was actually so poor that it produced pitch-black images in a lighting condition where you can actually read a book without difficulty and without any artificial lighting. Noticing this I quickly realised that the problem was due to the noise generation of the CMOS sensor that desperately tried to cope with the lighting conditions. The noise in the mask area was enough to exceed the sensitivity amount and trigger the alarm. I decided to increase the mask area a bit and make the detection less sensitive to avoid the flood of false alarm email messages coming from myself.

The next day was sunny, too. This time no email messages. I was happy at the results until I came home (where the webcam was) and found out that walking (and even jumping!!) past the mask area triggered no alarm at all until I took a flashlight and pointed it straight towards the webcam.

By that time the second problem came from behind and stabbed me in the back: In the pictures that the webcam took, the person moving in the picture was "interleaved" (or in laymans term: looking like a ghost) in such a way that visual recognition of the person was next to impossible. I found no way of getting past this problem and after two weeks of adjusting, frustration and failure after another, I made the disappointing decision to uninstall the program.

I welcome this feature if it is implemented, but I wish it is far more advanced than the one mentioned above, or that it would work only on webcams with CCD sensors, for they are more capable of producing decent images even in places where a CMOS-webcam can't see anything anymore.

This also may mean that the picture quality in Yawcam is forced to be set at 100% in motion detection mode, so that compression doesn't trigger the alarm.