Feature Request: Ability to distinguish/tag motions depending on location in the image
When surveilling a larger area, the image can often be divided
into two or three major different areas of interest. A "gardenview"
could have for example the regions "patio-area" in the front left
of the picture, a "lawn-area" at mid-distance and a "dont mess
around with my beloved wooden blockhouse - area" in larger distance...
Is there a way to tag detected motions depending on the area
the motion has happened in? E.g. instead of saving *m.jpg saving
*_a1_m1.jpg, *_a2_m.jpg files. (The areas could be defined by the user
in separate files analogous to the fixed mask file.)
This way it would be possible for an automatic program to distinguish
between truly important motion ("I send a message to the owners cell
phone right now") and not so urgent, but still important motion, which
the owner just wants to be recorded.
- 23 May 2006
What you are asking for is quite a big feature and one which another open source program called Zone Manager has been designed to do as the main feature.
There is another option.
Motion is made to be a program that easily integrates with other programs and can communicate with other programs without being leaving a huge footprint (RAM and CPU time).
If you combine two features Motion has with a small script program of your own you could make something very advanced.
First look at the on_picture_save
and note how it works and note the use of conversion specifiers
Now look at the Conversion Specifiers
. And pay special attention to %f, %K, %L, %D, %i, %J and %Q.
Your own script can use this data to decide if something is important or not.
- 24 May 2006
Great! Thank you very much for the quick response!! This might actually solve the problem.
- 24 May 2006
I think that this is a great idea. The on_picture_save method is OK, but I
think a better approach is this: Using a mask file (normally just black and
white) define a mask using black and white and shades of gray, where the
darkness or lightness of the gray indicates the "importance/priority" when
determining if the pixel threshold has been met during motion detection.
A mask with black on it will mask all pixels , a mask area with 30% gray
(more white than black) would multiply the number of motion detected changed
pixels by 0.70 . THUS when the total total number of "weighted" changed
pixels is compared to the pixel threshold (to determine if real motion had
occurred) the result would only yield a positive hit if enough of the area
did have motion in it. A mask file with black areas indicates to ignore
100% of the pixels, gray areas that are almost black are low priority ,
gray areas that are light gray (almost white) are high priority areas, areas
of the mask that are 100% white get all of the motion pixels counted (highest
- 29 Jun 2006
Greg. This is exactly how Motion works. The mask file is not only black and white. Grey means lower sensitivity exactly as you describe it.
- 29 Jun 2006
Ken -- I knew it was a good idea. Silly me for not reading the doc. Sorry.
- 01 Jul 2006