I have large glass windows (30m long and 3m high) at a new house that is being built. The windows are reflecting the surrounding forest and small birds are hitting the windows and dying. I'd like to prevent this, of course.
I'm looking into ways to detect the flying birds so as to be able to trigger a strobe light or a flash to scare incoming birds. I am wondering if I might be able to use Zone Minder to detect the birds and then pass a command to external hardware to activate.
2 questions:
Can anyone comment on the likely response time for detecting objects? How quickly is Zone Minder likely to detect the motion?
How sensitive can Zone Minder be? The birds are small. What size in pixels is a practical minimum size for the birds images to ensure detection?
Any guidance would be appreciated!
Detecting flying birds
Re: Detecting flying birds
A fascinating idea!
I have several ip cameras that catch and trigger on birds flying in front of them - entrances to barns etc and usually Swallows this time of year, which are very fast. Zoneminder will catch them, but not with 100% effectiveness, and you need a high contrast background and fairly fine modect zone settings. FPS plays large part, as does the size of the bird in the frame - if it's a long way away, fewer pixels, harder to detect.
But, what I think will be the main sticky point is latency. IP cam -> ZM -> strobe would be hard to get under a second, possibly under 2 seconds, there's so much processing involved it would be hard to make it faster. An analogue camera maybe, I have no experiece of those.
I'm guessing you'll need about 1/10th or 2/10ths of a second reaction time to scare them off, and also a garden may be too wide an approach angle to work reliably.
Maybe netting, some sort of awning to restrict reflection or a less technical solution may help?
I have several ip cameras that catch and trigger on birds flying in front of them - entrances to barns etc and usually Swallows this time of year, which are very fast. Zoneminder will catch them, but not with 100% effectiveness, and you need a high contrast background and fairly fine modect zone settings. FPS plays large part, as does the size of the bird in the frame - if it's a long way away, fewer pixels, harder to detect.
But, what I think will be the main sticky point is latency. IP cam -> ZM -> strobe would be hard to get under a second, possibly under 2 seconds, there's so much processing involved it would be hard to make it faster. An analogue camera maybe, I have no experiece of those.
I'm guessing you'll need about 1/10th or 2/10ths of a second reaction time to scare them off, and also a garden may be too wide an approach angle to work reliably.
Maybe netting, some sort of awning to restrict reflection or a less technical solution may help?
Re: Detecting flying birds
Hi Flasheart.
Thanks for the response. I'm pleased to hear that capturing birds is feasible, but you confirm my fears about response time for such a solution.
The architectural idea for the large windows was to reflect the surrounding forest when viewed from the outside. There are even glass panels in frnt of walls to make this seem like a seemless panel and thus create an even reflection. So anything that reduces the reflection is a non-starter from a aestetics perspective.
So I need to design some form of invisible barrier that, once broken by a bird, triggers the equipment to scare the birds. If using motion detection will not be fast enough, other ideas that I am toying with are:
- using a narrow beam passive infrared detector facing upwards some distance away in the garden (but I fear it will not be fast enough or react to the small birds (4" or 10cm in size)).
- using ultrasound to detect the distance, but that would be very directional and birds could be flying in at any height.
Sounds like I'll have to explore these options.
Thanks for the response. I'm pleased to hear that capturing birds is feasible, but you confirm my fears about response time for such a solution.
The architectural idea for the large windows was to reflect the surrounding forest when viewed from the outside. There are even glass panels in frnt of walls to make this seem like a seemless panel and thus create an even reflection. So anything that reduces the reflection is a non-starter from a aestetics perspective.
So I need to design some form of invisible barrier that, once broken by a bird, triggers the equipment to scare the birds. If using motion detection will not be fast enough, other ideas that I am toying with are:
- using a narrow beam passive infrared detector facing upwards some distance away in the garden (but I fear it will not be fast enough or react to the small birds (4" or 10cm in size)).
- using ultrasound to detect the distance, but that would be very directional and birds could be flying in at any height.
Sounds like I'll have to explore these options.
Re: Detecting flying birds
Either of those ideas sound quicker than an image processing one. Good luck, and if it works out it'd be great to hear what did and how!
Re: Detecting flying birds
How about many of the low tech solutions that exist today? In particular CollidEscape film seems to be very effective.
There are lots of good suggestions for all sorts of issues if you google around. Kudos for trying to save our feathered friends!
There are lots of good suggestions for all sorts of issues if you google around. Kudos for trying to save our feathered friends!