 |
|
 |
|||||||||||||||||
| ||||||||||||||||||||
Why do it?
In early 2005 we put up a nesting box, home made with a small hole. It wasn't used that year, maybe because it was up too late or too new. We saw blue tits visiting from early spring in 2006, then bringing some grass and other nest material. It went quiet until we saw the parents making many trips a day feeding the youngsters. They fledged successfully and the box was cleared out in the autumn but we missed a lot of the day to day detail. If they used the box in 2007, something had to be done. You see broadcasts on live TV from nestboxes, how hard can it be?
What to use?
The idea is that we want to see in the box without disturbing the birds, so
we start by checking what's available. Commercial nest boxes with cameras
are available at around £150. Web sites suggest 'board cameras' with TV
feeds and supplementary infra red lighting, the price is lower at around
£50 but do we want a TV feed? We're not going to watch it all day. It
would be good to record a few pictures and put them on a web site to view
when away from the house. A TV capture board for the PC? And editing
software? We can do this much more easily. How about a webcam? Cheap and
easy to set up because the nest box is on the kitchen wall within USB
range of the kitchen window. Effectively the only thing I had to buy, that
is the hardware I didn't already have or recycle was the webcam and
a USB extension lead.
Webcam
After emptying the box in late 2006 I was looking for a webcam. In the review pages of a magazine I saw the MSI Starcam 370i. It was relatively cheap at around £12 and has built in 'white' light leds, infra red (IR) leds and a microphone. Mounted on a stalk, it could also be fairly adjustable and the spec says it can focus down to 3cm. It claimed Windows 98SE support, which allows use of an old laptop. The software also promises 'motion detection', It is aimed at security camera applications but might be useful. So I bought one to try.
What was good was the low light ability. Some forum entries say that the colour works best with energy saving lights indoors. That's part of being IR sensitive; daylight or tungsten bulbs throw a lot of IR which drowns out the colour. The focus was also pretty easy to adjust by trial and error screwing the lens mount in and out. The flexible stalk didn't look secure enough to be left without attention for months and the stand was too big to do anything with so off it came. The 'white' leds lit up as soon as a PC was attached, so to avoid surprises I disabled them by cutting the track on the circuit board.
In practice the IR leds are bright enough to act as an intruder camera
with objects 3-4m away so relatively very bright when the bird is 10cm
away. I tried turning the IR on in daytime at first, leaving it on
overnight. Humans (and birds) should not be able to see IR but the leds
have a red glow and at that intensity the birds clearly can see something.
The Blue Tit came in early in the morning and there are several seconds of
peering warily at the camera to see what has changed.
I could try dimming the IR leds, but it's not under software
control so thats a soldering job to experiment with. In the meantime
we're making use of the low light capability of the webcam and using only
natural light.
Mounting
The commercial boxes I have seen have a camera mounting in a false ceiling. The box was home made from timber I had left over from another job; sizing was originally to suit the timber but relatively generous compared with basic models in the pet shop. The basic design is similar to the BTO guide. The existing box has enough space to insert a ceiling partition with a mounted webcam on so that's the way to go. Inside the existing box I nailed a small batten around the top above the hole and made two plywood ceilings, one as a blank I could leave all winter, one with a webcam mounted on it for a simple spring 'swap'
The packaging in the 'ball' case looked to give some protection so I left that on rather than using the bare circuit board. The easiest and most compact mounting was to drill a hole for the lens and strap the webcam to the board with cable ties. Cut a slot for the IR leds and a hole for the microphone. There is a limited amount of 'wriggle and clamp' room using the cable ties which gives some capacity to point the camera. The foot and lower stalk wouldn't come off without cutting the cable so a simple block connector was added to the board for the cable connection. A 5m USB cable was long enough to pass into the box at one end and the kitchen window at the other so no connections need to be exposed.
Since I had some more spare timber I built an 'open front' nestbox with identical camera mounting and put that on the wall on the other side of the kitchen window to see if robins might be interested. And I added some off cuts of shed felt to the roof of both boxes to make it a little more water resistant.
Initial mounting was with the cameras view as hole at the top,
back of the box at the bottom. A slight tilt was built into the ceiling to
aim toward the back of the box where the nest cup is usually built. This is
ideal for the eggs in the nest cup but the 3x4 view of the camera misses the
front of the box near the hole. When the chicks grow, flatten the nest cup and
crowd around the hole they are off the top of the picture. In 2009 I rotated
the cameras to have the hole to the left of the picture, this may
make the eggs harder to count but should help with the chicks just before fledging.
Blackbird nests appeared in cover where pyracantha and clematis montana
have been trained up the fence or the kitchen wall. Where the nest is close enough,
a simple USB extension can reach the kitchen window or the french windows.
Another web cam of the same type was quickly mounted with cable ties to a convenient branch.
The flexible mount made placing it fairly easy. The camera and plug joints
were lightly protected with a black polythene and electrical tape cover leaving just the
lens and the lead exposed before mounting. Later Blackbird nests are more distant
and have needed a 'USB extender' to reach, this carries USB over a UTP Cat5 cable
for up to 50 metres (only 20 metres here). Being outside, there are less joints
to protect than if we chained 'active' extensions and it covers a greater distance.
We're also using a powered USB hub since we're driving both the webcam and extender.
Monitoring
The webcam has a mini-CD of drivers for Win98 onwards and software for
displaying a live feed and motion detection. The supplied viewer is
limited and appears based on DirectX 9 SDK, it can save the streamed feed
as AVI but is that useful? The separate motion detection app works, but
the interface is strange and while it can save images or ftp them, it's
not that adaptable.
What I did find was Dorgem, a simple open source (GPL) freeware monitor that works on Win98. A lot of the commercial or even shareware programs expect Win2000. It is generally ideal for the purpose and generous of Frank to make it available. There are some limitations, the motion detection is not tunable and it gets confused with two identically named cameras. Until that can be sorted (in Dorgem or something else) we monitor one at a time. What Dorgem is good at is comparing snapshots at set intervals and if differences are detected it can either save them locally, ftp them, wrap it in html or run a program. Dorgem can have more than one of these 'store' actions active but the movement comparison is limited to the last image for any action. Practically then, the ftp upload is a separate scheduled script uploading the last image in a period from the local image store.
There is a motion detected capture locally at intervals of a few seconds
which can be tuned as a compromise between missing action and storage space.
These files are named with a timestamp, another Dorgem capability.
At longer intervals of a few minutes a command
script uses ftp to put the latest snapshot on this web site for the
'latest pictures' page. The advantage of a script over a 'fill in the
blanks' ftp routine is that we can customise it. While it uploads only one
image of around 15KB, the script serially renames the old versions to give
a number of previous pictures. This gives a better chance of something
being visible at any point in the day other than a few cloud shadows.
The webcam was originally plugged into an old Win98 laptop on the kitchen worktop.
Nothing fancy, it dated from around 2000/2001. 500MHz Celeron, 32MB
memory, 4GB HDD. The display was 800x600 and the battery was dead.
Inherently low power consumption, which is good if it is on all day. You
could probably find something similar for £20 or less on ebay. The only
addition was a PCCard wireless network adaptor; the card came bundled with
a router and is currently otherwise 'spare'. It saves an ethernet cable. The older laptop
is still being used as standby although a later laptop has been used when available as it is
much more responsive in use.
Website
This website is 'free' web space at 110MB, so it's probably limited in traffic capacity although I've not seen that written anywhere. If the account changes it may have to move at some point so a precaution is to redirect my 'owned' URL to wherever the space is.
First year I left the timestamp off the picture and used SSI calls to state the time, but sorting more than one nest over a longer period the timestamp helps so I have set Dorgem to add a discreet one bottom right. The 'latest pictures' page is actually static in the sense that the code doesn't change. The daily pages are a few selected images from the ones captured locally copied to a date directory. The html to frame them is generated with a simple command and the whole date directory copied to the web site along with any updates to the diary page.
Videos are captured as AVI using Amcap, supplied with the
webcam but originally a sample from the Microsoft Direct Show 9.0. These are
converted to flash movies using Riva utilities
and displayed using Flowplayer.