My goal is to create a remote webcam activated by motion detection. When powered on, it will:
- 1. Autorun a script to wait for motion, for (up to) a specified amount of time.
- 2. On motion detection, shoot video of up to a specified duration.
- 3. Remain powered on long enough for the Eye-Fi to upload the video.
- 4. Enter a low-power state for a few hours, turn itself off (possible?), or power off via cutting the external power.
- 5. After a few hours, external power is restored, and the camera powers on. (Either by keeping its power button pressed, or by using USB wakeup.)
- 6. Return to step 1.
The above cycle might repeat for weeks at a time. It must work reliably without requiring any local interaction.
One thing I don't like about this is that the power-on/off approach would unnecessarily and repeatedly retract and extend the lens. This is one obvious point of potential failure. It would also leave the camera lens open and exposed to dust and the environment. However, I don't know of another way to do this (within my limited abilities as a software developer), and from what I understand, leaving the camera powered on for long periods introduces reliability issues and is a bad idea (not to mention wastes power, however little).
Can anyone comment on any obvious showstoppers with this idea, or whether any aspect of it is ill-advised or likely to be unreliable? I have reviewed the
Fast Motion Detection script, and it seems fairly straightforward to me. However, I doubt I could modify or add to the CHDK firmware (my skill with C and assembler is way too rusty), or implement hardware hacks (using external relays or servos, etc.). I know at least a few other forum posters have tried something like this project, but I haven't seen much on how well it worked for them, nor how reliably.
I'm planning to use an SX200IS for this project. From what I can tell of the
SX200 CHDK wiki and
CHDK SX200 forum thread, it should support the needed functionality (ie. autorun CHDK, autostart scripts, video recording & quality control, USB wakeup, etc.). Am I mistaken? If I am, can you recommend a similar camera model which might work?
Also, here are some specific questions I haven't been able to find answers to yet:
- Can the lens cover be controlled via script? (I assume not.)
- Can the camera power itself off via script?
- Can motion detection be performed while recording video? (I'd love to halt recording if motion stops for a certain amount of time, and maybe even decide to delete the video file if motion was only detected for a very brief time. From the scripts I've seen, I doubt this is doable.)
- To anyone with Eye-Fi experience: is there any way to tell (from script) when the Eye-Fi has completed uploading a video? For example, is it possible to configure the Eye-Fi to delete videos when upload is complete, and have the script detect that the file is gone, then power itself down? I am aware of Eye-Fi's "Endless Mode", but not sure exactly how it works (when it deletes files).
- What is the cheapest camera still available for purchase, which would work for this project?
My sincere thanks for any helpful responses, and my apologies if I asked questions which I could have answered myself with more research. I've spent a couple of days getting to know CHDK, scripting, and this forum, and I felt it was time to ask the experts (ie. you).
I haven't yet purchased the equipment for this project, because I'm not sure about how feasible this is, and I don't have any hands-on experiece with CHDK yet.
