Critter Radar and Picture Capture...

rwgast_logicdesign · 2012-09-22 16:46

So this may seem really goofy to some people but when I was a kid I use to watch movies like independence day and sneakers, there labs/vans always had little sonar/radar screens with the green sweeping line. I always ALWAYS wanted one of these I thought it would be so cool. I had an idea to put alot of concepts togather to build a system that would detect animals in my yard and capture a picture of them then stream it along with data points for a radar screen in my room via wifi. I live in the desert so theres always animals running around.

I know exactly what to do for most of this system except the acuall radar part and being quick enough to snap pictures of animals like rabits. I figure to store pictures I would use an SD card, probably a little cheap cmos camera to take them, a microchip ENC28jblah for ethernet connected to a wireless bridge outside.

Im thinking what I would need for the acuall mechanical part would be a base, this is np. On that base maybe 4 ping sensors connected to a 180 servo and two cameras also connected to a seperate 180 servo. m thinking you would put a ping sensor close to the ground and another one above it this would be attached to a servo, then you would connect two more pings identically facing the oppisite way so when the servo does a 180 it can get 360 degrees of ping data. Then you would throw the cameras back to back on a second servo. Once a ping sensor detected something it would rotate the camera servo to the same spot and take a picture. The reason for using two of everything back to back on a 180 servo would be so the cameras or sonars never have to physically take the time to travel more than 180 degrees since critters and cyotes are quick.

I figured that I could acually maybe try to take the cheap cmos camera and conver the images to binary then try to run them against other common animals so the radar could send back jack rabbit and plot a special symbol on the radar for jack rabbits, but I have a feeling that kind of logic may be a bit over my head.

I know this all sounds really goofy but I really want a radar and i may even implement the radar screen on a propeller connected to a VGA monitor just to throw a few more concepts in the pot!

Duane Degn · 2012-09-22 16:56

This sounds like a fun project.

It sounds like you want to take still images, right?

I don't have any suggestions for still cameras but if you're interested in video cameras the ELEV-8 Camera thread has several relatively inexpensive options.

Ragtop has a bird feeder camera project that has a few things in common with what you want to do. It might be a good idea to check out his project thread.

ElectricAye · 2012-09-22 17:45

Phil's Boe-Dar?

http://www.parallax.com/tabid/345/Default.aspx

http://www.parallax.com/dl/docs/cusapps/Boe-dar.pdf

Project 10 on this: http://www.parallax.com/portals/0/Downloads/docs/prod/sic/Web-122-28100-SicMiniProjects-v1.0.pdf

http://forums.parallax.com/showthread.php?86110-Ping)))Dar-a-Radar-Style-Display

http://www.youtube.com/watch?v=t2tzBEOAtjg

varnon · 2012-09-22 21:44

That sounds like a pretty cool project.

I'm not sure how well the pings will work. The range for big solid objets is about 8 feet, but smaller objects have to be much closer to be detected. For some of my projects, pigeons become invisible at about 2 feet. They are also hard to detect at certain angles were the flight feathers reflect the sound away from the sensor.
Furry objects (mammals) also tend to be a little harder to pick up than household objects.

Since I also do a lot of animal detection I would love to hear how this goes. Keep the thread updated.

rwgast_logicdesign · 2012-09-22 22:24

well this project is gonna be slow going as i dont have alot of the materials..

i was acually just looking for input on if this could work before i cemented and plans and it seems as though a ping may be a bad choice... i thought the biggest issue would be spinning the camera to the posistion a ping sensor found an animal before the animal was gone. it seems i may not even want pings at all.

i will read through phils posts and think about this more. the goal of the project is to sweep a 30yard radius and display anything moving on a radar screen and its posistion.. maybe machine vision algorythms would be better but i dont wanna go down that path

rwgast_logicdesign · 2012-09-22 22:34

i saw this hadnt listend to the sound its pretty inspirational though this is very close to what i want the end product to be like

http://www.google.com/url?sa=t&source=web&cd=8&ved=0CFAQtwIwBw&url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DsfBOC980KYA&ei=Q51eUNy-K8nOyQGFlYDQCw&usg=AFQjCNGWW3Z5iF1vHhZkdG_-TIO5KJ4k_w&sig2=HgB0QLWQqiTWkDOCTKoHNQ

maybe it would be possible to build a sonar ping with a massive amount of range relative to the ping module, and a bigger detection area

Mark_T · 2012-09-23 04:09

rwgast_logicdesign wrote: »

maybe it would be possible to build a sonar ping with a massive amount of range relative to the ping module, and a bigger detection area

Possible but likely to keep at least some of the critters away - just because humans can't hear 40kHz doesn't mean other creatures can't. Comparing successive frames from the camera ought to produce interesting information about moving creatures.

RS_Jim · 2012-09-23 07:06

A couple of ideas to throw out. Look into Parallax's x-band motion detector. Also there are motion detector algorithms for video that might work using a fixed camera for detection and a servoed camera for pan and scan. Consider a 180* sweep as it makes for easier mechanicals. A fixed camera with ir lighting works well with video md, do the radar scan in software with detection along the scan line. Use location to servo pan flash camera and trip shutter. Send camera video direct via wifi. Prop does not need to see the video, just the servo commands and shutter controls via the network.
Jim

Duane Degn · 2012-09-23 09:15

There are some machine vision options with the Propeller. I've listed some of the ones I've seen in post #4 of my index.

If I were to try something like this, I'd probably use Hanno's method since I have some experience with it and I already have the hardware. For someone without experience with any of the machine vision applications, I'd suggest a TSL1401 Linescan Imaging sensor. I think Phil has several threads about using this sensor. Here's a link to one of them.

I'd think a single dimension imager would be enough to detect motion.

rwgast_logicdesign · 2012-09-24 08:58

so with a bit of research iv figured out what im trying to build is fairly out of sonars capability, ive seen the x10 and i think its pretty neat but theres no way to use it to get a location just detect motion.

so im looking for whats called continous wave doppler radar. mit has a whole open course on a diy radar from coffee cans and a microwave over.. theres is so advancex you can get SAR images from it. Im going to look into this but i have a feeling even if i can build it my math skills will limit me from truly understanding it.

Critter Radar and Picture Capture...

Comments