Asked
— Edited
Hi,
A couple of questions. Has anyone been able to run 2 cameras at one time?
I would like to use one camera fro EZ Face and the other for whatever...
Has anyone been able to come up with a way to recognize sound... maybe an alarm or breaking glass?
I have a glass breakage sensor I may use but I would like to be able have the robot recognize sounds also.
I guess EZ face still can only recognize faces which is working great and is great by itself, but how about objects?
Thanks for any input!
Upgrade to ARC Pro
Experience early access to the latest features and updates. You'll have everything that is needed to unleash your robot's potential.
Related Content
Synthiam
Based on your post activity, we found some content that may be interesting to you. Explore these other
tutorials and community conversations.
Sound sensing/ sound level is something I am currently working on. I don't know how great a sound sensor via a PC mic would be for recognizing sounds specifically. I really don't believe it would prove to be an accurate method. With sound level we could achieve the robot having an interest in the sound, such as "Hey, that's loud - is it the car alarm, house alarm, smoke detector?" Then if the robot had further logic to investigate the sound, like navigate towards the smoke alarm, car, house alarm, door bell, ect, based on the sound level increasing the robot could determine with a higher rate of accuracy the cause of the sound.
The highest accuracy for detect alarm events would be to have to the robot tied into those systems for direct signal integration, but that is obviously not always possible or practical.
You are right about accuracy in detection... it really needs a connection somehow. My crummy ADT system makes a lot of noise and calls the police but there is no way to connect any interface.
I have played with RoboRealm a bit last year and if it had some API's I might be able to play with it in a .NET app.
Have you published any API's for EZ face?
EZ robots basic interface is great for testing but if I want to move farther along I will need to write something more complex and expandable. When I get the next arm built I will upload code for the EZ robot interface.
It takes about 100 hours to print, build and assemble one these arms for InMoov. I intend to build several for different purposes.. one will be security...
Thanks for the input. You did a great job on EZ-Face.
Alan
Is it that you want your robot to be aware of humans moving around your house (other rooms to where it is actually in)?
An idea I had, and I will try out soon is battery (RF) PIR units in the rooms you want covered, the advantage here is that no new wiring or alarm system is required, you just have to stick/screw the PIR unit to a suitable location.
www.electricalcounter.co.uk/products/Security+Systems/Lynteck+Alarms/PIR+Sensors/Lynteck+LS1500X+(LY52-028-45)+Wireless+PIR/2740527446?gclid=CImS3LPR7r0CFUoJwwodBGAA9g
They have 2 years operation on Alkaline batteries, and probably 5 years operation on lithium types, so low maintenance. I have bought a couple of these, and when I get the time I will hack the RF transmission so I can decode it on a small PIC microcontroller. This could then just flag the EZ-B what areas have humans moving around in, if the script knows that it is late at night or that you are out of the house, then it can deduce a possible alarm condition. This idea is also useful in non-security as it would be useful for the robot to be aware of where humans are, as it could then move to different locations (rooms) to find them for interactions etc.
If this sounds of interest, then when I get round to doing the code, I can program for you a couple of PICs that you can use in your own system.
You would also need a simple super-regen receiver module, but they are only a couple of bucks. The licensed exempt frequency will be different for you in the States, here the LS1500x PIR it is 433Mhz.
Tony
@DJ RoboRealm does indeed have object recognition function that you can manually teach object. In my experiments I was able to teach RoboRealm to identify a telephone, hand sanitizer, water bottle and a lotion bottle.
I'm working an "EZ" style interface for RoboRealm so folks can take advantage of its visual processing.
Alan
It may be possible to do if you use a different driver which sets the camera's name as something else or if you can edit the driver to do that. It's something I started looking in to but it wasn't a necessity for me so I didn't spend too long on it.
ok thanks.. that is the problem they are both name the same. 2 eyes .. same cameras.
I just tested with 2 cameras that are the same driver and they both show up. They have the same name of course, but they're different devices. I am able to use 2 cameras with the same driver - but having to select the different one for each CAmera Control
I can't say for sure if the problem was ever fixed or not, I thought it had been but may have been mistaken since shortly after the issue I scrapped the whole 2 camera idea since it didn't add much to the project.
This was my topic, if I get chance I'll check through the release notes of updates shortly after the topic date which should point us in the right direction.