gunshot sound detecting

[Reloadron]

I do not understand why some people in your country shoot rare and wild animals in a protected area.

Would people hunt Moose in Canada if the Moose could shoot back?

Ron

 

[Reloadron]

hi,

I want to start project about gunshot sound detecting, i search for it with google and read a lot of article.
but i don't know how can i start it.i mean at first which type of microphones i need and then what is the circuit for this high frequency sound to detect.
i can process sound in Matlab but i don't know how can i do this in micro-controller such as Arm and Avr.
i want to hear sound with microphones and then send data to micro-controller and then the micro analys the sound at last if the sound is gunshot sound , micro produce sound to alert.

if anyone have any article or anything to guide me in this project please give me.
and i hear your advice and of course if do that.

with regard
AAAK

You may want to start with the microphone being directional or omni-directional? Do you want to just for example analyze a single gunshot and then another or using multiple microphones isolate a gunshot location and the type of gunshot such as rifle, shotgun, handgun?

There are systems out there for detecting gun shots but you need to decide what you want to do with the data.

Ron

 

[audioguru]

Moose hunting is done in Canada because there are millions of Mooses all over the place, they are not rare and are not in a protected area. I guess hunter guys compete with each other to shoot the largest heaviest moose. Maybe some people eat or sell the meat.

 

[Reloadron]

Moose taste good. Still wonder if they would be hunted if they could shoot back? Elk is also pretty yummy. Actually I oppose hunting the exotic animals like lion and tiger. Drawing an animal in (baiting) like a lion or tiger just plain sucks. They have decimated herds of elephants just for the damn ivory.

Back on topic, Chicago did some experiments with gun shot detection which makes sense because Chicago is like a war zone. Apparently the initial testing was expensive.

Ron

 

[nsaspook]

Reliable gunshot detection looks to be much easier than location.
https://www.ece.cmu.edu/~ee551/projects/F09/Group9FinalReport.pdf

https://www.freesound.org/search/?q=gun

 

[AAAK]

So many variables! The report of a gun is similar to that of a thunder clap, shunting rail cars, two books slapped together and so on. How large is the area to be protected? What kind of terrain? Open land or wooded? How far are the monitoring stations apart from each other? Is there a common grazing or watering area? Is there a large body of water? Is there a manned (24/7) monitoring station? What is the budget?
I could go on and on....E

it's large area, about 10 km * 10 km.at first in empty area we test it and then in jungle if it work.
and i found a lot of article in internet about characterization of gun sound and i know it's properties. and i know it's sound can detect 2 km far away from the sound location in quiet area.
and don't worry about the budget, i want to create the example model of this for proposing it to them and then we talk about budget.

 

[AAAK]

Reliable gunshot detection looks to be much easier than location.
https://www.ece.cmu.edu/~ee551/projects/F09/Group9FinalReport.pdf

https://www.freesound.org/search/?q=gun

ohhh, thank you.
your links are very helpful.

 

[PatM]

Milwaukee police use this system.
https://www.shotspotter.com/

 

[AAAK]

Milwaukee police use this system.
https://www.shotspotter.com/

shotspotter wasn't helpful. :\\

 

[AAAK]

You may want to start with the microphone being directional or omni-directional? Do you want to just for example analyze a single gunshot and then another or using multiple microphones isolate a gunshot location and the type of gunshot such as rifle, shotgun, handgun?

There are systems out there for detecting gun shots but you need to decide what you want to do with the data.

Ron

Aha.good questions.

I split it to 2 part. first recognizing the gunshot sound and second the location of sound.
at first i want to recognizing the gunshot sound by micro-controller, the second part is for the future if the first part done.

 

[cowboybob]

Thus far it looks like you're getting some pretty good advice.

I might point out that the Arduino/msgeq7 system I suggested is capable of handling stereo inputs.

If the Arduino had two of the msgeq7 units, four microphones could be placed facing four different quadrants and if configured correctly, they might also be able to assess the general direction of the gunshot location. Of course, each microphone would have to be appropriately aimed and "coned" to only "listen" to one, each, of the quadrants.

Sound amplitudes should allow for this differentiation.

 

[Reloadron]

My best guess with gunshots and I have heard plenty of gunshots is different types of guns are unique. They all go "bang" but not the same bang. For example there is a distinct difference between the sound of a rifle and a shotgun when they discharge. I have never bothered to record or analyze the actual sounds but they are different and then we get into handguns. Might upset my neighbors if I tried recording some gunshots.

Ron

 

[crutschow]

...................
I might point out that the Arduino/msgeq7 system I suggested is capable of handling stereo inputs.

If the Arduino had two of the msgeq7 units, four microphones could be placed facing four different quadrants and if configured correctly, they might also be able to assess the general direction of the gunshot location. Of course, each microphone would have to be appropriately aimed and "coned" to only "listen" to one, each, of the quadrants.

Sound amplitudes should allow for this differentiation.

You can, of course, also triangulate by the difference in time each microphone detects the gunshot sound.
Since sound travels only about 1 foot per millisecond, you should be able to triangulate with reasonable precision using normal speed microprocessors.

 

[nsaspook]

Aha.good questions.

I split it to 2 part. first recognizing the gunshot sound and second the location of sound.
at first i want to recognizing the gunshot sound by micro-controller, the second part is for the future if the first part done.

A wise decision. The shot location technique used in the paper (they say near and far) is a form of Multilateration (direction) with Trilateration (distance) so that's why they need at least 4 mics and accurate clocks for a basic system.
https://www.isis.vanderbilt.edu/sites/default/files/49.pdf

 

[RadioRon]

hi,

I want to start project about gunshot sound detecting, i search for it with google and read a lot of article.
but i don't know how can i start it.i mean at first which type of microphones i need and then what is the circuit for this high frequency sound to detect.
i can process sound in Matlab but i don't know how can i do this in micro-controller such as Arm and Avr.
i want to hear sound with microphones and then send data to micro-controller and then the micro analys the sound at last if the sound is gunshot sound , micro produce sound to alert.

if anyone have any article or anything to guide me in this project please give me.
and i hear your advice and of course if do that.

with regard
AAAK

This project will be mostly about DSP code. So perhaps one way to get started quickly is to use a general purpose PC with a good sound card with extended frequency range capability (like up to 50 KHz?).

Your block diagram would include one or more general purpose microphones feeding amplifiers and A/D converters to allow the audio to be digitized with good bandwidth. Once digitized, you can choose any computing platform that supports the algorithm that best helps detect the sound. All of your detection, filtering and other signal processing should be done in DSP, not in hardware. As a result, your hardware can be any general purpose audio A to D front end.