Trail cam - newby

General project help for Adafruit customers

Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.
Locked
User avatar
trex777
 
Posts: 13
Joined: Wed Jan 18, 2017 11:43 am

Trail cam - newby

Post by trex777 »

Hey I'm trying to build a trail cam. I have the raspberry pi 3, i have a list of parts i need to get pretty much. I am wondering if anyone has tried to make one and how well did it work out. I will post on here as I get going on this project. I know I want to be able to capture day and night pics and be able to capture video both day and night, would love to be able to track schedules/patterns/frequency of the animals. I am also building a weather station and would love to be able to pull data from there and have that data captured with the pics and video. I have a food plot that this will be monitoring. As of right now WIFI will be the main source of interaction.

User avatar
trex777
 
Posts: 13
Joined: Wed Jan 18, 2017 11:43 am

Re: Trail cam - newby

Post by trex777 »

Ok, so question to anyone passing by, i am starting this project with trying to locate the best possible motion detector, so far i am running in to a very limited selection and the best distance is around 12m. trying to find one that will go closer to 30-50m if possible, the greater the distance the better.

next purchase will be the camera, has to be able to shoot day and night (will be using IR LED's), great quality and fastest trigger times. May end up using 2.

When building this unit i have asked around and found the same issue from several people, most all trail cams in the market place do not catch all the animals that pass in front of their cams. So what i am thinking is maybe having more than 1 motion sensor so the trigger times can be faster, was also thinking maybe having a second camera for the same reason and/or build in an option to be able to trigger 1 camera to take pics while the other takes video, just a thought.

User avatar
adafruit_support_mike
 
Posts: 67459
Joined: Thu Feb 11, 2010 2:51 pm

Re: Trail cam - newby

Post by adafruit_support_mike »

You're going to run into problems with the response from any sensors that operate to 50m.

All sensor detection systems have to make a tradeoff between false positives.. cases where the sensors say they've detected something, but it wasn't what you wanted.. and false negatives.. cases where the thing you wanted was there but wasn't detected by the sensors.

In general, anything you do to reduce the false positive rate will increase the false negative rate, and vice versa. A fly passing 1m from a sensor will produce as much response as a turkey 50m away. Sometimes you can reduce one set of false readings more than you increase the other, and those are changes worth making.

Finding the false positive/negative response rates for a given sensor in a given environment is a big part of the design process, and is nearly impossible to predict. I know of a roller coaster with break-beam laser sensors that worked really well in the lab and on the test bench, but would generate false readings several times a day in deployment. Eventually a mechanic was standing next to a sensor he'd just replaced when a butterfly landed on the track, flexed its wings, and broke the beam. A few months later they linked an unusually regular pattern of failures in a specific sensor to a squirrel jumping across the track at just the wrong place.

For the kind of system you want to build, it will be much easier to identify false positives than false negatives. I'd suggest you start by taking pictures every time the sensors detect anything, then looking for patterns in the false positives, and trying ways to eliminate those.

User avatar
trex777
 
Posts: 13
Joined: Wed Jan 18, 2017 11:43 am

Re: Trail cam - newby

Post by trex777 »

It can be hard to find that happy medium between everything triggering the system and nothing triggering the system. Makes me wonder if having different types of "motion" sensors would be better, like a PIR, sound, etc... Each one has pros and cons but working together may add more pro than con or at least help limit the false reading. may end up building a few of these and seeing what each one catches and misses to see what system works best and if they could work together. That way if anyone of them detect something it takes a pic, i would rather it take to many pics instead of not enough. With storage space being as cheap as it is now that is not an issue only thing would be battery life, its not cheap and even less efficient and the more bells and whistles you add the worse it gets.

User avatar
adafruit_support_mike
 
Posts: 67459
Joined: Thu Feb 11, 2010 2:51 pm

Re: Trail cam - newby

Post by adafruit_support_mike »

The more sensors you have, the more complex your decision process can be. Of course that also leads to more decisions when the sensors give ambiguous results.

Building, testing, and looking for identifiable patterns in the behavior you don't want is the best way to approach the problem.

User avatar
trex777
 
Posts: 13
Joined: Wed Jan 18, 2017 11:43 am

Re: Trail cam - newby

Post by trex777 »

I agree, so now the waiting game begins, ordering parts here and there and then build and test, will update as things move along, thanks for the advice and thoughts.

User avatar
KandRForensic
 
Posts: 5
Joined: Wed Apr 04, 2018 7:09 am

Re: Trail cam - newby

Post by KandRForensic »

What is the frequency of light most sensitive to the IR Break Beam Sensor - 3mm LEDs. I need distance so want to employ a laser diode of same frequency.

User avatar
adafruit_support_mike
 
Posts: 67459
Joined: Thu Feb 11, 2010 2:51 pm

Re: Trail cam - newby

Post by adafruit_support_mike »

The laser works at DC, and the photodetector has a response time of about 2ms. It isn't frequency-modulated like an IR transmitter.

Locked
Please be positive and constructive with your questions and comments.

Return to “General Project help”