Published June 23, 2016 © GPL3+

IoT Tweeting Plants with Raspberry Pi

In this project, you'll build a device which tweets pictures and conditions of an IKEA VÄXER hydroponic garden for remote monitoring.

BeginnerProtip5 hours4,485

Things used in this project

Hardware components

Raspberry Pi Zero

Any other model could be used, if you like.

USB hub with Ethernet and Micro USB connector

Use one with a full-size USB if you're not using a Pi Zero. You could also use a normal USB hub and a WiFi adapter, if you'd like.

Pimoroni Enviro pHAT

USB Webcam - most will work

See the above link for compatible cameras.

Adafruit SD Card with Raspbian

You can also flash your own.

IKEA VÄXER Hydroponics Kit

This is what I mounted mine to. If you want to monitor something else, then you don't need one of these.

Adafruit 2x20 GPIO header

Software apps and online services

Raspberry Pi Raspbian

Tweepy

Twitter

Hand tools and fabrication machines

Monitor

Keyboard and Mouse

Pi Zero Adapters

Soldering iron (generic)

Story

Intro

This project is based of Pimoroni's project, Pimoroni Plants. It monitors the conditions in a garden (I'm using an IKEA VÄXER, because it's cheap and hydroponic), and tweets them along with a picture. However, their project uses their Flotilla modules (not available to the general public at the time of writing, and kinda over-complex), and a Raspberry Pi camera module (much more expensive than that old USB webcam we all have laying around). They've just released the Enviro pHAT, so I thought that it would be perfect to set up my own version. You can see it tweeting right now here.

You don't really need any experience with electronics, Raspberry Pi, Python or anything else to do this project. There's no hardcore wiring (just a spot of light PCB soldering), and it's easy to copy and paste all the code, if you so wish.

How-to

This project is pretty easy. First, you'll need to solder a 2x20 header into your Pi Zero (not required if you're using a different Pi), and Enviro pHAT. There are already a great many tutorials available on how to do this. A good one is available here. Next, you'll need to set up your Pi Zero with a keyboard, mouse, monitor, and network connection. You will need a mini HDMI to HDMI/VGA (whatever your monitor uses) adapter, and the USB hub. You'll also need an Ethernet cable, or a WiFi dongle, your SD card, and of course a Micro USB power adapter. Plug everything in, and power up.

Beware! The power must go into the right most micro USB port on the Pi Zero, and the USB hub into the left most. Otherwise, things will fry!

It's also possible to use a different model of Raspberry Pi for this, and transfer your SD card later on. Once you have booted into the Raspbian desktop, then open up a terminal window by clicking the little icon:

Click this to open the terminal window!

Then, you'll need to install all the necessary software. If you're using WiFi, you should set this up now by following this tutorial. Next, you'll need to enter a few commands to install the software.

Install the fswebcam package, which we'll use to take pictures with our USB webcam:

sudo apt-get install fswebcam

You can test this to make sure it's working properly, by asking it to take a test picture. Just plug in your webcam, and run the command:

fswebcam your_image_name_here.jpg

Then, open the file manager and double click the image you just made. You should see a picture of whatever the webcam was pointing at!

Now we need to install the software needed for the Enviro pHAT. Luckily, the nice folks from Pimoroni made a handy one-liner for this. Just enter:

curl -sS get.pimoroni.com/envirophat | bash

(That little line is a vertical bar by the way, not a capital I or lower case l).

You can run some tests on this to see if it's working too. These are described in this excellent tutorial. We need to install the tweepy module, which allows our Python program to talk to Twitter. Just run:

sudo pip install tweepy

That's all the software installs done. You'll need to make a copy of the code below to your Raspberry Pi. I put it on a memory stick and transferred it. Put this code on the desktop, and make doubly-sure it's called 'plants.py'.

You will also need to set up your twitter account at this point. Create a new account, and verify it with your phone (or else you can't have API access, which we need). Then, go to and create a new app. Note down the Consumer Key, Consumer Secret, Access Token, and Access Token Secret.

Double-click the plants.py file on the desktop of your Raspberry Pi, and click 'edit file' in the dialogue box which comes up. Then, enter your data into the spaces at the top of the program where it says to.

You can test to see if this is working, by running this command:

sudo python /Desktop/plants.py

It should output some readings, and tweet a picture at the next quarter-hour interval (xx:00, xx:15, xx:30 or xx:45). Once that's all working, then you need to set up an auto-run script. This is so that you don't have to manually click the file yourself, and it'll run automatically when the Pi is turned on. Enter:

cd Desktop
nano launcher.sh

This should get you a blank text. Add this to the script:

#!/bin/sh
# launcher.sh
null
null
cd /
cd home/pi/bbt
sudo python bbt.py
cd /

This code is also available in the 'code' section of this project. Then hit Ctrl-X, Y and Enter to save it. Make the script an executable:

chmod 755 launcher.sh

And test it with this command:

sh launcher.sh

Then add a error log in case something goes wrong:

cd
mkdir logs

We now need to create the actual autorun script. Enter the command:

sudo crontab -e

You might need to tell it which text editor to use - if it asks, choose Nano. You should now have a file filled with comments (starting with a #). After these lines, add this one:

@reboot sh /home/pi/Desktop/launcher.sh >/home/pi/logs/cronlog 2>&1

Reboot the Pi to check it's all working. If it is, you should see the Enviro pHAT's LEDs (one either side of the light/colour sensor) flash on and off during the boot sequence.

You're done! Now, go and attach your Raspberry Pi to your IKEA VÄXER, or whatever you want to monitor with it. You could make a remote monitoring for your chickens' coop, dog kennel, kitchen or anything else. Just remember that as the images are on Twitter, they're publicly accessible so don't set it up to tweet anything you're not happy for anyone to see (like your bedroom, or anything to do with security).

I attached my Pi Zero with an awesome Pibow Zero case and a sticky foam pad, and mounted the webcam near the plants with tape.

Going further

There's no limit to how far you can go with this. Check out Sandy J Macdonald's tutorial on the Enviro pHAT to see what other data you could extract. All the code is commented, so you should be able to find your way around it easily enough. It's by no means pretty - just what I came up with. You could put the data into a spreadsheet, or attach some extra hardware (LEDs or a screen) to show it visually. You can attach four more sensors to the Enviro pHAT's analogue input pins to get more data, or use a USB sensor (there are two spare ports on that hub). Personally, I will be working on adding some more intelligent reactions (like the "it's hot in here" one, but for other data), and randomly selecting them. I'd also like to have a flask page on my network to broadcast the code's status to me. Have fun!

#imports the time module (used for waiting)
import time
#imports the datetime module (used to get the current mintue)
from datetime import datetime
#imports all parts of the envirophat module
from envirophat import *
#imports the OS module, which we use for executing the command to take the picture
import os

#blinks leds to show working
leds.on()
time.sleep(1)
leds.off()

def SetupTweepy():
    #this imports the tweepy module (twitter for python)
    import tweepy
    #this sets up all the variables tweepy needs to use
    consumer_key = 'yours_goes_here'
    consumer_secret = 'yours_goes_here'
    access_token = 'yours_goes_here'
    access_token_secret = 'yours_goes_here'
    #this runs the authentication process
    auth = tweepy.OAuthHandler(consumer_key, consumer_secret)
    auth.set_access_token(access_token, access_token_secret)
    #this allows us to tweet from within a different function
    global api
    api = tweepy.API(auth)
    #lets us know it's all running well
    print("Tweepy Setup Done!")

def TweetText(texttotweet, medialoc):
    #tells us if we've asked it to tweet something to big, to avoid tweepy errors
    if len(texttotweet) > 144:
        print("text too long for a tweet!")
        return False
    else:
        #this actually tweets the text
        api.update_with_media(medialoc, status=texttotweet)
        #tells us what it's tweeted
        print" \" %s \" tweeted!" % (texttotweet)
        return True

def TweetConditions(addlmessage):
    #sets up an empty variable for the text to tweet to be stored in
    tweet = str()
    #adds the additional text (like motion detected)
    tweet += addlmessage
    #adds the light level to the tweet
    tweet += ("Light level: " + str(light.light()))
    #adds the temperature
    tweet += (", Temperature: " + str(round(weather.temperature(), 1)) + " C")
    #adds the pressure
    tweet += (", Pressure: " + str(round(weather.pressure(), 0)) + " hPa")
    '''If you want to add additional analogue sensors connected
    to the analogue header on the Enviro pHAT, then add this line:
    tweet += ("<<your value name here>>: " + your_analog_value_here)
    You will need to map your value from between 0 and 3.3v to a
    value that people will understand (like degrees C for example)
    more info on this in Pimoroni's tutorial, at learn.pimoroni.com'''

    #executes a command line command to take a picture
    os.system("cd Desktop\nsudo fswebcam image.jpg")
    #sleep for half a min to give it time to do all its picturey stuff
    time.sleep(30)

    #sets the media variable to the location of the media
    media = "/home/pi/Desktop/image.jpg"    
    #sends this to the function that sends and checks the tweet
    TweetText(tweet, media)

#sets all needed variables to initial values:
#the threshold for detecting whether something was moved (change this if it seems
#to be moved all the time)
threshold = 0.3
#an empty list to store acceleroneter values in
readings = []
#defaults the last Z reading to 0
last_z = 0
#sets up minutes as a global so it can be accessed anywhere
global mins
#sets minutes to the current mintue
mins = (datetime.now().strftime('%M'))
#defaults motion detected to false
motiondetected = False
#splits the RGB values into 3 separate variables
r, g, b = light.rgb()

#runs the tweepy setup code
SetupTweepy()

#do the following forever
while True:
    #if it's at one of the 15 minute intervals:
    if mins == "00" or mins == "15" or mins == "30" or mins == "45":
        print("Reached an interval:, ", mins)
        #if it has been moved, add that to the tweet and send it
        if motiondetected:
            TweetConditions("Ow! Someone moved me! ")
        #if the temp is above 32C, then add that to the tweet and send it
        elif (round(weather.temperature(), 1)) > 32:
            TweetConditions("Woah - it's hot in here! ")
        #if the blue light is greater than all others, add that to the tweet and send it
        elif b > r and b > g:
            TweetConditions("There seems to be a lot of blue light at the moment. ")
        #if the red light is greater than all others, add that to the tweet and send it
        elif r > b and r > g:
            TweetConditions("There seems to be a lot of red light at the moment. ")
        #if the green light is greater than all others, add that to the tweet and send it
        elif g > r and g > b:
            TweetConditions("There seems to be a lot of green light at the moment. ")
        #if nothing notable has happened, just send a plain old tweet
        else:
            TweetConditions("")
        #wait a minute to avoid more than one tweet going out
        time.sleep(61)
        #update the minutes again to avoid a neverending loop
        mins = "1"
        mins = (datetime.now().strftime('%M'))
        #reset motiondetected back to false
        motiondetected = False
    #if it's not time to tweet:
    else:
        #this next part is adapted from Pimoroni's example. It detects if the plants were disturbed.
        readings.append(motion.accelerometer().z)
        readings = readings[-4:]
        z = sum(readings) / len(readings)
        if last_z > 0 and abs(z-last_z) > threshold:
            motiondetected = True
        last_z = z
        time.sleep(0.01)
        #updates the RGB light values
        r, g, b = light.rgb()
        #sets minutes to the current mintue
        mins = (datetime.now().strftime('%M'))