PICT

A new an improved version of PICT (Droissart et al., 2021), based on Bookworm OS.

For building and installation instructions see Hardware and Installation.

For field checklist see Field protocol.

For more details on how to operate the PICT, see How to.

Field protocol

This is intended to be a quick list to check order of operations while in the field.

First setup

1. Position the camera looking at the object of interest
2. On your smartphone (or tablet), start the WiFi hotspot
3. Wait (up to 3 minutes!) for the Raspberry Pi(s) to connect to your hotspot
4. Find out the IP address of the Raspberry Pi you want to connect to
5. Open your browser and type [your IP address]:8000, hit Enter
6. This will open the live preview
   • adjust the physical position of the camera to get a good shot of your object
   • adjust the settings from the menus (i.e. resolution, focus, framerate)
7. Open your SSH app and connect to the Pi
8. Check the recording schedule:
   • type crontab -l, hit Enter
   • if it needs changing, type crontab -e and use the special buttons on your SSH app to navigate the text and edit it; exit with Ctrl + X, save file ("Save modified buffer? "Y) and overwrite ("File Name to Write: ..." Enter).
9. Check the recording parameters for the camera:
   • type nano videos_v6.1.py, hit Enter
   • With the special buttons in your SSH app, scroll down to the section "### Parameter declaration";
   • Check that the parameters are the same as your selected ones in the live preview
   • If not, edit them directly here (respect the correct formatting!)
   • Exit with Ctrl + X, save file ("Save modified buffer? "Y) and overwrite ("File Name to Write: ..." Enter).
10. Start the recording manually: type ./start_video_v6.1.sh
11. Check that the recording has started:
    • type ls -lh record/videos, hit Enter
    • look at the size of the most recent file at the bottom
    • Hit ⬆ on your SSH app, this will bring up the last command (ls -lh record/videos), hit Enter
    • The file size should slowly grow over time if the camera is recording.
12. Finally, disconnect from the Pi using your SSH app.

Returning to the field

1. On your smartphone (or tablet), start the WiFi hotspot
2. Wait (up to 3 minutes!) for the Raspberry Pi(s) to connect to your hotspot
3. Find out the IP address of the Raspberry Pi you want to connect to
4. Open your SSH app and connect to the Pi
5. [OPTIONAL] Check the recordings has started:
   • type ls -lh record/videos, hit Enter
   • take note of the last file name (useful to troubleshoot if recording has failed etc)
6. Stop the recording: type sudo python killall
7. Turn off the Pi: sudo halt
8. Disconnect from battery
9. Take out the SD card with the videos from the Pi
10. [OPTIONAL] move to another object / plant
11. Place a fresh SD card in
12. Replace battery with fresh one if necessary
13. Reconnect battery to Pi
14. Position camera
15. Check shot and settings from live preview
16. If no settings changes necessary, start recording; if changes necessary see above.
17. Disconnect from Pi

Hardware

In this guide I assume you are using a Raspberry Pi Zero W and a Raspberry Pi Camera v3 (with IR filter, non-wide).

For more detailed hardware specifications, take a look at 'How to build PICT' guide, most of the recommendations still apply.

Installation

If you are on Windows, you will need to install and run Ext2 volume manager to be able to read the data off of SD cards once flashed. There will be two partitions, the one with the data is rootfs.

Quickstart

The easiest way to get started is to use one of the pre-built images I provide.

You can download image ISOs here (requires permission): https://drive.google.com/drive/u/2/folders/136YHJ19of67geJv1VAqPn2JNwlM1lIbM

Use a software like Raspberry Pi Imager or Balena Etcher to flash one of these images to an SD card.

Note

If using Raspberry Pi Imager, when prompted write the image without modifying the configuration.

Once flashing is complete, insert the SD card in your Pi ad connect to power; you might have to wait for up to 10 minutes for the first boot to complete.

Note

The OS images provided should auto-expand to fill all available space on the SD card; is that doesn't happen for some reason, you can use the raspi-config utility to do that on the Pi directly.

Installing the OS

Alternatively, you can also start from a stock Raspberry Pi image and follow the steps below.

Using Raspberry Pi Imager v1.9.4 - https://downloads.raspberrypi.org/imager/imager_latest.exe

Selecting these values in RPi Imager:

• Model: RPi Zero
• OS version: Pi OS lite (32-bit) - no desktop, no dependencies

Selecting additional settings before flashing (input appropriate values for your WiFi network and region):

• Username: pi ; password: raspberry
• WiFi SSID: PICT_network_1 ; password: pollinators1
• wireless LAN country: CH
• locale: EU/Zurich In "Services" tab:
• Enable SSH; use password authentication

How to

Connect to the Pi

In the following section I presume the configuration outlined above, that assumes you are using a laptop or a smartphone as hotspot with this parameters:

• WiFi SSID: PICT_network_1
• password: pollinators1
• band: 2.4GHz

The Pi is already configured to look for this network and connect to it.

How you can verify if the Pi is connected to your hotspot varies depending on your device, but the steps are generally something like:

• Windows: Wi-Fi settings > Mobile hotspot
• Android: Settings > W-Fi & Internet > Hotspot & tethering > Wi-Fi hotspot
  • Alternatively, you can download NetAnalyzer and use the "LAN scan" function

What you are after is the IP address assigned to your Pi, as you will need this for the next steps.

Access live preview server

At boot, the Pi should automatically start a server showing the camera stream, which is useful to check framing and adjust parameters.

To access the server, in any browser (both from laptop or smartphone) type in the IP address of your Pi, followed by port 8000. For example, in my case, I would need to type (the IP address will be different for you):

192.168.137.71:8000

Press Enter and you should be presented with a screen like this:

There are a few preset values for resolution, autofocus and manual focus you can choose from; click apply and wait for the server to restart.

Connect via SSH

Currently, to modify configuration and manually start recording, you need to connect to the Pi via SSH.

To do so, you will need to open a terminal:

• Windows: press "Start", type cmd, press enter
• Android: download JuiceSSH or Termius, open that app

Type ssh pi@[yourIPaddress], for example in my case:

ssh pi@192.168.137.71

type password raspberry when prompted (unless you changed at configuration time).

You should see a greeting message like this:

The basic steps to get to this point are the same across platforms, but on smartphone apps the interface might be somewhat different.

Record video

crontab scheduling

To change when you want to start and stop the recording, you should use crontab.

First, connect via SSH, then type:

crontab -e

and you should be presented with a text interface; scroll all the way to the bottom (on smartphone apps there often are special buttons for arrow keys), and you will see this snippet:

0 21 * * * sudo killall python
1 21 * * * sudo ifconfig wlan0 down
55 5 * * * sudo reboot
0 6 * * * /home/pi/start_video_v6.1.py

Brief explanation line by line:

• 0 21 * * * sudo killall python means: at 21:00, stop all python processes (i.e. any recording in progress)
• 1 21 * * * sudo ifconfig wlan0 down means: at 21:01, switch off Wi-Fi
• 55 5 * * * sudo reboot means: at 5:55, reboot the system
• 0 6 * * * /home/pi/start_video_v6.1.py means: at 6:00, start recording using script start_video_v6.1.sh

If you want to change the recording times, simply modify the corresponding lines, respecting the format mm HH * * * command as shown above.

When done, exit by pressing Ctrl + X (or equivalent buttons on smartphone apps).

Change recording parameters

For now, the only way to change recording parameters like resolution, focus mode, etc is to directly modify the Python script.

The scripts are located in the root of this repo, you should use the most recent one, in this case videos_v6.1.py.

To open the file, type:

nano videos_v6.1.py

You will be presented with a text interface; navigate down using the arrow keys (or equivalent button on smartphone apps) and find this code block (starts at line 13):

### Parameters declaration
video_duration = 1800
video_number = 336
resolution = (1536, 864)  # Set the desired resolution as (width, height), WIDE format
target_fps = 10
focus = "Manual"  # Set to "Auto" or "Manual"
focus_distance = 4  # Only used if Focus is "Manual"; 0 (infinity) and 10.0 (approx. 10cm); default is 0.5 (focus at ~1m))
video_dir = "/home/pi/record/videos/"
os.makedirs(video_dir, exist_ok=True)

This is where you can modify the recording parameters; most variables are self-explanatory, but video_duration is expressed in seconds (e.g. 1800 second = 30 minutes); video_number is how many videos to record (e.g. 336 / 2 / 14 = 12 days of recording 14 hours per day, videos of 0.5 hours); focus_distance is expressed in dioptres, 1/focus_distance (e.g. 1/4 = 25 cm).

Modify the desired parameters, then pres Ctrl + X (or equivalent buttons on smartphone apps), and confirm overwriting with the same name.

Start recording manually

If you want to start recording manually, you can do so by running the dedicated .sh scripts; use the one named the same as the python file you want to use, in this case start_video_v6.1.sh.

You can run these files by typing:

./start_video_v6.1.sh

These scripts will stop the live server, if active, log the output of the python script and prevent another recording to automatically start. crontab scheduling will take place as instructed.

Autostart recording

There is a service running on the Pi that monitors user activity, and if none is detected (i.e. no connection to SSH, no open web browser with live preview server, not running commands), after 10 minutes of inactivity the recording will start using the latest shell script, in this case start_video_v6.1.sh.

If there is already a recording in progress (e.g. it was started manually), a new one will not be triggered.

Changelog

v3.0.3 - cleaning up repo, writing instructions

v3.0.2 - implementing managing of preview server and recording via system services

v3.0.1 - implementation of live preview server based on uvicorn.

v3.0.0 - re-hash of older software based on Pi OS Buster; the reason is that RPi camera v3 is incompatible with the picamera stack...