Summary

Hello dear reader! In this post I will be discussing how to control a DSLR camera using an Android app, Raspberry Pi, gphoto2 and some code. I found this capability very useful for my endeavours and hope that you may as well for yours. I will go over motivations, how to setup, linked code and an example use case. If you are restless, at the end I have a gif of a time-lapse I was able to create with the help of this app.

Here is a view of the deployed system:

  • gphoto2 is a free application that can be used to control a variety of different cameras via commandline.

  • The Raspberry Pi can be replaced with another computer. Raspberry Pi is chosen as it is leightweight and cheap.

  • Android app is useful because it allows for wireless, remote control and a nice user interface.

Here are reasons why you may find the setup useful:

Cameras come with limitations

In this blog I will be using the Canon M100.

In the case of the M100, the camera has a limitation on how long you can take time-lapse photos for. In the image below, three screenshots from the M100 configuration screen show that you can set to take photos every 2 to 4 seconds and up to 900 shots. Combining the two upper limits you can only takes a time-lapse spanning an hour in time.

What if you want to take a time-lapse of a full day? With the use of a battery cable and this setup you can overcome such limitations. The Android acts as user interface enhancement instead of using gphoto2 commandline directly on the Raspberry Pi.

M100 battery cable:

The camera manufacturer provided Android app is not sufficient

The Android app provided here can be modified to monitor for events for which regular manufacture provided Android apps for your camera will not have. Want to take picture at a specific time? Take a picture only when you detect something in the field of focus of cell phone camera? Take a picture when you say “action”? You will have to code this yourself for which the Android app introduced in this post provides a great starting point.

The programmatic freedom allows one to integrate your camera into a variety of applications.

Broken camera screen

May sound like a stretch but I had this happen. Camera menu screen was busted. The app can be useful here for providing a basic screen interface.

How to Setup

Install gphoto2 on Raspberry Pi

You will have to install gphoto2 software. This is a good repository on how to do this.

Connect the Camera to the Raspberry Pi

Connect the camera to the Raspberry Pi. In the case of the M100, I use the USB to mini USB B. Cable below:

The mini USB B port on the M100 is pictured below.

The result will look something like this:

Turn your camera on and see if your camera has a setting to stay on without sleeping. On the M100 the setting looks like:

In the Raspberry Pi terminal run

ps aux | grep gphoto

and run kill on the non grep results. Next run

gphoto2 --shell
>trigger-capture

Sometimes taking pictures will fail. In the image above, the failure was due to camera not focusing while in auto focus mode. For the second picture, I had to make sure I had better lighting.

gphoto2 supports many cameras to varying degrees. Your debug process and figuring out how to run your camera will thus vary. Be patient. By searching your error on a search engine you can usually find a thread on gphoto2’s repository or some stackoverflow post discussing a potential workaround. Here are some links that I found useful:

…happy debugging!

Connecting Raspberry Pi to Android Phone

Connect your phone and Raspberry Pi together. There are many instructions online on how to do this like here.

Setting up the Bluetooth server on Raspberry Pi

On the Raspberry Pi you will have to run

sudo python server.py

The python script can be found here.

Apply a setting to your Raspberry Pi so it does not fall asleep and stays on.

If you experience errors: make sure to install the required packages. For potential debugging when failing to advertise I found this to be helpful.

Build the Android app

You must be able to build the Android app using Android Studio. This can get a bit involved for a first time builder. I recommend starting here to get introduced to the basics.

The location of the app code is here.

Intermission

If you followed instructions above you now have

  • Bluetooth-paired the Android phone and Raspberry Pi
  • Launched the Bluetooth server on the Raspberry Pi
  • Connected your camera to the Raspberry Pi
  • Built and installed the application on your Android.

For illustration purposes I have setup the Raspberry Pi and camera.

The camera is connected to the Raspberry Pi and the battery is cable powered.

The test subject image here is a map of Winnipeg:

The Bluetooth server on the Raspberry Pi is running and waiting for a connection.

Running the Android Phone App

Now let us launch the Android application which is called “Bluetooth”. The landing page will look like

Here is me holding the phone:

After pressing connect you will be redicted to

And the Raspberry Pi will accept the connection:

When selecting “Single picture” and pressing the camera button you will see:

The Bluetooth server will reflect these pictures taken:

The picture is also saved to the SD card of the camera.

The other option here is to take a time-lapse:

The “Shutdown Signal” is for when you want to disconnect and quit the Bluetooth server. The app will then redict you back to the landing page.

Time-lapse example

Here is an example of the time-lapse feature. After the pictures were taken I combined them to form a gif. Earlier, I suggested that the M100’s time-lapse feature could be extended for a whole day. I was a bit lazy to do that for this post and did something that is quicker and would take up less memory:

Now the skeptical reader might ask: “Why do you need an Android app to send a single command to start executing a time-lapse or even a single photograph? Why not modify server.py and run the code straight on the Raspberry Pi?”

It is true. In such cases the phone is unnecessary. I could just run gphoto2 --shell in the terminal. However, for applications where you need more advance features, the Android phone is very versatile. In this post I describe such an advanced application that builds on top of the logic introduced here by using the phone’s camera and a Tensorflow model. This post is supposed to provide starting code and general introduction to combining the phone, camera and Raspberry Pi together. Happy hacking!