Today’s blog post will take a short diversion from our recent trend of Deep Learning tutorials here on the PyImageSearch blog and instead focus on a topic that I’ve been receiving a ton of emails about lately — common errors when using the Raspberry Pi camera module.
I want to start this post by mentioning Dave Jones, the maintainer and chief contributor to the picamera library. Dave is one of the most active open source contributors that I’ve had the privilege to interact with (and he’s a hell of a nice guy too).
A few months ago, I was using the (at the time) latest
picamera==1.11 library and was running in to a few errors. After checking the
picamera GitHub, I noticed an Issue had been posted regarding my problem. I confirmed the existence of the bug, which Dave then sought out — and fixed before the day was over, even releasing a new, updated version to the PyPI repository.
It goes without saying that without Dave, computer vision and OpenCV on the Raspberry Pi wouldn’t be nearly as fun — or half as accessible.
Over the past few years I’ve encountered a few “common” errors when using the Raspberry Pi and the
picamera library. My goal here today is to document some of these errors so you can easily fix them.
In fact, most of the issues I’m documenting here today are not real “errors” at all — they are simply misunderstandings on how the
picamera library works in conjunction with your Raspberry Pi setup.
Common errors using the Raspberry Pi camera module
Before we can look at common errors when using the Raspberry Pi camera module, let’s first discuss how we can simply access the
picamera video stream.
How to access your picamera video stream
To start, I am going to assume that you’ve already followed the instructions in Accessing the Raspberry Pi Camera with OpenCV and Python and installed the
picamera library on your Pi.
If you haven’t installed
picamera , this can be accomplished using
$ pip install "picamera[array]"
We add the
[array] portion to the command to ensure we can read frames as NumPy arrays, thus making the module compatible with OpenCV.
pip has finished installing
picamera , you can check the version number using the following command:
$ pip freeze
The reported version of picamera should be at least 1.12.
A quick note on Python virtual environments
If you’re a frequent reader of the PyImageSearch blog, you’ll know that I use Python virtual environments a lot — and because of this, you likely do as well.
Before we continue, take a second to see if you are using Python virtual environments by
source ‘ing your
~/.profile file and listing all the available virtual environments on your system:
$ source ~/.profile $ lsvirtualenv
If you get an error related to the
lsvirtualenv command not being found, then you are not utilizing Python virtual environments (or you have potentially made a mistake editing your
~/.profile file). If you’re not using Python virtual environments, then you can skip the next paragraph and move to the next sub-section.
Assuming you are using Python virtual environments, you can execute the
workon command to access each of the individual Python virtual environments on your system. In most install tutorials on the PyImageSearch blog, we name our Python virtual environment
cv , short for “computer vision”:
$ workon cv
A Python template for accessing your Raspberry Pi camera module
In order to access the
picamera video stream, I’ve created a simple, extendible template which I’ll detail for you below.
Open up a new file, name it
test_video.py , and insert the following code:
# import the necessary packages from picamera.array import PiRGBArray from picamera import PiCamera import time import cv2 # initialize the camera and grab a reference to the raw camera capture camera = PiCamera() camera.resolution = (640, 480) camera.framerate = 32 rawCapture = PiRGBArray(camera, size=(640, 480)) # allow the camera to warmup time.sleep(0.1)
Lines 2-5 handle importing our required Python packages.
We then initialize our
camera object on Line 8, which allows us to access the Raspberry Pi camera module. We’ll define the resolution of the video stream to be 640 x 480 with a maximum frame rate of 32 FPS (Lines 9 and 10).
From there, we initialize our
PiRGBArray object on Line 11, passing in the original
camera object and then explicitly re-stating the resolution as well. This
PiRGBArray object allows us to actually read frames from the Raspberry Pi camera module in NumPy format, thereby making the frames compatible with OpenCV.
Finally, we wait 0.1 seconds to allow the Raspberry Pi camera sensor to warmup.
Our next code block handles actually capturing frames from our Raspberry Pi camera sensor:
# capture frames from the camera for frame in camera.capture_continuous(rawCapture, format="bgr", use_video_port=True): # grab the raw NumPy array representing the image - this array # will be 3D, representing the width, height, and # of channels image = frame.array # show the frame cv2.imshow("Frame", image) key = cv2.waitKey(1) & 0xFF # clear the stream in preparation for the next frame rawCapture.truncate(0) # if the `q` key was pressed, break from the loop if key == ord("q"): break
On Line 17 we start looping over frames captured from the
camera using the
capture_continuous function. We pass three parameters into this method.
The first is
rawCapture , the format in which we want to read each frame. We then specify the
format to be
bgr since OpenCV expects image channels to be in BGR order rather than RGB. Finally, the
use_video_port boolean indicates that we are treating the stream as video.
Once we have the
frame , we can access the raw NumPy array via the
.array attribute (Line 20).
We display the frame to our screen on Lines 23 and 24 using OpenCV GUI functions.
But before we can move on to the next frame, we first need to prepare our stream by calling the
.truncate method on the
rawCapture object. If you do not do this, your Python script will throw an error — the exact error we’ll review later in this guide.
Finally, if the
q key is pressed (Lines 30 and 31), we break from the loop.
To execute the
test_video.py script, just open up a terminal/command line prompt and execute the following command:
$ python test_video.py
Note: If you’re using Python virtual environments, you’ll want to use the
workon command to switch to the Python environment that has your OpenCV + picamera library installed.
If all goes well, you should see the Raspberry Pi video stream displayed to your feed:
Otherwise, if you get an error — keep reading. I’ve detailed the most common error messages that I run in to below.
Can’t connect to your picamera module?
Does the following error message look familiar?
$ python test_video.py mmal: mmal_vc_component_create: failed to create component 'vc.ril.camera' (1:ENOMEM) mmal: mmal_component_create_core: could not create component 'vc.ril.camera' (1) Traceback (most recent call last): File "test_video.py", line 11, in <module> camera = PiCamera() File "/home/pi/.virtualenvs/cv2/local/lib/python2.7/site-packages/picamera/camera.py", line 488, in __init__ self.STEREO_MODES[stereo_mode], stereo_decimate) File "/home/pi/.virtualenvs/cv2/local/lib/python2.7/site-packages/picamera/camera.py", line 526, in _init_camera "Camera is not enabled. Try running 'sudo raspi-config' " picamera.exc.PiCameraError: Camera is not enabled. Try running 'sudo raspi-config' and ensure that the camera has been enabled.
If you are getting this error message, then you likely forgot to (1) run
raspi-config , 92) enable the camera, and (3) reboot your Pi.
If you are still getting an error message after running
raspi-config , then your camera is likely installed incorrectly. In this case, I would suggest giving this install video a watch and then trying to install your Raspberry Pi camera module again (be sure to power down your Pi first!)
Truncation errors are fairly easy to identify since they always end with the text
Incorrect buffer length for resolution . An example of such an error message can be found below:
$ python test_video.py Traceback (most recent call last): File "test_video.py", line 20, in <module> for frame in camera.capture_continuous(rawCapture, format="bgr", use_video_port=True): File "/home/pi/.virtualenvs/cv/local/lib/python2.7/site-packages/picamera/camera.py", line 1851, in capture_continuous if not encoder.wait(self.CAPTURE_TIMEOUT): File "/home/pi/.virtualenvs/cv/local/lib/python2.7/site-packages/picamera/encoders.py", line 850, in wait self.stop() File "/home/pi/.virtualenvs/cv/local/lib/python2.7/site-packages/picamera/encoders.py", line 881, in stop self._close_output() File "/home/pi/.virtualenvs/cv/local/lib/python2.7/site-packages/picamera/encoders.py", line 795, in _close_output output.flush() File "/home/pi/.virtualenvs/cv/local/lib/python2.7/site-packages/picamera/array.py", line 285, in flush self.array = bytes_to_rgb(self.getvalue(), self.size or self.camera.resolution) File "/home/pi/.virtualenvs/cv/local/lib/python2.7/site-packages/picamera/array.py", line 174, in bytes_to_rgb 'Incorrect buffer length for resolution %dx%d' % (width, height)) picamera.exc.PiCameraValueError: Incorrect buffer length for resolution 640x480
If you are getting this error message, you likely forgot to call
.truncate after you were done processing your frame. Line 27 of our
test_video.py script above demonstrates how to use the
In short: Go back to your script and ensure you have called
.capture_continuous is called again.
The picamera==1.11 and Python 3 specific error
The v1.11 release of the
picamera library introduced a Python 3 specific error. It was particularly hard to diagnose, but as Dave Jones points out, the issue was due to a mis-configuration in his test suite, leading to no Python 3 tests being performed.
This error is often diagnosed by seeing the text
TypeError: startswith first arg must be bytes or a tuple of bytes, not str as the final line in the error message. A full example of this error message is shown below:
$ python test_video.py Traceback (most recent call last): File "test_image.py", line 15, in <module> camera.capture(rawCapture, format="bgr") File "/home/pi/.virtualenvs/cv/local/lib/python3.4/site-packages/picamera/camera.py", line 1371, in capture camera_port, output_port, format, resize, **options) File "/home/pi/.virtualenvs/cv/local/lib/python3.4/site-packages/picamera/camera.py", line 652, in _get_image_encoder self, camera_port, output_port, format, resize, **options) File "/home/pi/.virtualenvs/cv/local/lib/python3.4/site-packages/picamera/encoders.py", line 1049, in __init__ parent, camera_port, input_port, format, resize, **options) File "/home/pi/.virtualenvs/cv/local/lib/python3.4/site-packages/picamera/encoders.py", line 534, in __init__ if not resize and format != 'yuv' and input_port.name.startswith('vc.ril.video_splitter'): TypeError: startswith first arg must be bytes or a tuple of bytes, not str
The easiest solution is to either upgrade or downgrade your
picamera module by one point version.
To upgrade to v1.12 (or whatever the current version of
picamera is), use the following command:
$ pip install --upgrade picamera
To downgrade to v1.10, just use this commands:
$ pip uninstall picamera $ pip install "picamera[array]"==1.10
Make sure you are being mindful of whether you’re using Python virtual environments or not. If you are, execute the
workon command before running
pip to ensure
picamera is installed in to your virtual environment.
Blank and/or black frame
The blank/black frame is a particularly strange problem — frames are being read from the video stream, they are just not being decoded and displayed properly.
First, run the
rpi-update to grab the latest firmware updates for your Raspberry Pi:
$ sudo rpi-update
After your Pi reboots, try re-executing your Python script.
If the frame retrieved by the Raspberry Pi camera is still blank/black, then downgrade your
picamera installation to v1.10:
$ pip uninstall picamera $ pip install "picamera[array]"==1.10
I’ve encountered situations where I’ve only had to run
rpi-update to resolve the issue. And I’ve also needed to both upgrade my firmware and downgrade my
picamera version. I’m not sure what the common underlying thread is (I think it’s related to using the newer version of the Raspberry Pi camera hardware +
picamera versioning, but I haven’t been able to nail that down yet).
What's next? I recommend PyImageSearch University.
25 total classes • 37h 19m video • Last updated: 7/2021
★★★★★ 4.84 (128 Ratings) • 10,597 Students Enrolled
I strongly believe that if you had the right teacher you could master computer vision and deep learning.
Do you think learning computer vision and deep learning has to be time-consuming, overwhelming, and complicated? Or has to involve complex mathematics and equations? Or requires a degree in computer science?
That’s not the case.
All you need to master computer vision and deep learning is for someone to explain things to you in simple, intuitive terms. And that’s exactly what I do. My mission is to change education and how complex Artificial Intelligence topics are taught.
If you're serious about learning computer vision, your next stop should be PyImageSearch University, the most comprehensive computer vision, deep learning, and OpenCV course online today. Here you’ll learn how to successfully and confidently apply computer vision to your work, research, and projects. Join me in computer vision mastery.
Inside PyImageSearch University you'll find:
- ✓ 25 courses on essential computer vision, deep learning, and OpenCV topics
- ✓ 25 Certificates of Completion
- ✓ 37h 19m on-demand video
- ✓ Brand new courses released every month, ensuring you can keep up with state-of-the-art techniques
- ✓ Pre-configured Jupyter Notebooks in Google Colab
- ✓ Run all code examples in your web browser — works on Windows, macOS, and Linux (no dev environment configuration required!)
- ✓ Access to centralized code repos for all 400+ tutorials on PyImageSearch
- ✓ Easy one-click downloads for code, datasets, pre-trained models, etc.
- ✓ Access on mobile, laptop, desktop, etc.
In this blog post, I reviewed common error messages that you may encounter when using your Raspberry Pi camera module and the picamera library.
These errors can be quite frustrating to debug, especially if you’re just getting started. My hope is that this guide helps point you in the right direction if you run in to any of these problems — I’ve certainly had my share of long nights trying to track these errors down!
Finally, I would like to mention again the wonderful work and contribution of Dave Jones, the maintainer and chief contributor to the picamera library. Without him, we wouldn’t be able to have near as much computer vision fun with our Pi’s!
Download the Source Code and FREE 17-page Resource Guide
Enter your email address below to get a .zip of the code and a FREE 17-page Resource Guide on Computer Vision, OpenCV, and Deep Learning. Inside you'll find my hand-picked tutorials, books, courses, and libraries to help you master CV and DL!