{"id":79062,"date":"2024-03-21T23:03:53","date_gmt":"2024-03-22T03:03:53","guid":{"rendered":"https:\/\/www.engineersgarage.com\/?p=79062"},"modified":"2024-03-22T14:01:07","modified_gmt":"2024-03-22T18:01:07","slug":"opencv4-5-raspberry-pi-image-video-access-recording","status":"publish","type":"post","link":"https:\/\/www.engineersgarage.com\/opencv4-5-raspberry-pi-image-video-access-recording\/","title":{"rendered":"How to use the Raspberry Pi camera with OpenCV"},"content":{"rendered":"

OpenCV is a popular cross-platform framework for image processing and real-time computer vision. The software framework is useful for efficiently developing real-time computer vision applications that would be extremely time-consuming to create from scratch.\u00a0<\/span><\/p>\n

The power of OpenCV\u2019s framework is that it runs seamlessly on embedded platforms like microcomputers. This means developing and running embedded computer vision applications in real-time can be done with little hassle. We covered how to install OpenCV 4.5 and QT5 on Raspberry Pi in this article<\/a>.<\/span><\/p>\n

Once you\u2019ve installed OpenCV, it\u2019s time to capture any desired images and\/or videos. The system can then process them for useful insights and applications. What\u2019s nice about using Raspberry Pi is the images and videos can be captured by a Raspberry Pi camera \u2014 which easily installs on a webcam or CSI port.\u00a0<\/span><\/p>\n

In this project, we\u2019ll use a Raspberry Pi camera and record the captured images and videos on a Micro SD card (which hosts Raspberry Pi OS). Let’s get started.\u00a0<\/span><\/p>\n

Components required\u00a0<\/span><\/strong><\/p>\n

1.<\/strong> Raspberry Pi 4B\/3B x1
\n<\/span>2.<\/strong> Raspberry Pi camera x1
\n<\/span>3.<\/strong> Power adaptor for Raspberry Pi x1
\n<\/span>4.<\/strong> HDMI-To-Micro-HDMI cable x1
\n<\/span>5.<\/strong> Display monitor x1
\n<\/span>6.<\/strong> Keyboard and mouse x1
\n<\/span>7.<\/strong> Raspberry Pi case with camera case x1 (optional)<\/span><\/p>\n

The camera module
\n<\/span><\/strong>To begin, attach the camera module to Raspberry Pi\u2019s CSI port. Raspberry Pi must be installed with an operating system like Raspbian or Debian Linux \u2014 and OpenCV must be installed on the Raspberry Pi OS. Review the\u00a0<\/span>installation process of OpenCV 4.5 on Raspbian Bullseye <\/span>here<\/a>.\u00a0<\/span>For tips on attaching the camera module to RPi\u2019s CSI port, <\/span>click here<\/span><\/a>.\u00a0 \u00a0<\/span><\/p>\n

Once the camera module is set, enable the camera module in Raspberry Pi OS. Open the terminal and execute the following command.<\/span><\/p>\n

$ sudo raspi-config\u00a0<\/strong><\/span><\/p>\n

In the configuration window, select ‘Interface Options,’ and then ‘Legacy Camera.’ Select ‘yes’ to enable the camera module and reboot the system.\u00a0<\/span><\/p>\n

\"\"<\/a><\/p>\n

\"\"<\/a>Test the camera
\n<\/strong>It\u2019s important to test the camera module to ensure it works before starting the OpenCV code. To do so, open the terminal and execute the below command.<\/p>\n

$ raspistill -o test.jpg<\/span><\/strong><\/p>\n

\"\"<\/a><\/p>\n

You can tell the camera module is working if an image is clicked and saved on Raspbian. If there\u2019s no image, either the camera module is faulty or it\u2019s not installed correctly on the CSI port.\u00a0<\/span><\/p>\n

Installing Picamera
\n<\/span><\/strong>Next, install the Picamera library for use with the RPi camera. You\u2019ll install the package in the virtual environment we created when installing OpenCV 4.5 on Raspberry Pi.\u00a0<\/span><\/p>\n

Open the terminal and activate the virtual environment by executing this command:\u00a0<\/span><\/p>\n

$ workon cv<\/strong><\/span><\/p>\n

Note that ‘cv’ is the name of the virtual environment created in the previous tutorial. You can install the Picamera package within the virtual environment with the following:<\/span><\/p>\n

$ pip install “picamera[array]”\u00a0<\/strong><\/span><\/p>\n

\"\"<\/a><\/p>\n

Picamera is necessary to work with the camera module. The OpenCV array sub-module is required for the images. This is because OpenCV represents images as NumPy arrays. The array sub-module lets us obtain the NumPy arrays from the camera module.\u00a0<\/span><\/p>\n

The images
\n<\/span><\/strong>There are two ways to record and access images using the RPi camera via OpenCV. One way is to capture an image using the Picamera module and then use OpenCV to read and display the image. The other way is to directly write an image as a Numpy array, and then display it.\u00a0<\/span><\/p>\n

First, import the required modules:\u00a0<\/span><\/p>\n

from picamera import PiCamera
\nimport time
\nimport cv2\u00a0<\/strong><\/span><\/p>\n

To write an image as an OpenCV array, the \u2018picamera.array\u2019 module must be imported in the Python script as follows.\u00a0<\/span><\/p>\n

from picamera.array import PiRGBArray<\/strong><\/span><\/p>\n

Whether capturing an image using Picamera or the picamera.array, it\u2019s necessary to instantiate a Picamera object.\u00a0<\/span><\/p>\n

camera = PiCamera()<\/strong><\/span><\/p>\n

To capture an image using the Picamera module, use this code:\u00a0<\/span><\/p>\n

time.sleep(0.1)
\ncamera.capture(‘test.jpg’)<\/strong><\/span><\/p>\n

A delay of 100 milliseconds is provided, allowing the camera module to warm up. The camera.capture() method takes a shot and saves it on a MicroSD card by a filename that\u2019s specified as a parameter.\u00a0<\/span><\/p>\n

To capture an image using the picamera.array module, use this code:\u00a0<\/span><\/p>\n

rawImage = PiRGBArray(camera)
\ntime.sleep(0.1)
\ncamera.capture(rawImage, format = “rgb”)
\nimage = rawImage.array<\/strong><\/span><\/p>\n

An object of the PiRGBArray class is instantiated to store an image array. The time delay is provided for the camera module to warm up. The image is captured as a NumPy array using the camera.capture() method. The captured image array is stored in a variable ‘image.\u2019\u00a0<\/span><\/p>\n

OpenCV\u2019s imread() function is used to read an image. The function has the following prototype.\u00a0<\/span><\/p>\n

imread(filename, flags)<\/span><\/strong><\/p>\n

The imread() function takes two arguments. The first argument is a fully qualified path to the image file. The second argument is an optional flag specifying how the image is represented. It can take any one of the following values.\u00a0<\/span><\/p>\n