scicamera

Make every frame count.

Scientific imaging often calls for control over exposure, sensor modes, and the path from sensor to array. scicamera began as a fork of picamera2 and focuses on high-quality, timing-sensitive still images with minimal processing.

From sensor to array.

A capture is more than a picture. The stream configuration determines what arrives, and metadata helps describe how it was acquired.

1. Choose a stream

   Build a still-capture configuration for the camera and the image format you need.

2. Set available controls

   Configure settings such as exposure time and analogue gain when the hardware exposes them.

3. Capture a frame

   Start the camera and obtain image data through an asynchronous capture result.

4. Keep the context

   Use capture metadata alongside your arrays or buffers in the next stage of your application.

scicamera runs on a Linux host with supported camera hardware and libcamera bindings. This page describes the capture workflow; it does not access your browser’s camera.

Prepare your camera host.

The project documents 64-bit Raspberry Pi OS (Bullseye or later) and x86 Ubuntu. The Python package depends on libcamera and its Python bindings, which must be installed separately for your platform.

Follow the project’s libcamera setup instructions, then install scicamera into a Python environment that can see those bindings.

python -m pip install scicamera

Before opening a camera, confirm that Python can import both packages:

python -c "import libcamera; import scicamera; print(scicamera.__version__)"

Capture your first array.

from scicamera import Camera, CameraConfig

with Camera() as camera:
    config = CameraConfig.for_still(camera)
    camera.configure(config)
    camera.start()

    camera.discard_frames(2).result()
    pixels = camera.capture_array().result()
    print(pixels.shape, pixels.dtype)

    camera.stop()

The context manager handles the camera lifetime, while capture methods return results you can wait for. Available resolutions, formats, and controls depend on the camera and its libcamera driver.

See array-capture examples in the project tests →

Build for the camera you have.

For low-bandwidth real-time video streaming, the project points to picamera2. scicamera’s focus is still-image acquisition for scientific applications.