IDS uEye Camera Acquisition

A lightweight Python/PyQt6 acquisition software for IDS uEye USB3 cameras (tested on UI-3040CP-M), developed for scientific brightfield imaging.

The codes and the github wrapping have been created using Anthropic Claude Sonnet 4.6.

Features

• Live preview with adjustable exposure and gain
• Hardware ROI selection
• Single frame grab, timelapse, and burst acquisition
• 16-bit TIFF output with embedded JSON metadata (exposure, gain, ROI, timestamp)
• Non-destructive preview processing: grayscale, threshold/binarization (binary or Otsu)
• Clean camera disconnect via UI button or window close

Requirements

System

• Ubuntu 22.04 / 24.04 (64-bit)
• IDS peak SDK ≥ 2.17 with uEye transport layer installed → Download: https://en.ids-imaging.com/downloads.html (IDS peak, Linux, USB)
• libxcb-cursor0 for Qt xcb platform:

  sudo apt install libxcb-cursor0

Python

Python 3.10+ recommended. Install dependencies in a virtual environment:

python3 -m venv ~/envs/camera
source ~/envs/camera/bin/activate
pip install -r requirements.txt

Installation

git clone https://github.com/YOUR_USERNAME/ids-ueye-acquisition.git
cd ids-ueye-acquisition
source ~/envs/camera/bin/activate
python main.py

Repository structure

ids-ueye-acquisition/
├── main.py            # Entry point
├── main_window.py     # PyQt6 GUI and threading logic
├── ids_camera.py      # IDS peak API wrapper (device, stream, parameters)
├── acquisition.py     # Frame saving: single, timelapse, burst (16-bit TIFF)
├── processing.py      # Preview pipeline: grayscale, threshold
├── requirements.txt
└── README.md

Usage

source ~/egit remote add origin https://github.com/FattaccioliLab/ids_uEye_camera.git
git branch -M main
git push -u origin mainnvs/camera/bin/activate
python main.py

1. The camera is detected and connected automatically on launch.
2. Click Start Preview to begin live streaming.
3. Adjust Exposure (spinbox or logarithmic slider) and Gain on the fly.
4. Set a hardware ROI and click Apply ROI (stream restarts automatically).
5. Enable Threshold overlay for preview-only binarization (does not affect saved files).
6. Select acquisition mode (Single / Timelapse / Burst), set save directory and prefix, then click Acquire.
7. Click Stop stream & Quit or close the window to properly release the camera.

Output format

Frames are saved as 16-bit TIFF files. Metadata is embedded in the ImageDescription tag as JSON:

{
  "timestamp": "2025-04-03T14:32:10.123456",
  "frame_index": 0,
  "exposure_us": 5000.0,
  "gain": 1.0,
  "roi_x": 0,
  "roi_y": 0,
  "roi_width": 2048,
  "roi_height": 1536
}

Read metadata in Python:

import tifffile, json
with tifffile.TiffFile("frame_00000.tiff") as tif:
    meta = json.loads(tif.pages[0].description)
    img  = tif.asarray()

Notes

• Exposure and gain are adjustable while streaming without interrupting the preview.
• ROI changes require a brief acquisition restart (buffer reallocation), handled automatically.
• The DataStream is opened once per session and reused across stop/start cycles to avoid GC_ERR_RESOURCE_IN_USE errors.
• Saved TIFFs are always 16-bit regardless of the camera pixel format (uint8 frames are upscaled to preserve dynamic range headroom).

To add in the future

• Possibility to add a scale manually, with a scalebar on the stream

License

MIT