caproto-image-viewer

DEPRECATED and unsupported. OpenGL support is broken on the latest numpy, and the basic image viewer does not have many features.

caproto/PyQt5-based EPICS AreaDetector image viewer with OpenGL/GLSL color mapping

Provides two widgets with varying support for image types (see next section):

1. ImageViewerWidgetGL
    a. Requires OpenGL 4.1-compatible hardware
    b. Optional color-mapping (using look-up tables from matplotlib)
    c. Bayer demosaic is done fully on the GPU
2. ImageViewerWidget 
    a. Raw image display - OpenGL not required
    b. No color mapping yet
    c. Limited image format support (on the TODO list)

Supported Image Formats

Format	ImageViewerWidget	ImageViewerWidgetGL
Monochrome	Yes	Yes
RGB1	Yes	Yes
RGB2	Yes	Yes
RGB3	Yes	Yes
Bayer - RGGB		Yes
Bayer - GBRG		Yes
Bayer - GRBG		Yes
Bayer - BGGR		Yes
YUV444
YUV422
YUV421

Notes:

RGB2 is much slower than RGB1, RGB3. This can be improved at some point.

Requirements

Python 3.6+
caproto or pyepics
PyQt5
numpy

Optional

Matplotlib (only for additional color maps)

Installation

$ pip install -e .

Usage

If you have the ADSimDetector example IOC running, run the following for the OpenGL viewer:

$ caproto-image-viewer 13SIM1: --acquire --gl

Note that the --acquire part starts acquisition of the detector, and this may not be necessary or desirable if you use it with an existing camera/detector.

Then press the spacebar to enable color-mapping:

Cycle through the colormaps by pressing either [ or ]. Or, press p to preview other colormaps in realtime - while still acquiring:

You can cycle through this list similarly with [ and ], and exit the preview with p.

Alternatively, use the basic non-OpenGL version:

$ caproto-image-viewer 13SIM1: --acquire

Speed

The OpenGL display widget is very fast. Basic testing has shown:

Minimal IOC-to-screen latency
Displays 200Hz, 1MP 8-bit images with real-time color-mapping
Utilizes ~40% of one core in a 4-core system

Though the viewer is capable of doing this with ease, you probably shouldn't saturate your network viewing things at such a high frame rate unnecessarily. That's up to you, of course!

klauer / caproto-image-viewer Goto Github PK