A simple, modern C#/.NET wrapper for GigE Vision / GenICam industrial cameras.
Grab frames, read/write camera parameters, and handle events in a few lines of idiomatic C#. Built for machine-vision engineers who want to stop fighting vendor SDKs.
⚠️ Status: early preview. The public API is taking shape. Star/watch to follow along.
Most GigE Vision access from .NET means either a heavy vendor SDK (Basler, IDS, Allied Vision — each different) or a C/C++ library you have to P/Invoke yourself. SharpGigE gives you one small, camera-agnostic surface:
using SharpGigE;
// Discover cameras on the network
foreach (var info in CameraDiscovery.Enumerate())
Console.WriteLine($"{info.Model} @ {info.IpAddress} ({info.Serial})");
// Open the first one, configure, grab
using var cam = Camera.Open(CameraDiscovery.Enumerate().First());
cam.SetParameter("ExposureTime", 5000.0); // microseconds
cam.SetParameter("PixelFormat", "Mono8");
cam.Set("Width", 1920);
cam.Set("Height", 1080);
cam.FrameReady += (_, frame) =>
Console.WriteLine($"Frame #{frame.Id}: {frame.Width}x{frame.Height}, {frame.Data.Length} bytes");
cam.StartGrabbing();
Console.ReadLine();
cam.StopGrabbing();That's the whole loop. No callbacks-of-callbacks, no manual buffer pools to start.
- 🔍 Discovery — enumerate GigE Vision devices on the local network
- 🎥 Grab — start/stop streaming, single-frame acquisition
- ⚙️ Parameters — typed read/write of GenICam features (
ExposureTime,Gain,PixelFormat, ROI…) - 📢 Events —
FrameReady,Disconnected,Error - 📦 Formats — Mono8, Mono16, BayerRG8, RGB8 → convert to
byte[]/Span<byte> - 🧪 Built-in simulator — develop & test your whole pipeline with no camera (patterns, noise, custom frames)
- 💾 Frame export — save frames to PGM/PPM with zero image dependencies
- 🧩 Targets — .NET 6 / .NET 8, Windows & Linux
- 🔗 Multi-camera hardware sync (PTP / trigger chaining)
- ⚡ Zero-copy high-throughput buffer pipeline
- 🎯 Hardware trigger helpers & strobe control
- 🛠️ Priority support & private issue tracker
Pro is sold directly (not through this repo). Interested? Open a Discussion or watch this space.
dotnet add package SharpGigERequires a GigE Vision compatible camera on the same subnet. See docs/setup.md for jumbo-frame / firewall tips.
See the runnable sample in samples/SharpGigE.Grab:
cd samples/SharpGigE.Grab
dotnet runWith no camera present it runs against the built-in virtual test camera so you can try the API immediately.
SharpGigE ships a first-class device simulator so you can build and unit-test a vision pipeline before any hardware arrives — or run it in CI where no camera exists.
| Moving-bar pattern (+ noise) | SMPTE color bars |
|---|---|
 |
 |
Actual output from the sample, saved via FrameWriter — no image library required.
using SharpGigE;
var sim = Camera.OpenSimulated(new SimulatorOptions
{
Width = 640,
Height = 480,
Format = PixelFormat.Mono8,
Pattern = TestPattern.MovingBar, // Gradient | Checkerboard | MovingBar | Flat | ColorBars
NoiseStdDev = 8.0, // add reproducible Gaussian noise
FrameRate = 60.0
});
var frame = sim.GrabSingle();
// Save it — no ImageSharp / System.Drawing needed.
FrameWriter.Save(frame, "frame"); // -> frame.pgm (mono) / frame.ppm (color)Or feed your own image bytes so tests run against known input:
var sim = Camera.OpenSimulated(new SimulatorOptions
{
Width = w, Height = h, Format = PixelFormat.Rgb8,
CustomFrame = myRgbBytes // exactly w*h*3 bytes
});Everything else (discovery, parameters, StartGrabbing, events) behaves identically
to a real device, so your code doesn't change when the camera shows up.
- Discovery + parameter model (preview)
- Real GigE Vision transport (GVCP/GVSP) — in progress
- USB3 Vision support
- NuGet 1.0 release
- Pro tier launch
Issues and PRs welcome. This is an early project — API feedback is especially valuable. See CONTRIBUTING.md.
MIT © 2026 — see LICENSE.
The MIT-licensed core is free forever. Commercial Pro modules are separately licensed.