Edit

Part 1: Vision Pipeline

Goal: Get a computer vision pipeline working.

Skills: Connect a machine to Viam, configure components in the Viam UI, use fragments to add preconfigured services.

Time: ~10 min

Prerequisites

Before starting this tutorial, you need the can inspection simulation running. Follow the Gazebo Simulation Setup Guide to:

Build the Docker image with Gazebo Harmonic
Create a machine in Viam and get credentials
Start the container with your Viam credentials

Once you see “Can Inspection Simulation Running!” in the container logs and your machine shows Live in the Viam app, return here to continue.

What you're working with

The simulation runs Gazebo Harmonic inside a Docker container. It simulates a conveyor belt with cans (some dented) passing under an inspection camera. viam-server runs on the Linux virtual machine inside the container and connects to Viam’s cloud, just like it would on a physical machine. Everything you configure in the Viam app applies to the simulated hardware.

1.1 Verify Your Machine is Online

If you followed the setup guide, your machine should already be online.

Open app.viam.com (the “Viam app”)
Navigate to your machine (for example, inspection-station-1)
Verify the status indicator shows Live
Click the CONFIGURE tab if not already selected

Machine page showing the green Live status indicator next to the machine name.

Ordinarily, after creating a machine in Viam, you would download and install viam-server together with the cloud credentials for your machine. For this tutorial, we’ve already installed viam-server and launched it in the simulation Docker container.

1.2 Locate Your Machine Part

Your machine is online but empty. To configure your machine, you will add components and services to your machine part in the Viam app. Your machine part is the compute hardware for your robot. This might be a PC, Mac, Raspberry Pi, or another computer.

In the case of this tutorial, your machine part is a virtual machine running Linux in the Docker container.

Find inspection-station-1-main in the CONFIGURE tab.

1.3 Configure the Camera

You’ll now add the camera as a component.

What's a component?

In Viam, a component is any piece of hardware: cameras, motors, arms, sensors, grippers. You configure components by declaring what they are, and Viam handles the drivers and communication.

The power of Viam’s component model: All cameras expose the same API—USB webcams, Raspberry Pi camera modules, IP cameras, simulated cameras. Your application code uses the same GetImages() method regardless of the underlying hardware. Swap hardware by changing configuration, not code.

Add a camera component

To add the camera component to your machine part:

Click the + button and select Component or service
Click Camera
Search for gz-camera
Select gz-camera:rgb-camera
Click Add module
Enter inspection-cam for the name

Why were two items added to my machine part?

After adding the camera component, you will see two items appear under your machine part. One is the actual camera hardware (inspection-cam) that you will use through the Viam camera API. The other is the software module (gz-camera) that implements this API for the specific model of camera you are using. All components that are supported through modules available in the Viam registry will appear this way in the CONFIGURE tab. For built-in components, such as webcams, you will not also see a module appear in the configuration.

Configure the camera

To configure your camera component to work with the camera in the simulation, you need to specify the correct camera ID. Most components require a few configuration parameters.

In the ATTRIBUTES section, add:
```
{
  "id": "/inspection_camera"
}
```
Click Save in the top right

What happened behind the scenes

You declared “this machine has a camera called inspection-cam” by editing the configuration in the Viam app. When you clicked Save, viam-server loaded the camera module, added a camera component, and made the camera available through Viam’s standard camera API. Software you write, other services, and user interface components will use the API to get the images they need. Using the API as an abstraction means that everything still works if you swap cameras.

1.4 Test the Camera

Verify the camera is working. Every component in Viam has a built-in test card right in the configuration view.

Open the test panel

You should still be on the CONFIGURE tab with your inspection-cam selected
Look for the TEST section at the bottom of the camera’s configuration panel
Click TEST to expand the camera’s test card

The camera component test card uses the camera API to add an image feed to the Viam app, enabling you to determine whether your camera is working. You should see a live video feed from the simulated camera. This is an overhead view of the conveyor/staging area.

Checkpoint

Your camera is working. You can stream video and capture images from the simulated inspection station.

1.5 Add a vision pipeline with a fragment

Now you’ll add machine learning to run inference on your camera feed. You need two services:

ML model service that loads a trained model for the inference task
Vision service that connects the camera to the model and returns detections

Components versus services

Components are hardware: cameras, motors, arms
Services are capabilities: vision (ML inference), motion (arm kinematics), custom control logic

Services often use components. A vision service takes images from a camera, runs them through an ML model, and returns structured results, detections with bounding boxes and labels, or classifications with confidence scores.

The ML model service loads a trained model (TensorFlow, ONNX, or PyTorch) and exposes an Infer() method. The vision service handles the rest: converting camera images to tensors, calling the model, and interpreting outputs into usable detections.

Instead of adding each service manually, you’ll use a fragment. A fragment is a reusable block of configuration that can include components, services, modules, and ML models. Fragments let you share tested configurations across machines and teams.

The try-vision-pipeline fragment includes an ML model service loaded with a can defect detection model and a vision service wired to that model. The fragment accepts a camera_name variable so it works with any camera.

Add the fragment

Click + next to your machine part
Select Insert fragment
Search for try-vision-pipeline
Select it and click Insert fragment

Set the camera variable

The fragment needs to know which camera to use for inference.

In the fragment’s configuration panel, find the Variables section
Set the camera_name variable to inspection-cam
```
{
  "camera_name": "inspection-cam"
}
```
Click Save in the upper right corner

What the fragment added

The fragment added two services and their dependencies to your machine:

model-service: An ML model service running TensorFlow Lite with the can-defect-detection model from the Viam registry. This model classifies cans as PASS or FAIL.
vision-service: A vision service that takes images from your camera, runs them through the ML model, and returns structured detection results.

The fragment also added the TFLite CPU module and the ML model package. Everything is wired together and ready to use.

Fragments and reuse

This fragment works with any camera. If you were using a USB webcam instead of the simulation camera, you’d set camera_name to whatever you named your webcam component. The ML pipeline stays the same. This is how fragments enable reuse across different hardware setups.

Test the vision service

Find the TEST section at the bottom of the vision-service configuration panel
Expand the TEST card
If not already selected, select inspection-cam as the camera source
Set Detections/Classifications to Live
Check that detection and labeling are working

Vision service test panel showing a can detected with a bounding box and FAIL label.

What you've built

A complete ML inference pipeline. The vision service grabs an image from the camera, runs it through the TensorFlow Lite model, and returns structured detection results. This same pattern works for any ML task: object detection, classification, segmentation. Swap the model and camera, and the pipeline still works.

Checkpoint

You added a camera component manually and used a fragment to add a complete ML vision pipeline. The system can detect defective cans. Next, you’ll set up continuous data capture so every detection is recorded and queryable.

Explore the JSON configuration

Everything you configured through the UI is stored as JSON. Click JSON in the upper left of the Configure tab to see the raw configuration. You’ll see your camera component, the fragment reference, and how the fragment’s services connect to your camera. As configurations grow more complex, the JSON view helps you understand how components and services connect.

Continue to Part 2: Data Capture →

Was this page helpful?

Glad to hear it! If you have any other feedback please let us know:

We're sorry about that. To help us improve, please tell us what we can do better:

Thank you!