Getting Started

First launch, runtime setup, hardware connection, and creating your first analysis

Installation

Step 1: Install the Application

• Download the latest Csmart Digit Desktop installer from the releases page.

• Locate the downloaded installer file and double-click it to begin the installation.

• Ignore any Windows or antivirus warnings and proceed with the installation.

Important: Select the "Only Me" option during installation to avoid requiring admin privileges for software use.

• Once installation is complete, a Csmart Digit Desktop shortcut will be created on your desktop.

First Launch

When you open Digit Desktop for the first time, you need to install the Python runtime and verify your hardware connection before creating your first analysis.

Step 2: Install the Python Runtime

Digit Desktop requires a Python runtime to execute image processing, AI inference, and data export operations. The runtime is a self-contained Python distribution that does not interfere with any existing Python installation.

1. Navigate to Runtime Environment from the sidebar.

2. If the runtime is missing, click Install Runtime.

3. The download and extraction progress is displayed in real time with a progress bar and log output.

4. Once complete, the status updates to Success with the installed version number.

The runtime is approximately 1-2 GB in size. An internet connection is required for the initial download. The sidebar is locked during installation to prevent navigation.

Step 3: Verify GPU Availability

If you have an NVIDIA GPU installed, verify that it is correctly detected:

1. On the Runtime Environment screen, click Check GPU.

2. The application runs a diagnostic script that checks CUDA availability.

3. The result shows your GPU name and CUDA version, or indicates that no GPU was found.

AI inference without a GPU is possible but slower. If your GPU is not detected, ensure that NVIDIA drivers are properly installed and up to date.

Step 4: Connect the Digit Machine

Before recording, ensure your Digit machine hardware is connected:

1. Connect the Basler camera via USB.

2. Connect the ESP32 controller via USB serial.

3. Navigate to Acquisition Settings to verify the camera feed and hardware controls.

The Home Screen

After setup, the Home screen is your starting point. It displays six action cards:

Card	Description
Create	Create a new HDF5 analysis file
Open	Open an existing analysis file
Browse	View all analysis files in a folder
AI	Run AI classification on the current file
Blend & Compare	Compare multiple lots or simulate blends
Knowledge Base	Open the online documentation

The bottom of the screen shows two counters: Total AI Analyses run and Recording Working Hours accumulated.

Some cards require an open analysis file. Start by clicking Create to make a new file, or Open to load an existing one.

Creating a New Analysis

1. From the Home screen, click Create.

2. On the New Analysis screen, click Select Folder to choose where the HDF5 file will be saved.

3. Optionally enter a filename. If left blank, a timestamp-based name is generated automatically.

4. Click Create.

The folder path must not contain special characters (spaces, accents, symbols). If an invalid path is detected, a warning dialog appears asking you to rename the folder or move the file.

The application creates the HDF5 file and navigates you to the Record Analysis screen to begin capturing seed images.

Opening an Existing Analysis

1. From the Home screen, click Open.

2. In the file dialog, navigate to your .hdf5 file and select it.

3. The application reads the file summary and loads it into the current session.

4. You can now navigate to any screen (Dashboard, AI, Export, etc.) to work with the loaded data.

Browsing Multiple Analyses

1. From the Home screen, click Browse.

2. Select a folder containing multiple .hdf5 files.

3. The Browse screen displays cards for each analysis file with key metrics (defects, color, entropy, analysis date).

4. Sort by any field (ascending or descending) using the sort controls at the top.

5. Click any card to load that analysis.

Your First Complete Workflow

A typical first analysis follows these steps:

1. Create a new analysis file.

2. Record seed images using the Digit machine camera.

3. Run AI classification with your ONNX model.

4. Review results on the Dashboard and Image Mosaic screens.

5. Export a PDF report or raw data for further analysis.

Record AnalysisAI Classification