- 1. Overview
- 2. Requirements
- 3. Object Detection Workflow
- 4. Setup Instructions
- 5. Download the Model from Qualcomm AI Hub
- 6. Prepare the Application
- 7. Run the Object Detection Application
This demo showcases the edge AI capabilities of the Arduino® UNO Q using a trained model from Qualcomm AI Hub. This application enables real-time object detection from a live video feed captured by a USB webcam.
⚠️ Important: This demo must be run in Network Mode or SBC within the Arduino App Lab.
- Arduino® UNO Q
- USB camera (x1)
- USB-C® hub adapter with external power (x1)
- A power supply (5V, 3A) for the USB hub (e.g., a phone charger)
- Personal computer with internet access
flowchart LR
A[Start] --> B[Setting Up Arduino UNO Q Device]
B --> C[Download Required Models]
C --> D[Create the Object Detection Application]
D --> E[Run the Object Detection Application]
E --> F[✅ Object Detection Active]
Before proceeding further, please ensure that all the setup steps outlined below are completed in the specified order. These instructions are essential for configuring the various tools required to successfully run the application.
Each section provides a reference to internal documentation for detailed guidance. Please follow them carefully to avoid any setup issues later in the process.
Arduino App Lab enables you to create and deploy Apps directly on the Arduino® UNO Q board, which integrates both a microcontroller and a Linux-based microprocessor. The App Lab runs seamlessly on personal computers (Windows, macOS, Linux) and comes pre-installed on the UNO Q, with automatic updates. Please follow the setup instructions carefully to ensure smooth development and deployment of Apps.
For detailed steps, refer to the documentation: Set up Arduino App Lab
Arduino Flasher CLI provides a streamlined way to flash Linux images onto your Arduino UNO Q board. Please follow the setup instructions carefully to avoid flashing errors and ensure proper board initialization.
For detailed steps, refer to the documentation: Arduino Flasher CLI
Arduino UNO-Q must be properly configured to ensure reliable communication with the host system and accurate sensor data acquisition. Please follow the setup instructions carefully to avoid hardware conflicts and ensure seamless integration with the software stack.
For detailed steps, refer to the documentation: Set up Arduino UNO-Q.
Qualcomm AI Hub empowers you to discover pre‑optimized models, optimize and profile your own AI models, and seamlessly deploy them across Qualcomm‑powered devices for efficient on‑device AI.
Open Models → IoT, select Object Detection from the left‑hand category list, locate YOLOX, enter its model page, and download the TFLite → float version of the model.
This section will guide you on how to create a new application from an empty project, write your own components—including Python code, HTML code, and imported official bricks—set up the Python environment, and build a complete App from scratch. Starting from a clean project gives you full control over the structure and design of your application; however, for first‑time users, we recommend trying the built‑in examples first to understand the workflow before returning to this guide.
Arduino App Lab allows you to start from a clean, empty project and build your application entirely from scratch, giving you full flexibility over its structure and functionality. This section will guide you through creating a brand‑new application, adding your own Python code and HTML code, importing official bricks, setting up the Python environment, and assembling everything into a functional App ready for deployment on the Arduino UNO Q.
In this example, we are starting with a completely new project to demonstrate how to build an application step‑by‑step.
For detailed steps, refer to the documentation: Create a New App
After creating a new application in App Lab, you can extend its functionality by adding a Web UI interface. To do this, open your application and click the “Bricks” button located in the upper-left corner of the App Lab interface. From the available brick categories, select WebUI-HTML to insert an HTML brick into your project.
This HTML brick will serve as the entry point for your user interface and will automatically link to the index.html file you created earlier under the assets directory.
For detailed steps, refer to the documentation: Understanding the Brick Concept
This step guides you through downloading the required application files from Qualcomm’s official GitHub repository. The repository contains multiple demo projects, so we use Git sparse checkout to fetch only the files needed for this demo, keeping the workspace clean and lightweight.
-
Create your working directory:
mkdir my_working_directory cd my_working_directory -
Download Your Application:
git clone -n --depth=1 --filter=tree:0 https://github.com/qualcomm/Startup-Demos.git cd Startup-Demos git sparse-checkout set --no-cone /CV_VR/IoT-Robotics/Object_detection_via_Tflite_model/ git checkout
Once you have obtained the optimized model in the previous step, transfer it to the Arduino UNO Q to enable real‑time inference and seamless integration within your application. Upload both the model file (e.g., yolox-float.tflite) and the require file (e.g., coco_labels.txt, main.py) to the application directory on the device.
You can upload the files from your host PC using either ADB or SCP, depending on your setup:
# Host PC → UNO Q (via ADB)
adb push <path/to/file> <destination/path/on/device>
# Host PC → UNO Q (via SCP)
scp <path/to/file> arduino@<UNOQ_IP>:<destination/path/on/device>
Upload location:Make sure to upload the model file to:/home/arduino/ArduinoApps/<app name>
For detailed steps, refer to the documentation: File Transfer
Before running the application, verify that all required files are present in your project directory. If you followed this guide, your folder should include (you may have additional files, but these are mandatory):
/home/arduino/ArduinoApps/<app name>/
├─ python/
│ ├─ main.py
│ ├─ yolox-float.tflite
│ ├─ coco_labels.txt
│ └─ requirements.txt
└─ assets/
└─ index.html
Once your application is configured and built in Arduino App Lab, it can be deployed and executed directly on the Arduino UNO Q. This section will guide you through launching the application, verifying camera input, and observing real‑time object recognition.
