Grove Vision AI v2 Kit User Manual

1. Introduction

The Grove Vision AI v2 Kit is an advanced MCU-based vision AI module designed for edge computing and AI applications. Powered by the Himax WiseEye2 HX6538 processor, it features a dual-core Arm Cortex-M55 and an integrated Arm Ethos-U55 neural network unit, offering significant boosts in on-device machine learning performance with low power consumption. This module supports popular AI frameworks like TensorFlow and PyTorch, and is compatible with various development platforms including Arduino, Raspberry Pi, and Seeed Studio XIAO ESP32C3.

2. Setup Guide

2.1 Unpacking the Kit

Carefully unpack your Grove Vision AI v2 Kit. Depending on your purchase, the kit may include:

• Grove Vision AI v2 Kit:
  • Grove Vision AI v2 Module: 1 unit
  • OV5647 Camera Module: 1 unit
  • Seeed Studio XIAO ESP32 C3: 1 unit
• Grove Vision AI v2 (Module Only):
  • Grove Vision AI v2 Module: 1 unit
  • Cable: 1 unit

2.2 Hardware Connection

Follow these steps to connect the components:

1. Connect the Camera: The Grove Vision AI v2 module is compatible with all Pi cameras through a standard CSI interface. Connect the OV5647 Camera Module to the CSI connector on the Grove Vision AI v2 module.
2. Connect to a Host Controller (e.g., XIAO ESP32 C3): Use the provided cable or appropriate connectors to link the Grove Vision AI v2 module to your chosen host controller (e.g., Seeed Studio XIAO ESP32 C3, Arduino, Raspberry Pi). The module features various interfaces like IIC, UART, SPI, and a Type-C port for connectivity.
3. Power Supply: Ensure your host controller or the Grove Vision AI v2 module itself is powered correctly via its Type-C port or through the host controller's power supply.

2.3 Software Setup

The Grove Vision AI v2 supports various software environments:

• SenseCraft AI: For no-code/low-code model deployment and immediate visualization of identification results. This web-based platform simplifies the process of deploying pre-trained or custom AI models.
• Development Environments: Compatible with Arduino IDE, Micropython, CircuitPython, and PlatformIO for integration with XIAO series and other microcontrollers.
• AI Frameworks: Supports TensorFlow and PyTorch for advanced model development and deployment.

3. Operating Instructions

3.1 Deploying AI Models with SenseCraft AI

SenseCraft AI provides a streamlined 4-step process for quick model deployment:

1. Connect Device: Connect your Grove Vision AI v2 module to your computer.
2. Select Model: Choose from off-the-shelf or your custom AI models available on the SenseCraft AI platform.
3. Click Deploy: Initiate the deployment process to transfer the selected model to your device.
4. Check Inference Results: Visualize the identification results directly on the SenseCraft AI website to evaluate model performance.

3.2 Programming with Microcontrollers

For more customized applications, you can program the Grove Vision AI v2 module using various development environments:

• Arduino: Integrate with Arduino IDE for projects requiring C/C++ programming.
• Micropython/CircuitPython: Utilize these Python-based environments for rapid prototyping and development, especially when paired with XIAO boards.
• PlatformIO: A versatile ecosystem for embedded development that supports multiple platforms and frameworks.

The module's rich interfaces (IIC, UART, SPI) allow for flexible communication with host controllers and other peripherals.

4. Maintenance

To ensure the longevity and optimal performance of your Grove Vision AI v2 Kit, follow these maintenance guidelines:

• Handle with Care: Electronic components are sensitive. Avoid dropping the module or exposing it to physical shock.
• Keep Dry: Protect the module from moisture and humidity. Do not operate it in wet environments.
• Cleanliness: Keep the module free from dust and debris. Use a soft, dry brush or compressed air for cleaning. Avoid using liquids or harsh chemicals.
• Storage: When not in use, store the module in an anti-static bag in a cool, dry place.

5. Troubleshooting

If you encounter issues with your Grove Vision AI v2 Kit, consider the following troubleshooting steps:

• No Power/Not Responding:
  • Check all power connections. Ensure the Type-C cable is securely connected and the power source is adequate.
  • Verify that the host controller (e.g., XIAO, Raspberry Pi) is powered on and functioning correctly.
• Camera Not Detected/No Image:
  • Ensure the CSI camera cable is correctly and securely inserted into both the camera module and the Grove Vision AI v2 module.
  • Check if the camera module itself is functional.
• Model Deployment Issues:
  • Verify your internet connection if using SenseCraft AI.
  • Ensure the correct device is selected and recognized by the SenseCraft AI platform or your development environment.
  • Check for any error messages during the deployment process and consult the documentation for specific error codes.
• Communication Errors (IIC, UART, SPI):
  • Double-check wiring between the Grove Vision AI v2 module and your host controller.
  • Verify that the communication protocols and addresses are correctly configured in your software.

6. Specifications

6.1 Hardware Overview Diagram

7. Applications

The Grove Vision AI v2 Kit is suitable for a wide range of applications, particularly in edge computing and AI, including but not limited to:

• Industrial Automation: Quality inspection, predictive maintenance, voice control.
• Smart City: Equipment monitoring, energy management.
• Transportation: Status monitoring, location tracking.
• Intelligent Agriculture: Environmental monitoring.
• Mobile IoT Devices: Wearable devices, handheld devices.
• Advanced Vision AI: Facial Keypoint Detection, Human Pose Estimation, Object Detection (e.g., using Yolo v8n model).

8. User Tips

• Compatibility: The Grove Vision AI v2 module is highly compatible with popular development boards such as Arduino, ESP32, and Raspberry Pi, making it versatile for various projects.
• SenseCraft AI for Beginners: For users new to AI vision, SenseCraft AI offers an intuitive no-code/low-code platform to quickly get started with model deployment and testing.
• Power Efficiency: With its low power consumption (0.35W), the module is ideal for battery-powered endpoint AI applications.

9. Documentation and Support

For detailed technical information and further assistance, please refer to the following resources:

• HX6538 Datasheet: Consult the official datasheet for in-depth specifications of the Himax WiseEye2 HX6538 processor.
• Himax examples for Seeed Grove Vision AI Module V2: Explore example code and projects to get started with your development.
• User Manual (PDF): A comprehensive user manual in PDF format is available for download. You can access it via this link: Grove Vision AI v2 User Manual (PDF).

For additional support, visit the official Seeed Studio website or community forums.