MQTT Data Interaction and Reporting

This document primarily describes how to configure the MQTT/MQTTS function of NeoEyes NE301, use the MQTTX client to receive data reported by the device, and provides a detailed analysis of the reported data protocol format. This will help you better understand the data sent by the device for upper-layer application development.

Introduction

NeoEyes NE301 supports reporting AI inference results and device-related information to a specified MQTT Broker via the MQTT protocol. Developers can subscribe to the corresponding Topic to obtain real-time data for secondary development or system integration.

Preparation

Before starting, please ensure the following preparations are complete:

1. NE301 Device: Powered on and connected to the network (ensure the device allows network communication with the MQTT Broker).
2. MQTT Broker: An MQTT server (such as EMQX, Mosquitto, etc.) is deployed, or use a public test server.
3. MQTT Client Tool: We recommend using MQTTX for debugging and data viewing.

If you are not familiar with MQTT protocol, please refer to MQTT Protocol for learning.

1. Configure NE301 MQTT Parameters

Login to the NE301 Web management interface, go to the "Application Management" -> "MQTT/MQTTS" page, and configure the following parameters:

• Broker Address: MQTT server IP address or domain name.
• Port: Server port (default is 1883).
• Client ID: Device unique identifier (optional, usually defaults to the device serial number. Please ensure it does not duplicate with other client IDs to avoid conflicts).
• Username / Password: If the server requires authentication, enter the username and password.
• Data Receiving Topic: Topic for data downlink (e.g., ne301/2A207D/down/control), used for the device to receive control commands. The device generates one by default; modify if customization is needed.
• Data Reporting Topic: Topic for data reporting (e.g., ne301/2A207D/upload/report), used for the device to report data. The device generates one by default; modify if customization is needed.

After configuration, click "Save", and the device will attempt to connect to the MQTT server.

Switch to MQTTS configuration if encryption is required.

2. Configure MQTTX Client

Open the MQTTX software and click "+" to create a new connection:

1. Name: Custom connection name (e.g., NE301_Monitor).
2. Host: Enter the same MQTT Broker address configured on the device.
3. Port: Enter the port (e.g., 1883).
4. Click the "Connect" button to connect to the server.

After successful connection, add a subscription:

1. Click "New Subscription".
2. Topic: Enter the reporting Topic configured on the device (e.g., ne301/device/data).
3. Click "Confirm".

3. Trigger Capture and Receive Data

Trigger Methods

You can trigger device capture and inference in the following ways:

• Web UI: Click the capture button on the "Live" page.
• Physical Button: Press the camera button on the side of the device.
• Sensor/IO Trigger: Trigger capture via the IO interface.
• Scheduled Trigger: Set up scheduled tasks.
• Remote Control: Trigger capture remotely via MQTT control commands.

Receiving Data

When the device completes inference, the MQTTX client will receive a message in JSON format. As shown below:

4. Data Protocol Format Analysis

The data reported by the device uses standard JSON format, containing device information, timestamps, and AI inference results.

Example Data

{
  "metadata": {
    "image_id": "cam01_1766132582",
    "timestamp": 1766132582,
    "format": "jpeg",
    "width": 1280,
    "height": 720,
    "size": 61880,
    "quality": 60
  },
  "device_info": {
    "device_name": "NE301-2A207D",
    "mac_address": "44:9f:da:2a:20:7d",
    "serial_number": "SN202500001",
    "hardware_version": "V1.1",
    "software_version": "1.0.1.1064",
    "power_supply_type": "battery",
    "battery_percent": 41,
    "communication_type": "wifi"
  },
  "ai_result": {
    "model_name": "YOLOv8 Nano Object Detection Model - COCO 80 Classes (Int8)",
    "model_version": "1.0.0",
    "inference_time_ms": 50,
    "confidence_threshold": 0.5,
    "nms_threshold": 0.5,
    "ai_result": {
      "type": 1,
      "detections": [
        {
          "index": 0,
          "class_name": "bottle",
          "confidence": 0.57646054029464722,
          "x": 0.11801555752754211,
          "y": 0.052199184894561768,
          "width": 0.27234357595443726,
          "height": 0.93958532810211182
        }
      ],
      "detection_count": 1,
      "poses": [],
      "pose_count": 0,
      "type_name": "object_detection"
    }
  },
  "image_data": "data:image/jpeg;base64,...",
  "encoding": "base64"
}

Field Description

Field	Type	Description
metadata	Object	Image metadata
metadata.image_id	String	Unique Image ID
metadata.timestamp	Number	Capture Timestamp (Unix)
metadata.format	String	Image Format
metadata.width	Number	Image Width
metadata.height	Number	Image Height
metadata.size	Number	Image Size (Bytes)
metadata.quality	Number	Image Compression Quality
device_info	Object	Device Information
device_info.device_name	String	Device Name
device_info.mac_address	String	MAC Address
device_info.serial_number	String	Device Serial Number
device_info.hardware_version	String	Hardware Version
device_info.software_version	String	Software Version
device_info.power_supply_type	String	Power Supply Type
device_info.battery_percent	Number	Battery Percentage
device_info.communication_type	String	Communication Type
ai_result	Object	AI Inference Result Information
ai_result.model_name	String	Model Name (includes quantization details)
ai_result.model_version	String	Model Version
ai_result.inference_time_ms	Number	Inference Time (ms)
ai_result.confidence_threshold	Number	Confidence Threshold
ai_result.nms_threshold	Number	NMS Threshold
ai_result.ai_result	Object	Specific Inference Data
ai_result.ai_result.type	Number	Result Type ID
ai_result.ai_result.type_name	String	Task Type Name
ai_result.ai_result.detection_count	Number	Count of Detected Objects
ai_result.ai_result.detections	Array	List of Detections
detections[].index	Number	Object Index
detections[].class_name	String	Object Class Name
detections[].confidence	Number	Confidence Score (0-1)
detections[].x	Number	Normalized Center X
detections[].y	Number	Normalized Center Y
detections[].width	Number	Normalized Width
detections[].height	Number	Normalized Height
image_data	String	Base64 Encoded Image Data
encoding	String	Encoding Method

By parsing this JSON data, you can easily obtain information about objects in the image and the current status of the device.

5. FAQ

1. The data reported by the NE301 device is not what I want, what should I do?

Currently, the NE301 device does not support modification of the data format. If you need to customize the reported data format or related application logic, you may consider secondary development or contact us for custom development.

2. MQTT is configured, but the device is not connected to the MQTT server, what should I do?

Please check if the device is configured with the correct MQTT parameters, including Broker address, port, username, and password. If the configuration is correct but the device still cannot connect to the MQTT server, please check if the MQTT server is running normally and ensure the device can access the network.

3. MQTT is connected, but no data is reported, what should I do?

Please troubleshoot whether the device's current network is stable and if it can access the internet. Also, check if the MQTT server is running normally. After confirmation, disconnect the MQTT connection and try again.