Getting Started

Operation Steps

1. Set device address;
2. Set output mode;
3. Version Check;
4. Configure hardware settings;
5. Configure algorithm settings;
6. Read detected result.

Set Device Address

Set the device address of MU sensor via DIP switch, refer to address selection.

Set Output Mode

Set the commincation mode via DIP switch, refer to Output mode selection.

Each time you change the output mode or device address, you must restart the sensor by pressing the reset button.

Version Check

Read the PROTOCOL_VER and FIRMWARE_VER registers to confirm that the firmware version matches the version documented in the user manual.
If there is a mismatch, register functions and algorithm behaviors may differ.

Setup Hardware

Set hardware-related registers per the actual application and requirements.
Key registers include:

• CAMERA_CONF1
• LED1
• LED2

Setup Vision

Algorithm for each VISION_ID has its own configuration registers. Set VISION_ID before setting the algorithm parameters.

Get results Data via Registers

List of included Vision Algorithms

Detection Result Data:

(1) id: Vision ID

(2) num: the number of detection results [0 or 1]

(3) x: normalized horizontal center coordinate [0...100]

(4) y: normalized vertical center coordinate [0...100]

(5) width: the normalized width of the object [0...100]

(6) height: the normalized height of the object [0...100]

(7) label: classification label number

Precautions for Using Vision Algorithms

Image recognition performance is highly affected by environmental factors such as lighting, color, and background. To achieve more accurate detection results, please follow these guidelines:

• Avoid using the vision sensor in dark environments, under a single direct light, or facing strong backlight (e.g., window light).
• Do not point the camera toward a bright light source.
• Prevent strong light or direct sunlight from shining on the detection target to avoid specular highlights.
• Avoid colored lights or rapidly changing lighting; a stable, uniformly distributed white light source is recommended.
• Do not use color-sensitive algorithms against backgrounds with similar colors (e.g., avoid placing a green tennis ball on a green carpet).
• Avoid environments with visually similar objects (e.g., avoid detecting a ping-pong ball near an orange).