Component guide · Beginner · 4 min read
AI Vision Sensors

DFRobot HuskyLens and HUSKYLENS 2 Microscope, MU Vision, Pixy2.1 and Pan-Tilt, Seeed reCamera, Sipeed Maix, RP2040 and Artemis Camera Kits
By Geekobot Learning Team · Updated 29 Oct 2026
On this page
AI vision sensors do the recognition on the camera and send simple results to your board: what was seen and where. HuskyLens learns faces, objects, colours, lines and tags with one button; Pixy2.1 tracks coloured objects and lines fast; the MU Vision sensor suits classrooms; and the Seeed reCamera and Sipeed Maix boards run Linux AI models. The RP2040 and Artemis kits are for TinyML experiments.
Key specs at a glance
- Recognition
- Faces, objects, colours, lines, tags, QR (model dependent)
- Interface
- I2C and UART (HuskyLens, Pixy, MU); network (reCamera)
- Supply
- 3.3 to 5 V (HuskyLens, Pixy)
- Display
- Built-in screen on HuskyLens
- Pan-tilt
- Pixy kits with servos
- Boards
- Arduino, micro:bit, ESP32, Pi
Models in this guide
These are alternatives. Pick the one that fits your project; you only need one.
Option 1DFRobot SEN0305 Gravity HuskyLens K210 AI Vision Sensor w/ Display₹4,009
Option 2DFRobot SEN0683 HUSKYLENS 2 Microscope Camera Module 30x Magnification₹1,461
Option 3DFRobot SEN0314 MU Vision Sensor ESP32 AI Camera₹7,279
Option 4Pixy 2.1 Smart Vision Sensor Object Tracking Camera₹10,048
Option 5Pixy Pan and Tilt Mechanism for Pixy2.1 Smart Vision Camera₹3,289
Option 6Pixy Pan and Tilt Mechanism for Pixy 1.0 & CMUcam5₹2,900
What the HuskyLens, Pixy2 and AI vision sensors is
An AI vision sensor is a camera with its own processor that recognises faces, colours, lines or tags and sends only the result to your board.
With HuskyLens, point it at an object and press the learn button; your Arduino then reads ID and box position. That makes object-following robots possible in an afternoon.
Pixy2.1 runs at 60 fps for colour and line tracking and drives its pan-tilt servos itself.
HuskyLens, Pixy2 and AI vision sensors pinout
The pins on the HuskyLens, Pixy2 and AI vision sensors and what each one does.
| Pin | Function |
|---|---|
| T / R | UART TX and RX (HuskyLens) or I2C SDA and SCL |
| + / - | Power |
| ICSP / I2C (Pixy) | Arduino connection cable |
- T / R
- UART TX and RX (HuskyLens) or I2C SDA and SCL
- + / -
- Power
- ICSP / I2C (Pixy)
- Arduino connection cable
Wiring the HuskyLens, Pixy2 and AI vision sensors
Set the protocol on HuskyLens to I2C for this wiring.
| HuskyLens, Pixy2 and AI vision sensors pin | HuskyLens pin | Arduino Uno | ESP32 |
|---|---|---|---|
| T (SDA) | A4 | GPIO 21 | |
| R (SCL) | A5 | GPIO 22 | |
| + | 5 V | 3V3 or 5 V | |
| - | GND | GND |
- T (SDA)
- HuskyLens pin:
- A4
- Arduino Uno:
- GPIO 21
- R (SCL)
- HuskyLens pin:
- A5
- Arduino Uno:
- GPIO 22
- +
- HuskyLens pin:
- 5 V
- Arduino Uno:
- 3V3 or 5 V
- -
- HuskyLens pin:
- GND
- Arduino Uno:
- GND
Example code
Pixy2 uses the Pixy2 library (pixy.ccc.getBlocks()).
#include "HUSKYLENS.h"
#include <Wire.h>
HUSKYLENS hl;
void setup() {
Serial.begin(115200);
Wire.begin();
while (!hl.begin(Wire)) { Serial.println("HuskyLens not found"); delay(1000); }
}
void loop() {
if (hl.request() && hl.available()) {
HUSKYLENSResult r = hl.read();
Serial.printf("ID %d at x=%d y=%d w=%d h=%d\n", r.ID, r.xCenter, r.yCenter, r.width, r.height);
}
delay(100);
}Setup and tuning
use I2C or UART and set the same mode in the sensor's menu and in your code.
train objects or faces under the lighting where the robot will run.
fix the sensor rigidly at the height and angle you will use; moving it changes what it sees.
HuskyLens draws up to about 300 mA. Power it from the board's 5 V, not a weak 3.3 V pin.
read the box x, y, width and height to steer, for example keep the object in the centre.
What you need
Shop this list13 of 14 found in our store · ₹53,712
DFRobot SEN0305 Gravity HuskyLens K210 AI Vision Sensor w/ Display: the part this guide coversDFRobot SEN0305 Gravity HuskyLens K210 AI Vision Sensor w/ Display₹4,009
DFRobot SEN0683 HUSKYLENS 2 Microscope Camera Module 30x Magnification: another version this guide coversDFRobot SEN0683 HUSKYLENS 2 Microscope Camera Module 30x Magnification₹1,461
DFRobot SEN0314 MU Vision Sensor ESP32 AI Camera: another version this guide coversDFRobot SEN0314 MU Vision Sensor ESP32 AI Camera₹7,279
Pixy 2.1 Smart Vision Sensor Object Tracking Camera: another version this guide coversPixy 2.1 Smart Vision Sensor Object Tracking Camera₹10,048
Pixy Pan and Tilt Mechanism for Pixy2.1 Smart Vision Camera: another version this guide coversPixy Pan and Tilt Mechanism for Pixy2.1 Smart Vision Camera₹3,289
Pixy Pan and Tilt Mechanism for Pixy 1.0 & CMUcam5: another version this guide coversPixy Pan and Tilt Mechanism for Pixy 1.0 & CMUcam5₹2,900
Seeed Studio Studio reCamera 2002w 8GB RISC-V AI Vision Camera: another version this guide coversSeeed Studio reCamera 2002w 8GB RISC-V AI Vision Camera₹11,499Closest match, check the specs
Sipeed Maix I Binocular Camera Module Dual GC0328 30k Pixels: another version this guide coversSipeed Maix I Binocular Camera Module Dual GC0328 30k Pixels₹1,719
Sipeed Maix II M2 Dock AI Linux Vision Dev Board w/ 1.3-Inch IPS Screen: another version this guide coversSipeed Maix II M2 Dock AI Linux Vision Dev Board w/ 1.3-Inch IPS Screen₹5,339
Waveshare RP2040 Camera Dev Board w/ HM01B0 Grayscale Sensor & 1.14-Inch IPS LCD: another version this guide coversWaveshare RP2040 Camera Dev Board w/ HM01B0 Grayscale Sensor & 1.14-Inch IPS LCD₹1,449
SparkFun Artemis Development Kit with Himax Camera: another version this guide coversSparkFun Artemis Development Kit with Himax Camera₹3,889- Uno R3 development board: the controller used in the wiring tableFind this part
ESP32-WROOM-32 38-pin development board: for Wi-Fi logging and dashboardsESP32-WROOM-32 38-Pin Development Board₹389
TowerPro SG90 micro servo: for pan and tiltTowerPro SG90 9g Micro Servo (180°)₹442Closest match, check the specs
Alternatives and when to choose them
- OpenMV: programmable vision.
- ESP32 camera boards: Wi-Fi streaming.
- TinyML vision kits: train your own models.
Common mistakes to avoid
- Training objects under different light from where the robot runs.
- Setting I2C in code but UART in the sensor menu.
- Powering it from a pin that cannot supply enough current.
- Expecting it to identify many object types like a large AI model.
Frequently asked questions
Can HuskyLens recognise any object I show it?
It learns simple objects, faces, colours and tags with a button press, but not complex scenes like a full detection model.
Does it need a computer to work?
No. Recognition runs on the sensor itself, and your board only reads the results.
Which is better for line following, HuskyLens or Pixy2?
Both have line tracking. Pixy2 is very fast at colour and lines; HuskyLens is easier to teach and has more modes.
Can I use it with a micro:bit or ESP32?
Yes. HuskyLens and Pixy2 have libraries for Arduino, micro:bit, ESP32 and Raspberry Pi.
Why does recognition fail outdoors?
Strong sunlight and changing shadows confuse colour and object learning. Retrain in the real light or add shade.
Parts for this guide
Keep learning
Sources
Written by the Geekobot team from manufacturer datasheets. Always check the datasheet for the exact part and variant you buy.

