How to Control the Serial Servo Motor Using the OpenCV and ESP32 Microcontroller

YouTube Video:

https://youtube.com/shorts/yfd6txKc3kE

Step 1:

GET YOUR ESP32:

Here i am using an ESP32-Wroom board as microcontroller that Serial Data from the Computer and send that data to the Servo motor to move through the Servo driver Below,

Step 2:

This is the main linker that help the servo to work and helping sending the signal to correct servo motor based on its unique #ID by default the ID of all servos are set to #ID 1 so if you want to control the multiple servo motors then please assign different #IDs to them.

this code here will help you change the ID of the servos easily here :

#include <SCServo.h>

SMS_STS st;

int ID_ChangeFrom = 1; // Change the original servo ID, and the factory default is 1

int ID_Changeto = 2; // new ID

void setup(){

Serial1.begin(1000000);

st.pSerial = &Serial1;

while(!Serial1) {}

st.unLockEprom(ID_ChangeFrom); //Unlock EPROM-SAFE

st.writeByte(ID_ChangeFrom, SMS_STS_ID, ID_Changeto);//Change ID

st.LockEprom(ID_Changeto); // EPROM-SAFE is locked

}

void loop(){

}

PINS Connection:

Tx of servo driver ---------> RX2 of Esp32 or GPIO 16

RX of servo Driver ---------> TX2 of Esp32 or GPIO 17

VCC of servo driver ---------> VCC of Esp32

GND of servo driver --------> GND of Esp32

here in ESP32 there are two serial pins so be careful to connect the RX and TX to serial1 or GPIO 16 (RX) and GPIO 17 (TX) are used for connection of the servo driver not the primary ones.

once done:

Step 3:

UPLOAD THIS CODE:

#include <SCServo.h>

SMS_STS st;

#define S_RXD 16

#define S_TXD 17

#define SERVO1_ID 2

#define SERVO2_ID 1

void setup() {

Serial.begin(115200); // USB serial to PC (Python)

Serial1.begin(1000000, SERIAL_8N1, S_RXD, S_TXD); // Servo serial bus

st.pSerial = &Serial1;

delay(1000);

Serial.println("SCServo Control Ready");

}

void loop() {

if (!Serial.available()) return;

String data = Serial.readStringUntil('\n');

data.trim();

int commaIndex = data.indexOf(',');

if (commaIndex < 0) return;

int angle1 = data.substring(0, commaIndex).toInt();

int angle2 = data.substring(commaIndex + 1).toInt();

// Limit safety angle

angle1 = constrain(angle1, 0, 180);

angle2 = constrain(angle2, 0, 180);

// Convert degrees → SCServo position

int pos1 = map(angle1, 0, 180, 0, 2000);

int pos2 = map(angle2, 0, 180, 0, 2000);

// Send high-resolution move command

st.WritePosEx(SERVO1_ID, pos1, 1500, 50);

st.WritePosEx(SERVO2_ID, pos2, 1500, 50);

}

Step 4:

Upload these Code here:

Then Copy Past This Python Code here and Run it in any Code Editors you wanna use :

also Download these Dependencies here:

📦 Python Libraries (Install using pip)

1. OpenCV

For webcam access & image processing

pip install opencv-python

2. Mediapipe

For face mesh tracking

pip install mediapipe

3. NumPy

For mathematical calculations

pip install numpy

4. PySerial

For sending servo data to ESP32 over USB

pip install pyserial

before that select the correct serial port and and exact baud rate you have selected for esp32 to communicate it with.

import cv2

import mediapipe as mp

import numpy as np

import serial

import time

# ✅ Serial to ESP32 (update port if needed)

ser = serial.Serial('/dev/ttyUSB0', 115200, timeout=1)

time.sleep(2)

# Initialize Mediapipe Face Mesh

mp_face_mesh = mp.solutions.face_mesh

mp_draw = mp.solutions.drawing_utils

drawing_spec = mp_draw.DrawingSpec(thickness=1, circle_radius=1)

face_mesh = mp_face_mesh.FaceMesh(min_detection_confidence=0.8,

min_tracking_confidence=0.8)

cap = cv2.VideoCapture(1)

ws, hs = 600, 600

cap.set(3, ws)

cap.set(4, hs)

if not cap.isOpened():

print("Camera error")

exit()

def calculate_head_movement(nose, left_eye, right_eye):

face_center_x = (left_eye[0] + right_eye[0]) / 2

yaw = np.interp(nose[0], [face_center_x - 50, face_center_x + 50], [180, 0])

eye_level_y = (left_eye[1] + right_eye[1]) / 2

pitch = np.interp(nose[1], [eye_level_y - 50, eye_level_y + 90], [0, 180])

return yaw, pitch

def smooth(current, target, factor=0.1):

return current*(1-factor) + target*factor

current_yaw = 180

current_pitch = 180

while cap.isOpened():

success, img = cap.read()

if not success:

break

img = cv2.flip(img, 1)

rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)

results = face_mesh.process(rgb)

img = cv2.cvtColor(rgb, cv2.COLOR_RGB2BGR)

if results.multi_face_landmarks:

face = results.multi_face_landmarks[0].landmark

nose = [face[1].x * ws, face[1].y * hs]

left_eye = [face[33].x * ws, face[33].y * hs]

right_eye = [face[263].x * ws, face[263].y * hs]

target_yaw, target_pitch = calculate_head_movement(nose, left_eye, right_eye)

current_yaw = smooth(current_yaw, target_yaw)

current_pitch = smooth(current_pitch, target_pitch)

yaw_i = int(np.clip(current_yaw, 0, 180))

pitch_i = int(np.clip(current_pitch, 0, 180))

# ✅ Send servo commands to ESP32

cmd = f"{yaw_i},{pitch_i}\n"

ser.write(cmd.encode())

cv2.putText(img, f"Yaw:{yaw_i} Pitch:{pitch_i}", (20,50),

cv2.FONT_HERSHEY_SIMPLEX, 1, (0,255,0),2)

cv2.imshow("Head Tracking Control", img)

if cv2.waitKey(1) & 0xFF == ord('q'):

break

cap.release()

cv2.destroyAllWindows()

ser.close()

and thats it when your web cam starts you will see you servos responding accordingly based on your head yaw and roll..

if you like this project please left and heart here ..

its my first post so i am very open for suggestions and typos

i am vikas

i Love Robotics... ; )

byee....