Raspberry Pi Robot Car: Python GPIO & Camera Vision Build

Building a Raspberry Pi robot car with Python GPIO and camera vision is one of the most rewarding projects for electronics hobbyists in India. You combine real-time motor control, sensor integration, and live computer vision into a single mobile platform — all programmed in Python. Whether you’re a student exploring robotics or a maker wanting a capable autonomous bot, this guide walks you through every step: hardware selection, wiring, Python code, and OpenCV-based obstacle avoidance.

Hardware Requirements

Before writing a single line of Python, gather all the hardware components. For a Raspberry Pi robot car with camera vision you will need:

• Raspberry Pi 4B or 3B+ — the main compute board running Python
• 2WD or 4WD robot chassis — acrylic or aluminium frame with DC gear motors
• L298N or L293D motor driver module — bridges the Pi’s 3.3 V GPIO to 6–12 V motors
• Raspberry Pi Camera Module v2 or a compatible USB webcam
• HC-SR04 ultrasonic sensor — for distance measurement and obstacle detection
• 18650 Li-Ion battery pack or 4× AA battery holder — 7.4 V for motors, 5 V USB power bank for the Pi
• Jumper wires, breadboard, and M3 standoffs
• MicroSD card (16 GB minimum) with Raspberry Pi OS (Bullseye or Bookworm)

The total cost for a beginner build in India lands between ₹2,500 and ₹5,000 depending on chassis quality and whether you use a Camera Module or a USB webcam.

2WD Mini Round Double-Deck Smart Robot Car Chassis DIY Kit

Compact two-wheel-drive chassis with dual DC gear motors, encoder wheels, and a double-deck acrylic frame — perfect for mounting a Raspberry Pi and camera module on top.

View on Zbotic

Chassis Assembly & Motor Wiring

Most Indian robot car kits come with all screws and standoffs included. Follow these assembly steps:

1. Attach the two DC gear motors to the bottom plate using M3 screws.
2. Press-fit the rubber wheels onto the motor shafts or use the included hex couplings.
3. Mount the caster wheel at the rear for three-point stability.
4. Secure the L298N motor driver on the second deck with M3 nylon standoffs.
5. Place the Raspberry Pi on the top deck, ensuring the camera slot faces forward.
6. Run motor cables from the chassis to the L298N OUT1–OUT4 terminals.

L298N Wiring to Raspberry Pi GPIO:

Important: use a separate power supply for the motors. Never power the L298N from the Pi’s 5 V rail — high motor current will brown-out the Pi and corrupt the SD card.

Raspberry Pi GPIO Motor Control with Python

Install the RPi.GPIO library (pre-installed on Raspberry Pi OS) and create a motor driver class:

import RPi.GPIO as GPIO
import time

# BCM numbering
IN1, IN2, IN3, IN4 = 17, 27, 22, 23
ENA, ENB = 18, 25

GPIO.setmode(GPIO.BCM)
GPIO.setwarnings(False)
GPIO.setup([IN1, IN2, IN3, IN4, ENA, ENB], GPIO.OUT)

pwm_a = GPIO.PWM(ENA, 1000)  # 1kHz PWM
pwm_b = GPIO.PWM(ENB, 1000)
pwm_a.start(0)
pwm_b.start(0)

def set_motors(left_speed, right_speed):
    """speed: -100 to 100 (negative = reverse)"""
    # Left motor
    GPIO.output(IN1, left_speed > 0)
    GPIO.output(IN2, left_speed  0)
    GPIO.output(IN4, right_speed < 0)
    pwm_b.ChangeDutyCycle(abs(right_speed))

def forward(speed=70):  set_motors(speed, speed)
def backward(speed=70): set_motors(-speed, -speed)
def turn_left(speed=60):  set_motors(-speed, speed)
def turn_right(speed=60): set_motors(speed, -speed)
def stop():              set_motors(0, 0)

PWM speed control lets you fine-tune torque — vital for smooth turning and line following. Duty cycle 70 works well for most 200 RPM gear motors at 7.4 V.

Camera Module Setup

Enable the camera interface in raspi-config → Interface Options → Camera (for legacy camera support) or use the new libcamera stack on Bookworm.

Install OpenCV:

pip3 install opencv-python-headless picamera2
# Or for legacy picamera:
pip3 install picamera opencv-python-headless

Capture a frame with picamera2:

from picamera2 import Picamera2
import cv2

cam = Picamera2()
cam.configure(cam.create_preview_configuration(
    main={"size": (320, 240), "format": "RGB888"}
))
cam.start()

while True:
    frame = cam.capture_array()
    # frame is a NumPy array — ready for OpenCV
    cv2.imshow("Robot View", frame)
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break

Use 320×240 resolution for real-time processing at 30+ fps on a Pi 4. Higher resolutions drop frame rate and increase latency.

OpenCV Line Following & Obstacle Avoidance

Line following with OpenCV uses HSV colour thresholding to detect a black line on a white surface:

import cv2
import numpy as np

def get_line_error(frame):
    # Crop bottom third for near-ground view
    h, w = frame.shape[:2]
    roi = frame[int(h*0.6):h, :]

    gray = cv2.cvtColor(roi, cv2.COLOR_RGB2GRAY)
    _, thresh = cv2.threshold(gray, 60, 255, cv2.THRESH_BINARY_INV)

    M = cv2.moments(thresh)
    if M["m00"] > 500:
        cx = int(M["m10"] / M["m00"])
        error = cx - w // 2   # positive = line is to the right
        return error
    return None

Kp = 0.3  # proportional gain
base_speed = 60

while True:
    frame = cam.capture_array()
    error = get_line_error(frame)

    if error is not None:
        correction = int(Kp * error)
        set_motors(base_speed - correction, base_speed + correction)
    else:
        stop()  # line lost — halt

This proportional controller steers the robot to keep the line centred in the camera view. Tune Kp — too high causes oscillation, too low causes slow correction.

Ultrasonic Sensor Integration

The HC-SR04 provides distance data for obstacle stopping:

TRIG, ECHO = 24, 25
GPIO.setup(TRIG, GPIO.OUT)
GPIO.setup(ECHO, GPIO.IN)

def get_distance_cm():
    GPIO.output(TRIG, False)
    time.sleep(0.01)
    GPIO.output(TRIG, True)
    time.sleep(0.00001)
    GPIO.output(TRIG, False)

    t1 = time.time()
    while GPIO.input(ECHO) == 0:
        t1 = time.time()
    t2 = time.time()
    while GPIO.input(ECHO) == 1:
        t2 = time.time()

    return (t2 - t1) * 17150  # distance in cm

In your main loop, call get_distance_cm() and stop the motors if the reading falls below 20 cm.

60MM-K Mecanum Wheel (Pack of 4) – Black

Upgrade your robot car to omnidirectional movement. These 60mm mecanum wheels let the Pi robot strafe sideways — ideal for advanced maze-solving projects.

View on Zbotic

Complete Python Code — Autonomous Robot Car

Combining all modules into a single script:

#!/usr/bin/env python3
import RPi.GPIO as GPIO, cv2, time
from picamera2 import Picamera2

# --- GPIO setup (as above) ---
# ... (motor and ultrasonic setup)

cam = Picamera2()
cam.configure(cam.create_preview_configuration(
    main={"size": (320, 240), "format": "RGB888"}
))
cam.start()
time.sleep(1)

Kp, base_speed = 0.3, 60

try:
    while True:
        dist = get_distance_cm()
        if dist < 20:
            stop()
            time.sleep(0.5)
            backward(50)
            time.sleep(0.4)
            turn_right(60)
            time.sleep(0.35)
            continue

        frame = cam.capture_array()
        error = get_line_error(frame)
        if error is not None:
            correction = int(Kp * error)
            set_motors(base_speed - correction,
                       base_speed + correction)
        else:
            stop()
except KeyboardInterrupt:
    stop()
    GPIO.cleanup()
    cam.stop()

4mm Hex Coupling for Robot Smart Car Wheel (30mm)

Securely couple DC motor shafts to robot wheels. These 4mm hex couplings fit most standard 2WD/4WD chassis motors and ensure zero slippage at full speed.

View on Zbotic

Troubleshooting Tips

• Motors not spinning: Check ENA/ENB jumpers on L298N — they must be removed if you are using PWM, or shorted for full-speed mode.
• Pi keeps rebooting: Motor power draw is causing USB power supply brownout. Always power Pi and motors separately.
• Camera not detected: Run vcgencmd get_camera (legacy) or libcamera-hello. Ensure flat-flex ribbon is seated firmly.
• OpenCV slow on Pi 3: Drop resolution to 160×120 and convert to grayscale before processing. Use THRESH_BINARY instead of Canny edge detection.
• HC-SR04 giving erratic readings: Use a 1kΩ + 2kΩ voltage divider on the ECHO pin — 5 V logic can damage Pi GPIO.

ACEBOTT ESP32 Tank Robot Car Expansion Pack (QD001–QD004)

Take your robot to the next level with this ESP32-powered tank chassis expansion. Add camera vision, IR sensors, and WiFi remote control right out of the box.

View on Zbotic

Frequently Asked Questions