Raspberry Pi Projects in Bangalore for ECE

A Raspberry Pi 4B runs a face recognition attendance system using OpenCV for face detection and dlib's ResNet-based 128-dimensional face descriptor for recognition. Enrolled faces are stored as descriptor vectors in a SQLite database. The system processes a 720p camera frame in 280 ms, achieving real-time recognition at 3–4 fps on the Pi 4. On a successful match, the employee name and timestamp are logged to the database and displayed on a 7-inch HDMI touchscreen. A Flask web interface allows administrators to enrol new faces, view attendance logs, and export reports as CSV.

A Raspberry Pi Zero 2W controls a two-wheeled robot that follows a black line on a white surface using five IR sensors and avoids obstacles using an HC-SR04 ultrasonic sensor. The line-following algorithm uses a weighted centroid calculation to generate proportional steering corrections sent to two N20 gear motors via an L298N driver. When an obstacle is detected within 20 cm, the robot executes a pre-programmed avoidance manoeuvre. The entire control loop runs at 50 Hz in a Python asyncio event loop. A Pi Camera streams a live MJPEG feed to a browser for remote monitoring.

A MobileNet-SSD model quantised to INT8 and converted to TensorFlow Lite format runs on a Raspberry Pi 4B with a Pi Camera Module 3. The model detects 80 COCO object classes at 8 fps on the CPU and 22 fps with a Coral USB Accelerator. Detected objects are annotated with bounding boxes and confidence scores on a live preview window. Detection events are logged to a CSV file with timestamps and object counts. A Flask API endpoint returns the latest frame with annotations as a JPEG, enabling integration with a web dashboard or mobile app.

A Raspberry Pi 3B+ runs a home automation hub using Home Assistant OS with a custom MQTT integration. Four ESP8266 nodes control lights, fans, and a door lock relay, publishing state updates and subscribing to command topics. A Snowboy hotword detector running on the Pi listens for "Hey Pi" and routes voice commands to a local Vosk speech recognition engine, converting speech to MQTT commands without cloud dependency. A Node-RED dashboard provides a web UI for manual control and automation rule configuration. The system operates entirely on the local network with no internet dependency.

A Raspberry Pi 4B running Suricata IDS monitors network traffic on a home or lab network by operating in inline IPS mode on a USB Ethernet adapter. Custom Suricata rules detect port scans, SSH brute-force attempts, and known malware C2 communication patterns. Alerts are forwarded to an ELK stack (Elasticsearch, Logstash, Kibana) running on the same Pi via a Docker Compose deployment. A Kibana dashboard visualises alert frequency, source IP geolocation, and rule hit counts over time. The project demonstrates network security monitoring, IDS rule writing, and log analytics.

A Raspberry Pi Zero W drives a 32×16 LED matrix display using the rpi-rgb-led-matrix library. Content — text messages, scrolling announcements, and simple graphics — is managed through a Flask web application accessible on the local network. Administrators log in, type a message, and it appears on the display within 2 seconds. A cron job fetches weather data from the OpenWeatherMap API every 30 minutes and displays current temperature and conditions between announcements. The project demonstrates GPIO-based display driving, web application development on embedded Linux, and API integration.

A Raspberry Pi 4B with two Pi Camera modules runs a motion-detection CCTV system using OpenCV's background subtraction (MOG2 algorithm). When motion is detected in a defined region of interest, a 10-second video clip is saved to a USB drive and a thumbnail is sent to a Telegram bot. Clips older than 7 days are automatically deleted to manage storage. A live MJPEG stream is accessible on the local network via a Flask server. The system boots automatically on power-up via a systemd service and recovers from camera disconnection errors without manual intervention.

A Raspberry Pi 3B+ reads temperature, humidity (DHT22), barometric pressure (BMP280), wind speed (anemometer via GPIO pulse counting), and rainfall (tipping bucket rain gauge) every minute. Data is stored in a SQLite database and visualised on a Plotly Dash web application showing real-time gauges and 30-day trend charts. A daily summary email is sent via SMTP at 8 AM with min/max values and a 7-day forecast fetched from the Open-Meteo API. The project demonstrates multi-sensor data acquisition, time-series storage, and web-based data visualisation on a single-board computer.