This project implements a distributed image processing system using a Raspberry Pi 3, combining parallel filtering algorithms, network communication, hardware-based visualization, and classic encryption for data security. The system applies filters such as Sobel, Blur, Grayscale, and RGB enhancement (Red, Green, Blue), and computes the Frobenius norm before and after processing. The architecture integrates C-based software with OpenMPI, a custom kernel-space driver, and a real-time GPIO I/O interface for hardware interaction.
-
🔧 Custom Linux Kernel Driver in C
The system includes a character driver (gpio_driver.c) that enables:- Writing 3-digit numbers to 7-segment displays via GPIO.
- Reading a 3-bit binary code from GPIO input pins, which selects the filter to apply.
- Modular and efficient control of multiple displays and input lines from user space.
-
🖼️ Distributed Image Processing with OpenMPI
Using MPI, the BMP image is split into horizontal segments and processed in parallel:- Available filters: Sobel, Blur, Grayscale, Red, Green, Blue.
- Results are gathered and the Frobenius norm is computed in a distributed fashion.
-
🔐 Caesar Cipher for BMP Image Encryption
Prior to distributing data, a custom Caesar-based encryption with per-byte shift is applied to ensure image integrity during transmission. -
📤 Client-Server Image Transfer
Images are transferred using TCP/IP from a client to the processing server, avoiding shared file dependencies. -
📺 Embedded GPIO Interface
- Final Frobenius norm is displayed on the 7-segment hardware.
- User can interact with the system using physical buttons to choose the desired filter.