Single-Pixel Camera

A low-cost single-pixel camera prototype built with an RGB sensor, Arduino, and stepper motors, paired with a MATLAB simulation for image reconstruction using compressive sensing.

Overview

Conventional cameras use millions of pixels to capture an image simultaneously. A single-pixel camera does the opposite — it captures one measurement at a time using structured light patterns, then reconstructs the full image mathematically. This makes it viable for imaging beyond the visible spectrum where traditional sensors are expensive or impractical.

How it works

1. A sequence of optimised light patterns is projected onto the scene
2. A single RGB sensor records the reflected intensity for each pattern
3. Lasso regression (compressive sensing) reconstructs the original image from the raw measurements
4. A MATLAB simulation validates the approach before physical construction

Results

• Simulation achieved near-perfect image reconstruction accuracy
• Physical prototype successfully captured recognisable low-resolution images within hardware budget constraints

Stack

MATLAB · C · Processing · Arduino

Structure

• simulation/ — MATLAB simulation code and image reconstruction pipeline
• hardware/ — 3D printed component files for the physical prototype
• docs/ — Full project report

Background

BEng Final Year Project — University of Kent (Distinction)