Back to Portfolio Source

Demo · Embedded Imaging

Automated Bacterial
Colony Counter

An end-to-end embedded imaging system on Raspberry Pi: controlled LED illumination, a custom power PCB, and an OpenCV pipeline that segments and counts colonies with two-layer anomaly detection (statistical filtering + a Random Forest classifier). Below is an interactive walkthrough of the touch GUI: capture, quantify, and review, using real annotated plate output.

PythonOpenCVRaspberry Piscikit-learnFlaskKiCad
Interactive mockup of the operator GUI. The live system runs headless on a Raspberry Pi with a camera and LED ring. This recreates the capture → quantify → review flow with genuine pipeline output.
Plate Imaging System · localhost:5000
Live Preview Live
Agar plate under LED illumination
Ready
Backlight
Brightness
80%
Image
Contrast
1.1×
Saturation
1.0×
Recent
No captures yet captured thumbnail
Results
Capture a plate, then run Quantify Selected.
Total Count
39
Anomalies
4
Mean Area
2.14 mm²
Confluence
7.8 %
#92.9 mm² · circ 0.71Irregular
#293.4 mm² · circ 0.63ML anomaly
#202.1 mm² · circ 0.92Clean
#51.8 mm² · circ 0.95Clean