📏 OMS — Object Measurement System

A standalone desktop application for measuring physical objects using a webcam and a quarter as a calibration reference. Built with Python, PyQt5, and OpenCV.

---

Features

• Live camera feed at ~30fps via OpenCV
• Automated calibration using a US quarter (24.26mm diameter) as a reference object
• Automatic object detection using Otsu thresholding and contour analysis
• Per-object measurements — width, height, area, and angle
• Annotated result images with measurement overlays
• Running measurement table showing all captures in the current session
• Auto-save after every capture — latest data is always on disk
• Session naming — user-defined session name with full timestamp retained in data
• Calibration file management — save, load, and review past calibrations

---

Requirements

Python

Python 3.8 or higher

Dependencies

PyQt5
opencv-python
numpy
pandas

Install all dependencies with:

pip install PyQt5 opencv-python numpy pandas

---

Folder Structure

OMS desktop app\
├── oms_app.py
├── images\
│   ├── bq_logo.png
│   └── ruler.png
├── data\
│   ├── calibration\
│   └── results\
└── README.md

Data folders are created automatically on first run if they do not exist.

---

Running the App

cd "C:\Users\james\OneDrive\Documents\Object Measurement System\OMS desktop app"
python oms_app.py

---

Packaging as a Desktop App

To build a standalone .exe:

pip install pyinstaller
pyinstaller --onefile --windowed --icon=images\ruler.ico oms_app.py

The executable will be created at dist\oms_app.exe. Right click it and select Send to > Desktop (create shortcut) to add a desktop icon.

---

Workflow

Step 1 — Calibration

1. Click Detect cameras to scan for available cameras
2. Select your camera from the dropdown and click Start camera
3. Enter a session name when prompted (defaults to today's date)
4. Either load an existing calibration from the list or click Run new calibration

Note: OMS should be recalibrated for each new measurement setup. To maximize measurement quality, ensure even lighting and minimize vertical distance between the camera lens and the measurement object.

Step 2 — Calibration Capture (new calibration only)

1. Place a US quarter at each of the 5 indicated positions on screen
2. Click Capture at each position — 5 frames are captured automatically per location
3. OMS detects the quarter in each frame and calculates pixels per mm
4. Review the result at each position — accept or retake

Step 3 — Calibration Summary

• Displays mean pixels per mm, standard deviation, and coefficient of variation (CV)
• CV below 2% is shown in green; above 2% is shown in amber as a quality warning
• Calibration is saved automatically to data/calibration/

Step 4 — Capture Measurements

1. Place your object in the camera's field of view
2. Click Capture — OMS detects all objects in the ROI and measures them
3. Results are displayed in the table and the annotated image is shown alongside the live feed
4. Repeat for as many measurements as needed
5. Click Done when finished

---

Key Constants

Constant	Value	Description
QUARTER_DIAMETER_MM	24.26	Diameter of a US quarter in mm
MIN_CONTOUR_AREA	100	Minimum contour area in pixels to detect as an object
DECIMAL_PLACES	3	Decimal places shown in measurements
CAPTURES_PER_LOCATION	5	Frames captured per calibration location
CAPTURE_DELAY_MS	400	Delay between calibration frames in ms
CV_WARNING_THRESHOLD	2.0	CV% above which a calibration quality warning is shown
PREVIEW_W / PREVIEW_H	640 / 480	Live feed preview resolution

---

Output Files

All output is saved to data/results/session_<name>\:

File	Description
results.csv	All measurement data with timestamps
raw_<n>.png	Raw captured image for each measurement
result_<n>.png	Annotated result image for each measurement

Calibration files are saved to data/calibration/ as .json files containing timestamp, pixels per mm, and CV.

---

Troubleshooting

Camera not detected

• Click Detect cameras again
• Try a different camera index in the dropdown
• Ensure no other application is using the camera

High calibration CV

• Ensure the quarter is flat and fully visible
• Improve lighting — avoid shadows and glare
• Reduce distance between camera and surface

Objects not detected

• Ensure good contrast between objects and background
• A plain white or black background works best
• Check that objects are within the green ROI boundary shown on screen

---

Credits

Developed by BrokenQuality