Segment a colored part with an HSV mask

What you need

Use Python, OpenCV, saved images and a controlled-light dataset including shadows and empty backgrounds.

The calculation

coverage = selected_pixels / ROI_pixels

Coverage is a fraction within a defined region of interest. It is a diagnostic statistic, not proof that the selected region is the correct part.

Worked example

Try it step by step

1. Save representative images and convert with cvtColor using the correct BGR-to-HSV conversion for OpenCV-loaded images.
2. Choose bounds from sampled part pixels and account for hue wraparound, especially for red.
3. Use inRange to create the mask, then inspect false detections on background-only images before contour extraction.
4. Store threshold settings with the lighting and camera configuration and validate on images not used to choose bounds.

Offline starter code

This snippet processes local data only; it sends no robot commands.

import cv2
image = cv2.imread("part.png")
if image is None:
    raise FileNotFoundError("part.png")
hsv = cv2.cvtColor(image, cv2.COLOR_BGR2HSV)
mask = cv2.inRange(hsv, (35, 70, 40), (85, 255, 255))
print("Selected fraction:", (mask > 0).mean())
# Example green bounds only; inspect and tune on your images.

How to check the result

Measure missed parts and background detections across part orientations. Check that selected regions remain spatially plausible, not only large enough.

Common mistake to avoid

OpenCV’s common 8-bit HSV hue range is 0–179, not 0–360. Do not mix thresholds from different representations.

Reference reading

Primary references for the underlying models, APIs or application context. The worked numbers and plots above are educational calculations, not results reported by these sources.

• OpenCV — inRange and color segmentation

Read our methods, limitations and safety notes.