🧠 ObjectSegmenter – Modular Object Segmentation with OpenCV (C++)

ObjectSegmenter is a modular, reusable C++ library built on top of OpenCV for industrial-grade object segmentation and geometric analysis.

It is designed to be plugged into real-time computer vision pipelines (e.g., inspection, counting, pick-and-place, defect detection) and supports:

• Preprocessing (grayscale, blur, thresholding, morphology)
• Contour-based object detection
• Centroid, area, perimeter, and bounding box extraction
• Annotated output for visual inspection
• Easy integration into threaded pipelines or embedded systems

---

🚀 Features

• 🔲 Robust Binary Segmentation: Otsu thresholding with Gaussian blur and optional inversion
• 📦 Contour-Based Object Detection: Uses cv::findContours() in RETR_EXTERNAL mode
• 🧮 Geometric Measurements: Area, perimeter, bounding box, centroid
• 🖼️ Visual Output: Contours, bounding boxes, centroids, object ID overlays
• 🧩 Modular Design: Clean C++ class with no global state — plug-and-play friendly

---

📁 File Structure

/project-root
  │
  ├── ObjectData.h # Struct to hold per-object data
  ├── ObjectSegmenter.h # Header for segmentation module
  ├── ObjectSegmenter.cpp # Implementation of segmentation logic
  ├── main.cpp # Example usage entry point 
  └── README.md # This file

---

🛠️ Dependencies

• OpenCV 4.x
• C++17 standard compiler (tested with g++ 11+)
• Linux/macOS/Windows compatible

Install OpenCV with:

sudo apt install libopencv-dev

🔧 Build & Run

g++ main.cpp ObjectSegmenter.cpp -o object_segmenter `pkg-config --cflags --libs opencv4` -std=c++17
./object_segmenter

Ensure your image file (e.g., objects.jpg) is in the project directory.

🧪 Example Use Case

ObjectSegmenter segmenter;
std::vector<ObjectData> results;
cv::Mat output;

segmenter.process(inputImage, output, results);

for (const auto& obj : results) {
    std::cout << "Object " << obj.id << ": Area = " << obj.area << ", Centroid = " << obj.centroid << std::endl;
}

🔄 Future Extensions (WIP Ideas)

• ObjectFilter module (by area, shape, orientation)
• Data export to CSV/JSON/SQL
• Integration with cv::dnn for defect classification
• Camera calibration and metric measurement (mm)

📄 License

MIT License — use it freely in commercial or academic projects.