Vision Angle Detector 📐✨

A computer vision-powered Streamlit web application that isolates schematic or diagrammatic regions within scanned documents or images, mathematically extrapolates intersecting line segments, and computes precise interior and exterior angles.

Built with OpenCV, Streamlit, and Matplotlib, this tool is explicitly engineered to handle real-world scanned artifacts, page borders, and low-contrast lines dynamically via an interactive calibration interface.

Below is the link for the Streamlit App: https://vision-angle-detector.streamlit.app/

---

🚀 Key Features

• Document-Agnostic Input: Seamlessly upload and process multi-page PDFs, PNGs, JPGs, or JPEGs.
• Dimension-Agnostic Processing: Utilizes array-flattening strategies to remain fully compatible across varying local and cloud versions of OpenCV (opencv-python vs opencv-python-headless).
• Advanced Geometry Pipeline:
  • Locates and isolates the primary diagram using automated ellipse-fitting contour segmentation.
  • Automatically filters out page margins, text blocks, and document scanning frames.
  • Extrapolates line segments mathematically to resolve hidden vertex intersection points.
• Reflex Angle Analytics: Visualizes both the acute interior sweep and the full-circle exterior remainder (Exterior Angle = 360° - Interior Angle).
• Cascade Fallback Engineering: If severe artifacting or noise prevents line segment parsing, the app gracefully falls back to displaying your localized boundary ellipse overlay instead of throwing a fatal execution error.
• Interactive Calibration Panel: Real-time sidebar sliders for tuning edge-detection thresholds, morphological line-stitching, and boundary size limits.
• Integrated Data Logs: Features an on-the-fly table displaying evaluated segment lengths and normalized headings alongside optional Tesseract-powered OCR text blocks.

---

🛠️ System Requirements & Prerequisites

Because this application relies on low-level system bindings for rendering documents and processing text, you must ensure the following non-Python binary dependencies are available on your host environment:

1. Poppler: Required by pdf2image to unpack and rasterize PDF pages.
2. Tesseract OCR (Optional): Required by pytesseract to view document text layouts.

Local Environment Setup

🍏 macOS (via Homebrew)

brew install poppler tesseract