CCV (Chenguang Computer Vision) is a comprehensive computer vision library that provides foundational mathematics, kinematics, dynamics, estimation, and core CV algorithms.

The core algorithmic implementation is designed to have no third-party dependencies (no OpenCV, Eigen, or PCL in core logic), ensuring it is highly portable and cross-platform. Alongside the core library, this repository contains a variety of wrappers, demos, and GUI applications (C++, Python, Java, C#) demonstrating integrations with major industry tools like OpenCV, PCL, OpenGL, Pangolin, Qt, and ROS.

• Zero Dependencies: Core modules are written from scratch for maximum portability.
• Modular Design: Compile only what you need using CMake flags.
• High Performance: Uses a FLOAT macro to toggle between single and double precision.
• Robust Testing: Validated against industry standards (Eigen, OpenCV) using GTest.
• Well Documented: Mathematical formulas are explicitly annotated in Doxygen comments with LaTeX.

Basic math methods, matrices, and vectors.

• Random Number Generation
• Matrix & Vector Structures
• Interpolation (Linear, Bilinear)
• SVD Decomposition

Rotation matrices, Hamilton quaternions, and Euler angles.

• Rotation Matrix
• Quaternion (Hamilton)
• Euler Angle (12 conventions)
• Converters (with small-angle approximations)

State estimation for SLAM, filters, and optimization.

• Filters (EKF, Particle Filter)
• MAP (Gauss-Newton, Levenberg-Marquardt)
• Bundle Adjustment (Sparse Hessian)
• Solver Interfaces (Ceres, G2O, GTSAM)

2D/3D data structures and core vision algorithms.

• Basic Structures (Size, Point2, Point3)
• 2D Image Processing (Copy, Gaussian Filter, Pyramids)
• 2D Features (FAST)
• 2D Features (ORB)
• Camera Models & Binocular Stereo Vision

# Configure the project
# Flags: -DBUILD_TEST=ON, -DBUILD_DOCS=ON, -DWITH_OPENCV=ON, -DWITH_PCL=ON, -DWITH_GL=ON
cmake -S ./ -B build -DBUILD_TEST=ON

# Build and install
cmake --build build --target install/strip --parallel $(($(nproc) / 4))

# (For CMake >= 3.15)
# cmake --install build --prefix /opt/ccv

catkin build -j$(nproc) -DWITH_ROS=ON -DWITH_PCL=ON ccv

• CVStudio: A comprehensive GUI application for computer vision using Qt.
• MedAI: Medical Artificial Intelligence project integrating computer vision and deep learning.
• CamKit: A collection of camera utilities (Android, V4L2, MIPI, RealSense, ROS).
• Language Wrappers:
  • Python: Available via pip install libccv. See python/libccv/img_match_graph.py for match visualization.
  • Java/C#: Desktop Image Processing (DIP) demos included.
• Industry Demos: Integration examples with OpenCV, PCL, OpenGL, Pangolin, and FFTW.

C# DIP Demo on Ubuntu

The codebase follows the Google C++ Code Style. For detailed guidelines on contributions, testing, and documentation, see CONTRIBUTING.md.

• Unit Tests: Located in unit_test/, powered by GTest.
• Documentation: Generate Doxygen docs with cmake --build build --target docs.

This project is licensed under the MIT License - see the LICENSE file for details.