Blue Sign Location & Recognition

Kenneth Dawson-Howe's Computer Vision course.

Description

The problem is to develop a program in C++ to locate and recognise 81 information signs from Dublin train stations in the 33 provided images.

Environment Setup

• This project uses openCV v3.2. Unix download is available from OpenCV.
• CMake available from CMake.
• In order to #include filesystem with Clang compiler you wil need Boost library available from Boost.
• You may have to add export OpenCV_DIR=/usr/local/share/OpenCV to your bash_profile so that CMake will know where to find OpenCV libraries.

Building and running

1. Make a build directory and navigate to it to keep your workspace clean.

   mkdir build

   cd build

2. cmake ..

3. Build the binary executable.

   make

4. To run the program, you must be in the base directory (i.e. the one containing the "Blue Signs" folder)

   cd ..

   build/blueSignRecognition

Class Diagram

• ObjectAndLocation::compareObjects() perfroms recognition of blue signs.
• ImageWithBlueSigns::LocateAndAddAllObjects() performs detection of blue signs.

Theory

Summary of procedure for detection of blue signs.

Summary of procedure for recognition of blue signs.

Results