1_10

Building C++ Libraries for SLAM

This tutorial covers how to build and use C++ libraries - an essential skill for SLAM development where you'll work with Eigen, OpenCV, PCL, g2o, GTSAM, and more.

How to build

Dependencies: None (pure C++17)

Local build:

mkdir build
cd build
cmake ..
make -j

Docker build:

docker build . -t slam_zero_to_hero:1_10

How to run

Local:

./build/use_static_lib
./build/use_shared_lib
./build/use_header_only

Docker:

docker run -it --rm slam_zero_to_hero:1_10

---

Topics Covered

1. Library Types

Type	Extension	Linking	Use Case
Static	.a (Linux), .lib (Windows)	Compile-time	Eigen, small libs
Shared	.so (Linux), .dll (Windows)	Run-time	OpenCV, PCL
Header-only	.hpp / .h	Include only	Eigen, Sophus

2. CMake Concepts

• add_library(name STATIC/SHARED sources...) - Create a library
• target_include_directories() - Specify header locations
• target_link_libraries() - Link libraries together
• find_package() - Find installed libraries
• PUBLIC/PRIVATE/INTERFACE - Visibility of dependencies

3. Project Structure

project/
├── CMakeLists.txt
├── include/
│   └── mylib/
│       └── mylib.hpp      # Public headers
├── src/
│   └── mylib.cpp          # Implementation
└── examples/
    └── main.cpp           # Usage example

---

Why This Matters for SLAM

• Eigen: Header-only library, just include and use
• OpenCV: Shared library, needs find_package(OpenCV) and linking
• g2o/GTSAM: Can be static or shared, need proper CMake setup
• Your own SLAM modules: Often organized as libraries for reusability