Dependencies Management

PiTrac relies on several external libraries, some requiring builds from source for compatibility with Raspberry Pi hardware and OS versions.

Core Dependencies

PiTrac’s dependencies as defined in Software/LMSourceCode/ImageProcessing/meson.build:

Dependency Required Version Source Notes
OpenCV ≥4.9.0 Built from source Debian provides 4.6.0
Boost ≥1.74.0 System package Threading, logging, filesystem modules
ActiveMQ-CPP Any Built from source Not in Debian repos
lgpio Any Built from source Pi 5 GPIO compatibility
libcamera System System package Kernel-dependent
msgpack-cxx Any Built from source Serialization library
yaml-cpp Any System package Configuration parsing
fmt Any System package String formatting
OpenSSL Any System package Security/encryption
APR-1 Any System package ActiveMQ dependency

Build Scripts

All dependency build scripts are located in packaging/scripts/:

  • build-opencv.sh - Builds OpenCV 4.11.0 for ARM64
  • build-activemq.sh - Builds ActiveMQ-CPP 3.9.5
  • build-lgpio.sh - Builds lgpio library
  • build-msgpack.sh - Builds msgpack-cxx
  • build-all-deps.sh - Orchestrates all dependency builds

Building All Dependencies 

cd packaging
./scripts/build-all-deps.sh

This creates artifacts in packaging/deps-artifacts/:

  • opencv-4.11.0-arm64.tar.gz
  • activemq-cpp-3.9.5-arm64.tar.gz
  • lgpio-0.2.2-arm64.tar.gz
  • msgpack-cxx-6.1.1-arm64.tar.gz

Docker-Based Build System

PiTrac uses Docker containers to build dependencies reproducibly:

OpenCV Build

The packaging/Dockerfile.opencv builds OpenCV 4.11.0 with:

  • DNN support for YOLO models
  • TBB for parallelism
  • V4L/libcamera support
  • Selected modules: core, imgproc, imgcodecs, calib3d, features2d, highgui, videoio, photo, dnn, objdetect

ActiveMQ-CPP Build

The packaging/Dockerfile.activemq builds ActiveMQ-CPP 3.9.5 with APR dependencies.

lgpio Build

The packaging/Dockerfile.lgpio builds lgpio from the joan2937/lg repository for Pi 5 GPIO support.

msgpack-cxx Build

The packaging/Dockerfile.msgpack builds msgpack-cxx from the cpp_master branch as a header-only library.

Build System

Primary: Meson/Ninja

PiTrac uses Meson as its primary build system:

cd Software/LMSourceCode/ImageProcessing
meson setup build --buildtype=release
ninja -C build pitrac_lm

Package Building

The main packaging orchestrator:

cd packaging
./build.sh build     # Build PiTrac binary
./build.sh all       # Build everything including dependencies

APT Package Creation 

cd packaging
./build-apt-package.sh

Creates a .deb file in packaging/build/package/.

Platform Differences

Raspberry Pi 4 vs Pi 5

Aspect Pi 4 Pi 5
Config location /boot/config.txt /boot/firmware/config.txt
Camera tools libcamera-* rpicam-*
GPIO chip 0 4
Camera package libcamera-apps rpicam-apps

Architecture Support

  • Primary: ARM64 (aarch64)
  • Secondary: ARMv7 (32-bit ARM)
  • Development: x86_64

Dependency Resolution

PKG-Config

Ensure libraries are discoverable:

export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig:$PKG_CONFIG_PATH
pkg-config --modversion opencv4

Library Loading

Runtime library path setup:

export LD_LIBRARY_PATH=/usr/local/lib:$LD_LIBRARY_PATH
ldd /usr/bin/pitrac_lm  # Verify library linking

Troubleshooting

OpenCV Not Found 

# Check installation
pkg-config --modversion opencv4

# Set environment
export OpenCV_DIR=/opt/opencv/lib/cmake/opencv4
export PKG_CONFIG_PATH=/opt/opencv/lib/pkgconfig:$PKG_CONFIG_PATH
sudo ldconfig

Missing Dependencies 

# Check what's missing
meson setup build  # Will report missing dependencies

# Debug library loading
LD_DEBUG=libs ./pitrac_lm 2>&1 | head -20

Build Failures 

# Clean rebuild
rm -rf build
meson setup build --buildtype=release

# Check for architecture mismatches
file build/*.o  # Should all be same architecture

Package Installation

The APT package installs:

  • Binary: /usr/lib/pitrac/pitrac_lm
  • CLI: /usr/bin/pitrac
  • Config: /etc/pitrac/
  • Test images: /usr/share/pitrac/test-images/
  • Services: systemd units for pitrac and tomee

Summary

PiTrac’s dependency management uses Docker for reproducible builds, creating reusable artifacts that speed up development. The combination of system packages and custom-built libraries ensures compatibility across different Raspberry Pi models while maintaining optimal performance.