Monocular 3D perception pipeline detecting humans and estimating metric (X, Y, Z) coordinates

ACME Monocular Perception & 3D Spatial Tracking Module

Engineered a production-grade C++ monocular perception library for ACME Robotics fusing YOLOv5 2D object detection with the ‘Depth Anything’ foundation model. Delivers real-time 3D human obstacle detection and metric spatial localization from a single RGB camera stream, supported by a comprehensive CMake, CTest, Cppcheck, and Doxygen CI/CD suite.

October 18, 2025