Back

Radar Soccer Goal Detection

Product Manager·Sep 2024 – Dec 2024·GitHub ↗

Built a radar data pipeline in MATLAB and Python to process 500+ simulated .mat files, deployed as a modular .h5 inference component. Implemented a ResNet50 classifier on composite 2D radar matrices — RCS and angle profiles merged — achieving 76% goal-line detection accuracy validated with F1-score and confusion matrix.

76% accuracy · 500+ .mat files · .h5 inference component

Stack & Methods

PythonMATLABPyTorchResNet50ML