Deteksi Charging Port CCS Type 2 Menggunakan Metode Yolo-V2 Deep Learning untuk Kendaraan Listrik

The presentation "Deteksi Charging Port CCS Type 2 Menggunakan Metode YOLO-V2 Deep Learning pada MATLAB" showcases the development of an automatic detection system for CCS Type 2 charging ports using the YOLO-v2 algorithm. High-resolution images (1,331 total), captured with Intel RealSense D455 and iPhone 14 cameras, were processed to 416x416 pixels, labeled, and augmented to enhance training. Anchor box estimation achieved a mean IoU of 0.8411, and training was performed using SGDM for 20 epochs.
The model achieved a mean Average Precision (mAP) of 85% and IoU values above 0.5, demonstrating high accuracy and robustness. Precision remained strong across recall thresholds, with depth data improving localization accuracy. This real-time detection system effectively identifies CCS Type 2 charging ports, making it ideal for automated EV charging applications.