Small-Object Focused Augmentation for Electric Vehicle Charger Socket Detection and Tracking

Objective

Built upon the previous Semi-Supervised Video Object Segmentation for Electric Vehicle Charger Socket Tracking project by applying small-object-focused data augmentation to address missed detections when charger sockets appeared smaller or farther from the camera. This improved the model’s ability to detect small and hard-to-detect sockets while reducing missed detections in real-world scenarios.

Stack

Delivery highlights

• Built upon the previous Semi-Supervised Video Object Segmentation for Electric Vehicle Charger Socket Tracking project to address missed detections in real-world scenarios, where charger sockets could be detected reliably at close range but became difficult to detect when they were farther from the camera and appeared smaller. To improve this, the training dataset was enhanced with data augmentation focused on small objects, increasing the number of training samples that reflected challenging real-world conditions. This improvement helped the model learn better representations of small and hard-to-detect sockets, resulting in improved small-object detection performance and fewer missed detections compared with the original training setup.