A Fast Algorithm for Foil Recognition and Positioning Based on Monocular Stereo Vision

Foil is a passive jamming device composed of metal foil, metal wire or metal-coated medium. Due to its low cost, convenient use and wide frequency band, it has been widely applied in modern military fields such as radar jamming, missile penetration and electronic warfare. However, as the foil is thin and overlaps with each other during the transmission on the conveyor belt, some edges are occluded, thereby interfering with the recognition effect. Therefore, the recognition and positioning of foil in a dynamic environment by machine vision confronts significant challenges. The rapid and accurate recognition and positioning of foil is a key technology to ensure the precise grasping of the robotic manipulator. Although traditional methods such as manual, mechanical and sensor positioning have their respective advantages, they have limitations in aspects such as recognition accuracy, reliability and environmental adaptability. To address the issue of the robotic manipulator’s rapid and accurate grasping of foil during its transmission on the conveyor belt, this paper compares three object detection algorithms: Faster R-CNN, SSD and YOLOv8s.