A Steel Wire Rope Defect Detection Method for Gantry Crane in Gravity Energy Storage Systems

Gravity energy storage is a sustainable energy storage technology widely used in power systems and the energy industry. Steel wire ropes are critical components of gravity energy storage systems, and their safety and reliability are vital for the operation of the energy storage system. In recent years, significant progress has been made in defect detection in the field of deep learning-based object detection methods. This study presents an enhanced YOLOv7 model-based approach for wire rope defect detection. Firstly, this paper applies mosaic data augmentation to a portion of the image data, followed by mixup data augmentation on a subset of the mosaic-augmented data, to strengthen the resilience and generalizability of the enhanced model. Additionally, this paper adopts feature-level cascaded reuse in the backbone of the YOLOv7 model to address the issue of information loss from lower-level networks to higher-level networks. Experimental results demonstrate that in contrast to the original YOLOv7 model, the enhanced model achieves a 4.3% increase in mAP@0.5 and improves the detection of small defects in steel wire ropes.