Abstract:This paper presents a novel algorithm for non-destructive damage detection for steel ropes in high-altitude environments (aerial ropeway). The algorithm comprises two key components: First, a segmentation model named RGBD-UNet is designed to accurately extract steel ropes from complex backgrounds. This model is equipped with the capability to process and combine color and depth information through the proposed CMA module. Second, a detection model named VovNetV3.5 is developed to differentiate between normal and abnormal steel ropes. It integrates the VovNet architecture with a DBB module to enhance performance. Besides, a novel background augmentation method is proposed to enhance the generalization ability of the segmentation model. Datasets containing images of steel ropes in different scenarios are created for the training and testing of both the segmentation and detection models. Experiments demonstrate a significant improvement over baseline models. On the proposed dataset, the highest accuracy achieved by the detection model reached 0.975, and the maximum F-measure achieved by the segmentation model reached 0.948.

What problem does this paper attempt to address?

This paper attempts to solve the problem of non - destructive testing of wire ropes in high - altitude environments (such as aerial ropeways). Specifically, traditional optical detection methods have difficulties in detecting subtle anomalies, and face significant challenges in detecting wire ropes at high altitudes or in hard - to - reach locations. Therefore, the paper proposes a new dual - model algorithm, which combines a segmentation model and a detection model to improve the accuracy and efficiency of wire rope surface damage detection. ### Main contributions of the paper include: 1. **Proposing a new U - Net architecture network**: This network can process RGB and depth images as input, and effectively segment wire rope objects by introducing the CMA (Cross - Modal Attention) module. 2. **Designing a detection model**: This model is based on the VovNet architecture which is friendly to small objects, and enhances performance through the DBB (Diverse Branch Block) module to improve the ability to detect wire rope surface anomalies. 3. **Introducing a special wire rope data set**: This data set contains approximately 700 wire rope images and their corresponding labels for training and testing the segmentation and detection models. 4. **Proposing a new background enhancement method**: This method improves the performance of the segmentation model in actual scenarios, especially when facing unfamiliar natural backgrounds. ### Specific problem - solving methods: - **Segmentation model**: Use the RGBD - UNet model, and fuse color and depth information through the CMA module to effectively extract wire ropes from complex backgrounds. - **Detection model**: Use the VovNetV3.5 model, and enhance feature diversity through the DBB module to improve detection ability. - **Background enhancement**: Generate diverse training samples by randomly embedding wire rope targets into different natural backgrounds to improve the generalization ability of the model. ### Experimental results: - **Segmentation model**: Experiments on the Rope - Seg data set show that the RGBD - UNet model outperforms the baseline model in multiple evaluation metrics, with the highest F - measure reaching 0.948. - **Detection model**: Experiments on the Rope - Detect data set show that the VovNetV3.5 model outperforms other advanced models in multiple evaluation metrics, with the highest accuracy reaching 0.975. ### Conclusion: This research proposes a new dual - model algorithm, which combines segmentation and detection models, and effectively solves the problem of non - destructive testing of wire ropes in high - altitude environments. Future research can consider applying the CMA module and depth information to other U - structure models to further refine the classification of wire rope anomalies.

A new method for optical steel rope non-destructive damage detection

Steel Wire Rope Surface Defect Detection Based on Segmentation Template and Spatiotemporal Gray Sample Set

Visual Sensing Inspection for the Surface Damage of Steel Wire Ropes With Object Detection Method

Damage Object Detection of Steel Wire Rope-Core Conveyor Belts Based on the Improved YOLOv5

Detection of External Defects in Steel Wire Rope Based on Image Processing and Neural Network

Surface damage detection for steel wire ropes using deep learning and computer vision techniques

A Review of Non-Destructive Damage Detection Methods for Steel Wire Ropes

Non-destructive Testing of Wire Rope Algorithm Based on Lightweight YOLOv7-tiny

Quantitative Nondestructive Testing of Broken Wires for Wire Rope Based on Magnetic and Infrared Information

Study of Noise Reduction and Identification of Internal Damage Signals in Wire Ropes

Fault classification based on computer vision for steel wire ropes

Intelligent Visual Detection Method for the Early Surface Damage of Mine Hoisting Wire Ropes

Wire Rope Damage Detection Based on Magnetic Leakage and Visible Light

Quantitative detection of wire rope damage based on local structural characteristics

A Framework for Fault Detection Method Selection of Oceanographic Multi-Layer Winch Fibre Rope Arrangement

A New Signal Processing Method for the Nondestructive Testing of a Steel Wire Rope Using a Small Device

Visualization and Quantification of Broken Wires in Steel Wire Ropes Based on Induction Thermography

Quantitative Nondestructive Testing for Wire Rope Based on Multi-Source Information Fusion

Structure and Character Analysis of a New Type of Steel Wire Rope NDT Detector Apparatus

A more efficient method for wire rope surface defect detection based on fusing cross-stage features and lightweight convolution modules

Nondestructive Detection of Wire Rope Damage Using Leakage Magnetic Technique based on Dual-Layer Sensors