The High-Precision Detection Method for Insulators’ Self-Explosion Defect Based on the Unmanned Aerial Vehicle with Improved Lightweight ECA-YOLOX-Tiny Model

Abstract

Aiming at the application of the overhead transmission line insulator patrol inspection requirements based on the unmanned aerial vehicle (UAV), a lightweight ECA-YOLOX-Tiny model is proposed by embedding the efficient channel attention (ECA) module into the lightweight YOLOX-Tiny model. Some measures of data augmentation, input image resolution improvement and adaptive cosine annealing learning rate are used to improve the target detection accuracy. The data of the standard China power line insulator dataset (CPLID) are used to train and verify the model. Through a longitudinal comparison before and after the model improved, and a cross-sectional comparison with other similar models, the advantages of the proposed model are verified in terms of multi-target identification for normal insulators, localization for small target defect areas, and the parameters required for calculation. Finally, the comparative analysis between the proposed ECA-YOLOX-Tiny model and YOLOV4-Tiny model is given by introducing the visualization method of class activation mapping (CAM). The comparative results show that the ECA-YOLOX-Tiny model is more accurate in locating the self-explosion areas of defective insulators, and has a higher response rate for decision areas and some special backgrounds, such as the overlapping small target insulators, the insulators obscured by tower poles, or the insulators with high-similarity backgrounds.

Research on Manufacturing Equipment Operation State Evaluation Technology Based on Fractional Calculus

“…Liu et al [28] proposed a target detection algorithm based on the improved RetinaNet, which is suitable for transmission-line defect detection and improved the intelligent detection accuracy of UAV in power systems. Cheng et al [29] proposed a lightweight ECA-YOLOX-Tiny model by embedding an efficient channel attention (ECA) module into the lightweight YOLOX-Tiny model, which has a higher response rate for decision areas and special backgrounds, such as overlapping small target insulators, insulators obscured by tower poles, or insulators with high-similarity backgrounds. Liu Wenqiang et al [30] introduced a point cloud segmentation and recognition method based on three-dimensional convolutional neural networks (3-D CNNs) to determine the different components of the catenary cantilever devices.…”

High Accuracy Detection Technique for Information Data Under the Influence of Multiple Factors based on Fractional Partial Differential Equations

“…From the analysis of the amplitude frequency characteristics of fractional differential operators, we can know that in the signal enhancement stage, when the influencing factors reach a certain value, the fractional differential operator can fuse the differences between the detection data caused by the influencing factor effectively, so, the fractional calculus is used to describe many phenomena in engineering and science [13,14]. The fractional differential equations have been used enormously for last two decades because of their varied applications in many spheres of physical and biological sciences [15][16][17], but it can only be applied to processing one-dimensional signal. In the process of information data collection, the detection information will be affected by many factors, such as testing equipment performance, equipment working environment, et al Therefore, we need to apply the fractional partial differential equations to expand the fractional differential operators in multi-dimensional space.…”

“…Jun Liu et al [15] proposed a target detection algorithm based on the improved RetinaNet which is suitable for transmission lines defect detection, improved the intelligent detection accuracy of UAV in power system. Ru Chengyin et al [16] proposed a lightweight ECA-YOLOX-Tiny model by embedding the efficient channel attention (ECA) module into the lightweight YOLOX-Tiny model, which has a higher respons rate for decision areas and some special backgrounds, such as the overlapping small target insulators, the insulators obscured by tower poles, or the insulators with high-similarity backgrounds. Liu Wenqiang et al [17] introduced a point cloud segmentation and recognition method based on three-dimensional convolutional neural networks (3-D CNNs) to determine the different components of the catenary cantilever devices.…”

Research on Manufacturing Equipment Operation State Evaluation Technology Based on Fractional Calculus

“…Liu et al [ 29 ] proposed a target detection algorithm based on improved RetinaNet, which is suitable for transmission-line defect detection and improves the intelligent detection accuracy of UAV in power systems. Cheng et al [ 30 ] proposed a lightweight ECA-YOLOX-Tiny model by embedding an efficient channel attention (ECA) module into it, which has a higher response rate for decision areas and special backgrounds, such as overlapping small target insulators, insulators obscured by tower poles, or insulators with high-similarity backgrounds. Liu Wenqiang et al [ 31 ] introduced a point cloud segmentation and recognition method based on three-dimensional convolutional neural networks (3-D CNNs) to determine the different components of the catenary cantilever devices.…”