| 基于红外图像边缘检测算法的风电机组叶片裂纹视觉检测技术 |
 点此下载全文 |
| 引用本文：杨德海.基于红外图像边缘检测算法的风电机组叶片裂纹视觉检测技术[J].计算技术与自动化,2025,(3):117-122 |
| 摘要点击次数: 331 |
| 全文下载次数: 211 |
|
|
| 中文摘要:风电机组叶片损伤无损检测过程容易受到外界环境干扰，细节处理效果较差，F1值较低。因此，提出了基于红外图像边缘检测算法的风电机组叶片裂纹视觉检测技术。将实时采集的风电机组叶片红外图像变换为灰度图像，再利用引导滤波器进行线性移变滤波，得到边缘细节增强的图像。应用最大类间方差法选取最优图像分割阈值，从原始图像中分割出待检测的叶片目标。引入熵分块算法和最优阈值选取算法，设计改进基于改进Canny算法的边缘检测算法，通过图像自动分块、子图像高低阈值计算等操作，获取红外图像边缘信息，基于此提取叶片裂纹图像的周长、长轴、短轴等特征。运用改进萤火虫算法优化后的支持向量机模型，对特征向量进一步分析，得出叶片裂纹视觉检测结果。实验结果表明：新研究技术应用后得出的裂纹检测结果F1值超过了0.9，满足了风电机组巡检要求。 |
| 中文关键词:红外图像 风电机组 叶片 边缘检测 裂纹检测 支持向量机 |
| |
| Visual Detection Technology for Wind Turbine Blade Cracks Based on Infrared Image Edge Detection Algorithm |
|
|
| Abstract:The non-destructive testing process of wind turbine blade damage is prone to external environmental interference, with poor detail processing effect and low F1 value. Therefore, a visual detection technology for wind turbine blade cracks based on infrared image edge detection algorithm is proposed. Transform the real-time collected infrared images of wind turbine blades into grayscale images, and then use a guided filter for linear shift filtering to obtain images with enhanced edge details. Apply the maximum inter class variance method to select the optimal image segmentation threshold and segment the leaf target to be detected from the original image. Introducing the entropy partitioning algorithm and the optimal threshold selection algorithm, designing and improving an edge detection algorithm based on the improved Canny algorithm. Through automatic image segmentation, sub image high and low threshold calculation, and other operations, infrared image edge information is obtained. Based on this, features such as perimeter, major axis, and short axis of the blade crack image are extracted. Using the optimized support vector machine model using the improved firefly algorithm, further analyze the feature vectors and obtain the visual detection results of blade cracks. The experimental results show that the F1 value of the crack detection result obtained after the application of the new research technology exceeds 0.9, meeting the inspection requirements of wind turbines. |
| keywords:infrared image wind turbines blades edge detection crack detection support vector machine |
| 查看全文 查看/发表评论 下载pdf阅读器 |
位访问者,今日访问
