基于人工智能的风电叶片破损识别及破损面积计算的算法设计

Algorithm Design of Wind Power Blade Damage Identification and Damage Area Calculation Based on Artificial Intelligence

针对风电工程中对设备保持措施中叶片破损区域的监控与识别,本文提出了一种基于YOLO神经网络识别算法以实现对航拍图片中的风场中的风机叶片破损区域进行自动识别。本文还配套设计了另外一种基于传统计算机视觉算法的破损面积提取和计算算法。通过本文提出的算法可以实现对风机叶片破损区域及其变化智能化、自动化的识别、跟踪及管理。文中算法已进行了测试实验,实验结果表明了算法具有有效性和精确性。

For monitoring and identification of the wind turbine blade damaged areas in equipment maintenance measures of the wind power projects, one kind of recognition algorithm based on YOLO neural network is proposed to automatically identifies the wind turbine blade damaged areas by the aerial photos of the wind field. Another kind of algorithm which is based on traditional computer vision algorithms has been designed to match the befor one to extract and calculate the damaged areas of the wind turbine blade. The algorithms proposed in this paper can intelligently and automatically identify,track and manage the damaged areas and their changes of the wind turbine blade. The proposed algorithms also have been tested, the experimental results showed that the designed algorithms are effective and accurate.