An Image Processing Method for Dynamic Monitoring of Wind Turbine Blade Operation Based on Ground Multi‐Synchronous Camera Capture

With the increase in the capacity of wind turbine units, the length of their blades has significantly grown. Existing machine vision‐based image acquisition methods are unable to capture the full view of the blades, leading to the inability to accurately monitor the operational status of wind turbine blades dynamically. Therefore, this study proposes an image processing method for dynamic monitoring of wind turbine blade operation based on point and line features from images captured by ground multi‐synchronous cameras. After image dehazing filtering and edge detection, the method utilizes line detection, so as to extract line features. Additionally, on the basis of improved Harris and Scale‐Invariant Feature Transform (SIFT) registration, constraints such as line feature constraints, parallel constraints, and equidistant intercept constraints are incorporated for blade stitching. This process involves stitching fragmented images into a complete image, along with image enhancement, and assessing seam smoothness using root mean square error. Results indicate that this method can effectively capture and retain complete blade information, particularly blade edge information and blade damage information. The method proposed in this paper integrates machine vision, image recognition, and trajectory tracking technologies to construct a database of blade operating conditions and images. It is capable of operating in hazy environments, enabling accurate monitoring of blade operating conditions, supporting intelligent maintenance of wind farms, improving resource utilization, and reducing operating costs. Compared with existing methods, it has certain advantages. The research findings are expected to provide valuable data support for detecting potential defects or damages on the blades using machine vision‐based methods.