Research On Wind Blades Based On Multispectral Images

As a critical part of wind turbine generator, wind blade affects wind turbines security to a large extent. With the development of wind power generation, the detection and maintenance of wind blades raise much more concern. Now, many wind power plants use high definition cameras to inspect the defect of a wind blade, which only gives surface information. Infrared thermograph has been widely used for its high efficient, noncontact and can be applied to detect a large area on a single time. Infrared images make up for visual images for it can detect internal damages which are barely visible at the surface. This paper proposed a matching method for visible image and infrared image of wind blade to give a complete impression for defect detection.Because of the incompleteness of contour curve and limitation of feature points of individual wind blade, an image matching technique combined point features with linear features is proposed. First, the edge information of visual image and infrared image is detected, and then the main contour of the wind blade is tracked and represented by Freeman chain code. In order to get the feature points on the tip part of the blade, difference code and cumulative difference code are constructed. The matching points of contours can be then captured through calculating the curvature and angle of each reserved contour point. Then, the transformation model between the visual image and infrared image of wind blade is calculated by the fitting lines of the leaf and the matching points on the tip part of the blade. Furthermore, by the study of three image fusion algorithms, an image fusion algorithm based on IHS color space is taken for the fusion of visual images and infrared images of wind blades.Finally, a wind blades detection and repair management system is established to deal with the digital information of wind blades.