Texture Direction Detection for Side of Stacking Solar Cells

The stripe texture features for the side of stacking solar cells are rich, and the texture direction detection is an important research content of image segmentation and quantitative calculation for stacking solar cells. However, the complex background and noise interference bring difficulties to the high-precision and efficient texture direction detection. Therefore, based on the texture direction consistency for the side of stacking solar cells, the existing direction detection algorithms based on slicing and gradient are improved in this paper. In addition, according to the gray distribution, Hough transform, Radon transform, and spectrum analysis for stacking solar cell images, four direction detection methods of image texture for the side of stacking solar cells are proposed. The four proposed methods based on gray distribution, Hough transform, Radon transform, spectrum analysis, and the two improved methods based on slicing and gradient analysis are used to extract the texture direction of stacking solar cells. The comparison experiment results show that, compared with the two improved methods, the average errors of texture direction detection for the side of stacking solar cells by the four proposed methods is reduced by 1.49°, 1.47°, 1.51°, 1.55°, and 1.34°, 1.32°, 1.36°, 1.4° respectively. The proposed methods have great advantages in accuracy, speed, complexity, and noise resistance performance. It lays a foundation for accurate and fast image segmentation of stacking solar cells based on image texture direction.