Image Processing and Splicing Method for 3D Optical Scanning Surface Reconstruction of Wood Grain

Based on environment compensation, scanning image processing technology was employed to investigate point cloud data and space matching method for wood grain. Collision avoidance recognition algorithm was used to collocate mark points, when remarkably reduced the error matching of distance coincidence mark points. The proposed method used marking of flag sample points based on weight to compensate for the marking points ambiguity of distinguishing information in scanning environment, and select the optimal path for the weighted results. The same splicing points in different images was identified, solving the problem of fuzzy splicing by distance matching. Experimental results and three-dimensional (3D) printing wood cross-section model reconstructed by surface fitting were compared. Results showed that the 3D scanning image mosaic of wood growth texture at the cross-section had no obvious stereo characteristics. The proposed method has improved the accuracy of surface mosaic in reverse scanning imaging for wood grain. This method can be applied to support the application needs of reverse surface reconstruction.