Abstract: In order to effectively extract pattern borders consistent with tire structure design characteristics from point cloud data, an extraction method of tire pattern border based on the point cloud mapping matrix is proposed. Firstly, the tire tread point cloud acquired by optimizing segmentation is set as the object, and the grid is treated as the elementary unit, so as to realize the transformation from 3D tire tread point cloud into point cloud mapping matrix. Secondly, pattern structure knowledge unit is gained via the connective growth mode of column chain codes, and a structure knowledge base able to characterize pattern structure information is built. Then, pattern clustering in the knowledge base is realized by calculating the similarity of knowledge units, and matching between clustering results and the basic pattern design types is realized via pattern recognition. Finally, pattern border characteristic points are extracted with the classification mechanism of basic pattern border characteristics. The illustrative results indicate that this method can extract pattern border characteristics in semi-steel radial tire point cloud efficiently, stably and accurately.