Research On 3d Model Symmetry Detection Method Based On Image

Symmetry is very common in nature. Whether it is cellular, the particle microstructure, or solar system and other stars in the universe,it has symmetry of different degrees. The human perception to the symmetry is strong that the big man-made objects are symmetric, and symmetry in psychology is considered as a basic principle of human perception. According to the study, when people identify the object "symmetry", corresponding peak values in the brain activity occurs. Thus, the symmetry affects the people's initial attention mechanisms, and guides the variety post-processing on objects.While objects in nature is difficult to meet the strict symmetry, so it is not enough to detect the precision relying on mathematical definition to symmetry. In addition, because of various reasons such as the transformation of objects angles, noise caused by graphics digitization, the processed graphics lost the original symmetry features, which also brought the inaccurate test results. Symmetry detection plays an important role not only in locating and identifying objects, but also in the reconstruction of 3D objects. With the constantly intellectualization of industrial and medical equipment, symmetry plays a significant role in texture object detection and guidance grab machine. Therefore, extraction the symmetry from complex scenes is of great significance to the pattern recognition and image processing.The key issues on symmetry detection of three-dimensional model is how to quickly and accurately identify the symmetry axis. This article a large number of domestic and foreign literature is consulted, symmetry detection is studied extensively. First, the appropriate point is selected to do the projection of the three-dimensional model to ensure the symmetry of projection, and then marginalization to the image using Sobel operator is done. On this basis, the detection to symmetry of three-dimensional model based on image method is focused on. And feasibility of the method is proved through the relevant definitions and theories, and the validity of the method is verified by experimental results. The main innovation of this paper is as follows:(1)To the use of the respective characteristics of implicit polynomial method and principal component analysis, the two methods is combined for the detection to projection of three-dimensional model. The method reduces the computation and improve the detection efficiency. (2) The mirror symmetry curve detection algorithm is proposed based on the use of geometric properties of curves. The algorithm is more effective to the higher symmetry degree of mirror model.The main work of this paper is as follows:(1) The importance of symmetry is pointed out by introducing the research background of symmetry detection. While today, the more intelligent days, symmetry detection is widely used in various fields and play an important role. After an overview of the existing symmetry detection algorithm, the advantages and disadvantages of each method is compared and analyzed.(2) Implicit polynomial curves can accurately describe the objects with a little coefficients and less computation. With full basis, symmetry detection theory based on principal component analysis has less requirements to the quality of sample sets, and it do not treat the specific shape of the object detected on prior knowledge. With the advantages of integrating the two methods, first high implicit polynomial curve projection of three-dimensional model to fit the boundary contour is used, the implicit polynomial curves is calculated. And the zero curvature and the curvature for the most value of the point is calculated, and then set as the main element of these points were analyzed to calculate point set covariance matrix. After obtaining eigenvalues, the final direction of the resulting principal component is analyzed to determine the axis symmetry. Experimental results show that the combined method can fast detect symmetry axis, and has good detection efficiency.(3) The geometric properties of the curve can reflect the features of shape on the image more intuitive. The mirror symmetry curve detection algorithm based on differential properties is proposed. First, the edge of the three-dimensional model of projector is extracted to get a closed curve, a array of long sequence is abtained by recording the curve point by point, and then the central and crucial point of curves is found. The connection between these key points and closed curves center is the candidate image symmetry axises, and finally set a specific threshold, according to verification based on the nature of the closed curve on the candidate image symmetry, an accurate mirror symmetry axis is abtained.