Based On PDE Sub-section Defect Image Veneer Repair

Texture is an important feature of veneer, The defects of the veneer in the natural section of sub-structure could damage the uniformity and integrity of the texture, which reduce its quality and greatly reduce the grade of the veneer. It affects not only the visual nature of the surface veneer and processing properties, but more importantly it affects its value, which is not conducive to the effective use of the veneer. Repairing texture of the veneer has a very high industrial value. It not only can be used to detect the texture of wood, but also widely used for wood-based panels production, the decoration of the texture of the veneer in all aspects of the industry. This will become an important part of the automatic production line and this will help improve accuracy of industrial production and productivity. Based on partial differential equations of image processing and vision technology of the machine, combining the characteristics of the veneer of the past, finishing the restoration algorithm, the author proposes a new algorithm to achieve an effective method of image restoration of peeling veneer. Paper's main work is as follows:Firstly, this paper makes a summary of the mainstream of today's image restoration. The author points out that the algorithm that based on partial differential equations BSCB repair and algorithms based on sample blocks Criminisi repair algorithms are only suitable for general image of a single structure or texture part of the repair, but for peeling veneer surface defect image restoration,it has its obvious shortcomings. The BSCB algorithm can only restore smaller part of the image and sample-based Criminisi algorithms can only repair the part of the texture which did not change significantly. It can not repair the inner structure and texture in a good way.According to the weaknesses of BSCB, the paper puts forward a new and improved BSCB algorithm, the improved algorithm can repair defects in large peeling veneer, but for some of its texture, the repairing effect is not satisfactory. Therefore, this paper also proposed an improved BSCB algorithm coupled with new algorithm of the restoration based on sample surface defect in the peeling veneer. It is effective to apply this new algorithm to sub-section defect of the image on structure and the texture of the veneer at the same time. But when the board texture and structure of part of the change is not obvious, repairing effect of the algorithm are also affected.Finally, according to the limitations of the single method of restoration, the paper also proposes a new algorithm of restoration based on image decomposition. The first application of TV algorithm is decomposed into structural elements and the texture part of the structural part of the improved algorithm with the BSCB to repair. Part of the texture, the paper also proposes Criminisi algorithm as an expansion of based on the sample block to repair, and then fix a good synthesis of the two superimposed images, and through the experimental comparison, the effect is good, basically to meet the requirements of image restoration of peeling veneer.