Texture image synthesis method based on tag database

Abstract

The invention discloses a texture image synthesis method based on tag database. The texture image synthesis method is characterized by comprising the following steps of: 1, establishing a tag databaseL, wherein the tag database L is used for storing image data of an image texture block and tag data I thereof; 2, searching an image texture block matched with the edge by using the tag data I, and sewing the image texture block to obtain a complete texture image. The texture image synthesis method has the beneficial effects that the information of the edge area and the corner area of each imagetexture block is extracted, and the comprehensive tag data of the image texture block is formed according to the position classification in the image texture block in the different areas, the tags on,the corresponding position only needs to be screened when searching, a large amount of computation data can be reduced, at the same time, time information of each image texture block is also stored in the tag data, when searching, the image texture blocks with the same time can be found, or when a clear image in the same time can not be found, the image texture blocks with the near time can be found to ensure the image clarity.

Description

technical field [0001] The invention relates to the technical field of texture block image synthesis, in particular to a texture image synthesis method based on a label database. Background technique [0002] Texture image synthesis is an important part of image processing, and its synthesis effect directly affects the accuracy of image processing, especially in dynamic images that are updated over time, the synthesis speed and accuracy of texture images determine real-time image analysis, such as the process of driving a car image recognition of vehicle conditions. [0003] Existing texture image synthesis technology is usually completed by stitching discrete image texture blocks, and the selection of each image texture block becomes the key to synthesis speed and accuracy. Image texture blocks are smaller squares, which include four sides and Four corners, in the existing method of selecting image texture blocks, by extracting the side information of each image texture bl...

AI Technical Summary

Problems solved by technology

However, the prior art usually ignores that the matching method of the image texture block has particularity in the stitching position. For example, if the region to be stitched is the left region image, then only the image texture block whose side information on the left side matches can be selected for stitching, that is, only need to consider Whether the side information on the left meets the stitching conditions can satisfy the texture image synthesis, which can omit a large amount of other side information data and improve the synthesis speed; at the same time, the angle information that is often ignored in the prior art can also be used as a matching information source, because The amount of information of the corner is much smaller than that of the side, so it can save more time in the extraction speed; moreover, the dynamic image also contains time sequence, so the image texture blocks must also be combined according to the time sequence, which is the existing Another element overlooked in technology

Images