Feature Point Extraction Method of Vector Graphics Based on Orientation Feature and Local Uniqueness

Abstract

The invention discloses a vector graphic feature point extraction method based on direction feature and local uniqueness, which is divided into two parts: direction feature value calculation and local uniqueness screening. In the direction feature value calculation part, the edge feature is calculated through a sliding window; in the local The uniqueness screening part calculates the uniqueness between each region, combines the direction feature and its uniqueness and scores each window region, and sorts in descending order to obtain the top n optimal feature points. A number of alignment feature points can be quickly extracted from the vector map, and the feature points have good uniqueness. At present, this algorithm has been maturely applied to automatic alignment equipment.

Description

technical field [0001] The invention belongs to the field of image processing, in particular to a vector graphics feature point extraction method based on directional features and local uniqueness. Background technique [0002] When the products manufactured according to the two-dimensional plane CAD / CAM are subjected to automatic optical inspection (AOI), the detection and positioning must rely on the key technology of feature recognition. Through feature recognition, we can get the type, position (rotation and offset) of the product, and even calculate the deformation variable (stretch, distortion). Such applications are concentrated in the fields of digital storage of CAD drawings, PCBAOI, and film AOI. With the continuous development of science and technology, image processing technology also keeps pace with the times and keeps innovating. However, the current mainstream research is still concentrated in the field of natural images and medical images, and there is littl...

AI Technical Summary

Problems solved by technology

The target graphics for feature point extraction are CAM vector graphics. In CAM vector graphics, all graphics are standard geometric patterns, such as straight lines and polygons composed of arcs. These graphics can be easily converted into bitmap images; but different For bitmaps, since the CAM graphics are very standard, there will be a large number of shapes with the same structure, which will lead to the found feature points in a certain area, which is not unique, which will cause misalignment in the later feature matching. Happening

Using traditional feature extraction algorithms, such as SIFT, SUFT, Opencv Fast algorithm, etc. cannot solve this problem well

Images