Method for decoding coding mark points of measuring stick based on skeleton extraction

Abstract

The invention relates to a method for decoding coding mark points of a measuring stick based on skeleton extraction. The method comprises the following steps: segmenting an image by an self-adaptive threshold method and extracting foreground regions which comprise non-coding mark point regions; eliminating a noise region by utilizing the area of each foreground region to acquire a region to be extracted; performing local sub-pixel edge extraction on the region to be extracted in the original image and then positioning an oval outline; fitting an oval by oval outline points, extracting a coding mark point region and intercepting an image of the coding mark point region; performing inverse affine on the oval in the image to enable the oval to become a circle, so as to acquire a image to be decoded; performing partial differential processing and extracting the skeleton of a coding ring region; decoding the coding mark points and performing rotate right on 10 binary codes, wherein the minimal decimal value is the coding mark point decoding value; performing one-to-one correspondence on determined feature point codes on the measuring stick of left and right images, namely matching point pair. The method has the advantages that the defect that three-dimensional matching calculation is complex in the traditional algorithm is overcome, the matching precision is improved, and the important application value is achieved in the field of binocular vision three-dimensional measurement.

Description

technical field [0001] The invention belongs to the technical field of binocular stereo vision measurement, and in particular relates to a skeleton extraction-based coding and marking point decoding method of a measuring stick, which can be used for binocular stereo vision measurement of difficult points such as deep holes and hidden points. Background technique [0002] Binocular stereo vision is an important branch of computer vision. Its principle is to calculate the relative displacement of space objects in the left and right images by using two cameras in different positions, or one moving camera to view the same scene. The passive ranging technology based on binocular stereo vision is the most researched and most widely used distance measurement technology. [0003] The complete stereo vision can be divided into four modules: camera calibration, image acquisition, stereo matching and 3D information recovery. Stereo matching is the most important and difficult problem ...

AI Technical Summary

Problems solved by technology

Since the code points have their own geometric characteristics, if the image resolution is not high or the illumination is not ideal, the code information may be damaged due to the change of the image gray value, which will affect the correct rate of code recognition

Images