CN108846403B

CN108846403B - Method for extracting facet gem image features

Info

Publication number: CN108846403B
Application number: CN201810562783.0A
Authority: CN
Inventors: 杨琇明; 雷自力; 刘晓军
Original assignee: Hubei Emers Intelligent Testing Equipment Co ltd; Wuhan Hengyu Scientific And Educational Instrument And Equipment R&d Co ltd
Current assignee: Hubei Emers Intelligent Testing Equipment Co ltd; Wuhan Hengyu Scientific And Educational Instrument And Equipment R&d Co ltd
Priority date: 2018-06-04
Filing date: 2018-06-04
Publication date: 2023-04-07
Anticipated expiration: 2038-06-04
Also published as: CN108846403A

Abstract

The invention discloses a method for extracting the characteristics of a faceted gem image. The optical system collects a facet gem image, and denoising and binaryzation pretreatment are carried out on the facet gem image; obtaining the outline characteristics of the gem by adopting an edge extraction algorithm and a fitting algorithm; establishing a distance interval for screening angular point features; obtaining characteristic angular points of the gem by using an angular point algorithm; and calculating the distance from the characteristic corner points to the circular points of the gem outline, and taking the characteristic corner points in the distance interval as the strong corner point characteristics of the faceted gem. The method fully combines the geometric properties of the faceted gem, screens strong angular points representing the characteristics of the faceted gem by using the distance intervals, eliminates the interference of redundant angular points and provides scientific reference data for the identification of the faceted gem.

Description

Method for extracting facet gem image features

Technical Field

The invention belongs to the field of gem identification.

Background

The facet gem is a circulation finished product with higher commercial value obtained after the natural gem is cut by people, and the gem identification is an essential link in the gem commercial circulation, but at present, the gem identification mainly depends on manual visual inspection and qualitative analysis, the error rate is high, the standardization degree is low, and the development of the gem industry is influenced. The method is characterized in that a facet gem image is acquired by using an optical imaging system, a standard feature extraction method is researched, the geometric features such as outlines, angular points and the like of the facet gem are automatically acquired from the image, the geometric feature data of the gem are established, a more scientific reference basis is provided for gem identification, and the method has important significance for improving the facet gem circulation supervision capability in the jewelry industry.

Disclosure of Invention

The invention discloses a method for extracting facet gem outline characteristics and strong corner point characteristics according to the requirements on characteristic data in gem identification. Collecting a facet gem image by using an optical system, and performing denoising and binaryzation pretreatment on the facet gem image; obtaining the outline characteristics of the gem by adopting an edge extraction algorithm and a fitting algorithm, and establishing a distance interval for screening the corner point characteristics; obtaining a characteristic corner set of the gem by using a corner algorithm; and calculating the distance from the characteristic angular points to the circular points of the gem outline, wherein the characteristic angular points in the distance interval form the strong angular point characteristics of the faceted gem.

A method for extracting the characteristics of a faceted gem image comprises the following steps:

step one, carrying out non-local mean value denoising processing on the facet gem image.

And step two, carrying out Otsu method binarization processing on the image.

And step three, extracting a gem edge point set in the image through a canny algorithm.

And fourthly, performing least square method circumference fitting on the edge point set, calculating to obtain a center point and a radius value R, and taking the fitting circumference as the profile characteristic of the faceted gem.

And fifthly, according to the R value, defining 4 distance intervals of 0.18R to 0.27R,0.38R to 0.42R,0.58R to 0.62R and 0.78R to 0.82R respectively.

And step six, calculating characteristic corner points of the image by adopting a harris corner point algorithm.

And step seven, calculating the distance from the characteristic angular point to the fitting central dot.

And step eight, taking the characteristic corner points with the distance within the distance interval defined in the step five as the characteristic of the strong corner points of the faceted precious stone.

According to the method, the outline and the strong corner point characteristics are taken as characteristic extraction objects, on the basis of fully considering the geometrical properties and the corner point distribution rule of the faceted gem, the strong corner points which can represent the characteristics of the faceted gem are screened by using the distance intervals, the interference of redundant corner points is eliminated, and the characteristic data for identifying the faceted gem is established.

Drawings

FIG. 1 is an image of a faceted gemstone for use with an embodiment of the present invention.

FIG. 2 is a diagram illustrating a profile of a faceted gemstone and a distribution of features of strong corners extracted according to an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

The embodiment of the invention provides a method for extracting the characteristics of a facet gem image, which comprises the following steps:

step one, performing non-local mean denoising treatment on the faceted gem image. Redundant information in the image is fully utilized, the detail characteristics of the image can be furthest maintained while denoising is carried out, and the algorithm processing process is as follows,

the de-noised gray value at x position in the image containing noise is set as u (x),

/>

(1)

in the formula, V (y) is the gray value at y position in the image, w (x, y) is the similarity between x and y representing pixel points, and the value is determined by the distance between the adjacent domains of V (x) and V (y) of the rectangle taking x and y as the center

/>

(2)

In the formula, Z (x) is a normalized coefficient, and the larger h is a smoothing coefficient, the more gradual the change of the Gaussian function is, the higher the denoising level is, but the image is blurred. The smaller h, the more edge detail components remain, but too many noise points remain. The specific value of h is based on the degree of grinding and polishing of the facet gem, and is generally 0.5 to 0.7.

And step two, carrying out Otsu binaryzation processing on the image.

And step three, extracting a gem edge point set in the image through a standard canny algorithm, and in a double-threshold selection link of the canny algorithm, selecting a proportion of a high threshold value in the image gradient histogram, wherein the proportion is generally determined by the reflectivity of the faceted gem, and a low threshold value is 20% -30% of the high threshold value.

And fourthly, performing least square method circumference fitting on the edge feature point set, calculating to obtain a central point and a radius value R, and taking the fitting circumference as the profile feature of the facet gem.

Step six, calculating characteristic angular points of the image by adopting a harris angular point algorithm,

/>

(3)

/>

（4)

in the formulas (3) and (4), P is an image gradient distribution matrix calculated by a three-canny algorithm, v _xy Is the corner response value at image (x, y), k is the corner decision threshold, C _v Is a characteristic corner distribution map of the image. In general, the value of k is such that the total number of corner points obtained from equation (4) is greater than 1.5 times the number of actual polished corner points of the faceted gemstone.

And step eight, taking the characteristic corner points with the distances within the distance interval defined in the step five as the strong corner point characteristics of the faceted gem.

Claims

1. A method for extracting the characteristics of a faceted gem image is characterized by comprising the following steps:

firstly, carrying out non-local mean value denoising treatment on a facet gem image;

step two, carrying out Otsu method binarization processing on the image;

step three, extracting a gem edge point set in the image through a canny algorithm;

performing least square method circumference fitting on the edge point set, calculating to obtain a central point and a radius value R, and taking the fitting circumference as the profile characteristic of the faceted gem;

establishing distance intervals for screening corner point characteristics according to the R value, and defining 4 distance intervals of 0.18-0.27R, 0.38R-0.42R, 0.58R-0.62R and 0.78R-0.82R;

step six, calculating characteristic angular points of the image by adopting a harris angular point algorithm;

step seven, calculating the distance from the characteristic angular point to the fitting center round point;

and eighthly, taking the characteristic corner points with the distances within the distance interval defined in the step five as the strong corner point characteristics of the faceted gem image.