CN114994242B

CN114994242B - Group discrimination method for concentric circle cutting half landmark point complement based on fish otolith

Info

Publication number: CN114994242B
Application number: CN202210501019.9A
Authority: CN
Inventors: 王英俊; 宋爱环; 邹琰; 吴莹莹; 赵文溪; 刘莹; 刘童; 于超勇; 张伟; 官曙光
Original assignee: Shandong Academy Of Marine Sciences Qingdao National Marine Science Research Center
Current assignee: Shandong Academy Of Marine Sciences Qingdao National Marine Science Research Center
Priority date: 2022-05-10
Filing date: 2022-05-10
Publication date: 2023-08-11
Anticipated expiration: 2042-05-10
Also published as: CN114994242A

Abstract

The invention relates to a group discrimination method for supplementing points of semi-landmark points based on concentric circle cutting of fish otoliths, which belongs to the field of fish otolith discrimination, wherein intersection points of a series of concentric circles and otolith auditory grooves and edges are used as landmark points, and intersection points of the concentric circles, the otolith auditory grooves and contours are selected to be used as landmark points for statistical analysis; repeatedly marking a point on the landmark points on the approximate positions of the missing otoliths for the non-common intersection points, and then inputting all the landmark points into the tpsrelw32 software to perform group judgment; according to the invention, on one hand, the number of selectable landmark points is increased, and on the other hand, the integrity of otolith feature extraction is ensured and the discrimination rate of the group is improved by increasing the complement points reflecting the base page features. Moreover, all selected points are objective points, so that the influence of subjective factors on the point selection position is avoided, the standardization of the tpsrelw32 software point selection process is realized, and a foundation is laid for building a landmark point database and subsequent deep application.

Description

Group discrimination method for concentric circle cutting half landmark point complement based on fish otolith

Technical Field

The invention belongs to the technical field of fish otolith identification, and particularly relates to a group identification method based on concentric circular cutting half landmark point supplement of fish otolith

Background

The fish otolith is a hard tissue with balance and hearing function in the inner ear of teleosts, mainly consists of calcium carbonate, has morphological characteristics regulated and controlled by genetic factors, has species specificity and is often used for classifying fish species or populations. Early studies on otoliths focused on metric measurements, describing the smoothness, degree of protrusion, etc. of the otoliths. In the 80 s of the 20 th century, morphological analysis methods and multivariate statistical analysis tools tended to be mature, and geometric morphology measurements developed. Geometry measurement mainly includes two methods: contour line method and landmark method. The punctuation method (landmark methods) is a shape statistical method based on cartesian landmarks, and is characterized in that landmark points are points which have obvious characteristics and are easy to distinguish by acquiring coordinate data of landmark points X, Y on a two-dimensional image, performing Relative distortion (Relative Warp) and thin plate spline analysis, drawing a grid deformation map, analyzing otolith morphological variation and performing multiple statistical analysis. Currently, in biological applications, landmark points are often divided into three categories: the I-shaped landmark points are contact points and intersection points among different tissues, such as the intersection points of veins and leaves, the intersection points of the fish fins and the fish body, and the like. Type II landmark points are depressions or protruding points in the structure, such as sharp curvatures of teeth, protruding bones, etc. Type III landmark points are the endpoints or poles of the longest, narrowest, widest, etc. points in the structure. The landmark method firstly obtains X and Y coordinate values of the otolith landmark point, also called Cartesian coordinate (Cartesian coordinate data), and the coordinate values are expressed as the pure morphological information of the otolith through centering, rotation and scaling (scaling), so that the non-shape effects such as otolith position, scale, angle and shooting multiple are eliminated, and the morphology of the study object is quantized. The removal of these non-shape-variant disturbances is often referred to as overprinting. The principle is that a minimum distance between coordinate points is found out through a least square criterion (least-square criterion), and then centering, rotation, scaling and the like are carried out to obtain an average form (mean shape) of an otolith sample. The thin-plate spline analysis theory utilizes knowledge of a distortion energy matrix (bending energy matrix) in the material science, deformation changes generated by an object are reflected in conversion of corresponding functions, and the function conversion involves a series of complex steps including principal distortion (primary warp), relative distortion (relative warp), partial distortion (partial warp) and the like, and corresponding data files are generated and are subjected to differential comparison by using statistical software. In addition, morphological changes of the sample can be directly visually demonstrated by thin plate spline analysis. In the prior art, a tpsDig2 software is used for marking landmark points on otoliths to generate a 'tps file', then the generated 'tps file' is input into a tpsrelw32 software to extract otolith characteristic information, and then a statistical method is combined to complete group analysis based on otolith characteristics.

However, the landmark point method in the prior art is used for classifying the fish otolith groups and has the characteristics of strong point selection main appearance and few selectable landmark points.

Disclosure of Invention

The invention aims to solve the technical problem of providing a group judgment method for supplementing points of semi-landmark points by concentric circle cutting based on fish otoliths.

The invention is realized by the following technical proposal

A group discrimination method based on fish otolith concentric circle cutting half landmark point supplement, the method adopts a series of concentric circles and intersection points of otolith auditory grooves and edges as landmark points, the outermost circle of the series of concentric circles takes the longest diameter of the otolith as the diameter, the radius of the outermost circle is equally divided into 8-10 parts in the inside of the series of concentric circles, the concentric circles of the outermost circle are constructed, and the intersection points of the concentric circles, the otolith auditory grooves and the outlines are selected to be used as landmark points for statistical analysis; repeatedly marking a point on a landmark point on the approximate position of the missing otolith as the non-common intersection point, and inputting all landmark points into tpsrelw32 software to perform group judgment; the non-common intersection points refer to landmark points that do not appear on all otoliths.

Further, the radius was equally divided into 10 parts at the inner part of the outermost circle.

Compared with the prior art, the invention has the beneficial effects that:

according to the method, the points of the boundary points of the otolith, the auditory canal and the concentric circles are selected as the landmark points by a method of cutting the landmark points and supplementing the points by the concentric circles, so that the number of selectable landmark points is increased on one hand, and on the other hand, the integrity of otolith feature extraction is ensured and the discrimination rate of the group is improved by adding the supplement points reflecting the basic page features. Moreover, all selected points are objective points, so that the influence of subjective factors on the point selection position is avoided, the standardization of the tpsrelw32 software point selection process is realized, and a foundation is laid for building a landmark point database and subsequent deep application.

Drawings

Fig. 1 is a diagram of hexagrammos otakii otolith: the right side is the front end of the otolith, the left side is the rear end of the otolith, the upper side is the back side, the lower side is the abdomen side, 1 the main grooves are auditory grooves, 2 the tail parts of the auditory grooves, 3 the wing leaves, 4 the base leaves;

FIG. 2 is a schematic diagram of original landmark point fetching;

FIG. 3 is a 10×1 concentric circle landmark point plot;

fig. 4 is a plot of concentric landmark point complements.

Detailed Description

The technical scheme of the present invention is further explained by examples below, but the scope of the present invention is not limited in any way by the examples.

Example 1

Selecting six-line-fish samples from 7 wild groups, namely Liyun gang, qingdao, wisea, tobacco stand, qinghai Royal island, dalian, dandong and 1 Wisea culture area, measuring information such as fish body length, body quality and the like, and picking up otoliths. The otoliths are cleaned by ultrapure water, then put into an oven for drying to constant weight, weighed by an electronic balance (graduation value 0.01 mg), recorded, and put left and right otoliths into a clean 1.5ml centrifuge tube for separate numbering and storage.

In the embodiment, the morphology difference of left and right otoliths is not obvious, so that left otoliths of the Hexagrammos otakii are uniformly selected as a research object, and an OLYMPUS SZ61 stereoscopic microscope is adopted for image acquisition on the inner side surface of the otoliths in the otolith morphology analysis. The individual left otoliths were replaced with right otoliths when missing. The otolith image is shown in fig. 1, wherein the right side of the image is the front end of the otolith, the left side is the rear end of the otolith, the upper side is the back side, the lower side is the abdomen side, 1 is the main groove, namely the auditory canal, 2 is the tail of the auditory canal, 3 is the wing leaf, and 4 is the base leaf.

The collected otolith images are processed according to an original landmark point method, a concentric circle landmark point method (without manual point filling) and a concentric circle landmark point method (manual point filling), and the specific method and the result are as follows:

1. original landmark point method

According to the requirements of the homology and the characteristics of the landmark points, 14 landmark points are selected from the external outline of the otolith and the outline of the auditory canal on the morphology of the otolith of the Hexagrammos otakii, the points are taken in the clockwise direction, the landmark points and the sequences of all samples are kept consistent, and the 14 landmark points are classified as shown in Table 1. The I-type landmark points are 6, 11, 12, 13 and 14, the II-type landmark points are 2, 4, 8 and 10, the III-type landmark points are 1, 3, 5, 7 and 9, and corresponding coordinate point value files are established. The last found landmark point is shown in fig. 2.

TABLE 1 types and definitions of landmark points

2. Concentric circle landmark point method (without artificial point supplement)

The intersection point of the concentric circles and the otolith auditory canal and the edge is used as a landmark point. The longest diameter of the otolith is taken as the diameter to be the outermost circle. Inside the outermost circle, radius 10 is equally divided, a series of concentric circles are constructed, and intersection points of the series of concentric circles, the otolith auditory canal and the outline are selected to be used as landmark points for statistical analysis, as shown in fig. 3. However, the base leaves and wing leaves of the Hexagrammos otakii are different in proportion, and the characteristic is one of important characteristics of group division, the number of wings She Jiaodian is inconsistent when concentric circles are uniformly distributed, and when landmark points are used for otolith morphological analysis, the uniform number of points obtained on the otolith of each individual is required to be ensured, so that non-common intersection points can be omitted during analysis. As shown in FIG. 4, not all otoliths at point 51 in a have this intersection point, and the previous analysis was selected to discard, and only 50 points were used for analysis, and 50 landmark points were input to the tpsrelw32 software for otolith characterization.

3. Concentric circle landmark point method (Manual complementary points)

The method performs manual point filling on non-common points, such as a in fig. 4 shows landmark point 51, but in the right diagram, there is no landmark point 51, by manually adding one more point to the position of point 8 as point 51, and in fact, point 51 and point 8 are overlapped, but in order to clearly show the figure, point 51 is specifically separated from point 8.

The three methods are used for carrying out group judgment on the greenling of different groups, and the results are shown in the following table 1.

TABLE 1.10X1 comparison of discrimination results of six-line fish swarm with concentric circles and non-complementary points

Example 2

In this embodiment, the radius of the concentric circle with the longest diameter of the otolith as the diameter is equally divided in different parts, and the influence of different parts on the discrimination result is checked. Also taking the otolith of Hexagrammos otakii as an example, the radius of the largest circle with the longest diameter as the diameter is divided into 8 parts and 10 parts respectively. The groups were discriminated by the concentric circles non-manual complement method, and the discrimination results are shown in table 2:

TABLE 2 comparison of discrimination results of six-line fish swarm with different equally divided concentric circles and without points

The group discrimination success rate of the 8 multiplied by 1 concentric circle landmark point method is 88%, the group discrimination success rate of the 8 multiplied by 1 concentric circle and complementary point landmark point method is slightly improved to 91.8%, the group discrimination success rate of the 10 multiplied by 1 concentric circle landmark point method is 98.4% which is greatly higher than the success rate of the 8 multiplied by 1 concentric circle, the accuracy of the 10 multiplied by 1 concentric circle and complementary point landmark point method is up to 99.5%, and the contribution rate of the added ground surface point to the group discrimination is limited, so that the 10 multiplied by 1 concentric circle and complementary point landmark point method is selected.

Claims

1. A group discrimination method based on fish otolith concentric circle cutting half landmark point supplement is characterized in that the method adopts intersection points of a series of concentric circles, otolith auditory grooves and edges as landmark points, the outermost circle of the series of concentric circles takes the longest diameter of the otolith as the diameter, the radius of the outermost circle is equally divided into 8-10 parts in the inside of the series of concentric circles, the concentric circles of the outermost circle are constructed, and the intersection points of the concentric circles, the otolith auditory grooves and the outlines are selected to be used as landmark points for statistical analysis; repeatedly marking a point on a landmark point on the approximate position of the missing otolith as the non-common intersection point, and inputting all landmark points into tpsrelw32 software to perform group judgment; the non-common intersection points refer to landmark points that do not appear on all otoliths.

2. The method for determining the population of the half landmark point complements by the concentric circle cutting of the fish otoliths according to claim 1, wherein the radius of the inner part of the outermost circle is equally divided into 10 parts.