Method for discriminating quality of snakehead fry

Abstract

The invention discloses a method for discriminating the quality of snakehead fry, and belongs to the field of aquaculture. According to the method, the behavioral parameters such as average speed, acceleration and motion trail uniformity of the snakehead fry are selected as fry quality correlation indexes, so that the quality of the snakehead fry is accurately judged. By means of the intelligent means, the detection speed of 2 minutes per batch is achieved, meanwhile, fry damage is reduced by 85.6%, early intervention of diseases can be achieved according to the recognition result, therefore, the efficiency and accuracy of snakehead fry quality judgment are improved, errors and damage caused by manual operation are reduced, and important application value and popularization prospects are achieved.

Description

Technical Field

[0001] The invention relates to the field of aquaculture, and in particular to a method for distinguishing the quality of snakehead fish fry. Background Art

[0002] Snakehead snakehead (Gibberish) is an important economic fish species, and the quality of its seedlings directly impacts aquaculture profitability. Currently, the quality of snakehead snakehead seedlings is primarily assessed by manual observation, based on characteristics such as their activity and body shape. This method is highly subjective, inefficient, and difficult to accurately quantify, failing to meet the demand for rapid and accurate seed quality assessment in large-scale aquaculture. Furthermore, the measurement and capture processes can easily cause stress and mechanical damage to the fry, impacting their growth and survival rates.

[0003] Therefore, there is an urgent need for a scientific and efficient method for judging the quality of snakehead fish seedlings. Summary of the Invention

[0004] The present invention aims to provide a method for distinguishing the quality of snakehead fish seedlings, which can accurately count the fish species, quickly detect the quality of the seedlings, and reduce the seedling damage rate.

[0005] In order to achieve the above-mentioned purpose, the technical solutions provided by the present invention are as follows:

[0006] A method for distinguishing the quality of snakehead fry is provided. The method selects behavioral parameters of snakehead fry, such as average speed, acceleration, and motion trajectory uniformity, as fry quality-related indicators to accurately distinguish the quality of snakehead fry. The specific steps are as follows:

[0007] (1) Real-time collection of fry movement data through a data acquisition module, including video streams or sensor data;

[0008] (2) Calculate core indicators such as individual or group acceleration, average speed, and motion trajectory uniformity based on the collected data. The motion trajectory uniformity includes parameters such as speed variance and spatial distribution dispersion.

[0009] (3) The Deepocsort algorithm was used to construct an evaluation model, and Python was used to complete the back-end software development to achieve intelligent judgment of the quality of black snakehead seedlings.

[0010] Furthermore, in step (1), the data acquisition module uses software such as DarkLabel to collect seedling marking data.

[0011] Furthermore, for video stream data collection, a high-definition camera was installed above the culture pond to shoot the seedling movement video in real time at a frame rate of 25 frames per second. The video was input into the DarkLabel software, and the seedlings were automatically labeled to obtain the initial position coordinates of each seedling.

[0012] For sensor data collection, acceleration sensors and position sensors are arranged in the breeding pond. The acceleration sensors collect acceleration data of seedling movement at a sampling frequency of 100 Hz, and the position sensors collect position coordinate data of seedlings at a sampling frequency of 50 Hz.

[0013] Furthermore, in step (3), the evaluation model is constructed using the DeepSORT algorithm by using the dynamic behavior data of fry collected in real time;

[0014] The motion state adopts the Kalman filter model, and the state vector includes position (x, y), velocity (v x，t , v y，t ), acceleration (a x，t , a y，t ), the state transfer equation is:

[0015]

[0016] Where: x t is the current position coordinate of the target in the horizontal direction (x-axis direction), t-1 The position coordinate of the target in the horizontal direction (x-axis direction) at the previous moment; y t is the vertical position coordinate of the target at the current moment (y-axis direction), t-1 The target's vertical (y-axis) position coordinate at the previous moment;

[0017] v x，t is the current target's horizontal speed (x-axis direction), v x，(t-1) v is the target's horizontal speed (x-axis direction) at the previous moment; y，t is the current target's vertical velocity (y-axis direction), v y，(t-1) is the target's vertical speed (y-axis direction) at the previous moment;

[0018] a x，t is the acceleration of the target in the horizontal direction (x-axis direction) at the current moment, a x，(t-1) is the acceleration of the target in the horizontal direction (x-axis direction) at the previous moment; a y，t is the acceleration of the target in the vertical direction (y-axis direction) at the current moment, a y，(t-1) It is the acceleration of the target in the vertical direction (y-axis direction) at the previous moment.

[0019] Furthermore, the quality of fry is graded based on the constructed algorithm model, and by setting different thresholds, the quality of fry is divided into different levels such as excellent, good, medium and poor.

[0020] Beneficial effects: Compared with the prior art, the advantages of the present invention are:

[0021] Through intelligent means, the detection speed of 2 minutes / batch is achieved, while the damage to seedlings is reduced by 85.6%. Earlier intervention of diseases can be achieved based on the identification results, thereby improving the efficiency and accuracy of the quality judgment of black snakehead seedlings, reducing the errors and damage caused by manual operations, and has important application value and promotion prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 Flowchart of the present invention. DETAILED DESCRIPTION

[0023] The technical solution of the present invention is described in detail below with reference to the accompanying drawings, but the protection scope of the present invention is not limited to the embodiments.

[0024] A method for distinguishing the quality of snakehead fry is provided. The method selects behavioral parameters of snakehead fry, such as average speed, acceleration, and motion trajectory uniformity, as fry quality-related indicators to accurately distinguish the quality of snakehead fry. The specific steps are as follows:

[0025] (1) The fry movement data, including video stream or sensor data, were collected in real time by using data acquisition modules such as DarkLabel. For video stream data collection, a high-definition camera with a resolution of 1920×1080 was installed above the culture pond to shoot the fry movement video in real time at a frame rate of 25 frames per second. The video was input into the DarkLabel software, and the fry were automatically marked to obtain the initial position coordinates of each fry. For sensor data collection, an acceleration sensor and a position sensor were arranged in the culture pond. The acceleration sensor collected the acceleration data of the fry movement at a sampling frequency of 100 Hz, and the position sensor collected the position coordinate data of the fry at a sampling frequency of 50 Hz. The measurement range of the acceleration sensor was ±10 g, and the positioning accuracy of the position sensor was ±1 cm.

[0026] (2) Based on the collected data, core indicators such as individual or group acceleration, average speed, and motion trajectory uniformity are calculated. The motion trajectory uniformity includes parameters such as speed variance and spatial distribution dispersion. The calculation method is as follows:

[0027] The calculation method of average speed is: for individual seedlings, the ratio of the total displacement of its movement to the total time within a period of time, that is:

[0028]

[0029] In formula (s1), v avg Indicates the average speed of seedlings over a period of time, (x i ,yi ) is the position coordinate at the i-th moment, t is the time interval, (x i-1 ,y i-1 ) is the position coordinate at the i-1th moment;

[0030] The calculation method of acceleration a is: by performing differential processing on the velocity data, that is:

[0031]

[0032] Where, v i is the velocity at the i-th moment, v i-1 The speed at the i-1th moment, Δt is the time interval;

[0033] Speed ​​variance The calculation formula is

[0034]

[0035] Spatial distribution dispersion σ xy It is measured by calculating the standard deviation of the seedling position coordinates, i.e.

[0036]

[0037] Where, is the average value of the position coordinates.

[0038] (3) The DeepSORT algorithm is used to construct an evaluation model based on the dynamic behavior data of fry collected in real time; the back-end software development is completed in combination with Python to achieve intelligent judgment of the quality of snakehead fry. The details are as follows:

[0039] The motion state adopts the Kalman filter model, and the state vector includes position (x, y), velocity (v x，t , v y，t ), acceleration (a x，t , a y，t ), the state transfer equation is:

[0040]

[0041] Where: x t is the current position coordinate of the target in the horizontal direction (x-axis direction), t-1 The position coordinate of the target in the horizontal direction (x-axis direction) at the previous moment; y t is the vertical position coordinate of the target at the current moment (y-axis direction), t-1 The target's vertical (y-axis) position coordinate at the previous moment;

[0042] v x，tis the current target's horizontal speed (x-axis direction), v x，(t-1) v is the target's horizontal speed (x-axis direction) at the previous moment; y，t is the current target's vertical velocity (y-axis direction), v y，(t-1) is the target's vertical speed (y-axis direction) at the previous moment;

[0043] a x，t is the acceleration of the target in the horizontal direction (x-axis direction) at the current moment, a x，(t-1) is the acceleration of the target in the horizontal direction (x-axis direction) at the previous moment; a y，t is the acceleration of the target in the vertical direction (y-axis direction) at the current moment, a y，(t-1) It is the acceleration of the target in the vertical direction (y-axis direction) at the previous moment.

[0044] (4) Based on the constructed algorithm model, the quality of fry is graded. By setting different thresholds, the quality of fry is divided into different levels such as excellent, good, medium and poor, as shown in the following table:

[0045]

[0046] As described above, although the present invention has been shown and described with reference to specific preferred embodiments, it should not be construed as limiting the present invention itself. Various changes may be made to it in form and detail without departing from the spirit and scope of the present invention as defined in the appended claims.

Claims

1. A method for determining the quality of snakehead fish seed, characterized in that: By selecting behavioral parameters such as average speed, acceleration, and motion trajectory uniformity of snakehead fry as seed quality-related indicators, accurate judgment of snakehead fry quality can be achieved. The specific steps are as follows: (1) Real-time collection of fry movement data through a data acquisition module, including video streams or sensor data; (2) Calculate core indicators such as individual or group acceleration, average speed, and motion trajectory uniformity based on the collected data. The motion trajectory uniformity includes parameters such as speed variance and spatial distribution dispersion. (3) The Deepocsort algorithm was used to construct an evaluation model, and Python was used to complete the back-end software development to achieve intelligent judgment of the quality of black snakehead seedlings.

2. The method for determining the quality of snakehead fish seed according to claim 1, wherein: In step (1), the data acquisition module uses software such as DarkLabel to collect seedling marking data.

3. The method for determining the quality of snakehead fish seed according to claim 2, wherein: For video stream data collection, a high-definition camera was installed above the culture pond to capture the seedling movement video in real time at a frame rate of 25 frames per second. The video was input into the DarkLabel software, and the seedlings were automatically labeled to obtain the initial position coordinates of each seedling. For sensor data collection, acceleration sensors and position sensors are arranged in the breeding pond. The acceleration sensors collect acceleration data of seedling movement at a sampling frequency of 100 Hz, and the position sensors collect position coordinate data of seedlings at a sampling frequency of 50 Hz.

4. The method for determining the quality of snakehead fish seed according to claim 1, wherein: In step (3), the evaluation model is constructed by using the DeepSORT algorithm based on the dynamic behavior data of fry collected in real time; The motion state adopts the Kalman filter model, and the state vector includes position (x, y), velocity (v x，t , v y，t ), acceleration (a x，t , a y，t ), the state transfer equation is: Where: x t is the current position coordinate of the target in the horizontal direction (x-axis direction), t-1 The position coordinate of the target in the horizontal direction (x-axis direction) at the previous moment; y t is the vertical position coordinate of the target at the current moment (y-axis direction), t-1 The target's vertical (y-axis) position coordinate at the previous moment; v x，t is the current target's horizontal speed (x-axis direction), v x，(t-1) v is the target's horizontal speed (x-axis direction) at the previous moment; y，t is the current target's vertical velocity (y-axis direction), v y，(t-1) is the target's vertical speed (y-axis direction) at the previous moment; a x，t is the acceleration of the target in the horizontal direction (x-axis direction) at the current moment, a x，(t-1) is the acceleration of the target in the horizontal direction (x-axis direction) at the previous moment; a y，t is the acceleration of the target in the vertical direction (y-axis direction) at the current moment, a y，(t-1) It is the acceleration of the target in the vertical direction (y-axis direction) at the previous moment.

5. The method for determining the quality of snakehead fish seed according to claim 1, wherein: The quality of fry is graded based on the constructed algorithm model. By setting different thresholds, the quality of fry is divided into different levels such as excellent, good, medium and poor.

Images