An intelligent screening and breeding identification method, system and device for tilapia fry

Abstract

This invention belongs to the field of automatic control optimization and provides a method, system, and device for intelligent screening, separation, and identification of tilapia fry. It collects behavioral image data of the fry and makes judgments based on the time-shifting differences corresponding to feeding frequency and fin movement amplitude. It also makes judgments based on the time-shifting differences corresponding to movement trajectory deviation and local density. To determine whether to use the time-shifting differences of feeding frequency, fin movement amplitude, movement trajectory deviation, and local density data across different time windows to generate a behavioral differentiation adjustment factor, which is then used to regulate water flow. This achieves more precise and real-time sorting control and effectively reduces equipment malfunctions and fry stress. It not only improves separation accuracy but also enhances the system's adaptability and stability.

Description

Technical Field

[0001] This invention belongs to the field of automated control optimization, specifically relating to a method, system, and equipment for intelligent screening, separation, and identification of tilapia fry. Background Technology

[0002] With the continuous development of aquaculture, intelligent and automated aquaculture technologies have become an important trend in the industry. Especially in the tilapia fry separation stage, the application of intelligent sorting technology is gradually replacing traditional manual screening and physical separation methods, improving aquaculture efficiency and accuracy. Currently, many aquaculture systems use intelligent sorting devices based on computer vision and sensors. These devices collect real-time behavioral data of the fry using high-frame-rate cameras, infrared sensors, and other equipment, and analyze characteristics such as body length, swimming speed, and feeding activity using algorithms such as deep learning. These technologies can already identify differences in fry size to a certain extent, enabling classification and sorting. For example, deep learning technology can be used to analyze behavioral characteristics such as feeding frequency and movement trajectory of fry to determine whether tilapia fry are suitable for allocation to a specific aquaculture environment. Simultaneously, many sorting devices have begun to apply automatic control systems, such as servo motors controlling screen aperture and water flow rate to adapt to the passage requirements of fry of different sizes. In particular, intelligent control systems based on mechanical devices, combined with automatic adjustment functions, can adjust equipment parameters according to real-time fry data, improving sorting efficiency and accuracy.

[0003] While existing technologies have made some progress in intelligent screening and rearing of fish fry, they still rely too heavily on manual or preset parameter control. Many existing rearing devices depend on manual intervention or fixed screen aperture and water flow velocity parameters for adjustment. Although some devices can dynamically adjust screen aperture and water flow velocity using sensors, such as the fish fry screening and counting device described in patent document CN109042453B, these adjustments are usually based on preset fixed parameters or simple rules, lacking intelligent analysis based on changes in fish fry behavior. Therefore, they cannot flexibly address the behavioral differences between different tilapia fry during aquaculture, resulting in limited improvement in rearing accuracy and farming efficiency.

[0004] While existing technologies include devices that sort fish fry based on behavioral data, they often fail to accurately capture the subtle behavioral changes caused by the squeezing of smaller fry by larger fry. For example, if a fish fry intelligent sorting system and its method described in patent document CN117502356A are used, smaller tilapia fry in a school may exhibit abnormal behaviors such as reduced feeding frequency and restricted movement under the squeezing of larger fry. These subtle changes are often difficult to identify through simple threshold judgments or traditional single-behavioral analysis methods, thus affecting the sorting accuracy.

[0005] While many current sorting devices possess automated functions, most rely on static behavior patterns and simple threshold judgments, lacking the ability to make real-time, dynamic adjustments based on differences in fish fry behavior. For example, the fish fry grading device for aquaculture described in patent document CN117859689A often fails to flexibly adjust the screen aperture, screen angle, or water flow rate according to the degree of change in fish fry behavior. It can only operate based on fixed rules or preset parameters, thus failing to adapt to the complex environmental changes and diverse fish fry behaviors in aquaculture.

[0006] In traditional fish fry sorting devices, fish fry are often subjected to mechanical and physical forces or water flow impacts during the sorting process, leading to increased stress responses. Damaged fish fry health not only affects growth rate but may also reduce sorting accuracy. Furthermore, existing technologies have limited means to alleviate fish fry stress responses, often relying on adjustments to single equipment parameters rather than fully utilizing dynamic monitoring data for precise adjustments.

[0007] Current intelligent sorting systems often lack adaptability when dealing with complex aquaculture environments such as changes in water quality, lighting, and differences in fish fry populations. Existing technologies typically rely on fixed parameters for automatic adjustments, which limits the system's flexibility and accuracy in responding to different environmental and aquatic conditions. Summary of the Invention

[0008] The purpose of this invention is to propose an intelligent screening, separation, and identification method, system, and device for tilapia fry, in order to solve one or more technical problems existing in the prior art, and at least provide a beneficial option or create conditions.

[0009] To achieve the above objectives, according to one aspect of the present invention, a method for intelligent screening and separate rearing of tilapia fry is provided, the method comprising the following steps:

[0010] Collect behavioral image data of fish fry, mark the area occupied by each fish fry in the image, and sort the fish fry according to the area size.

[0011] Data on feeding frequency, fin movement amplitude, movement trajectory deviation, and local density of fish fry were collected at different time windows.

[0012] The differences in the direction of time shift were calculated between different time windows based on data of feeding frequency, fin movement amplitude, movement trajectory deviation, and local density.

[0013] The behavior differentiation regulation factor is determined based on the time-shifting direction difference corresponding to feeding frequency and the time-shifting direction difference corresponding to fin movement amplitude, as well as the time-shifting direction difference corresponding to movement trajectory deviation and local density. This determination is used to identify whether it is necessary to use the time-shifting direction differences of feeding frequency, fin movement amplitude, movement trajectory deviation, and local density data in different time windows to generate a behavior differentiation regulation factor. If so, the behavior differentiation regulation factor is used to regulate water flow.

[0014] Furthermore, underwater camera equipment is used in conjunction with a multi-angle camera layout to collect real-time behavioral image data of the fish fry. Computer vision models are used to identify and mark each fish fry, and the area occupied by each fish fry in the image frame of the behavioral image data is marked.

[0015] Furthermore, the method for regulating water flow using the aforementioned behavior differentiation adjustment factor is as follows: water flow is regulated by controlling the adjustment of the screen aperture using the aforementioned behavior differentiation adjustment factor.

[0016] Furthermore, the method for regulating water flow using the aforementioned behavioral differentiation modulator may be as follows:

[0017] The water flow is regulated by controlling the screen tilt angle using the aforementioned behavior-differentiating adjustment factor.

[0018] Furthermore, the method for regulating water flow using the aforementioned behavioral differentiation modulator may also be:

[0019] Water flow is regulated by controlling the flow velocity using the aforementioned behavioral differentiation adjustment factor.

[0020] Furthermore, the method for calculating the time shift direction differences between different time windows based on data such as feeding frequency, fin movement amplitude, movement trajectory deviation, and local density is as follows:

[0021] The fish fry samples are sorted according to the area they occupy in the image. In each time window, based on the sorting of the fish fry samples, a feeding frequency vector is formed by the feeding frequency of each fish fry sample, a fin movement amplitude vector is formed by the fin movement amplitude of each fish fry sample, a movement trajectory deviation vector is formed by the movement trajectory deviation of each fish fry sample, and a local density vector is formed by the local density of each fish fry sample.

[0022] The cosine similarity between the feeding frequency vector of the previous time window and the feeding frequency vector of the subsequent time window is used as the difference in the feeding frequency time shift direction; the cosine similarity between the fin movement amplitude vector of the previous time window and the fin movement amplitude vector of the subsequent time window is used as the difference in the fin movement amplitude time shift direction; the cosine similarity between the movement trajectory deviation vector of the previous time window and the movement trajectory deviation vector of the subsequent time window is used as the difference in the movement trajectory deviation time shift direction; and the cosine similarity between the local density vector of the previous time window and the local density vector of the subsequent time window is used as the difference in the local density time shift direction.

[0023] Furthermore, the method for generating behavioral alienation modulators using data on feeding frequency, fin movement amplitude, movement trajectory deviation, and local density across different time windows is as follows:

[0024] The ratio of the exponentially expressed value of the difference in the direction of movement trajectory deviation combined with the difference in the direction of local density deviation to the exponentially expressed value of the difference in the direction of feeding frequency combined with the difference in the direction of fin movement amplitude deviation is the behavioral alienation regulation factor.

[0025] Furthermore, the specific steps involved in making conditional judgments include:

[0026] Determine if the absolute value G1 of the difference between the value of the feeding frequency in the time shift direction and the value of the fin movement amplitude in the time shift direction is greater than 1 / 2. If not, then there is no need to adjust the water flow.

[0027] If so, then continue to determine whether the absolute value G2 of the difference between the value of the deviation of the motion trajectory from the time-shift direction and the value of the local density in the time-shift direction is greater than 1 / 2. If not, no water flow regulation is required, but if so, the behavior alienation adjustment factor needs to be calculated.

[0028] This invention also provides an intelligent screening and separation identification system for tilapia fry. The system includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the intelligent screening and separation identification method for tilapia fry. The system can run on computing devices such as desktop computers, laptops, handheld computers, and cloud data centers. The runnable system may include, but is not limited to, processors, memory, and server clusters. The processor executes the computer program within the following system units:

[0029] The data acquisition unit is used to collect behavioral image data of fish fry, mark the area occupied by each fish fry in the image, and sort the fish fry according to the area size.

[0030] The feature calculation unit is used to acquire data on fish fry, including feeding frequency, fin movement amplitude, movement trajectory deviation, and local density, in different time windows.

[0031] The time-shift direction unit is used to calculate the time-shift direction difference between different time windows based on data such as feeding frequency, fin movement amplitude, movement trajectory deviation, and local density.

[0032] The judgment and adjustment unit is used to make judgments based on the time-shifting direction difference corresponding to feeding frequency and the time-shifting direction difference corresponding to fin movement amplitude, and based on the time-shifting direction difference corresponding to movement trajectory deviation and the time-shifting direction difference corresponding to local density, so as to determine whether to use the time-shifting direction differences of feeding frequency, fin movement amplitude, movement trajectory deviation and local density data in different time windows to generate a behavior differentiation adjustment factor. If so, the behavior differentiation adjustment factor is used to regulate water flow.

[0033] Correspondingly, the present invention also provides an electronic device, a readable storage medium, and a computer program product:

[0034] An electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the intelligent screening and separation identification method for tilapia fry and the methods for each step therein.

[0035] A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to execute the intelligent screening, separation, and identification method for tilapia fry and the methods for each step therein.

[0036] A computer program product includes a computer program that, when executed by a processor, implements the intelligent screening, separation, and identification method for tilapia fry and the methods for each step therein.

[0037] The beneficial effects of this invention are as follows: This invention provides an intelligent screening, separation, and identification method, system, and device for tilapia fry. It collects behavioral image data of the fry, judges based on the time-shifting differences corresponding to feeding frequency and fin movement amplitude, and further judges based on the time-shifting differences corresponding to movement trajectory deviation and local density. It then determines whether to use the time-shifting differences of feeding frequency, fin movement amplitude, movement trajectory deviation, and local density data across different time windows to generate a behavioral differentiation regulation factor, which is then used to regulate water flow. This achieves more precise and real-time sorting control and effectively reduces equipment malfunctions and fry stress responses. It not only improves separation accuracy but also enhances the system's adaptability and stability. Attached Figure Description

[0038] The above and other features of the present invention will become more apparent from the detailed description of the embodiments shown in conjunction with the accompanying drawings. In the accompanying drawings, the same reference numerals denote the same or similar elements. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without any creative effort. In the drawings:

[0039] Figure 1 The diagram shows a flowchart of an intelligent screening and separation method for tilapia fry.

[0040] Figure 2 The diagram shows the system structure of an intelligent screening and separation identification system for tilapia fry. Detailed Implementation

[0041] The following will provide a clear and complete description of the concept, specific structure, and technical effects of the present invention in conjunction with the embodiments and accompanying drawings, so as to fully understand the purpose, solution, and effects of the present invention. It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0042] In the description of this invention, "several" means one or more, "more than" means two or more, "greater than," "less than," and "exceeding" are understood to exclude the stated number, while "above," "below," and "within" are understood to include the stated number. The use of "first" and "second" in the description is merely for distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the order of the indicated technical features.

[0043] like Figure 1 The diagram shown is a flowchart of a method for intelligent screening, separate rearing, and identification of tilapia fry according to the present invention. The following is a combination of... Figure 1This invention describes a method, system, and device for intelligent screening, separate rearing, and identification of tilapia fry according to embodiments of the present invention.

[0044] This invention proposes an intelligent screening and separate rearing identification method for tilapia fry, the method specifically including the following steps:

[0045] Collect behavioral image data of fish fry, mark the area occupied by each fish fry in the image, and sort the fish fry according to the area size.

[0046] Data on feeding frequency, fin movement amplitude, movement trajectory deviation, and local density of fish fry were acquired within different time windows. Based on the data on feeding frequency, fin movement amplitude, movement trajectory deviation, and local density, the time shift direction differences between different time windows were calculated.

[0047] The absolute difference between the time-shift direction difference corresponding to the feeding frequency and the time-shift direction difference corresponding to the fin movement amplitude is used to determine whether the value is greater than 1 / 2; otherwise, the water flow regulation is not activated.

[0048] If so, the condition is determined based on the absolute difference between the time-shift direction difference corresponding to the deviation of the movement trajectory and the time-shift direction difference corresponding to the local density. If the difference is still greater than 1 / 2, the behavior differentiation regulation factor is generated by using the data of feeding frequency, fin movement amplitude, movement trajectory deviation and local density in different time windows. The behavior differentiation regulation factor is then used to regulate the water flow.

[0049] In some embodiments, specifically:

[0050] In the fish pond of the tilapia fry to be tested, set up the regulating equipment and its control and activation device;

[0051] Each sample of fish fry in the fish school was photographed, identified and marked, and the area occupied by each sample of fish fry in the image was determined.

[0052] For each fish population, data on feeding frequency, fin movement amplitude, movement trajectory deviation, and local density were obtained within different time windows, with the earlier time window designated as the front time window and the later time window as the back time window.

[0053] The fish fry samples are sorted according to the area they occupy in the image. Within each time window, a feeding frequency vector is formed based on the feeding frequency of each fish fry sample, a fin movement amplitude vector is formed based on the fin movement amplitude of each fish fry sample, a movement trajectory deviation vector is formed based on the movement trajectory deviation of each fish fry sample, and a local density vector is formed based on the local density of each fish fry sample.

[0054] The cosine similarity between the feeding frequency vector of the previous time window and the feeding frequency vector of the subsequent time window is taken as the feeding frequency time shift direction F; the cosine similarity between the fin movement amplitude vector of the previous time window and the fin movement amplitude vector of the subsequent time window is taken as the fin movement amplitude time shift direction A; the cosine similarity between the trajectory deviation vector of the previous time window and the trajectory deviation vector of the subsequent time window is taken as the trajectory deviation time shift direction T; and the cosine similarity between the local density vector of the previous time window and the local density vector of the subsequent time window is taken as the local density time shift direction D.

[0055] Then, perform conditional checks:

[0056] Determine if the absolute value G1 of the difference between the value of the feeding frequency in the time shift direction and the value of the fin movement amplitude in the time shift direction is greater than 1 / 2. If not, then there is no need to adjust the water flow.

[0057] If so, then continue to determine whether the absolute value G2 of the difference between the value of the deviation of the motion trajectory from the time-shift direction and the value of the local density from the time-shift direction is greater than 1 / 2. If not, then there is no need to adjust the water flow.

[0058] However, if it is still greater than that, then BDaF needs to be calculated:

[0059] BDaF = [Exp(Movement trajectory deviation in time shift direction T) * exp(Local density in time shift direction D)] / [exp(Feeding frequency in time shift direction F) * exp(Fin movement amplitude in time shift direction A)];

[0060] The formula simplifies to BDaF = [Exp(time shift direction of movement trajectory deviation T + time shift direction of local density D)] / [exp(time shift direction of feeding frequency F + time shift direction of fin movement amplitude A)];

[0061] In certain situations, the uncertain effect can be simplified to BDaF = (move trajectory deviation from time shift direction T + local density time shift direction D) / exp(feeding frequency time shift direction F + fin movement amplitude time shift direction A);

[0062] 1-BDaF was used as a weight to adjust the water flow.

[0063] Furthermore, underwater camera equipment is used in conjunction with a multi-angle camera layout to collect real-time behavioral image data of the fish fry. Computer vision models are used to identify and mark each fish fry, and the area occupied by each fish fry in the image frame of the behavioral image data is marked.

[0064] In some embodiments, high frame rate and high resolution underwater camera equipment can be used, preferably supplemented by a multi-angle and multi-view camera layout to acquire data from different perspectives and capture image data including motion details of the behavior of each fish fry. A pre-trained large-scale visual model, including but not limited to the YOLO model, is used to perform target detection of each fish fry in the images and videos, and to identify and label the fish fry.

[0065] In the identified and labeled video sequences, a 3D CNN model is used to directly capture fine-grained motion features in the spatial and temporal dimensions of the video frame sequence. This involves extracting both local and overall motion features of each fish fry, including their fins and tails. This includes, but is not limited to, detecting the amplitude of fin movements between frames in the video sequence as fin movement amplitude; calculating feeding frequency by detecting the proportion of times each fish fry was detected feeding behavior out of the total number of frames; calculating trajectory deviation by detecting the distance each fish fry moved from the initial frame to the final frame between frames in the video sequence; and detecting the density of each fish fry within a radius of 10-15 cm centered on the fry as local density between frames in the video sequence.

[0066] Preferably, for situations where there is insufficient light or turbidity in the underwater environment, infrared thermal imaging can be used to supplement visible light data.

[0067] Preferably, the four data points of each fish fry—feeding frequency, fin movement amplitude, movement trajectory deviation, and local density—are first normalized before being used in subsequent calculations.

[0068] Furthermore, the method for regulating water flow using the aforementioned behavior differentiation adjustment factor is as follows: water flow is regulated by controlling the adjustment of the screen aperture using the aforementioned behavior differentiation adjustment factor.

[0069] In some embodiments, the adjustment of the screen aperture can be achieved through automated control, including but not limited to a screening and rearing device as described in patent document CN116158395A. This device uses multiple screening valves and screens to form zones with different apertures, thereby enabling the screening of fry of different sizes. The screen aperture decreases sequentially to accommodate the passage requirements of fry of different sizes. Preferably, when the behavioral differentiation factor (BDaF) is higher than 1, it indicates that small-sized fry are behaving abnormally due to compression by larger-sized fry. In this case, the screen aperture needs to be reduced to ensure that small-sized fry are not mistakenly screened out. Generally, the screen aperture can be dynamically adjusted based on BDaF. For example, when BDaF is 1.23, the screen aperture can be reduced by a value of 1-BDaF to allow small fry to pass through smoothly. A value close to -0.23 can be used to round down to one-fifth, and the screen aperture is reduced approximately proportionally to prevent fry from being filtered out by larger screen apertures. A BDaF value below 1 indicates minimal behavioral variation in fish fry. In this case, the screen mesh size can be appropriately increased to improve sorting efficiency and reduce interference during fry selection. The screen mesh size adjustment is accomplished by a servo motor or automatic adjustment device. By controlling the motor and transmission system, the screen mesh size is automatically adjusted based on the BDaF value. For high BDaF, the motor control system reduces the mesh size to prevent fry from being crushed or mis-screened; for low BDaF, the mesh size is increased to improve sorting efficiency.

[0070] Furthermore, the method for regulating water flow using the aforementioned behavioral differentiation modulator may be as follows:

[0071] The water flow is regulated by controlling the screen tilt angle using the aforementioned behavior-differentiating adjustment factor.

[0072] In some embodiments, by changing the screen inclination angle, the speed and path of fish fry passing through the screen can be adjusted, thereby improving sorting accuracy. Changes in the screen inclination angle affect the water flow speed and the flow trajectory of the fish fry, thus affecting the stratification and sorting of the fish fry. When the BDaF value is greater than 1, it indicates a strong squeezing effect on the fish fry. Increasing the inclination angle helps smaller fish fry pass through the screen smoothly. For example, when BDaF is 1.09, the original inclination angle is adjusted to approximately 109% (rounded down) to reduce the residence time of the fish fry on the screen and avoid prolonged squeezing. When the BDaF value is not higher than 1, it indicates that the fish fry behavior is relatively uniform and the squeezing effect is weak. In this case, the screen inclination angle can be appropriately reduced to improve the natural flow of the fish fry and reduce the burden on the screen system. The screen inclination angle is controlled by a servo motor or hydraulic system and dynamically adjusted based on the specific BDaF value (rounded down). The system described in this invention adjusts the screen tilt angle based on the real-time monitored BDaF value, thereby optimizing the sorting path and speed of the fish fry.

[0073] Furthermore, the method for regulating water flow using the aforementioned behavioral differentiation modulator may also be:

[0074] Water flow is regulated by controlling the flow velocity using the aforementioned behavioral differentiation adjustment factor.

[0075] In some embodiments, adjusting the water flow velocity can effectively guide fry of different sizes to stratify according to their behavioral characteristics, thereby improving sorting efficiency. The water flow velocity directly affects the stratification, flow speed, and position of the fry. When BDaF is greater than 1, it indicates a significant squeezing effect, and the behavior of tilapia fry changes drastically. In this case, it is preferable to increase the water flow velocity to help larger fry move quickly while slowing down the flow of smaller fry, making them easier to stratify correctly. When BDaF is less than 1, it indicates a smaller change in fry behavior, and excessive flow velocity may affect the natural flow of the fry. In this case, the water flow velocity can be controlled to decrease to reduce interference with the fry. In one embodiment, the water flow velocity is regulated by a water pump control system, and the power or flow rate of the water pump can be automatically adjusted according to BDaF. At higher BDaF values, the power of the water pump increases proportionally to increase the flow velocity; at lower BDaF values, the water flow velocity decreases linearly to ensure normal flow of the fry.

[0076] Furthermore, the method for calculating the time shift direction differences between different time windows based on data such as feeding frequency, fin movement amplitude, movement trajectory deviation, and local density is as follows:

[0077] The fish fry samples are sorted according to the area they occupy in the image. In each time window, based on the sorting of the fish fry samples, a feeding frequency vector is formed by the feeding frequency of each fish fry sample, a fin movement amplitude vector is formed by the fin movement amplitude of each fish fry sample, a movement trajectory deviation vector is formed by the movement trajectory deviation of each fish fry sample, and a local density vector is formed by the local density of each fish fry sample.

[0078] The cosine similarity between the feeding frequency vector of the previous time window and the feeding frequency vector of the subsequent time window is used as the difference in the feeding frequency time shift direction; the cosine similarity between the fin movement amplitude vector of the previous time window and the fin movement amplitude vector of the subsequent time window is used as the difference in the fin movement amplitude time shift direction; the cosine similarity between the movement trajectory deviation vector of the previous time window and the movement trajectory deviation vector of the subsequent time window is used as the difference in the movement trajectory deviation time shift direction; and the cosine similarity between the local density vector of the previous time window and the local density vector of the subsequent time window is used as the difference in the local density time shift direction.

[0079] The area of ​​tilapia fry in an image reflects their actual position and relative size within the field of view. The size of tilapia fry is usually closely related to their body size, health status, and activity level. In behavioral analysis of tilapia fry, large and small fry may exhibit significant differences in behavior. For example, large tilapia fry typically swim faster and feed more frequently, while small fry may be relatively slower or have weaker feeding behavior. Sort the fry by area to help track behavioral patterns based on changes in body size, avoiding data noise caused by different fry distribution ratios. The sorted feature vectors allow for the placement of fry in order of size, effectively reducing interference from different fry proportions. This allows the analysis to focus only on behavioral changes in groups of similar body sizes, rather than simply mixing data from all fry. This data organization makes subsequent behavioral analysis more accurate and efficient. After sorting, when calculating differences in behavioral direction over time, time-series analysis can be performed on groups of fry of the same body size, avoiding interference from differences caused by mixing fry of different body sizes. During the behavioral changes of fish fry, smaller fry are usually the most affected by crowding. Sort them and then perform behavioral feature vector analysis to more accurately capture the behavioral abnormalities caused by crowding from larger fry. Because the behavioral differences among smaller fry are relatively small, sorting allows for more precise detection of behavioral abnormalities caused by crowding from larger fry, such as decreased feeding frequency. This makes the calculation of behavioral differentiation regulators more accurate and allows for automatic adjustment of water flow, screens, and other equipment to ensure precise separation of fry and avoid missorting and stress responses.

[0080] This method calculates the temporal shift direction of fry behavior by using data on feeding frequency, fin movement amplitude, movement trajectory deviation, and local density within different time windows to capture behavioral changes. Existing technologies typically rely on single behavioral features such as feeding frequency or movement trajectory deviation, while this invention introduces multi-dimensional features including feeding frequency, fin movement amplitude, movement trajectory deviation, and local density for comprehensive behavioral analysis. By comparing behavioral changes within different time windows, it can accurately capture behavioral differences in fry, especially the behavioral changes caused by larger fry squeezing smaller fry. This method can dynamically calculate and compare behavioral differences in fry in real time, offering higher real-time performance and flexibility compared to traditional static parameter-based screening methods. By monitoring the dynamic behavioral changes of fry within different time windows, it is better adapted to complex aquaculture environments and the diversity of fry behavior. Existing methods often rely on static thresholds or simple behavioral standards, failing to capture subtle behavioral changes; however, by calculating temporal shift direction differences, this invention can accurately assess subtle behavioral changes in fry over a short period, which is crucial for timely identification of behavioral abnormalities in smaller fry caused by squeezing from larger fry. The method described in this invention combines the directional differences of multiple key features across various dimensions of time, including feeding frequency, fin movement amplitude, movement trajectory deviation, and local density. Compared to a single indicator, it can more comprehensively reflect the behavioral state of fish fry. By calculating the temporal differences in behavioral features at different times, this invention can dynamically monitor changes in fish fry behavior and avoid misjudgments caused by static threshold limitations.

[0081] Furthermore, the method for generating behavioral alienation modulators using data on feeding frequency, fin movement amplitude, movement trajectory deviation, and local density across different time windows is as follows:

[0082] The ratio of the exponentially expressed value of the difference in the direction of movement trajectory deviation combined with the difference in the direction of local density deviation to the exponentially expressed value of the difference in the direction of feeding frequency combined with the difference in the direction of fin movement amplitude deviation is the behavioral alienation regulation factor.

[0083] In this embodiment of the invention, the behavioral differentiation regulation factor, denoted as BDaF, is a regulation coefficient based on the changes in the behavioral characteristics of fish fry within different time windows. It is used to precisely adjust the parameters of the separation device, including screen aperture, screen inclination angle, and water flow velocity, thereby optimizing the sorting effect of the fish fry. The calculation of BDaF is based on the feeding frequency, fin movement amplitude, movement trajectory deviation, and time-shifting differences in local density of the fish fry, representing the degree of behavioral change of the fish fry under environmental influences such as compression. The magnitude of the BDaF value directly affects the regulation strength of the separation device, used to automatically adjust the working state of the water flow, screen, and sorting device to ensure accurate grading and healthy growth of the fish fry.

[0084] Increased local density and trajectory deviation, coupled with decreased feeding frequency and reduced fin movement amplitude, indicate a suppression signal. Trajectory perturbation and increased density are represented in the numerator, while decreased feeding and fin activity are represented in the denominator. It represents the overall trend of behavioral change from one window to the next, not judging static values, but rather the structural relationship between behavioral trends; it does not require preset weights or rely on model training, but rather constructs a structural indicator that can measure the alienation effect.

[0085] The behavioral differentiation adjustment factor is used to dynamically adjust the working state of the fish fry screening device, including the screen aperture, screen tilt angle, and water flow velocity. Its value is based on the feeding frequency, fin movement amplitude, movement trajectory deviation, and time shift direction of local density of the fish fry within the preceding and following time windows. The differences in these behavioral characteristics are calculated using a cosine similarity algorithm, and a BDaF value is generated based on these differences.

[0086] The magnitude of the BDaF value determines the adjustment range of the separation device. A higher value indicates a greater degree of compression of the fry, requiring stronger adjustments to ensure sorting accuracy and fry health. The value of the behavioral alienation adjustment factor is calculated. Based on the real-time changes in BDaF, the system can automatically adjust the parameters of the separation device, such as reducing the screen aperture, increasing the screen angle, or adjusting the water flow rate, thereby achieving precise classification and stratification of the fry and effectively reducing the stress response caused by compression.

[0087] This invention generates a behavioral differentiation adjustment factor (BDaF) by calculating the temporal differences in feeding frequency, fin movement amplitude, movement trajectory deviation, and local density at different times. This factor is used to automatically adjust the screen aperture, screen inclination angle, and water flow velocity, thereby regulating the working state of the separation device. The automatic adjustment of BDaF based on behavioral differences is dynamically calculated according to the temporal differences in fry behavior, representing the degree of change in fry behavior. Adjusting equipment parameters through BDaF achieves automated adjustment and precise management of fry sorting. Existing technologies typically rely on fixed screen apertures or fixed sorting strategies, while this invention can adaptively adjust based on real-time data, greatly improving sorting accuracy. Traditional methods, which rely on manual control or fixed parameter adjustments, are easily affected by environmental factors and struggle to cope with the changing aquaculture environment. This invention, by reflecting the dynamic changes in fry behavior through BDaF, can precisely adjust the screen aperture, screen inclination angle, and water flow velocity, reducing stress responses caused by fry squeezing or over-sorting. By calculating BDaF and adjusting the working state of the separation device in real time, sorting accuracy can be improved and fry stress response can be reduced. This invention minimizes fry stress and ensures that smaller fry are not damaged or mis-sorted due to compression by larger fry. Compared with existing methods, this invention provides higher precision in fry classification and stratification. The BDaF adjustment mechanism, based on dynamic behavioral differences, reflects the dynamic trend of fry behavior changes, allowing equipment parameter adjustments to respond in real time to changes in fry behavior, thus ensuring accurate sorting. The adaptive mechanism of this invention automatically optimizes equipment parameters based on real-time collected behavioral data, improving sorting accuracy and reducing stress response.

[0088] Furthermore, the specific steps involved in making conditional judgments include:

[0089] Determine if the absolute value G1 of the difference between the value of the feeding frequency in the time shift direction and the value of the fin movement amplitude in the time shift direction is greater than 1 / 2. If not, then there is no need to adjust the water flow.

[0090] If so, then continue to determine whether the absolute value G2 of the difference between the value of the deviation of the motion trajectory from the time-shift direction and the value of the local density in the time-shift direction is greater than 1 / 2. If not, no water flow regulation is required, but if so, the behavior alienation adjustment factor needs to be calculated.

[0091] In this embodiment, 1 / 2 is preferably chosen to define a critical value for behavioral change, especially in the analysis of fish fry behavior patterns, such as feeding frequency and fin movement amplitude. Behavioral differences among fry often exhibit certain regularities. Using 1 / 2 as a dividing point ensures that the system only adjusts when the behavioral difference exceeds a certain critical value; if it is less, the difference is considered insufficient to affect separation, preventing the system from overreacting. If the threshold is set too low, such as 1 / 4 or less, it may lead to overly sensitive adjustments, causing even minor behavioral fluctuations to trigger system adjustments, resulting in unnecessary equipment adjustments and energy waste. Using 1 / 2 as a threshold reduces ineffective adjustments, ensuring that the system only responds when behavioral differences are significant.

[0092] The design employs a multi-layered filtering mechanism, with two separate checks targeting G1 and G2. The first check determines if G1 is greater than 1 / 2, ensuring the difference between feeding frequency and fin movement amplitude is significant. Then, it checks if G2 is greater than 1 / 2, further filtering out unnecessary adjustments. This layered approach enhances system robustness, ensuring precise adjustments even in complex environments. The first layer checks the difference between feeding frequency and fin movement amplitude, which typically reflect the fry's basic behavioral state, such as food availability and activity level. The second layer, if the first layer deems a significant behavioral difference, compares movement trajectory deviation and local density. This second layer strengthens the detailed analysis of behavior, preventing over-adjustment. Through these two checks, the system's response becomes more precise. For example, if only G1 > 1 / 2 while G2 <= 1 / 2, it indicates a small behavioral difference, requiring no comprehensive adjustment. These two checks effectively filter out unnecessary adjustments, making the system more accurate and real-time, avoiding over-responding to unimportant changes. Without this secondary judgment, the system might adjust its devices based on minute fluctuations, such as differences in feeding frequency and fin movement amplitude, leading to frequent changes, energy waste, and over-adjustment. Therefore, setting a 1 / 2 judgment threshold avoids this problem and improves the device's efficiency and stability.

[0093] In this embodiment, this step uses conditional judgments based on feeding frequency, fin movement amplitude, movement trajectory deviation, and temporal differences in local density to ensure that the control mechanism is activated only when adjustments to the separation equipment are needed, avoiding unnecessary operations. By effectively avoiding unnecessary equipment adjustments through conditional judgments G1 and G2, this invention can ensure the sorting effect of fish fry while avoiding frequent or unnecessary equipment adjustments. Traditional automated separation devices may adjust the equipment immediately after each data collection, while this invention, by first judging the significance of behavioral changes, only activates equipment control when necessary, reducing energy consumption and equipment wear. By comprehensively judging the temporal differences of multiple behavioral characteristics, this invention can ensure that adjustments are only made when there are significant changes in fish fry behavior, thereby avoiding erroneous equipment adjustments due to minor fluctuations. This design effectively reduces ineffective adjustment operations, improving the operating efficiency and stability of the equipment. At the same time, the operating cycle and service life of the equipment are extended, saving breeding costs. This invention, through multi-level conditional judgments, ensures that the equipment is adjusted only when needed, avoiding over-adjustment or unnecessary operations caused by errors. This method can accurately judge the degree of change in fish fry behavior, ensuring the accuracy of equipment adjustment and avoiding overreaction.

[0094] The intelligent screening and separation identification system for tilapia fry runs on any computing device, such as a desktop computer, laptop computer, handheld computer, or cloud data center. The computing device includes a processor, a memory, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps in the intelligent screening and separation identification method for tilapia fry. The runnable system may include, but is not limited to, a processor, a memory, and a server cluster.

[0095] An embodiment of the present invention provides an intelligent screening and separation identification system for tilapia fry, such as... Figure 2 As shown, an intelligent screening and separation identification system for tilapia fry in this embodiment includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps in the above-described embodiment of the intelligent screening and separation identification method for tilapia fry. The processor executes the computer program in the following system units:

[0096] The data acquisition unit is used to collect behavioral image data of fish fry, mark the area occupied by each fish fry in the image, and sort the fish fry according to the area size.

[0097] The feature calculation unit is used to acquire data on fish fry, including feeding frequency, fin movement amplitude, movement trajectory deviation, and local density, in different time windows.

[0098] The time-shift direction unit is used to calculate the time-shift direction difference between different time windows based on data such as feeding frequency, fin movement amplitude, movement trajectory deviation, and local density.

[0099] The judgment and adjustment unit is used to make judgments based on the time-shifting direction difference corresponding to feeding frequency and the time-shifting direction difference corresponding to fin movement amplitude, and based on the time-shifting direction difference corresponding to movement trajectory deviation and the time-shifting direction difference corresponding to local density, so as to determine whether to use the time-shifting direction differences of feeding frequency, fin movement amplitude, movement trajectory deviation and local density data in different time windows to generate a behavior differentiation adjustment factor. If so, the behavior differentiation adjustment factor is used to regulate water flow.

[0100] In order to better unify the linear relationship and probabilistic connection between physical quantities with different units of measurement, dimensionless processing can be performed on different physical quantities.

[0101] Preferably, all undefined variables in this invention, if not explicitly defined, can be manually set thresholds.

[0102] The aforementioned intelligent screening and rearing identification system for tilapia fry can run on computing devices such as desktop computers, laptops, handheld computers, and cloud data centers. This system includes, but is not limited to, a processor and a memory. Those skilled in the art will understand that the examples described are merely illustrations of an intelligent screening and rearing identification method, system, and device for tilapia fry and do not constitute a limitation on such a method, system, and device. The system may include more or fewer components, combinations of certain components, or different components. For example, the intelligent screening and rearing identification system for tilapia fry may also include input / output devices, network access devices, and buses.

[0103] The present invention also provides an electronic device, a readable storage medium, and a computer program product:

[0104] An electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the intelligent screening and separation identification method for tilapia fry and the methods for each step therein.

[0105] A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to execute the intelligent screening, separation, and identification method for tilapia fry and the methods for each step therein.

[0106] A computer program product includes a computer program that, when executed by a processor, implements the intelligent screening, separation, and identification method for tilapia fry and the methods for each step therein.

[0107] The term "electronic device" is intended to refer to various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic devices can also refer to various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0108] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0109] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0110] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0111] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0112] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0113] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other.

[0114] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete component gate circuits, transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. This processor is the control center of the intelligent tilapia fry screening, separation, and identification system, connecting various sub-regions of the system via various interfaces and lines.

[0115] The memory can be used to store the computer program and / or modules. The processor, by running or executing the computer program and / or modules stored in the memory, and by calling the data stored in the memory, realizes various functions of the intelligent screening and separation identification method, system, and device for tilapia fry. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created based on the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0116] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this invention can be achieved, and this is not limited herein.

[0117] This invention provides an intelligent screening, separation, and identification method, system, and device for tilapia fry. It collects behavioral image data of the fry and makes judgments based on the time-shifting differences corresponding to feeding frequency, fin movement amplitude, and movement trajectory deviation and local density. The method determines whether to use the time-shifting differences of feeding frequency, fin movement amplitude, movement trajectory deviation, and local density data across different time windows to generate a behavioral differentiation regulation factor. This behavioral differentiation regulation factor is then used to regulate water flow. This achieves more precise and real-time sorting control and effectively reduces equipment malfunctions and fry stress. It not only improves separation accuracy but also enhances the system's adaptability and stability.

[0118] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for intelligent screening and breeding identification of tilapia fry, characterized in that, The method includes: Behavioral image data of fish fry were collected, the area occupied by each fish fry in the image was marked, and the fish fry were sorted according to the area size; data on feeding frequency, fin movement amplitude, movement trajectory deviation and local density of fish fry were obtained in different time windows. Based on data on feeding frequency, fin movement amplitude, trajectory deviation, and local density, the cosine similarity between the feeding frequency vector of the previous time window and the feeding frequency vector of the subsequent time window is used as the difference in the time shift direction of feeding frequency; the cosine similarity between the fin movement amplitude vector of the previous time window and the fin movement amplitude vector of the subsequent time window is used as the difference in the time shift direction of fin movement amplitude; the cosine similarity between the trajectory deviation vector of the previous time window and the trajectory deviation vector of the subsequent time window is used as the difference in the time shift direction of trajectory deviation; and the cosine similarity between the local density vector of the previous time window and the local density vector of the subsequent time window is used as the difference in the time shift direction of local density. The behavioral differentiation regulation factor is determined by judging the time-shifting differences in feeding frequency and fin movement amplitude, as well as the time-shifting differences in movement trajectory deviation and local density. This is used to identify whether to generate a behavioral differentiation regulation factor based on the time-shifting differences in feeding frequency, fin movement amplitude, movement trajectory deviation, and local density data across different time windows. If so, the behavioral differentiation regulation factor is used to regulate water flow. The behavioral differentiation regulation factor is calculated by comparing the exponentially expressed value of the time-shifting difference in movement trajectory deviation combined with the time-shifting difference in local density with the exponentially expressed value of the time-shifting difference in feeding frequency combined with the time-shifting difference in fin movement amplitude.

2. The method according to claim 1, wherein, in, Using underwater camera equipment with a multi-angle camera layout, behavioral image data of fish fry is collected in real time. Computer vision models are used to identify and mark each fish fry, and the area occupied by each fish fry in the image frame of the behavioral image data is marked.

3. The method according to claim 1, wherein the method is characterized by, The method for regulating water flow using the aforementioned behavioral differentiation modulator is as follows: Water flow is regulated by controlling the adjustment of the screen aperture using the aforementioned behavior-differentiating adjustment factor.

4. The intelligent screening and separate rearing identification method for tilapia fry according to claim 1, characterized in that, The method for regulating water flow using the aforementioned behavioral differentiation modulator may be as follows: The water flow is regulated by controlling the screen tilt angle using the aforementioned behavior-differentiating adjustment factor.

5. The method according to claim 1, wherein the method is characterized by, The method for regulating water flow using the aforementioned behavioral differentiation modulator may be as follows: Water flow is regulated by controlling the flow velocity using the aforementioned behavioral differentiation adjustment factor.

6. The method according to claim 1, wherein the method is characterized by, The method for calculating the difference in time shift direction between different time windows based on data such as feeding frequency, fin movement amplitude, movement trajectory deviation, and local density is as follows: The fish fry samples are sorted according to the area they occupy in the image. Within each time window, a feeding frequency vector is formed based on the feeding frequency of each fish fry sample, a fin movement amplitude vector is formed based on the fin movement amplitude of each fish fry sample, a movement trajectory deviation vector is formed based on the movement trajectory deviation of each fish fry sample, and a local density vector is formed based on the local density of each fish fry sample.

7. The method according to claim 1, wherein the method is characterized by, in, The specific steps involved in performing conditional judgments include: Determine if the absolute value of the difference between the value of the feeding frequency in the time shift direction and the value of the fin movement amplitude in the time shift direction is greater than 1 / 2. If not, there is no need to adjust the water flow. If so, then continue to determine whether the absolute value of the difference between the value of the deviation of the motion trajectory from the time-shift direction and the value of the local density in the time-shift direction is greater than 1 / 2. If not, no water flow regulation is required, but if so, the behavior alienation adjustment factor needs to be calculated.

8. A smart screening and separation identification system for tilapia fry, characterized in that, The intelligent screening and separation identification system for tilapia fry operates on any computing device, such as a desktop computer, a laptop computer, or a cloud data center. The computing device includes a processor, a memory, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of the intelligent screening and separation identification method for tilapia fry as described in any one of claims 1 to 7.

9. An electronic device comprising: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 7.

Images