Automatic grading method for live mandarin fish suitable for factory farming

Abstract

The present invention relates to the technical field of factory farming, and discloses a method for automatically grading live mandarin fish suitable for factory farming. The method comprises: constructing a grading channel with a gradually expanding cross-sectional area and injecting a constant water flow to form a flow field with a decreasing flow velocity gradient, so that mandarin fish individuals with different swimming abilities can autonomously select an adaptive flow velocity segment to reside in, thereby achieving spatial stratification and subsequent partitioned collection. The present invention transforms traditional mechanical grading into a stress-free autonomous positioning process through the synergistic effect of flow field gradients and the behavioral instincts of fish. At the same time, the eddy current correction mechanism induced by the wedge-shaped grooves of the guide grid is used to ensure grading accuracy under high-density working conditions, and multiplexing flow field pulsation signals to achieve non-destructive evaluation of physiological activity, thereby providing a multi-dimensional grading decision-making basis for factory farming.

Description

Technical Field

[0001] The invention relates to an automatic grading method for live mandarin fish suitable for factory farming, belonging to the technical field of factory farming. Background Art

[0002] In factory-based mandarin fish farming, live grading is a key link that affects farming efficiency. The current mainstream technology mainly relies on mechanical screening (such as drum screens) or image recognition-assisted sorting, forcing fish schools to separate through physical constraints or external stimuli. This adversarial grading mechanism exposes significant defects in high-frequency operation scenarios: on the one hand, mechanical contact inevitably causes fish scales to fall off and mucous membranes to be damaged; on the other hand, the fish school's continuous stress response causes a sharp drop in food intake within 24 hours after grading, resulting in the hidden cost of periodic growth inhibition.

[0003] Specifically, the existing technology has three deep bottlenecks: 1. Under high-density conditions, fish schools shield each other to form active porous media, resulting in distortion of flow field signals and randomization of individual behaviors, and nonlinear attenuation of grading accuracy; 2. Stress-induced curling postures falsify the fluid characteristics of fish bodies, resulting in misjudgment of size; 3. Static physical indicators cannot capture the differences in physiological activity of individuals of the same size, restricting breeding optimization and precision feeding. Although some improvements have attempted to introduce behavioral stimulation, they have not solved the fundamental contradiction between flow field response distortion and physiological information blind spots. Therefore, how to construct a non-destructive sorting method based on fish behavioral instincts that can simultaneously achieve size grading and physiological activity evaluation has become the technical problem to be solved by the present invention. Summary of the Invention

[0004] The present invention provides a method for automatic grading of live mandarin fish suitable for factory farming. Its main purpose is to solve the problems of existing grading technology, such as fish damage and stress growth inhibition caused by antagonistic operations, as well as flow field response distortion and lack of physiological activity assessment under high-density working conditions.

[0005] To achieve the above object, the present invention provides a method for automatically grading live mandarin fish suitable for factory farming, comprising the following steps:

[0006] Step a, providing a grading channel having an inlet end and an outlet end, wherein the cross-sectional area of ​​the grading channel gradually increases from the inlet end to the outlet end, so as to form a gradient flow field in the channel in which the flow velocity gradually decreases from the inlet end to the outlet end;

[0007] Step b, constantly injecting water from the inlet end of the grading channel to maintain a gradient flow field;

[0008] Step c, introducing the shoal of mandarin fish to be classified into a gradient flow field, wherein individual mandarin fish in the shoal autonomously select and stay in a flow velocity section in the gradient flow field that matches their swimming ability, thereby achieving spatial stratification of shoals of mandarin fish of different sizes;

[0009] Step d: collecting the mandarin fish schools from the stratified areas where the mandarin fish schools of different sizes are located in the grading channel.

[0010] Preferably, in step c, individual mandarin fish with stronger swimming ability stay in the higher flow velocity section near the inlet end against the current, and individual mandarin fish with weaker swimming ability seek downstream and stay in the lower flow velocity section near the outlet end.

[0011] Preferably, a guide grid is provided at the bottom of the grading channel, and a plurality of wedge-shaped grooves arranged in an array are provided on the flow-facing surface of the guide grid; in step c, the wedge-shaped grooves correct the fluid shape of the curled-up mandarin fish passing through by inducing the formation of local vortices, so that the mandarin fish restores its stretched posture, thereby ensuring the accurate response of the mandarin fish to the gradient flow field.

[0012] Preferably, the wedge-shaped groove has an inclination angle ranging from 10 degrees to 20 degrees and a depth ranging from 4 mm to 12 mm.

[0013] Preferably, the method also includes the following steps: collecting the pressure pulsation signal of the water flow through a pressure sensor arranged in the grading channel; performing fast Fourier transform on the pressure pulsation signal to obtain a spectrum, and calculating the energy proportion within a predetermined frequency range in the spectrum; comparing the energy proportion with a preset energy proportion threshold, and when the energy proportion is higher than the preset energy proportion threshold, triggering the diversion operation of the high physiological activity group.

[0014] Preferably, the predetermined frequency range is 1 Hz to 5 Hz.

[0015] Preferably, the energy ratio ( ) is calculated using the following formula: ,in, Indicates the frequency The power spectrum density of the pressure pulsation signal, to Indicates the predetermined frequency range, to Indicates the total frequency range of the pressure pulsation signal.

[0016] Preferably, the cross-sectional shape of the grading channel is rectangular, circular or elliptical, and its width or diameter increases linearly or nonlinearly along the water flow direction.

[0017] Preferably, in step d, the lateral collection ports provided in different layered areas of the grading channel are opened to achieve separate collection of schools of mandarin fish of different sizes.

[0018] Preferably, the method further comprises the following steps: adjusting the feed amount and feeding frequency of factory farming, as well as the breeding density of the graded mandarin fish schools based on the grading results of mandarin fish schools of different sizes and the assessed physiological activity levels.

[0019] Compared with the prior art, the present invention has the following beneficial effects:

[0020] 1. By constructing a graded channel with gradually expanding cross-sectional area and a constant water flow, a flow field environment with a decreasing velocity gradient is formed within the channel. By utilizing the natural rheotaxis and swimming ability differences of mandarin fish, individuals of different sizes can autonomously choose to reside in the adaptive flow velocity section. This mechanism transforms traditional antagonistic grading into an autonomous positioning process that conforms to the biological habits of fish, avoiding fish damage and stress-induced growth inhibition caused by mechanical contact.

[0021] 2. The wedge-shaped grooves arranged in an array on the surface of the guide grid form a dynamic coupling with the fish's posture by inducing local vortices. When a curled-up individual passes by, the vortex pressure difference forces it to passively stretch its body; the individual that recovers the stretched posture naturally escapes the influence of the vortex. This mechanism converts the common posture interference in high-density fish schools into a self-correcting trigger signal, ensuring the accuracy of the flow field response and significantly improving the classification robustness under extreme working conditions; the water flow pulsation signal collected by the multiplexing pressure sensor is analyzed through spectral energy focusing to analyze the changes in the proportion of specific frequency bands. The energy intensity of this frequency band is strongly correlated with the physiological behaviors of the fish school, such as swimming, tail wagging, and gill cover opening and closing. Conventional monitoring equipment can simultaneously realize non-destructive evaluation of the physiological activity level of the fish school without adding new hardware, providing an implicit screening dimension for breeding optimization.

[0022] 3. The coupling of gradient flow field design, guide grid microstructure, and pulsation signal analysis forms a technical closed loop of environmental induction-behavioral response-state feedback. The channel geometry passively generates a velocity gradient through the principle of fluid continuity; the wedge-shaped groove automatically corrects posture interference using fluid mechanics; and the frequency domain analysis of the pressure signal mines the information byproducts of biological behavior. The three work together to achieve cross-domain linkage between physical structure, fluid properties, and biological instincts. The fusion of grading results and physiological activity data provides a two-dimensional decision-making basis for factory farming. Spatial stratification data guides specification-based farming in different ponds to avoid resource mismatches caused by growth differences. Physiological activity indicators dynamically optimize feed feeding strategies and density control thresholds, building a closed-loop management path of grading-assessment-control. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is a flow chart of the method for automatically grading live mandarin fish suitable for factory farming according to the present invention;

[0024] Figure 2 This is a schematic diagram of the flow field for automatic classification of live mandarin fish suitable for factory farming according to the present invention;

[0025] Figure 3 This is a comparison diagram of the frequency responses of high-activity and low-activity mandarin fish groups of the present invention.

[0026] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0027] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below. It should be understood that the specific embodiments described herein are only used to explain the present invention rather than to limit the present invention. Based on the embodiments of the present invention, all other methods obtained by ordinary technicians in this field without making any creative work are within the scope of protection of the present invention.

[0028] The present invention discloses a method for automatic grading of live mandarin fish suitable for factory farming. Its technical solution aims to construct an automated system integrating fluid dynamics induction, biological behavioral response, microscopic posture correction, and physiological signal analysis. This system mainly consists of a grading channel module with gradually expanding cross-sectional area, an embedded guide grid posture correction module, a physiological activity assessment module based on pressure pulsation signals, and a partitioned collection and closed-loop control module, which operate in collaboration. In a typical application scenario, such as processing a batch of mandarin fish with an average body length of 15 cm to be graded, efficient and lossless grading is achieved. The initial stage of the method is to construct and maintain a gradient flow field environment as the physical basis for subsequent autonomous grading. Given that in current farming practices, forced mechanical sorting methods can easily cause fish to curl up due to stress, resulting in a serious discrepancy between their fluid shape and actual specifications, leading to grading errors, the core of this solution is to transform the passive, antagonistic sorting process into an active, stress-free, autonomous environmental selection process through a fluid dynamics environment and the mandarin fish's biological instinct to follow currents and avoid harm.

[0029] Specifically, the grading channel is a rectangular water tank with a total length of 15 meters, a width of 0.4 meters at the inlet, a width of 2 meters at the outlet, and a depth of 1.0 meters. Its width increases linearly along the direction of water flow. At the inlet of the channel, an axial flow pump injects water at a constant flow rate of 0.3 cubic meters per second. The coupling of this geometric structure and the constant flow rate, based on the principle of fluid continuity, stably generates a gradient flow field inside the channel from about 0.75 meters per second at the inlet to a linear decrease of about 0.15 meters per second at the outlet. When the school of mandarin fish to be graded is introduced from the entrance of the channel, its flow-seeking instinct is activated, and the fish Each individual in the school will swim against the direction of the water flow and try to find a flow section where its own swimming ability can just overcome it and stay. Because swimming ability is positively correlated with fish size, larger and more capable mandarin fish will swim against the current and eventually stay in the higher flow section near the channel entrance, with a flow rate between 0.5 meters per second and 0.75 meters per second. Conversely, smaller and less capable individuals cannot hover stably in the high-speed water flow and will be gradually pushed backward by the water flow until they find a flow section near the channel exit, with a flow rate between 0.15 meters per second and 0.In this way, only through the synergy of flow field and biological instinct, the automatic stratification of mandarin fish of different sizes is realized in space, which provides a reliable premise for subsequent zoning collection. This process completely avoids mechanical contact and forced pursuit, fundamentally avoids physical damage and stress response, and ensures the health of the fish school and subsequent growth rate after grading. In other words, under high-density breeding conditions, interactions between individuals or environmental stress will cause some mandarin fish to temporarily curl up. This posture greatly changes the fluid dynamics shape of the fish body, distorting the fluid force it feels, which may cause an individual with strong swimming ability to passively retreat due to poor posture, resulting in misjudgment of size. In order to meet this challenge, this scheme specially sets a guide grid at the bottom of the grading channel. The guide grid has a plurality of wedge-shaped grooves arranged in an array on the frontal surface. In the specific calibration procedure, for mandarin fish with a body length of 10 to 20 cm, by calculating the fluid dynamics Mechanical simulations determined the optimal groove geometry parameters: a 15-degree inclination angle and an 8-mm depth. When a curled-up mandarin fish passes over the guide grid, its irregular shape interacts with the wedge-shaped grooves, inducing a pair of stable, oppositely directed local vortices. These vortices create a weak but effective pressure differential on either side of the fish. This pressure differential acts like an invisible pair of fluid dynamics tweezers, exerting a gentle flattening torque on the curled fish, forcing it to resume its stretched, streamlined posture. Once the mandarin fish regains its stretched posture, its smooth fluid shape no longer satisfies the conditions for inducing strong vortices, naturally escaping the corrective effect of these local vortices and continuing to precisely respond to the filtering of the main gradient flow field. This design transforms behavioral disturbances under high-density conditions into a self-triggering, self-terminating posture correction mechanism, ensuring that each individual participates in the classification process with its most realistic fluid characteristics, significantly improving classification accuracy and system robustness.

[0030] Furthermore, traditional physical grading completely ignores the differences in physiological activity between individuals of the same size, which are crucial for accurate feeding and excellent breeding. In view of this, this solution reuses flow field monitoring equipment while achieving physical grading to synchronously evaluate the physiological activity of fish schools in a non-destructive manner; the system embeds high-frequency pressure sensors in the side walls of each section of the grading channel to collect water flow pressure pulsation signals caused by the movement of fish schools, especially tail wagging and gill cover breathing. The sampling frequency is set to 50 Hz to capture sufficiently rich details; the signal processor then performs a fast Fourier transform on the collected 60-second pressure pulsation signal data block to obtain its power spectral density function , research shows that the energy generated by the effective swimming and breathing behavior of the mandarin fish group is mainly concentrated in a specific frequency range. In order to determine the specific value of this range, a calibration procedure needs to be implemented, that is, a group of healthy and active mandarin fish are selected, and their pressure pulsation signals are recorded at different flow rates. By analyzing their spectrum, it is found that the energy peak appears stably in the frequency band of 1 Hz to 5 Hz, so this frequency band is determined as the predetermined frequency range; then, the system uses the formula Calculate the energy proportion within the frequency band, where is 1 Hz, is 5 Hz, and and The total frequency range of the corresponding signal is 0 Hz to 25 Hz. The value intuitively reflects the overall activity intensity of the fish school; the system presets an energy percentage threshold, which is determined by measuring the reference fish schools that are known to be at different stress levels or health conditions, and taking the The statistical median of the value is used as the basis for judgment. When the value is above this threshold, the system determines that the physiological activity of the fish in the current area is high and can trigger subsequent diversion operations, such as collecting them as high-quality breeding fish candidates for special collection. This mechanism provides valuable physiological dimension information for breeding decisions without adding any invasive equipment.

[0031] Finally, when the mandarin fish school completes self-stratification in the grading channel and the physiological activity evaluation is completed, the system enters the collection and closed-loop control stage. On the side wall of the grading channel, corresponding to different flow rate sections such as high-speed zone, medium-speed zone and low-speed zone, there are lateral collection ports that can be controlled by pneumatic devices. After confirming that the fish school stratification is stable, the system opens the corresponding collection ports in sequence or synchronously according to the preset program, and gently guides the mandarin fish of different sizes to their respective temporary holding ponds, realizing efficient and lossless separation and collection; thus, the method of the present invention forms a complete management closed loop from grading, evaluation to regulation. The system integrates the size distribution data obtained by grading with the assessed physiological activity level data, providing a two-dimensional decision-making basis for the precise management of factory farming. If the system detects a medium-sized fish school, its physiological activity index If the index of physiological activity of a large-scale fish school is continuously higher than the threshold, the system will automatically adjust the subsequent feeding strategy in the breeding pond, appropriately increase the feeding frequency of high-protein feed, and match it with a relatively loose breeding density according to the number of batches to maximize its growth potential; on the contrary, if the physiological activity index of a large-scale fish school is low, the system may recommend reducing the breeding density and adjusting the feed formula to improve its health. In this way, the present invention expands the one-time grading operation into a data input terminal for dynamic optimization of the entire breeding process, significantly improving the intelligence level and economic benefits of factory breeding.

[0032] As well as the establishment of key physical parameters and physiological evaluation models of the present invention, it includes a complete set of standardized engineering debugging and calibration procedures, among which the optimal geometric dimensions of the wedge-shaped groove of the guide grid are determined by empirically comparing test plates of different size combinations in a transparent test tank. The specific steps are to arrange a series of test plates with inclination angles ranging from 10 degrees to 20 degrees and depths ranging from 4 mm to 12 mm in an array, and introduce sample fish that are curled up after short-term stress. A high-speed camera system is used to synchronously record and quantify the posture correction rate and average passing time corresponding to each group of test plates. Finally, the geometric parameter combination with the highest correction rate and no significant increase in the normal fish passing time is selected; and the energy proportion threshold for physiological activity evaluation , is calibrated by applying quantifiable environmental factor changes to representative sample fish (n≥30) in this optimized physical channel, that is, the benchmark energy ratio is measured under the optimal breeding parameters. , and the inhibition energy ratio measured under single stress conditions such as water temperature dropped to 18 degrees Celsius or dissolved oxygen dropped to 3.0 mg / L , the threshold That is, the arithmetic mean of the two is taken. This calibration process is repeated once in each new breeding cycle or when there is a fundamental change in environmental parameters.

[0033] The core of closed-loop control is the feeding density coordinated adjustment matrix, and the adjustment coefficients within it are and , are all products based on multi-factor orthogonal experiments and data regression modeling rather than preset empirical values. The construction method is to set up multiple parallel breeding experimental systems, with graded specifications, different feed adjustment coefficient levels and density adjustment coefficient levels as experimental factors. After a medium- to long-term experiment such as 30 days, the total weight gain, feed conversion rate and real-time physiological activity indicators of each experimental group are collected. and other data, and use the multivariate regression analysis method to establish a quantitative response surface model between breeding benefits, such as total weight gain per unit cost, and each experimental factor. The optimal adjustment coefficient corresponding to each combination of specification and activity level in the matrix is ​​the coefficient value corresponding to the maximum value of the benefit function output by the above-mentioned response surface model. In this way, the graded evaluation results are directly linked to the economic benefits of breeding, forming a set of quantifiable and optimal decision-making logic, which are all extended implementation methods known to ordinary technicians in this field.

[0034] Example 1: In this example, in a continuously operating high-density recirculating aquaculture system, a batch of mandarin fish that have just completed a three-day drug quarantine cycle are waiting to be graded and transferred to the late fattening pond. The objective working conditions of this batch of fish are extremely complex. Not only is the size span wide, covering individuals ranging from 12 cm to 20 cm, but due to individual differences in drug stress response, the physiological activity state of the fish is severely polarized. Some strong individuals have resumed normal swimming and feeding, while a large number of sensitive individuals are still in a state of stress curled up or activity inhibition. If traditional mechanical drum screening is used, It is inevitable that secondary damage will be caused to such physiologically fragile individuals, and a large number of specifications will be misjudged due to their stress-induced curling posture. If they are directly mixed, the weak will become weaker due to differences in their ability to grab food, and eventually cause large-scale losses. When this batch of mandarin fish with complex components is introduced into the grading system of the present invention, its inherent technical synergy mechanism is immediately activated. First, all mandarin fish individuals, regardless of their physiological state, enter the gradient flow field constructed by the grading channel. At this time, the technical dilemma recognized by the industry, that is, the contradiction between sorting efficiency and individual damage, is resolved in the framework of the present invention. The pressure under the frame is initially resolved. The gradient flow field uses the fish's instinct to move toward the flow to perform contactless screening, fundamentally eliminating mechanical damage and ensuring the basic safety of physiologically fragile individuals. On this basis, a deeper synergistic effect is reflected in the linkage between the guide grid posture correction module and the physiological activity assessment module. Under the action of water flow, larger individuals in a stressed and curled state should be mistakenly classified as small-sized fish due to their non-streamlined, pseudo-shaped appearance. However, when they pass through the guide grid at the bottom of the channel, the local vortex induced by the wedge-shaped groove array on it exerts a contactless posture correction force on the curled fish body, causing them to recover and stretch. This process ensures that even stressed fish can respond to the gradient flow field screening with their true fluid dynamic size, thereby achieving precise physical size stratification. This precise physical stratification, in turn, provides the prerequisite for high-quality signal acquisition for subsequent physiological activity assessment. Pressure sensors installed in different stratification areas are able to collect pressure pulsation signals in the environment of fish of uniform size. At this time, by performing a fast Fourier transform on the signal and calculating the energy proportion within the predetermined frequency range, the pressure pulsation signal is collected. The physiological activity index obtained has a precise comparative significance because it eliminates the huge interference variable of tail swing force caused by individual size differences. The value can clearly point to the high activity intensity of the uniform group, and vice versa. In this deduction, the entire system is not only operating as a grading tool, but also as a multi-dimensional diagnosis and decision-making platform. It does not directly solve the biological problem of how to soothe the stressed fish. Instead, through its unique architecture, it redefines the problem as how to accurately identify subgroups with different economic potentials in a group with complex physiological states. Finally, this batch of mandarin fish is automatically separated into at least four groups with clear later management directions, namely large-scale high-activity groups, large-scale low-activity groups, and small-scale Based on this, the breeding managers can immediately transfer the large-scale high-activity group to the high-density fattening process before listing, adopt a low-density rest and recovery strategy for the large-scale low-activity group, and adopt a nutrition-enhanced catch-up feeding program for the small-scale high-activity group. The small-scale low-activity group can be marked as a batch that needs further observation or elimination. This kind of refined grouping based on the dual dimensions of size and physiological activity will upgrade a simple grading operation to a strategic decision-making process for the optimal allocation of breeding resources, thereby maximizing breeding benefits.

[0035] Example 2: To confirm the quantitative accuracy of the physiological activity assessment method based on pressure pulsation signals of the present invention and to verify its application in establishing a refined feeding strategy, this example performs a control experiment. This experiment is conducted in a graded channel equipped with a precision temperature control system. The water temperature is used as a control variable that directly affects the metabolic rate of mandarin fish. The experiment simulates aquaculture scenarios with different physiological activity levels. The core of the experiment is to calibrate the pressure sensor sampling frequency and verify the physiological activity index. The sampling frequency setting of the pressure sensor is directly related to the signal fidelity and system computing load. Its value is based on the effective spectrum bandwidth of the measured signal, that is, the pressure pulsation signal caused by the physiological activities of the fish school. According to the Nyquist sampling theorem, in order to reconstruct the signal without distortion, the sampling frequency must be greater than twice the highest effective frequency of the signal. In view of the fact that the flow field energy generated by the large-scale physiological activities of the mandarin fish group is concentrated below 25 Hz, the sampling frequency is set to This setting ensures the complete capture of all valid signal components, including the fundamental frequency and key harmonics, while keeping the data processing load within an economically reasonable range.

[0036] The experimental process first selected a group of healthy mandarin fish with an average body length of 15 cm and uniform size and placed them in the medium-speed constant flow area of ​​the grading channel. The water temperature was stabilized at 22 degrees Celsius as the baseline state through the temperature control system and maintained for 2 hours. Then the system collected 60 seconds of pressure pulsation signals and calculated the energy proportion. Afterwards, the temperature control system dropped the water temperature to 18 degrees Celsius at a rate of 2 degrees Celsius per hour. A second measurement was performed after stabilizing for 2 hours. Finally, the water temperature was gradually raised to 26 degrees Celsius. A third measurement was performed after stabilizing for 2 hours. During the test, the activity status of the fish school was synchronously observed by the underwater camera system for cross-verification. The key data points recorded are as follows, see Table 1.

[0037] Table 1: Fish activity status and energy proportion under different water temperatures.

[0038]

[0039] The experimental data show that The value increases systematically with increasing water temperature. The mechanism of this association is that water temperature is the core variable affecting the metabolic rate of mandarin fish. Its increase will directly lead to an increase in the frequency and amplitude of the fish's heart beat, gill cover opening and closing, and swimming and tail wagging. The enhancement of these physiological activities is manifested as higher-energy pressure pulsations in the flow field, especially in the predetermined frequency range of 1 Hz to 5 Hz, which is strongly related to effective swimming and breathing behavior. Therefore, the energy proportion of this frequency band is The results of the experiment show that the physiological activity evaluation method proposed by the present invention can accurately and non-destructively quantify the collective physiological state of the mandarin fish group. This conclusion provides a reference for aquaculture managers to use real-time monitoring to evaluate the physiological activity of mandarin fish. The dynamic function relationship between the value and the feed feeding amount provides a data basis, for example, The value is compared with the preset activity threshold to trigger the corresponding adjustment of the feed amount, thereby expanding the grading operation from simple physical separation to a two-dimensional data collection process that integrates specification sorting and physiological status diagnosis, thereby realizing closed-loop precision feeding based on feedback from real biological needs, maximizing feed conversion efficiency and breeding economic benefits.

[0040] Example 3: This example combines Figures 1 to 3 , the automatic grading method of live mandarin fish suitable for factory farming is described, such as Figure 1As shown in the figure, first, the hierarchical control system is linked with the data fusion center to obtain and integrate relevant physiological activity data and perform data processing at the same time; in this process, the two-dimensional data enters the control decision database, and the relevant parameters are optimized based on the feedback; next, the behavioral response of the mandarin fish is analyzed according to the optimized parameters, and relevant operation instructions are generated in the mandarin fish management system. According to these instructions, the system controls the execution equipment and starts the effect monitoring module; in addition, the density management system and the control management system also cooperate with each other to monitor and manage the distribution of the mandarin fish population in real time, and adjust factors such as density and water flow rate. After these processes, the grading system completes the automatic stratification of mandarin fish individuals and feeds the results back to the effect monitoring part for further optimization and adjustment. This process is repeated to ensure the accuracy of grading and physiological activity evaluation.

[0041] like Figure 2 As shown in the figure, the water flow direction is divided from left to right into high flow rate zone, medium flow rate zone, low flow rate zone and extremely low flow rate zone. The water flow rate in the high flow rate zone is 0.5-0.7m / s, which is suitable for larger mandarin fish (18-20cm); the water flow rate in the medium flow rate zone is 0.3-0.5m / s, which is suitable for medium-sized mandarin fish (15-18cm); the water flow rate in the low flow rate zone is 0.15-0.3m / s, which is suitable for smaller mandarin fish (12-15cm); the water flow rate in the extremely low flow rate zone is lower than 0.15m / s, which is suitable for small-sized mandarin fish. According to the swimming ability of mandarin fish, the fish will choose the appropriate residence area in each flow rate zone according to their own ability, thereby realizing autonomous classification. The division of each zone and the relationship between its corresponding water flow rate and the swimming ability of mandarin fish ensure an efficient and accurate live classification process.

[0042] like Figure 3 As shown in the figure, the horizontal axis is frequency (Hz), ranging from 0 to 12.5 Hz; the vertical axis is power spectrum density (unit: corresponding value), which represents the intensity of water flow pulsation signal at each frequency. The solid line in the figure represents the power spectrum density curve of the high-activity mandarin fish group, and the dotted line represents the power spectrum density curve of the low-activity mandarin fish group. It can be seen from the figure that the power spectrum density of the high-activity mandarin fish group presents a higher peak in the range of 0.5 Hz to 5 Hz, especially in the frequency band of 1 Hz to 3 Hz, the power is significantly higher, and the spectrum curve is relatively smooth; while the power spectrum density of the low-activity mandarin fish group is relatively low, and does not change much in the entire frequency range, which indicates that the swimming and physiological activities of the high-activity mandarin fish group in the low-frequency band generate stronger water flow pulsation signals, while the water flow pulsation signals of the low-activity mandarin fish group are weaker, and the physiological activity differences are obvious, and the activity levels of the groups can be distinguished by frequency response.

[0043] Example 4: In order to convert the physiological activity evaluation results of the present invention into executable and quantitative breeding management instructions, and thus realize the complete technical implementation from hierarchical evaluation to closed-loop control, this embodiment elaborates on the core decision-making mechanism of the closed-loop control module, that is, the calibration procedure of the key threshold and the construction method of the control logic. At the beginning of a breeding cycle, the breeding manager needs to preset a set of precise feeding and density management plans for a batch of newly introduced mandarin fish seedlings of a specific specification, which can be adaptively adjusted according to their growth status and physiological activity changes. The traditional model that relies on manual experience has significant lag and uncertainty in this link, which can easily lead to feed waste or inappropriate density. To cope with the stress caused by the physiological activity, the system first performs an offline, standardized physiological activity benchmark calibration procedure to determine the energy ratio threshold used to distinguish different physiological activity levels. In operation, thirty healthy sample fish are randomly selected from the large group to be managed and placed in a test tank equipped with a precision dissolved oxygen control module. The water flow rate is stabilized at the critical swimming speed of the sample fish. First, the dissolved oxygen concentration in the water is maintained at an optimal 7.5 mg / L through the dissolved oxygen control module, and the fish are given 2 hours of sufficient adaptation time. This state is defined as the baseline active state. During this period, the system continuously collects pressure pulsation signals and calculates the average baseline energy ratio. Then, the dissolved oxygen concentration in the water was slowly reduced to a stress level of 3.0 mg / L within 30 minutes through the dissolved oxygen control module. This state was defined as a low inhibition state. Similarly, after stabilization for 2 hours, the average inhibition energy ratio was collected and calculated. , based on this, the judgment threshold of the high physiological activity group is determined as the arithmetic mean of the two state means, which is calculated as This procedure uses quantifiable environmental factors as a reference to establish a threshold setting path with clear physical meaning and repeatable verification, eliminating the ambiguity of human experience judgment and providing a solid data foundation for subsequent automated decision-making.

[0044] After the threshold is determined, the system loads a built-in feeding density collaborative adjustment matrix based on the dual-dimensional input of specification and physiological activity. This matrix constitutes the core of the closed-loop control logic. This matrix combines the specification levels corresponding to different sections of the grading channel with the energy proportion value. and threshold The activity levels determined are correlated and a set of quantitative adjustment coefficients are output for each combination, namely the feed adjustment coefficient Density adjustment factor When the classification process starts, a group of fish resides in the medium speed zone and is identified as medium size, and its calculated Value higher than When it is judged to be highly active, the system decision matrix is ​​triggered, automatically matching and outputting the adjustment coefficient corresponding to the specific combination. and This set of coefficients is then transmitted to the breeding management execution unit, which multiplies the basic feeding standard by the coefficient, thereby accurately increasing the amount of feed put into the pond on that day by 15%, and at the same time The coefficient maintains its stocking density at the standard level. If another group of large-sized fish residing in the high-speed area is assessed as low activity, the matrix will output another set of corresponding adjustment coefficients. and In this way, the present invention upgrades the one-time grading operation into a continuous dynamic optimization process based on objective data feedback, and constructs an intelligent management closed loop from precise perception of biological status to quantitative regulation of aquaculture behavior, thereby ensuring the efficient use of aquaculture resources.

[0045] Example 5: The establishment of the key physical parameters and physiological evaluation model of the present invention includes a complete set of standardized engineering debugging and calibration procedures. Among them, the optimal geometric dimensions of the wedge-shaped groove of the guide grid are determined by empirically comparing test plates of different size combinations in a transparent test tank. The specific steps are to arrange a series of test plates with inclination angles ranging from 10 degrees to 20 degrees and depths ranging from 4 mm to 12 mm in an array, and introduce sample fish that are in a curled-up posture after short-term stress. A high-speed camera system is used to synchronously record and quantify the posture correction rate and average passing time corresponding to each group of test plates. Finally, the geometric parameter combination with the highest correction rate and no significant increase in the normal fish passing time is selected; and the energy proportion threshold for physiological activity evaluation , is calibrated by applying quantifiable environmental factor changes to representative sample fish (n≥30) in this optimized physical channel, that is, the benchmark energy ratio is measured under the optimal breeding parameters. , and the inhibition energy ratio measured under single stress conditions such as water temperature dropped to 18 degrees Celsius or dissolved oxygen dropped to 3.0 mg / L , the threshold That is, the arithmetic mean of the two is taken. This calibration process is repeated once in each new breeding cycle or when there is a fundamental change in environmental parameters.

[0046] The core of closed-loop control is the feeding density coordinated adjustment matrix, and the adjustment coefficients within it are and , are all products based on multi-factor orthogonal experiments and data regression modeling rather than preset empirical values. The construction method is to set up multiple parallel breeding experimental systems, with graded specifications, different feed adjustment coefficient levels and density adjustment coefficient levels as experimental factors. After a medium- to long-term experiment (such as 30 days), the total weight gain, feed conversion rate and real-time physiological activity indicators of each experimental group are collected. and other data, and used the multivariate regression analysis method to establish a quantitative response surface model between breeding benefits, such as total weight gain per unit cost, and each experimental factor. The optimal adjustment coefficient corresponding to each combination of specification and activity level in the matrix is ​​the coefficient value corresponding to the maximum value of the benefit function produced by the above response surface model. In this way, the graded evaluation results are directly linked to the economic benefits of breeding, forming a set of quantifiable and optimal decision-making logic.

[0047] And, the internal coefficient values ​​of the feeding density co-adjustment matrix can be determined, for example, by a standardized procedure combining a multi-factor orthogonal experiment with a response surface regression analysis, wherein the experimental factors are first set to include A specification grade, B physiological activity grade, C feed adjustment coefficient, whose levels are set at five equal intervals between 0.8 and 1.2, and D density adjustment coefficient, whose level is set the same as C; secondly, the selected The experiment was arranged by orthogonal tables, and a complete 30-day test cycle was carried out in multiple parallel aquaculture systems with strictly consistent water environment parameters. Secondly, the response variable of the experiment was set as the total weight gain per unit feed cost, which was calculated by dividing the total weight gain of each experimental group during the cycle by the total cost of the feed consumed. Finally, using all the collected data, a multivariate quadratic regression model was established between the response variable and each experimental factor. For each specific combination of specification grade A and physiological activity grade B, the model was solved to obtain the C feed and D density coefficient values ​​when the response variable reached the maximum value. These coefficient values ​​were directly used to fill the corresponding cells of the matrix. The geometric dimensions of the groove and the morphological parameters of the nonlinear graded channel are derived from standardized experimental calibration. The determination of the geometric dimensions of the groove involves quantifying two core indicators on a series of test plates with different inclination angles and depth combinations, namely the attitude correction rate, which is defined as the percentage of the number of curled-up fish that recover their stretched posture after passing through the guide fence to the total number of curled-up fish at the beginning, and the increase rate of passing time, which is defined as the average increase in the percentage of time that the normal posture fish pass through the fence area compared to the non-barrier area. The parameter combination that can make the attitude correction rate not less than 90% and the increase rate of passing time not more than 15% is selected; for the nonlinear graded channel, one specific implementation method is the channel width With length The change follows the functional relationship ,constant By adjusting according to the target flow velocity gradient, this design can achieve a gradient distribution in which the flow velocity changes smoothly in the inlet section and aggravated flow velocity changes in the outlet section. It is suitable for fine separation of mixed fish schools with extremely different swimming abilities, and is an extended implementation method known to ordinary technicians in this field.

[0048] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0049] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for automatically grading live mandarin fish suitable for factory farming, characterized in that: The method comprises the following steps: Step a, providing a grading channel having an inlet end and an outlet end, wherein the cross-sectional area of ​​the grading channel gradually increases from the inlet end to the outlet end, so as to form a gradient flow field in the channel in which the flow velocity gradually decreases from the inlet end to the outlet end; Step b, constantly injecting water from the inlet end of the grading channel to maintain a gradient flow field; Step c, introducing the school of mandarin fish to be classified into a gradient flow field, wherein the individual mandarin fish in the school autonomously select and stay in a flow velocity section in the gradient flow field that matches their swimming ability according to their own swimming ability; Step d, collecting the mandarin fish schools from the stratified areas where the mandarin fish schools of different sizes are located in the grading channel; A guide grid is provided at the bottom of the grading channel, and a plurality of wedge-shaped grooves arranged in an array are provided on the flow-facing surface of the guide grid; in step c, the wedge-shaped grooves correct the flow shape of the curled-up mandarin fish passing through by inducing the formation of local vortices, causing them to restore to a stretched posture; The method also includes the following steps: collecting the pressure pulsation signal of the water flow through a pressure sensor arranged in the grading channel; performing a fast Fourier transform on the pressure pulsation signal to obtain a spectrum, and calculating the energy proportion within a predetermined frequency range in the spectrum; comparing the energy proportion with a preset energy proportion threshold, and when the energy proportion is higher than the preset energy proportion threshold, triggering the diversion operation of the high physiological activity group.

2. The method for automatically grading living mandarin fish suitable for factory farming according to claim 1, characterized in that: In step c, the mandarin fish with stronger swimming ability stays in the higher flow velocity section near the inlet end against the current, while the mandarin fish with weaker swimming ability seeks the downstream and stays in the lower flow velocity section near the outlet end.

3. The method for automatically grading living mandarin fish suitable for factory farming according to claim 1, characterized in that: The inclination angle of the wedge-shaped groove ranges from 10 degrees to 20 degrees, and the depth ranges from 4 mm to 12 mm.

4. The method for automatically grading living mandarin fish suitable for factory farming according to claim 1, characterized in that: The predetermined frequency range is 1 Hz to 5 Hz.

5. The method for automatically grading living mandarin fish suitable for factory farming according to claim 1, characterized in that: Energy ratio ( ) is calculated using the following formula: ,in, Indicates the frequency The power spectrum density of the pressure pulsation signal, to Indicates the predetermined frequency range, to Indicates the total frequency range of the pressure pulsation signal.

6. The method for automatically grading living mandarin fish suitable for factory farming according to claim 1, characterized in that: The cross-sectional shape of the grading channel is rectangular, circular or elliptical, and its width or diameter increases linearly or nonlinearly along the water flow direction.

7. The method for automatically grading living mandarin fish suitable for factory farming according to claim 1, characterized in that: In step d, the lateral collection ports arranged in different layered areas of the grading channel are opened to achieve separate collection of schools of mandarin fish of different sizes.

Images