Three-dimensional bullfrog breeding system and precise feeding control method

Abstract

The invention discloses a bullfrog three-dimensional breeding system and an accurate feeding control method, and the bullfrog three-dimensional breeding system is characterized in that the bullfrog three-dimensional breeding system comprises a modular breeding unit, a gradient water level adjusting device and an interlayer isolation water circulation system, the three-step type three-dimensional framework is divided into a bottom layer tadpole pond, a middle layer froglet net cage and a top layer adult frog breeding bin, a control module is installed on each layer, and anti-escape assemblies are installed on the middle layer froglet net cage and the top layer adult frog breeding bin; the gradient water level adjusting device comprises a bionic step and a water level sensor which are arranged and installed, and the water level gradient adjusting and behavior inducing functions are achieved through engineering design. The interlayer isolation water circulation system comprises a distributed sensor, a UV sterilization and biological membrane filtration dual purification assembly and an adjustable air curtain, the adjustable air curtain forms a directional barrier between different layers of the three-dimensional culture system by dynamically regulating and controlling airflow, and independent environmental parameters of each layer are maintained.

Description

Technical Field

[0001] The present invention relates to the technical field of bullfrog breeding, and particularly to a three-dimensional bullfrog breeding system and a precise feeding control method. Background Art

[0002] In recent years, with the rapid development of the bullfrog catering industry, bullfrog breeding has been increasing day by day. At present, bullfrogs play an important role in the special aquatic breeding varieties in China. Mainly, it occupies arable land for running-water breeding, and the breeding water is directly discharged without treatment, causing extremely serious waste water and waste gas pollution to the ecological environment. The adverse effects of freshwater pond breeding on the water environment mainly include the following four aspects: First, excessive feeding, improper feeding methods or poor-quality bait result in excessive residual bait and cause pollution. Second, the pollution caused by the accumulation of certain chemical agents. The use of chemical drugs in aquaculture will cause pathogens to develop drug resistance while killing beneficial microorganisms in the water, resulting in secondary pollution and causing water ecological imbalance. Third, the replacement of breeding tail water, such as controlling water quality replacement, fishing replacement, etc., is directly discharged without treatment and purification, thus polluting the external environment. Fourth, due to the lack of breeding layout and breeding mode of aquaculture ecology theory and related ecological regulation and other technical guidance, the water ecological balance is seriously damaged. If the breeding tail water discharge is large and the tail water is not reasonably treated, the tail water pollution will become increasingly serious. In the prior art, the floor area of bullfrog breeding is very large, which is not easy to manage and will cause waste of resource space. At the same time, there is a lack of reasonable scientific management and intelligent breeding for bullfrogs at different growth stages. It often still relies on the experience of farmers for breeding, which will cause problems such as increased labor costs and low efficiency, as well as waste caused by inaccurate feeding of feed, and the problem of how to control the cross-infection of bacteria in the bullfrog growth environment.

[0003] Therefore, it is necessary to provide a three-dimensional bullfrog breeding system and a precise feeding control method to solve the above technical problems. Summary of the Invention

[0004] The purpose of this part is to outline some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Simplifications or omissions may be made in this part, as well as in the abstract and title of the present application, to avoid obscuring the purpose of this part, the abstract, and the title. However, such simplifications or omissions shall not be used to limit the scope of the present invention.

[0005] To solve the above technical problems, the present invention provides the following technical solutions: A three-dimensional bullfrog breeding system, characterized in that it includes a modular breeding unit, a gradient water level adjustment device, and an interlayer isolation water circulation system. The modular breeding unit includes a three-tiered three-dimensional structure, which is divided into a bottom-layer tadpole pond, a middle-layer juvenile frog net cage, and a top-layer adult frog breeding warehouse. Control modules are installed on each layer, and anti-escape components are installed on the middle-layer juvenile frog net cage and the top-layer adult frog breeding warehouse; the gradient water level adjustment device includes a bionic step and a water level sensor installed, and realizes the functions of water level gradient adjustment and behavior induction through engineering design; the interlayer isolation water circulation system includes distributed sensors, a UV sterilization + biofilm filtration dual purification component, and an adjustable air curtain. The distributed sensors and the dual purification component are installed in each layer of the three-tiered three-dimensional structure. The adjustable air curtain forms a directional barrier between different levels of the three-dimensional breeding system through dynamic regulation of air flow, maintaining independent environmental parameters for each layer.

[0006] As a preferred solution of the three-dimensional bullfrog breeding system of the present invention, a temperature control module is installed in the three-dimensional bullfrog breeding system, and the temperature control module is connected to the control module.

[0007] As a preferred solution of the three-dimensional bullfrog breeding system of the present invention, the dual purification component includes two types: UV sterilization and biofilm filtration, and the dual purification is installed in the bottom-layer tadpole pond, the middle-layer juvenile frog net cage, and the top-layer adult frog breeding warehouse.

[0008] As a preferred solution of the three-dimensional bullfrog breeding system of the present invention, a plurality of bionic steps are provided, with different heights set according to different growth stages of bullfrogs. The surface of the bionic steps is provided with a coating, and the edge of the steps is set in an inverted T shape.

[0009] As a preferred solution of the three-dimensional bullfrog breeding system of the present invention, the anti-escape component includes a protective plate and ultrasonic waves. The protective plate is inclined, and the ultrasonic waves are arranged in a circumferential array.

[0010] As a preferred solution of the three-dimensional bullfrog breeding system of the present invention, the adjustable air curtain includes an air curtain generator, a servo motor, multi-parameter sensors, and an ultrasonic atomizer, and the adjustable air curtain is connected to the control module.

[0011] Advantages of the present invention: By setting up three-dimensional farming, not only space is saved, but also management is facilitated. At the same time, a gradient water level adjustment device is set up, which is more matched with the limb movement characteristics of bullfrogs through the setting of bionic steps. The steps adopt a curved surface design to avoid damage to the limbs of bullfrogs. A dual purification component is set up for the farming space to prevent bacterial infection. At the same time, an adjustable air curtain is installed for the farming environment. By dynamically regulating the air flow, a directional air barrier is formed between different levels of the three-dimensional farming system to block the cross-layer transmission of pathogenic microorganisms, dust and aerosols, and at the same time maintain independent environmental parameters for each layer. At the same time, an anti-escape component is installed on the middle and high-level farming bins to prevent bullfrogs from jumping out of the farming bins and improve the bullfrog farming efficiency.

[0012] In view of the problem that the bullfrog farming system can be further improved, a precise feeding control method is proposed.

[0013] To solve the above technical problems, the present invention provides the following technical solutions. A precise feeding control method includes the bullfrog three-dimensional farming system described in any one of the above embodiments, and further includes: establishing a three-dimensional decision-making model including growth stage, environmental parameters, and group behavior by fusing data through multi-modal sensors;

[0014] Extracting the neck, head and limb features of bullfrogs by using an improved YOLOv5s model through a multi-spectrum vision detection module;

[0015] Using a voiceprint feature extraction module, quantifying the feeding intensity based on MFCC coefficients, combining with a dynamic feeding decision-making model, and fusing the feeding demand index, environmental parameters and the output value of the prediction network;

[0016] Combining with a closed-loop feedback control system, and correcting the feeding parameters in real time through the amount of residual bait.

[0017] As a preferred solution of the precise feeding control method of the present invention, the improved YOLOv5s model adds a CA attention mechanism in the Neck module, and its mathematical expression formula is as follows:

[0018]

[0019] Among them, is the original parameter vector or matrix, F neck represents the features extracted from the input data, W CA is the attention weight matrix and convolution kernel, is the spliced feature, is the channel attention weight.

[0020] As a preferred solution of the precise feeding control method of the present invention, the dynamic feeding decision-making model is divided into 12 levels according to the growth stage of bullfrogs, and the mathematical expression formula of the feeding demand index is as follows:

[0021]

[0022] Among them, 0.4 and 0.3 represent weight coefficients, represents the proportion of active users, represents the signal strength ratio, represents the temperature influence term. When T w > 28 °C, the FBI weight is automatically adjusted.

[0023] As a preferred solution of the precise feeding control method described in the present invention, the residual bait amount is used to correct the feeding parameters in real time by using the threshold value W threshold dynamically set, and its mathematical expression formula is as follows: W threshold = 0.2·Q base + 0.05·(T w - 20)

[0024] Among them, W threshold represents the calculated threshold value, Q base is the base value, T w is the current working temperature, and 0.2 and 0.05 are weight coefficients used to adjust the influence degrees of the base value and the temperature on the final result.

[0025] The beneficial effects of the present invention: Establish a growth stage and a nutritional requirement matrix, monitor the activity intensity of the frog group through the Doppler vision detection module, judge the feeding intensity of the frog group by using the voiceprint feature extraction module, and control the feeding amount by monitoring the residual bait amount in real time through the closed-loop feedback control system, so as to improve the feeding accuracy, reasonably arrange the feeding of the frog groups at different growth stages, avoid overfeeding, underfeeding or diet mismatch, reduce the waste of feed, and effectively improve the feeding accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0027] Among them:

[0028] Figure 1 is a side view of the three-stage three-dimensional structure of the bullfrog three-dimensional breeding system and the precise feeding control method provided by the present invention;

[0029] Figure 2 is a schematic diagram of the breeding system of the bullfrog three-dimensional breeding system and the precise feeding control method provided by an embodiment of the present invention;

[0030] Figure 3 The structural schematic diagram of the bullfrog three-dimensional breeding system and the breeding system of the precise feeding control method provided by the present invention;

[0031] Figure 4 The schematic flow chart of the control method of the bullfrog three-dimensional breeding system and the precise feeding control method provided by the present invention. Detailed implementation manners

[0032] In order to make the above objects, features and advantages of the present invention more obvious and understandable, the following detailed description of the specific implementation manners of the present invention will be given with reference to the accompanying drawings of the specification.

[0033] In the following description, many specific details are set forth to facilitate a thorough understanding of the present invention. However, the present invention may be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0034] Secondly, the present invention will be described in detail with reference to the schematic diagrams. When detailing the embodiments of the present invention, for the sake of clarity, the cross-sectional views showing the device structures will be enlarged locally out of the ordinary scale, and the schematic diagrams are only examples and should not limit the scope of protection of the present invention herein. In addition, in actual production, three-dimensional spatial dimensions including length, width and depth should be included.

[0035] Furthermore, the so-called "one embodiment" or "embodiment" herein refers to a specific feature, structure or characteristic that can be included in at least one implementation manner of the present invention. The appearances of "in one embodiment" in different places in this specification do not all refer to the same embodiment, nor are they separate or alternative embodiments that exclude each other with other embodiments.

[0036] Embodiment 1

[0037] Refer to Figures 1-3, the first embodiment of the present invention provides a three-dimensional bullfrog breeding system, which is characterized in that it includes a modular breeding unit 100, a gradient water level adjustment device 200 and an interlayer isolation water circulation system 300. The modular breeding unit 100 includes a three-stage three-dimensional structure 101, which is divided into a bottom-layer tadpole pond 102, a middle-layer juvenile frog net cage 103 and a top-layer adult frog breeding warehouse 104. Control modules are installed on each layer, and the control modules are connected to the general control room of the breeding system. Anti-escape components 105 are provided and installed on the middle-layer juvenile frog net cage 103 and the top-layer adult frog breeding warehouse 104; the gradient water level adjustment device 200 includes a bionic step 201 and a water level sensor provided and installed, and realizes the functions of water level gradient adjustment and behavior induction through engineering design; the interlayer isolation water circulation system 300 includes distributed sensors, a dual purification component 301 of UV sterilization + biofilm filtration and an adjustable air curtain 302. The distributed sensors and the dual purification component 301 are installed in each layer of the three-stage three-dimensional structure 101. The adjustable air curtain 302 forms a directional barrier between different levels of the three-dimensional breeding system by dynamically regulating the airflow to maintain independent environmental parameters for each layer. Specifically, a multi-layer steel structure breeding tower is built and arranged on different floors according to the growth stage of bullfrogs. Each layer is equipped with an independent environmental control module, and a temperature control module is also installed. The temperature control module is connected to the control module. The control module selects the Siemens S7-1200 series PLC. Through the temperature control module, the temperature control accuracy can be controlled within ±0.5 °C and the humidity within ±3% RH. Dual purification components are installed in each layer of the three-stage three-dimensional structure. The dual purification component 301 includes two types: UV sterilization and biofilm filtration. Because the interlayer isolation water circulation system 300 is used to circulate or discharge the water quality of each layer, the water quality is sterilized by a UV sterilization lamp, and at the same time, the sterilized water quality is filtered by a biofilm to ensure that the circulating water is healthy. At the same time, the generated wastewater is discharged. How to discharge the wastewater during the bullfrog breeding process is already prior art, and the present invention will not elaborate on the related technical solutions. In order to make the bullfrog breeding environment more superior, an intelligent environmental control system is equipped and installed. By installing a distributed sensor array, temperature and humidity information can be collected more accurately. Through the collected information, the intelligent environmental control system can start or stop automatically. The distributed sensor is a three-in-one probe for temperature, PH and dissolved oxygen. Because it is installed in an array, every 5m 2The installed density ≥ 3 can provide more data support for the intelligent environmental control system. The intelligent control system includes an adjustable air curtain. The adjustable air curtain 302 includes an air curtain generator, a servo motor, a multi-parameter sensor, and an ultrasonic atomizer. The function of the adjustable air curtain 302 is to direct the air in the breeding environment to prevent cross-infection between layers. By dynamically regulating the air flow, a directional air barrier is formed between different levels of the three-dimensional breeding system to block the cross-layer transmission of pathogenic microorganisms, dust, and aerosols, while maintaining independent environmental parameters for each layer;

[0038] Dynamic adjustment: Automatically adjust the air curtain parameters according to real-time environmental data (such as wind speed, temperature and humidity difference, ammonia concentration).

[0039] Energy optimization: Only start the high-intensity air curtain when detecting potential pollution risks to reduce energy consumption.

[0040] Precise isolation: Achieve physical isolation of the air between layers through the coordinated control of the air flow angle and speed.

[0041] When the temperature difference between adjacent layers ≥ 2°C or the ammonia nitrogen concentration difference ≥ 0.5 mg / L, trigger the air curtain to start. When the particle counter detects that the microbial concentration in the upper layer air exceeds the threshold (>500 CFU / m 3 ), automatically increase the wind speed. Wind speed control: Dynamically adjust according to the pressure difference between layers; make the air flow form an angle of 15 - 30° with the horizontal plane through the guide vane to form a downward-sloping isolation barrier. When detecting specific markers of pathogens (such as Aeromonas hydrophila), start the atomization module to release fungicides.

[0042] Table 1 describes the hardware structure of the adjustable air curtain

[0043]

[0044] Set the relevant parameters. Set the parameters as the basic wind speed: 2 m / s (maintaining a pressure difference of 5 Pa between layers), and at the same time, the emergency mode is also turned on. When detecting germs, the wind speed will be increased to 4 m / s and hypochlorous acid mist (concentration 0.2 ppm) will be released.

[0045] The following table shows the comparison between the adjustable air curtain of the present invention and the traditional air curtain

[0046]

[0047]

[0048] Meanwhile, install a gradient water level regulating device in the breeding system, and install a bionic step 201 structure in the tadpole, juvenile frog, and adult frog breeding ponds. Set different multi-level gentle slope steps according to the growth stage of bullfrogs. Among them, set the morphological parameters. The height of a single step is set to 5 - 8 cm (matching the jumping height threshold of bullfrogs), and the bionic step 201 is set in an inclined style with an inclination angle of 8° - 12° (the best climbing angle measured in experiments). To increase the friction of the bionic step 201, the surface of the bionic step 201 is set to a textured shape and made into a lotus leaf micro-structure by laser etching; the bionic step 201 set in the juvenile frog stage is a low-step shallow water area (water depth 3 - 5 cm) to promote the transition from gill respiration to lung respiration, and the bionic step 201 set in the adult frog stage is a high-step deep water area (water depth 15 - 20 cm) to simulate the natural deep pool environment. Install an embedded water level sensor in the breeding areas of juvenile frogs and adult frogs, and install a dynamic water level regulating module and a pneumatic lifting component. The data collected by the sensor is sent to the dynamic water level regulating module, and the dynamic water level regulating module can control the start and stop of the pneumatic lifting component according to the data collected by the sensor, simulating the law of tidal ebb and flow, which is more suitable for the growth of bullfrogs. To enhance the anti-escape function of the bionic step 201, set an inverted T-shaped flange at the edge of the bionic step 201, with a height of 2 cm and an inclination angle of 60°. At the same time, coat the surface of the step with a nano-silica coating (the friction coefficient is reduced from 0.6 to 0.15).

[0049] The following table shows the comparison between the present invention and the traditional step structure

[0050]

[0051] Set artificial waterweeds (polyethylene bionic fibers, density 120 roots / m 2), a microporous aeration (pore diameter 0.2 mm, gas volume 0.5 L / min) is provided at the bottom of the breeding tank to simulate the dissolved oxygen environment of a stream. The bionic steps 201 can play a role in behavior induction. The height gradient of the bionic steps 201 triggers the vertical migration nature of bullfrogs, promoting exercise metabolism (experiments show that the daily exercise amount increases by 40%). The curved surface transition design avoids limb injuries (the joint stress is reduced by 65%). Different water depth gradient steps are matched according to different growth stages of bullfrogs; Tadpole stage: 0 - 2 steps (water body dissolved oxygen > 8 mg / L); Metamorphosis stage: 3 - 5 steps (the proportion of the land area is 30%); Adult frog stage: 6 - 8 steps (dynamic adjustment of water depth); At the same time, an anti-escape component 105 is installed on the breeding tanks of adult frogs and juvenile frogs. The anti-escape component 105 includes a protective plate and ultrasonic waves. The protective plate is a transparent polycarbonate protective plate with a light transmittance ≥ 85% and a surface friction coefficient ≤ 0.2. The protective plate is inclined with an outward inclination angle of 30°, which can effectively prevent bullfrogs from jumping out of the breeding tank. At the same time, an ultrasonic array is set at the top of the breeding tank, and the principle of auditory interference is used to drive away bullfrogs. The sensitive frequency range of bullfrog hearing is 0.1 - 5 kHz, and the ultrasonic system emits

[0052] 18 - 22 kHz high-frequency sound waves: (1) Beyond the bullfrog's hearing range but close to its physiological discomfort threshold, causing a sense of irritability and moving away from the sound source; (2) The sound waves are focused on the boundary of the breeding tank (such as the top of the guardrail) to form an "acoustic barrier"; (3) Pulsed ultrasonic waves (such as

[0053] 20 kHz / 0.5 s interval) can interfere with the transmission of bullfrog motor nerve signals and reduce the jumping desire (experimental data: the number of jumping attempts is reduced by 83%).

[0054] The following table shows the composition and deployment plan of the ultrasonic system of the present invention

[0055]

[0056] Workflow, (1) Dynamic monitoring: The infrared sensor scans the edge area of the breeding tank in real time; (2) Target recognition: When a bullfrog enters the warning area 0.5 m away from the boundary, the ultrasonic wave is triggered;

[0057] The hierarchical response is divided into a first-level warning and a second-level suppression. First-level warning: 18 kHz continuous wave (driving away inactive individuals); Second-level suppression: 22 kHz pulsed wave (0.2 s on / 0.3 s off, preventing jumping behavior). In terms of safety protection: The single continuous emission ≤ 10 minutes to prevent acoustic overload stress.

[0058] As an example: Provide an 8m×6m breeding tank, with 4 groups of ultrasonic transmitters arranged in a ring at the top, set the parameters of the ultrasonic waves, and set two modes: daytime and night; Daytime mode: 19kHz intermittent wave (working for 30min / sleeping for 15min), Night mode: 21kHz trigger type (linked with infrared induction);

[0059] Among them, the physiological monitoring data of bullfrogs is collected: the coefficient of variation of heart rate (HRV) fluctuates <5% (normal range), and the difference in food intake ≤3% (compared with the control group). In the experiment on the long-term impact of ultrasonic waves: continuous use for 6 months, and the pathological detection of the hearing organs (cochlea) of bullfrogs is normal; the reproductive rate has no significant difference from that of the control group (P>0.05).

[0060] To sum up, by setting the gradient water level adjustment device and the interlayer isolation water circulation system in the present invention, adding double purification components, bionic steps, anti-escape components and adjustable air curtains, the growth environment of bullfrogs can be effectively enhanced and the survival rate can be improved.

[0061] Example 2

[0062] Refer to Figure 4 , the second embodiment of the present invention, different from the first embodiment, the present embodiment provides a precise feeding control method, including:

[0063] Establish a three-dimensional decision-making model including growth stage, environmental parameters, and group behavior by fusing data from multi-modal sensors;

[0064] Extract the features of the neck, head and limbs of bullfrogs by using the improved YOLOv5s model through the multi-spectrum vision detection module;

[0065] Use the voiceprint feature extraction module to quantify the feeding intensity based on the MFCC coefficients,

[0066] Combine the dynamic feeding decision-making model, and fuse the feeding demand index, environmental parameters and the output value of the prediction network;

[0067] Combine the closed-loop feedback control system to correct the feeding parameters in real time through the residual bait amount.

[0068] Specifically, the improved YOLOv5s model adds a CA attention mechanism in the Neck module, and its mathematical expression formula is as follows:

[0069]

[0070] Among them, is the original parameter vector or matrix, F neck represents the features extracted from the input data, W CA is the attention weight matrix and convolution kernel, is a splicing feature, is the channel attention weight. Multi-modal data collection monitors the activity intensity of the frog group by deploying multiple infrared thermal imagings, and judges the feeding demand of bullfrogs through the monitored activity intensity data, where the sampling frequency of the thermal imaging is 10Hz. In order to feed more accurately, a 3D image recognition system is also installed in the breeding environment to measure the individual body length of bullfrogs, with the error controlled within ±0.5mm. At the same time, a water quality sensor obtains the ammonia nitrogen concentration in real time, and the detection lower limit of the ammonia nitrogen concentration is 0.01mg / L. Among them, closed-loop feedback control. The amount of residual bait (W remain ) real-time corrects the feeding amount (Q adjust ), and dynamically adjusts the threshold (W threshold ). Through visual perception, the improved YOLOv5s model extracts the neck, head, and limb features of bullfrogs; Feature extraction: Extract multi-scale feature maps through the CSPDarknet53 backbone network; Through CA attention embedding, embed the coordinate attention layer in the Neck module (PANet) and calculate the channel weight: W CA =σ(Conv 1×1 ([GAP(F x );GAP(F y )]))

[0071] Among them, W CA represents the attention weight matrix, σ is the activation function, GAP(F x ) and GAP(F y ) are global average pooling, and Conv 1×1 represents a 1x1 convolution operation. Feature fusion: The weighted feature map is used for head regression detection, where global average pooling:

[0072]

[0073] Feature splicing

[0074]

[0075] 1X1 convolution

[0076]

[0077] The dynamic feeding decision model is divided into 12 levels according to the growth stage of bullfrogs. The mathematical expression formula of the feeding demand index is as follows:

[0078]

[0079] Among them, 0.4 and 0.3 represent weight coefficients, represents the proportion of active users, represents the signal strength ratio, Denoted as the temperature influence term, when T w > 28 °C, the FBI weight is automatically adjusted.

[0080] As an example, in terms of weight adaptability: when T w > 28 °C, adjust the weights to α = 0.3, β = 0.2, γ = 0.5, and reduce the feeding amount at high temperatures;

[0081] The following is the division of the growth stages of bullfrogs:

[0082] Stage Weight range (g) Protein requirement (%) Feeding frequency (times / day)

[0083]

[0084] As an example, for the dynamic feeding decision model, the feeding demand index FBI: water temperature T w = 28 °C, the detected proportion of active fusion individuals = 0.6, the voiceprint intensity and water temperature influence = 0.5, and then calculate the FBI. The mathematical expression is:

[0085]

[0086] It is divided into three parts, and each part is calculated separately. The 12 - level growth stage division: dynamically match the nutritional requirements according to weight and age. Decision output: The TCN - LSTM prediction model reduces the feeding amount to 85% of the benchmark value. Feedback correction for the residual bait amount. The mathematical expression is:

[0087] W threshold = 0.2·Q base + 0.05·(T w - 20)

[0088] Among them, W threshold is the calculated threshold, Q base is the base value, T w represents the current temperature, and 0.2 and 0.05 are weight coefficients used to adjust the influence degrees of the base value and temperature on the final result. Dynamic threshold: For every 1 °C increase in temperature, the threshold increases by 0.05 g to adapt to the decreased appetite at high temperatures. Correction expression:

[0089]

[0090] Among them, Q adjust represents the adjusted value, Q base is the base value, represents the proportion of the remaining weight relative to the threshold. Set the feeding parameters through the dynamic threshold. For example, the parameter is set as Q base = 100 g, T w = 26 °C, and W remain= 18 g. The mathematical formula for calculating the threshold value:

[0091] W threshold = 0.2 × Q base + 0.05 × (T w - 20)

[0092] where Q base = 100 (base value), T w = 26 (current working temperature)

[0093] Modify the feeding amount. Feed the frog group according to the collected data, send the collected data to the feeding device, and feed the frog group through the pneumatic feeding device (positioning accuracy ±2 cm, feeding amount error ≤1 g). The pneumatic feeding device uses the PID control algorithm to adjust the throwing speed and angle. The TCN-LSTM hybrid network is constructed through deep learning prediction to predict the food intake in the next 2 hours, which can accurately control the feed amount. Finally, the residual bait amount W remain is monitored in real time by the weighing sensor, and the feeding amount is dynamically corrected.

[0094] In summary, through the above fully automatic and intelligent breeding, the feeding efficiency of bullfrogs can be improved, the accuracy can be increased, and feed waste can be reduced.

[0095] In addition, to provide a concise description of the exemplary embodiments, all features of the actual embodiments may not be described (i.e., those features that are not relevant to the currently considered best mode of implementing the present invention or those features that are not relevant to the implementation of the present invention).

[0096] It should be understood that in the development of any actual implementation, such as in any engineering or design project, a large number of specific implementation decisions may be made. Such development efforts may be complex and time-consuming, but for those of ordinary skill in the art who benefit from this disclosure, without excessive experimentation, such development efforts will be a routine task of design, manufacturing, and production.

[0097] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not restrictive. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and all of them should be covered by the scope of the claims of the present invention.

Claims

1. A three-dimensional bullfrog breeding system, characterized in that, It includes a modular aquaculture unit (100), a gradient water level adjustment device (200) and an interlayer isolation water circulation system (300). The modular aquaculture unit (100) includes a three - stage three - dimensional structure (101), which is divided into a bottom - layer tadpole pond (102), a middle - layer juvenile frog net cage (103) and a top - layer adult frog breeding warehouse (104). Control modules are installed on each layer, and anti - escape components (105) are provided on the middle - layer juvenile frog net cage (103) and the top - layer adult frog breeding warehouse (104); The gradient water level adjustment device (200) includes installed bionic steps (201) and water level sensors, and realizes the functions of water level gradient adjustment and behavior induction through engineering design; The interlayer isolation water circulation system (300) includes distributed sensors, a UV sterilization + biofilm filtration dual - purification component (301) and an adjustable air curtain (302). The distributed sensors and the dual - purification component (301) are installed in each layer of the three - stage three - dimensional structure. The adjustable air curtain forms a directional barrier between different levels of the three - dimensional aquaculture system by dynamically regulating the air flow, maintaining independent environmental parameters for each layer.

2. The bullfrog three-dimensional breeding system according to claim 1, wherein, A temperature control module is installed in the bullfrog three - dimensional aquaculture system, and the temperature control module is connected to the control module.

3. The bullfrog three-dimensional farming system according to claim 1, characterized in that, The dual - purification component (301) includes two types: UV sterilization and biofilm filtration, and the dual - purification component (301) is installed in the bottom - layer tadpole pond (102), the middle - layer juvenile frog net cage (103) and the top - layer adult frog breeding warehouse (104).

4. The bullfrog three-dimensional breeding system according to claim 1, characterized in that, The bionic steps (201) are provided in multiple numbers, with different heights set according to different growth stages of bullfrogs. The surface of the bionic steps (201) is provided with a coating, and the edge of the bionic steps (201) is set in an inverted T - shape.

5. The bullfrog three-dimensional breeding system according to claim 1, characterized in that, The anti - escape component (105) includes a protection plate and ultrasonic waves. The protection plate is inclined, and the ultrasonic waves are arranged in a circumferential array.

6. The bullfrog three-dimensional breeding system according to claim 1, characterized in that, The adjustable air curtain (302) includes an air curtain generator, a servo motor, multi - parameter sensors and an ultrasonic atomizer, and the adjustable air curtain is connected to the control module.

7. A precise feeding control method applied to the bullfrog three-dimensional breeding system described in any one of claims 1-6, characterized in that, It includes: Fusing data through multi - modal sensors to establish a three - dimensional decision - making model including growth stage, environmental parameters and group behavior; Using a multi - spectral vision detection module to extract the characteristics of the bullfrog's neck, head and limbs by adopting an improved YOLOv5s model; Utilizing a voiceprint feature extraction module to quantify the feeding intensity based on MFCC coefficients; Combining with a dynamic feeding decision - making model to fuse the feeding demand index, environmental parameters and the output value of the prediction network; Combining with a closed - loop feedback control system to real - time correct the feeding parameters through the residual bait amount.

8. The precise feeding control method according to claim 7, characterized in that, The improved YOLOv5s model adds a CA attention mechanism in the Neck module, and its mathematical expression formula is as follows: Among them, is the original parameter vector or matrix, F neck represents the features extracted from the input data, W CA is the attention weight matrix and convolution kernel, is the concatenated feature, is the channel attention weight.

9. The precise feeding control method according to claim 7, wherein The dynamic feeding decision - making model is divided into 12 levels according to the growth stage of bullfrogs, and the mathematical expression formula of the feeding demand index is as follows: Among them, 0.4 and 0.3 represent weight coefficients, represents the proportion of active users, represents the signal strength ratio, represents the temperature influence term. When T w > 28 °C, the FBI weight is automatically adjusted.

10. The precise feeding control method according to claim 7, wherein The real-time correction of the residual bait amount uses the threshold value W to set the feeding parameters dynamically, and its mathematical expression formula is as follows: threshold ​ W threshold = 0.2·Q base + 0.05·(T w - 20) Among them, W threshold represents the calculated threshold value, Q base is the base value, T w is the current working temperature, and 0.2 and 0.05 are weight coefficients used to adjust the influence degrees of the base value and the temperature on the final result.

Images