Intelligent feeding method, system and device based on fish and shrimp culture

Abstract

The invention belongs to the technical field of fish and shrimp feeding, and particularly relates to an intelligent feeding method, system and device based on fish and shrimp culture. Through real-time monitoring and analysis of water environment parameters, shrimp group activity states and dynamic changes in the feed putting process, accurate regulation and control of a feeding strategy are achieved, specifically, through calculation and prediction of shrimp group feeding activity indexes and correction of the amount of residual feed at the bottom of the pond, it is ensured that the feed putting amount is matched with the actual demand, and the feed putting efficiency is improved. In addition, due to the design of a segmented feeding mechanism and a dynamic optimization module, the dispersion uniformity of the feed in the water body is further improved, the risks of feed sedimentation and excessive deposition are reduced, it is guaranteed that shrimp groups can ingest in a proper environment, and the feed utilization rate is increased. It is ensured that the living water environment of the shrimp groups cannot deteriorate due to improper feeding, and correspondingly, healthy growth of the shrimp groups can be ensured.

Description

Technical Field

[0001] This invention belongs to the field of fish and shrimp feeding technology, specifically relating to an intelligent feeding method, system and device based on fish and shrimp farming. Background Technology

[0002] In traditional fish and shrimp farming, feeding relies heavily on manual experience, making it difficult to precisely control the amount of feed. This can easily lead to overfeeding or underfeeding, which in turn affects the growth rate of farmed organisms and the quality of the aquatic environment. In particular, improper feeding in fish and shrimp farming often leads to problems such as feed waste, water quality deterioration, and disease outbreaks. This, in turn, results in reduced feed utilization and increased farming costs, which undoubtedly contradicts the original intention of feeding fish and shrimp. To ensure the effectiveness of fish and shrimp feeding, optimizing the feeding process is clearly essential.

[0003] While some automatic feeding devices exist in the prior art, they mostly feed based on fixed times or simple environmental parameters, lacking the comprehensive perception and response capabilities to real-time shrimp behavior, feeding status, and dynamic changes in the water body. This makes it difficult to achieve precise feeding on demand. Compared to traditional manual feeding, they only automate the feeding process but fail to fundamentally solve the problem of improper feeding. Often, the amount of feed given does not match the actual demand, resulting in feed remaining in the water for too long or being over-ingested, which in turn exacerbates water quality fluctuations and aquaculture risks. Based on this, the present invention provides an intelligent feeding method and device for fish and shrimp farming to solve the above problems. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent feeding method, system, and device for fish and shrimp farming, which can dynamically optimize the feeding strategy based on data such as the real-time distribution density of shrimp populations, feeding activity, and dissolved oxygen levels in the water, so as to ensure the stability of the aquatic environment while achieving adequate feeding of the shrimp populations.

[0005] The specific technical solution adopted by this invention is as follows: A smart feeding method for fish and shrimp farming includes: Real-time acquisition of aquatic environmental parameters, images of shrimp activity, and the intensity of feeding sounds. The aquatic environmental parameters include water temperature, dissolved oxygen, pH value, and turbidity. The shrimp's body size and feeding behavior are identified from images of shrimp activity, and the feeding activity index is calculated by combining the intensity of feeding sounds. The feed demand of shrimp populations during future demand periods is predicted based on feeding activity index, and the feed demand is corrected based on the amount of feed remaining at the bottom of the pond to obtain the corrected actual feed demand. The dissolved oxygen rate during feed feeding is collected in real time. When the dissolved oxygen rate is lower than or equal to the preset dissolved oxygen rate threshold, a segmented feeding mechanism is triggered. When the dissolved oxygen rate is higher than the preset dissolved oxygen rate threshold, the settling speed and dispersion uniformity of the feed are determined by combining water temperature fluctuations and turbidity rise slope. The feeding parameters are dynamically optimized based on the settling speed and dispersion uniformity of the feed, and the feeding speed and distribution of feeding points are adjusted until the feed is fed in the current feeding cycle.

[0006] In a preferred embodiment, the step of real-time acquisition of aquatic environmental parameters, shrimp activity images, and feeding sound intensity includes: Multi-node sensor arrays are deployed in the surface, middle and bottom layers of the water body, and water temperature, dissolved oxygen, pH value and turbidity data are acquired in real time through the sensor arrays; A wide-angle camera was deployed underwater, and images of the shrimp's activities were captured at regular intervals using the wide-angle camera. At the same time, a directional underwater sonar device was used to collect the feeding sound wave signals of the shrimp during the feeding process. Water temperature, dissolved oxygen, pH, and turbidity data are aggregated into an aquatic environment data stream and time-stamped with shrimp activity images and feeding sound signals to form a related dataset.

[0007] In a preferred embodiment, the step of identifying the shrimp's body size and feeding behavior based on images of shrimp group activity, and calculating the shrimp group's feeding activity index in conjunction with the intensity of feeding sounds, includes: The images of shrimp activity are segmented to extract the outlines of individual shrimp, and the length and weight of the shrimp are estimated using a preset body shape template. Based on time series analysis of shrimp activity images, the frequency of mouthpart opening and closing and swimming trajectory of shrimp are identified. Combined with the dynamic changes in mouthpart opening and closing frequency and swimming trajectory, it is determined whether the shrimp group is in an active foraging state. Based on the energy intensity and frequency of the feeding sound wave signal, the intensity of the feeding sound is quantified and output as a feeding sound intensity index; The feeding sound intensity index is fused with the mouthpart opening and closing frequency and swimming activity, and output as a feeding activity index of shrimp groups.

[0008] In a preferred embodiment, the step of predicting the feed demand of shrimp populations during future demand periods based on feeding activity indicators includes: Collect feeding activity indicators during the current feeding cycle and summarize them into a historical feeding characteristic sequence; Time-series analysis of feeding activity indicators under historical feeding characteristic sequences is performed to output the trend of feeding activity changes; Obtain the demand forecast point for the future demand period, as well as the time interval between the demand forecast point and the real-time feeding node, and record it as the forecast time interval; Extrapolate the trend of feeding activity based on the predicted time interval, and construct a feed demand prediction function by combining the dynamic influence weight of aquatic environment data. The feeding activity trend, prediction time interval, and aquatic environmental parameters are input into the feed demand prediction function, and the output is recorded as feed demand.

[0009] In a preferred embodiment, the specific steps for outputting the dynamic influence weights of the water environment data are as follows: Acquire current water environment data and compare it with preset water environment adaptation thresholds to determine the real-time deviation of the water environment data; Based on the magnitude of the real-time deviation, basic influence coefficients are assigned to various environmental factors under the water environment data, and positive and negative values ​​are assigned to the basic influence coefficients according to the direction of deviation. The basic impact coefficients of each environmental factor, after being assigned a value, are weighted and fused with the corresponding time-series change rates to obtain the dynamic impact weights of each environmental factor.

[0010] In a preferred embodiment, the step of correcting the feed requirement based on the amount of residual feed at the bottom of the pond to obtain the corrected actual feed requirement includes: The distribution density of uneaten residual feed particles is identified by capturing the feed settling area at the bottom of the pool using a wide-angle camera. The total amount of residual feed at the bottom of the pond is calculated based on the distribution density of residual feed, and the time interval from the last feeding to the current time is obtained. The degradation amount of residual feed at the bottom of the pond is calculated in combination with the feed dissolution rate. Subtract the degradation amount from the total amount of feed remaining at the bottom of the pond to obtain the current effective residual amount. Then subtract the current effective residual amount from the feed demand. If the result is greater than zero, use it as the corrected actual feed demand; otherwise, set it to zero.

[0011] In a preferred embodiment, the step of determining the feed settling velocity and dispersion uniformity by combining water temperature fluctuations and turbidity rise slope when the dissolved oxygen change rate is higher than a preset dissolved oxygen rate threshold includes: The water temperature fluctuation sequence and turbidity change sequence during the current feeding period are collected using a sliding window mechanism. The water temperature fluctuation amplitude is calculated based on the water temperature fluctuation sequence, and the turbidity rise slope is calculated by combining the turbidity change sequence. Obtain the mapping relationship between feed settling velocity and water temperature fluctuation amplitude, and match the initial settling velocity based on the current water temperature fluctuation amplitude; The initial settling velocity was corrected based on the turbidity rise slope to obtain the corrected feed settling velocity. The suspension time of the feed particles in the water was then calculated by combining the water flow velocity in the pool with the distribution of the feeding points. The suspension time and the corrected feed settling velocity were coupled for analysis to evaluate the uniformity of feed dispersion in the water. If the uniformity of dispersion was lower than the preset standard, the spraying angle and frequency of the feeder were dynamically adjusted to optimize the spatial distribution of feed.

[0012] In a preferred embodiment, the step of dynamically optimizing the feeding parameters based on the feed settling velocity and dispersion uniformity, and adjusting the feeding speed and distribution of feeding points, includes: Based on the corrected feed settling velocity, the historical optimal feeding velocity under the same settling velocity conditions in the historical feeding data is matched. The historical optimal feeding velocity is the feeding velocity that maximizes the feeding rate of the farmed organisms while ensuring that the feed is fully dispersed and without excessive deposition. The baseline value of the current feeding speed is determined based on the historical best feeding speed, and dynamically corrected by combining the real-time dispersion uniformity evaluation results. When the dispersion uniformity is lower than the preset threshold, the feeding speed is reduced to prolong the feed diffusion time, and the spatial distribution density of the feeding points is optimized according to the cross-sectional area of ​​the pool and the direction of water flow. When the uniformity of dispersion is higher than the preset threshold, the feeding speed is increased to improve feeding efficiency, and the distribution angle of the feeding point is adjusted according to the water flow speed and the topography of the pond bottom to ensure that the feed covers the densely active areas of aquatic organisms.

[0013] This invention also provides an intelligent feeding system based on fish and shrimp farming, using the aforementioned intelligent feeding method for fish and shrimp farming, comprising: The data acquisition module is used to collect aquatic environmental parameters, images of shrimp activity, and the intensity of feeding sounds in real time. The aquatic environmental parameters include water temperature, dissolved oxygen, pH value, and turbidity. The feeding assessment module is used to identify the size and feeding behavior of shrimp based on images of shrimp activity, and to calculate the feeding activity index of the shrimp population by combining the intensity of feeding sounds. The feeding assessment module is used to predict the feed demand of shrimp groups during future demand periods based on feeding activity indicators, and to correct the feed demand based on the amount of feed remaining at the bottom of the pond to obtain the corrected actual feed demand. The segmented feeding module is used to collect the dissolved oxygen change rate during the feed feeding process in real time. When the dissolved oxygen change rate is lower than or equal to the preset dissolved oxygen rate threshold, the segmented feeding mechanism is triggered. When the dissolved oxygen change rate is higher than the preset dissolved oxygen rate threshold, the settling speed and dispersion uniformity of the feed are determined by combining water temperature fluctuations and turbidity rise slope. The feeding optimization module is used to dynamically optimize the feeding parameters based on the feed settling speed and dispersion uniformity, and adjust the feeding speed and distribution of feeding points until the feed is fed in the current feeding cycle.

[0014] And, an intelligent feeding device based on fish and shrimp farming, including a rearing tank, a guide platform in the middle of the rearing tank, and a feeder that moves along a sliding rail installed above the guide platform. The feeder integrates a processor connected to a wide-angle camera and sensor array; and a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which is then executed by the at least one processor to enable the at least one processor to perform the intelligent feeding method based on fish and shrimp farming as described in any one of claims 1 to 8.

[0015] The technical effects achieved by this invention are as follows: This invention achieves precise control of feeding strategies by real-time monitoring and analysis of aquatic environmental parameters, shrimp activity status, and dynamic changes during feed delivery. This not only effectively avoids feed waste but also significantly improves feed utilization, thereby reducing farming costs. By calculating and predicting shrimp feeding activity indicators and correcting for residual feed at the bottom of the pond, the invention ensures that the feed delivery amount matches actual needs, avoiding water quality deterioration caused by overfeeding. Furthermore, the segmented feeding mechanism and dynamic optimization module further improve the uniformity of feed dispersion in the water, reducing the risk of feed settling and excessive deposition. This ensures that shrimp can feed in a suitable environment, preventing the aquatic environment from deteriorating due to improper feeding, and consequently ensuring the healthy growth of the shrimp. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is a schematic diagram of the system modules of the present invention; Figure 3 This is a schematic diagram of the feeding device structure of the present invention; Figure 4 This is a schematic diagram of the processor of the present invention.

[0017] The attached diagram lists the components represented by each number as follows: breeding pond; Guide table; Slide rail; Feeder. Detailed Implementation

[0018] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0019] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0020] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in a preferred embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0021] Please see Figure 1 As shown, this invention provides an intelligent feeding method for fish and shrimp farming, comprising: S1. Real-time acquisition of aquatic environmental parameters, images of shrimp activity, and the intensity of feeding sounds. Among them, aquatic environmental parameters include water temperature, dissolved oxygen, pH value, and turbidity. In step S1, in the field of fish and shrimp farming, in order to ensure the healthy growth of shrimp populations and efficient use of feed, by capturing the body shape characteristics and feeding movements of shrimp populations, combined with the sound intensity produced during feeding, the feeding activity of shrimp populations can be effectively assessed, thereby providing effective data support for feed demand prediction. The steps of real-time acquisition of aquatic environmental parameters, images of shrimp population activity, and the intensity of feeding sound include: Multi-node sensor arrays are deployed in the surface, middle and bottom layers of the water body, and water temperature, dissolved oxygen, pH value and turbidity data are acquired in real time through the sensor arrays; A wide-angle camera was deployed underwater, and images of the shrimp's activities were captured at regular intervals using the wide-angle camera. At the same time, a directional underwater sonar device was used to collect the feeding sound wave signals of the shrimp during the feeding process. Water temperature, dissolved oxygen, pH, and turbidity data are aggregated into an aquatic environment data stream and time-stamped with shrimp activity images and feeding sound signals to form a related dataset. Specifically, when collecting aquatic environmental parameters, a multi-node sensor array is first deployed at different depths in the aquaculture pond to monitor changes in water temperature, dissolved oxygen, pH, and turbidity in real time. This stratified acquisition method allows for a comprehensive understanding of the dynamic changes in the aquatic environment, providing a reliable basis for subsequent data analysis. In addition, wide-angle cameras and directional underwater sonar devices are deployed in the shrimp's habitat to continuously capture the shrimp's activity trajectories and feeding sounds, thereby obtaining the shrimp's behavioral characteristics and feeding patterns under different environmental conditions. Subsequently, the water temperature, dissolved oxygen, pH, and turbidity data are time-stamped with the shrimp activity images and feeding sound signals to form a complete spatiotemporal correlation dataset, which provides corresponding data support for subsequent feeding decisions.

[0022] S2. Identify the body size and feeding behavior of shrimp based on images of shrimp activity, and calculate the feeding activity index of shrimp based on the intensity of feeding sounds. In step S2, after the shrimp activity images are acquired, the shrimp's body shape and feeding behavior are identified based on them. Simultaneously, the feeding activity level of the shrimp population is quantitatively assessed by combining the intensity of feeding sounds. This provides corresponding parameter support for subsequent feed demand prediction and feeding strategy optimization. The steps of identifying shrimp body shape and feeding behavior based on the shrimp activity images and calculating the shrimp population feeding activity index based on the intensity of feeding sounds include: The images of shrimp activity are segmented to extract the outlines of individual shrimp, and the length and weight of the shrimp are estimated using a preset body shape template. Based on time series analysis of shrimp activity images, the frequency of mouthpart opening and closing and swimming trajectory of shrimp are identified. Combined with the dynamic changes in mouthpart opening and closing frequency and swimming trajectory, it is determined whether the shrimp group is in an active foraging state. Based on the energy intensity and frequency of the feeding sound wave signal, the intensity of the feeding sound is quantified and output as a feeding sound intensity index; The feeding sound intensity index is fused with the mouthpart opening and closing frequency and swimming activity, and output as a feeding activity index of shrimp groups. Specifically, when assessing shrimp feeding activity indicators, the collected images of shrimp activity are first segmented to extract the outline information of each shrimp. Based on a pre-defined body size template, the outline dimensions are compared to estimate the shrimp's body length and weight, quantifying individual growth status. For example, by matching the aspect ratio and area parameters of the outline to a standard growth curve, the approximate developmental stage and weight range of the shrimp can be calculated, providing a basis for individualized adjustments to subsequent feeding amounts. Simultaneously, time-series analysis of shrimp activity is performed based on the images to capture mouthpart opening and closing frequency and swimming trajectory characteristics. The trend of mouthpart opening and closing frequency and the density of swimming trajectories are used to determine whether the shrimp are actively foraging. If mouthpart opening and closing are frequent and swimming trajectories show clustered diffusion, it is considered high-activity foraging behavior. Specifically, a threshold range can be set for mouthpart opening and closing frequency using a threshold judgment method. When the number of mouthpart opening and closing times per unit time exceeds a set baseline value and remains so... If the feeding continues for a certain period of time, it is considered a valid foraging behavior. Simultaneously, by combining the speed and direction consistency of the shrimp's swimming trajectory, the movement pattern of the shrimp group is identified. If most individuals swim at high speed in the same direction accompanied by high-frequency opening and closing of their mouthparts, it is determined to be collective foraging behavior. Furthermore, by combining the feeding sound wave signals captured by underwater sonar, its energy intensity and frequency are extracted to calculate the feeding sound intensity index. The feeding sound intensity index is calculated based on the average energy value and peak frequency of the sound wave signal per unit time. The specific calculation formula is: Feeding Sound Intensity Index = (Average Energy Value of Sound Wave Signal per Unit Time × Peak Frequency of Sound Wave) / Time Interval Coefficient. The time interval coefficient is dynamically adjusted according to the background noise level of the aquaculture environment to eliminate the influence of environmental interference on the index calculation. Finally, by weighting and fusing the feeding sound intensity index with the mouthpart opening and closing frequency and swimming activity, a multi-dimensional shrimp feeding activity index can be constructed, thus providing corresponding data support for subsequent feed feeding.

[0023] S3. Based on the feeding activity index, predict the feed demand of the shrimp population during the future demand period, and adjust the feed demand according to the amount of feed remaining at the bottom of the pond to obtain the adjusted actual feed demand. In step S3, after the shrimp population's feeding activity index is output, the feed demand for future demand periods can be predicted based on this, so as to plan feeding strategies in advance and provide the shrimp population with appropriate feed in a timely manner, avoiding growth inhibition and water quality deterioration due to insufficient or excessive feeding. The step of predicting the shrimp population's feed demand for future demand periods based on the feeding activity index includes: Collect feeding activity indicators during the current feeding cycle and summarize them into a historical feeding characteristic sequence; Time-series analysis of feeding activity indicators under historical feeding characteristic sequences is performed to output the trend of feeding activity changes; Obtain the demand forecast point for the future demand period, as well as the time interval between the demand forecast point and the real-time feeding node, and record it as the forecast time interval; Extrapolate the trend of feeding activity based on the predicted time interval, and construct a feed demand prediction function by combining the dynamic influence weight of aquatic environment data. Input the feeding activity change trend, prediction time interval and water environment parameters into the feed demand prediction function, and record the output as feed demand. Specifically, when predicting the feed demand of shrimp populations, it is first necessary to collect feeding activity indicators within the current feeding cycle and organize them into a continuous historical feeding characteristic sequence. If the current feeding cycle is insufficient to cover the minimum required sequence length, feeding activity indicator data of a corresponding duration from the previous feeding cycle is extracted to supplement it, ensuring sufficient data to support subsequent time-series analysis. When performing time-series analysis on the historical feeding characteristic sequence, a sliding window method or autoregressive model is used to fit the trend of the sequence, identifying the periodic fluctuations and growth or decline slopes of feeding activity, thereby predicting the direction of change in future demand periods. Taking the sliding window method as an example, first, the length of the sliding window is set, and then the sliding window is moved point by point on the historical feeding characteristic sequence. The mean and variance of the feeding activity indicator within each window are calculated, and the local change slope is obtained through trend line fitting, thereby judging the overall change trend and obtaining future... Once the demand forecast point for a given period is determined, the time interval between that point and the current real-time feeding node can be established, i.e., the forecast time interval. Based on this forecast time interval, the trend of feeding activity changes is extrapolated. Simultaneously, the influence of aquatic environmental parameters such as water temperature, dissolved oxygen, and pH on feed conversion ratio is considered, assigning dynamic influence weights to each parameter. Generally, higher water temperatures increase feed metabolism efficiency, sufficient dissolved oxygen promotes feeding activity, while pH deviations from the appropriate range inhibit feeding behavior. Therefore, when constructing the feed demand forecast function, the real-time monitoring values ​​of parameters such as water temperature, dissolved oxygen, and pH are multiplied by their corresponding dynamic influence weights, and then superimposed onto the extrapolated trend of feeding activity changes to obtain the corrected feed demand forecast value. Specifically, the constructed feed demand forecast function nonlinearly fuses the extrapolated trend of feeding activity changes with the weighted aquatic environmental parameters to estimate feed demand. In the formula, Indicates the amount of feed required. Indicates the predicted time interval. Extrapolated values ​​representing trends in feeding activity levels This represents an environmental correction factor, used to adjust the intensity of the impact of environmental parameters on feed demand. This indicates the dynamic impact weight, reflecting the degree of influence of aquatic environmental parameters on feed conversion. Vectorized representations of aquatic environmental parameters, including water temperature, dissolved oxygen, and pH value.

[0024] It should be noted that the specific steps for outputting the dynamic influence weights of water environment data are as follows: Acquire current water environment data and compare it with preset water environment adaptation thresholds to determine the real-time deviation of the water environment data; Based on the magnitude of the real-time deviation, basic influence coefficients are assigned to various environmental factors under the water environment data, and positive and negative values ​​are assigned to the basic influence coefficients according to the direction of deviation. The basic impact coefficients of each environmental factor after being assigned a value are weighted and fused with the corresponding time-series change rate to obtain the dynamic impact weight of each environmental factor. When outputting the dynamic impact weights of aquatic environmental data, it is first necessary to collect current aquatic environmental data, including water temperature, dissolved oxygen, pH value, etc., and then compare it with the preset aquatic environmental adaptation threshold. By calculating the difference between the real-time data and the threshold, the degree of deviation of each environmental factor is determined. Subsequently, based on the magnitude of the deviation, a basic impact coefficient is assigned to each environmental factor, and positive or negative values ​​are assigned according to the direction of deviation. For example, a positive value is assigned when the water temperature is higher than the suitable range, and a negative value is assigned when it is lower than the suitable range. Then, the assigned basic impact coefficients are combined with the time-series change rate of the environmental factors, and the dynamic impact weight of each environmental factor is obtained through weighted calculation. This ensures that the changes in aquatic environmental parameters can be quantified and reflected in feed demand forecasting, thereby improving the accuracy of feed feeding decisions.

[0025] Secondly, the steps for adjusting the feed requirement based on the amount of residual feed at the bottom of the pond to obtain the adjusted actual feed requirement include: The distribution density of uneaten residual feed particles is identified by capturing the feed settling area at the bottom of the pool using a wide-angle camera. The total amount of residual feed at the bottom of the pond is calculated based on the distribution density of residual feed, and the time interval from the last feeding to the current time is obtained. The degradation amount of residual feed at the bottom of the pond is calculated in combination with the feed dissolution rate. Subtract the degradation amount from the total amount of feed remaining at the bottom of the pond to obtain the current effective residual amount. Then subtract the current effective residual amount from the feed demand. If the result is greater than zero, use it as the corrected actual feed demand; otherwise, set it to zero. In actual feeding operations, some feed particles inevitably accumulate in areas where shrimp are active. These include uneaten feed particles and settled feed resuspended by water flow disturbance. Since these still provide feeding opportunities for the shrimp, they need to be taken into account when calculating the feed amount. Specifically, a wide-angle camera is used to monitor the feed settling area at the bottom of the pond in real time, identifying the distribution density of residual particles. Based on this density, the total amount of residual feed at the bottom of the pond is calculated. When identifying the distribution density of residual feed, image processing technology is used to analyze the images captured by the wide-angle camera. First, features of the feed particles in the image are extracted, including the shape, color, and size of the particles, and then compared with a preset standard feed particle model. The feed pellets are matched to distinguish them from other impurities. The distribution density is then calculated based on the number of feed pellets per unit area. Combined with the bottom area of ​​the aquaculture pond, the total amount of residual feed is estimated. Then, considering the time interval from the last feeding to the current moment and the environmental dissolution rate, the degradation and consumption of residual feed at the bottom of the pond is estimated. This part is then subtracted from the total residual amount to obtain the effective residual amount that still has feeding value. Based on this, the predicted feed demand is subtracted from the effective residual amount. If the result is positive, it is recorded as the corrected actual feed demand. If the result is less than or equal to zero, it means that the current residual feed is sufficient to meet the demand. In this case, the actual feed demand can be set to zero to avoid overfeeding, which would lead to resource waste and water pollution.

[0026] S4. Real-time collection of dissolved oxygen change rate during feed feeding process, and triggering segmented feeding mechanism when the dissolved oxygen change rate is lower than or equal to the preset dissolved oxygen rate threshold. When the dissolved oxygen change rate is higher than the preset dissolved oxygen rate threshold, the feed settling speed and dispersion uniformity are determined by combining water temperature fluctuation and turbidity rise slope. In step S4, the dissolved oxygen change rate is directly related to the shrimp's feeding metabolic intensity and feed utilization efficiency in the water environment where the shrimp are active. When the dissolved oxygen change rate is lower than the threshold, it indicates that the water's oxygen supply is insufficient to support high-intensity feeding. At this time, a segmented feeding mechanism needs to be activated, that is, by dividing the single feeding amount into multiple small dose periods for continuous feeding, to alleviate the feeding inhibition and feed waste caused by hypoxia in the shrimp, and to avoid anaerobic decomposition and water quality deterioration caused by the accumulation of feed at the bottom. Among them, when the dissolved oxygen change rate is higher than the preset dissolved oxygen rate threshold, the step of determining the feed settling speed and dispersion uniformity by combining water temperature fluctuations and turbidity rise slope includes: The water temperature fluctuation sequence and turbidity change sequence during the current feeding period are collected using a sliding window mechanism. The water temperature fluctuation amplitude is calculated based on the water temperature fluctuation sequence, and the turbidity rise slope is calculated by combining the turbidity change sequence. Obtain the mapping relationship between feed settling velocity and water temperature fluctuation amplitude, and match the initial settling velocity based on the current water temperature fluctuation amplitude; The initial settling velocity was corrected based on the turbidity rise slope to obtain the corrected feed settling velocity. The suspension time of the feed particles in the water was then calculated by combining the water flow velocity in the pool with the distribution of the feeding points. The suspension time and the corrected feed settling velocity were coupled for analysis to evaluate the uniformity of feed dispersion in the water. If the uniformity of dispersion was lower than the preset standard, the spraying angle and frequency of the feeder were dynamically adjusted to optimize the spatial distribution of feed. Specifically, when determining the corrected feed settling velocity and dispersion uniformity, the water temperature fluctuation sequence and turbidity change sequence are first collected based on a sliding window mechanism during the current feeding period. This refines the behavioral characteristics of the feed in the water. The calculation of the water temperature fluctuation sequence relies on the cumulative temperature difference per unit time. Combined with historical data of the aquaculture environment, short-term temperature change patterns are identified, thus providing a basis for subsequent settling velocity assessment. Simultaneously, the turbidity change sequence is monitored in real time by optical sensors to track the changing trend of suspended particle concentration in the water. The turbidity rise slope is calculated based on the increase in turbidity value per unit time. The turbidity rise slope reflects the diffusion speed of feed particles in the water and its impact on water transparency. After obtaining the water temperature fluctuation amplitude, the initial settling velocity is matched according to a preset mapping table. This mapping table is usually generated from experimental data or historical feeding records to ensure high accuracy in estimating the initial settling velocity. Then, the initial settling velocity is corrected by combining the turbidity rise slope to reflect the dynamic changes in the movement state of feed particles in the actual aquatic environment. For example, when the turbidity rise slope is large, it indicates that the feed particles diffuse faster in the water, and the settling velocity should be appropriately reduced. Conversely, the settling velocity needs to be increased to ensure that the feed can quickly reach the shrimp activity area. The corrected feed settling velocity will be combined with the water flow velocity in the pond and the distribution location of the feeding point to calculate the suspension time of feed particles in the water. The length of the suspension time directly affects the probability that the feed will be ingested by the shrimp. On this basis, the suspension time and settling velocity will also be coupled and analyzed to evaluate the uniformity of feed dispersion in the water. If the uniformity of dispersion is lower than the preset standard, it indicates that the feed particles may be concentrated in certain areas and fail to cover the entire shrimp activity range. At this time, the spraying angle and frequency of the feeding device will be adjusted to improve the spatial distribution effect of the feed, thereby improving the overall feeding efficiency of the shrimp.

[0027] S5. Dynamically optimize the feeding parameters based on the feed settling speed and dispersion uniformity, and adjust the feeding speed and distribution of feeding points until the feed feeding in the current feeding cycle is completed. In step S5, during the optimization of feeding parameters, the real-time feeding behavior characteristics of the shrimp population and the feedback data of the aquatic environment are comprehensively considered. The feeding speed is dynamically adjusted to match the feeding intensity at the current time. At the same time, the spatial layout of the feeding points is optimized in combination with the pond bottom topography and water flow field distribution to ensure the coverage of feed particles in the water. The steps of dynamically optimizing the feeding parameters based on the feed settling speed and dispersion uniformity, and adjusting the feeding speed and distribution of feeding points, include: Based on the corrected feed settling velocity, the historical optimal feeding velocity under the same settling velocity conditions in the historical feeding data is matched. The historical optimal feeding velocity is the feeding velocity that maximizes the feeding rate of the farmed organisms while ensuring that the feed is fully dispersed and without excessive deposition. The baseline value of the current feeding speed is determined based on the historical best feeding speed, and dynamically corrected by combining the real-time dispersion uniformity evaluation results. When the dispersion uniformity is lower than the preset threshold, the feeding speed is reduced to prolong the feed diffusion time, and the spatial distribution density of the feeding points is optimized according to the cross-sectional area of ​​the pool and the direction of water flow. When the dispersion uniformity is higher than the preset threshold, increase the feeding speed to improve feeding efficiency, and adjust the distribution angle of the feeding point according to the water flow speed and the topography of the pond bottom to ensure that the feed covers the densely active areas of aquatic organisms. In this implementation, when optimizing feeding parameters, the first step is to match the historical optimal feeding speed, based on the corrected feed settling velocity and relevant records from historical feeding data, to the optimal feeding speed that best matches the current conditions. The historical optimal feeding speed is the data that ensures the feed particles are sufficiently dispersed without excessive sedimentation at the bottom of the pond, allowing the shrimp population to reach peak feeding rates. This provides a benchmark value for the current feeding speed. The speed is then dynamically adjusted based on real-time monitoring of dispersion uniformity. For example, when dispersion uniformity is below a preset threshold, the feeding speed is automatically reduced to prolong the diffusion time of feed particles in the water, while simultaneously optimizing the spatial distribution density of the feeding points to ensure the feed covers a wider area. Conversely, when dispersion uniformity is above the preset threshold, the feeding speed is appropriately increased to improve feeding efficiency, and adjustments are made according to water flow. Adjusting the feeding speed and pond bottom topography to optimize the distribution angle of feed delivery points ensures that feed pellets cover areas where shrimp activity is high, maximizing feeding efficiency and minimizing waste. In dynamically optimizing feeding parameters, it's also necessary to consider the real-time feeding behavior of the shrimp and feedback from aquatic environmental data. For example, real-time monitoring of shrimp feeding activity and dissolved oxygen levels using sensors allows for further adjustments to the feeding speed and distribution strategy. If a decrease in shrimp feeding activity or a drop in dissolved oxygen levels is detected, the feeding speed should be reduced promptly to prevent water quality deterioration due to overfeeding. Simultaneously, combining pond bottom topography and water flow distribution optimizes the spatial layout of delivery points, ensuring even distribution of feed pellets in the water and reducing localized accumulation or blank areas, thus guaranteeing healthy shrimp growth.

[0028] Please see Figure 2 A smart feeding system based on fish and shrimp farming, using the aforementioned smart feeding method based on fish and shrimp farming, includes: The data acquisition module is used to collect aquatic environmental parameters, images of shrimp activity, and the intensity of feeding sounds in real time. The aquatic environmental parameters include water temperature, dissolved oxygen, pH value, and turbidity. The feeding assessment module is used to identify the size and feeding behavior of shrimp based on images of shrimp activity, and to calculate the feeding activity index of the shrimp population by combining the intensity of feeding sounds. The feeding assessment module is used to predict the feed demand of shrimp groups during future demand periods based on feeding activity indicators, and to correct the feed demand based on the amount of feed remaining at the bottom of the pond to obtain the corrected actual feed demand. The segmented feeding module is used to collect the dissolved oxygen change rate during the feed feeding process in real time. When the dissolved oxygen change rate is lower than or equal to the preset dissolved oxygen rate threshold, the segmented feeding mechanism is triggered. When the dissolved oxygen change rate is higher than the preset dissolved oxygen rate threshold, the settling speed and dispersion uniformity of the feed are determined by combining water temperature fluctuations and turbidity rise slope. The feeding optimization module is used to dynamically optimize the feeding parameters based on the feed settling speed and dispersion uniformity, and adjust the feeding speed and distribution of feeding points until the feed is fed in the current feeding cycle.

[0029] The execution process of the above-mentioned intelligent feeding system corresponds exactly to the process of the aforementioned method, so it will not be repeated here.

[0030] Please see Figure 3 A smart feeding device based on fish and shrimp farming includes a rearing pond 1, a guide workbench 2 in the middle of the rearing pond 1, and a feeder 4 that moves along a slide rail 3 installed above the guide workbench 2. The feeder 4 integrates a processor connected to a wide-angle camera and sensor array; and a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which is then executed by the at least one processor to enable the at least one processor to perform the intelligent feeding method based on fish and shrimp farming as described in any one of claims 1 to 8.

[0031] In the above-mentioned configuration, the rearing tank 1 is used to house the farmed shrimp population, the guide platform 2 is used to guide the feeder 4 along a predetermined path to facilitate precise feeding to different areas within the rearing tank 1, a wide-angle camera is used to capture real-time global images of the shrimp population's activities and is installed inside the rearing tank 1, and a sensor array is used to monitor water temperature, dissolved oxygen, pH value, and turbidity in real time, also installed inside the rearing tank 1, working in conjunction with a processor to achieve real-time acquisition and feedback of aquatic environmental parameters. The processor can be a central processing unit (CPU), graphics processing unit (GPU), or digital signal processor (DSP), etc., used to process the acquired images and aquatic environmental parameters. The data is analyzed and processed in real time, and the feeding activity and feed demand of the shrimp population are calculated. The memory may include high-speed random access memory and non-volatile memory, which are used to store feeding control algorithms and historical data to support optimization decisions. The feeder 4 is also equipped with an arithmetic logic unit (ALU) connected to the processor, an input interface and an output interface. The ALU can be an arithmetic logic unit, which is used to perform calculations such as feed demand calculation, dissolved oxygen change rate analysis and feeding parameter optimization. The input interface is used to receive external instructions and sensor data, and the output interface is used to send control signals to the feeder's drive device to control the feeding speed, etc.

[0032] The above description is merely a preferred embodiment of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention. Structures, devices, and operating methods not specifically described or explained in this invention are implemented according to conventional methods in the art unless otherwise specified or limited.

Claims

1. A smart feeding method for fish and shrimp farming, characterized in that: include: Real-time acquisition of aquatic environmental parameters, images of shrimp activity, and the intensity of feeding sounds. The aquatic environmental parameters include water temperature, dissolved oxygen, pH value, and turbidity. The shrimp's body size and feeding behavior are identified from images of shrimp activity, and the feeding activity index is calculated by combining the intensity of feeding sounds. The feed demand of shrimp populations during future demand periods is predicted based on feeding activity index, and the feed demand is corrected based on the amount of feed remaining at the bottom of the pond to obtain the corrected actual feed demand. The dissolved oxygen rate during feed feeding is collected in real time. When the dissolved oxygen rate is lower than or equal to the preset dissolved oxygen rate threshold, a segmented feeding mechanism is triggered. When the dissolved oxygen rate is higher than the preset dissolved oxygen rate threshold, the settling speed and dispersion uniformity of the feed are determined by combining water temperature fluctuations and turbidity rise slope. The feeding parameters are dynamically optimized based on the settling speed and dispersion uniformity of the feed, and the feeding speed and distribution of feeding points are adjusted until the feed is fed in the current feeding cycle.

2. The intelligent feeding method based on fish and shrimp farming according to claim 1, characterized in that: The steps for real-time acquisition of aquatic environmental parameters, shrimp activity images, and feeding sound intensity include: Multi-node sensor arrays are deployed in the surface, middle and bottom layers of the water body, and water temperature, dissolved oxygen, pH value and turbidity data are acquired in real time through the sensor arrays; A wide-angle camera was deployed underwater, and images of the shrimp's activities were captured at regular intervals using the wide-angle camera. At the same time, a directional underwater sonar device was used to collect the feeding sound wave signals of the shrimp during the feeding process. Water temperature, dissolved oxygen, pH, and turbidity data are aggregated into an aquatic environment data stream and time-stamped with shrimp activity images and feeding sound signals to form a correlated dataset.

3. The intelligent feeding method based on fish and shrimp farming according to claim 1, characterized in that: The steps of identifying shrimp body size and feeding behavior based on shrimp activity images and calculating shrimp feeding activity indexes by combining the intensity of feeding sounds include: The images of shrimp activity are segmented to extract the outlines of individual shrimp, and the shrimp length and weight are estimated using a preset body shape template. Based on time series analysis of shrimp activity images, the frequency of mouthpart opening and closing and swimming trajectory of shrimp are identified. Combined with the dynamic changes in mouthpart opening and closing frequency and swimming trajectory, it is determined whether the shrimp group is in an active foraging state. Based on the energy intensity and frequency of the feeding sound wave signal, the intensity of the feeding sound is quantified and output as the feeding sound intensity index; The feeding sound intensity index is fused with the mouthpart opening and closing frequency and swimming activity, and output as a feeding activity index of shrimp groups.

4. The intelligent feeding method based on fish and shrimp farming according to claim 1, characterized in that: The step of predicting the feed demand of shrimp populations during future demand periods based on feeding activity indicators includes: Collect feeding activity indicators during the current feeding cycle and summarize them into a historical feeding characteristic sequence; Time-series analysis of feeding activity indicators under historical feeding characteristic sequences is performed to output the trend of feeding activity changes; Obtain the demand forecast point for the future demand period, as well as the time interval between the demand forecast point and the real-time feeding node, and record it as the forecast time interval; Extrapolate the trend of feeding activity changes based on the predicted time interval, and construct a feed demand prediction function by combining the dynamic influence weight of aquatic environment data. The feeding activity trend, prediction time interval, and aquatic environmental parameters are input into the feed demand prediction function, and the output is recorded as feed demand.

5. The intelligent feeding method based on fish and shrimp farming according to claim 1, characterized in that: The specific steps for outputting the dynamic influence weights of the water environment data are as follows: Acquire current water environment data and compare it with preset water environment adaptation thresholds to determine the real-time deviation of the water environment data; Based on the magnitude of the real-time deviation, basic influence coefficients are assigned to various environmental factors under the water environment data, and positive and negative values ​​are assigned to the basic influence coefficients according to the direction of deviation. The basic impact coefficients of each environmental factor, after being assigned values, are weighted and fused with the corresponding time-series change rates to obtain the dynamic impact weights of each environmental factor.

6. The intelligent feeding method based on fish and shrimp farming according to claim 1, characterized in that: The step of correcting the feed requirement based on the amount of feed remaining at the bottom of the pond to obtain the corrected actual feed requirement includes: The distribution density of uneaten residual feed particles is identified by capturing the feed settling area at the bottom of the pool using a wide-angle camera. The total amount of residual feed at the bottom of the pond is calculated based on the distribution density of residual feed, and the time interval from the last feeding to the current time is obtained. The degradation amount of residual feed at the bottom of the pond is calculated in combination with the feed dissolution rate. Subtract the degradation amount from the total amount of feed remaining at the bottom of the pond to obtain the current effective residual amount. Then subtract the current effective residual amount from the feed demand. If the result is greater than zero, use it as the corrected actual feed demand; otherwise, set it to zero.

7. The intelligent feeding method based on fish and shrimp farming according to claim 1, characterized in that: The step of determining the feed settling velocity and dispersion uniformity by combining water temperature fluctuations and turbidity rise slope when the dissolved oxygen change rate is higher than a preset dissolved oxygen rate threshold includes: The water temperature fluctuation sequence and turbidity change sequence during the current feeding period are collected using a sliding window mechanism. The water temperature fluctuation amplitude is calculated based on the water temperature fluctuation sequence, and the turbidity rise slope is calculated by combining the turbidity change sequence. Obtain the mapping relationship between feed settling velocity and water temperature fluctuation amplitude, and match the initial settling velocity based on the current water temperature fluctuation amplitude; The initial settling velocity was corrected based on the turbidity rise slope to obtain the corrected feed settling velocity. The suspension time of the feed particles in the water was then calculated by combining the water flow velocity in the pool with the distribution of the feeding points. The suspension time and the corrected feed settling velocity were coupled for analysis to evaluate the uniformity of feed dispersion in the water. If the uniformity of dispersion was lower than the preset standard, the spraying angle and frequency of the feeder were dynamically adjusted to optimize the spatial distribution of feed.

8. The intelligent feeding method based on fish and shrimp farming according to claim 1, characterized in that: The step of dynamically optimizing the feeding parameters based on the feed settling velocity and dispersion uniformity, and adjusting the feeding speed and distribution of feeding points, includes: Based on the corrected feed settling velocity, the historical optimal feeding velocity under the same settling velocity conditions in the historical feeding data is matched. The historical optimal feeding velocity is the feeding velocity that maximizes the feeding rate of the farmed organisms while ensuring that the feed is fully dispersed and without excessive deposition. The baseline value of the current feeding speed is determined based on the historical best feeding speed, and dynamically corrected by combining the real-time dispersion uniformity evaluation results. When the dispersion uniformity is lower than the preset threshold, the feeding speed is reduced to prolong the feed diffusion time, and the spatial distribution density of the feeding points is optimized according to the cross-sectional area of ​​the pool and the direction of water flow. When the uniformity of dispersion is higher than the preset threshold, the feeding speed is increased to improve feeding efficiency, and the distribution angle of the feeding point is adjusted according to the water flow speed and the topography of the pond bottom to ensure that the feed covers the densely active areas of aquatic organisms.

9. An intelligent feeding system for fish and shrimp farming, characterized in that: The intelligent feeding method based on fish and shrimp farming according to any one of claims 1 to 8 includes: The data acquisition module is used to collect aquatic environmental parameters, images of shrimp activity, and the intensity of feeding sounds in real time. The aquatic environmental parameters include water temperature, dissolved oxygen, pH value, and turbidity. The feeding assessment module is used to identify the size and feeding behavior of shrimp based on images of shrimp activity, and to calculate the feeding activity index of the shrimp population by combining the intensity of feeding sounds. The feeding assessment module is used to predict the feed demand of shrimp groups during future demand periods based on feeding activity indicators, and to correct the feed demand based on the amount of feed remaining at the bottom of the pond to obtain the corrected actual feed demand. The segmented feeding module is used to collect the dissolved oxygen change rate during the feed feeding process in real time. When the dissolved oxygen change rate is lower than or equal to the preset dissolved oxygen rate threshold, the segmented feeding mechanism is triggered. When the dissolved oxygen change rate is higher than the preset dissolved oxygen rate threshold, the settling speed and dispersion uniformity of the feed are determined by combining water temperature fluctuations and turbidity rise slope. The feeding optimization module is used to dynamically optimize the feeding parameters based on the feed settling speed and dispersion uniformity, and adjust the feeding speed and distribution of feeding points until the feed is fed in the current feeding cycle.

10. An intelligent feeding device for fish and shrimp farming, characterized in that: Includes a feeding pool (1), a guide workbench (2) is provided in the middle of the feeding pool (1), and a feeder (4) is installed on the guide workbench (2) to move along the slide rail (3). The feeder (4) integrates a processor connected to a wide-angle camera and a sensor array; and a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which is then executed by the at least one processor to enable the at least one processor to perform the intelligent feeding method based on fish and shrimp farming as described in any one of claims 1 to 8.

Images