Intelligent feeding control method and system for aquaculture

Abstract

The invention relates to the technical field of intelligent fishery, in particular to an intelligent feeding control method and system for aquaculture. The method comprises the following steps: collecting original monitoring data, carrying out standardized preprocessing, and carrying out feature extraction on the preprocessed data to obtain multi-modal behavior features; on the basis of the multi-modal behavior characteristics and in combination with the pre-processed data, calculating the ingestion willingness index of the fish school, and judging whether a feeding process is started or not; when feeding is started, the feeding amount is calculated based on the fish school feeding willingness index, and feeding is conducted in combination with the feeding interval; after feeding, based on the multi-modal behavior characteristics, obtaining ingestion response data, and calculating a comprehensive feedback score; based on the comprehensive feedback score, the feeding amount and the feeding interval are updated in a self-adaptive mode, and closed-loop control is achieved. The problems that in aquaculture feeding, food intake state recognition is not accurate, water quality change response is lagged, a closed-loop feedback mechanism is lacked, artificial experience is relied on, and data credibility and tracing are insufficient are solved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent fishery technology, and in particular to an intelligent feeding control method and system for aquaculture. Background Art

[0002] Existing feeding technologies in the aquaculture sector have evolved from simple manual operations to semi-automated modes such as timing and quantitative feeding, and have gradually introduced auxiliary means such as video monitoring and water quality testing, achieving basic perception of fish activity status and water environment. Some systems combine simple control algorithms to automatically perform feeding tasks according to preset parameters, improving feeding efficiency and operational convenience. In terms of water quality management, existing technologies can obtain key parameters such as dissolved oxygen, temperature, and pH in real time through sensors, and use them as a reference for feeding adjustments. At the same time, some aquaculture management platforms have begun to try to digitize and platformize feeding records, realizing centralized storage and historical query of feeding data, providing basic data support for scientific aquaculture.

[0003] However, the above-mentioned aquaculture feeding systems generally have problems such as single data, centralized control, opaque algorithms, and insufficient scheduling optimization, making it difficult to meet the needs of modern aquaculture that requires intelligence, scalability, and traceability. Summary of the Invention

[0004] The present invention provides an intelligent feeding control method and system for aquaculture to solve technical problems such as inaccurate feeding status identification, delayed response to water quality changes, lack of closed-loop feedback mechanism, reliance on manual experience, and insufficient data credibility and traceability in aquaculture feeding.

[0005] The present invention provides an intelligent feeding control method and system for aquaculture, which specifically includes the following technical solutions: An intelligent feeding control method for aquaculture, comprising the following steps: S1. Collecting raw monitoring data and performing standardized preprocessing to obtain preprocessed data; extracting features from the preprocessed data to obtain multimodal behavioral features; and calculating the fish feeding willingness index based on the multimodal behavioral features and the preprocessed data. S2. Determine whether to start the feeding process based on the fish feeding willingness index. When feeding is started, calculate the feeding amount based on the fish feeding willingness index and feed the fish based on the feeding interval. After feeding, obtain feeding response data based on multimodal behavioral characteristics and calculate a comprehensive feedback score. Based on the comprehensive feedback score, adaptively update the feeding amount and feeding interval to achieve closed-loop control.

[0006] Preferably, the S1 specifically includes: The feeding willingness index of fish schools was calculated by taking a weighted sum of multimodal behavioral characteristics, combining the dissolved oxygen concentration in the preprocessed data, and introducing a group movement dynamic amplification factor.

[0007] Preferably, the S1 specifically includes: By performing optical flow analysis on the multimodal behavioral characteristics, the rate of change of the kinetic energy of the fish school in unit time is calculated, and the dynamic amplification coefficient is introduced to obtain the dynamic amplification factor of the group movement.

[0008] Preferably, the S2 specifically includes: The fish feeding willingness index is compared with the preset feeding willingness threshold. When the fish feeding willingness index is greater than the feeding willingness threshold, the feeding process is started. Combined with the exponential decay function, the feeding interval is set, and a mapping relationship is established between the feeding interval and the fish feeding willingness index.

[0009] Preferably, the S2 specifically includes: Target detection and motion trajectory analysis are performed on the image data in the preprocessed data to obtain the fish gathering center and swimming direction, and formulate feeding strategies.

[0010] Preferably, the S2 specifically includes: Based on the fish feeding willingness index, combined with the feeding willingness threshold and the preset single-mouth maximum feeding amount, the feeding amount is calculated and feeding is carried out in combination with the feeding strategy.

[0011] Preferably, the S2 specifically includes: After feeding is executed, normalization and weighted fusion calculations are performed based on the feeding response data and combined with the stability of fish aggregation to obtain a comprehensive feedback score. Based on the comprehensive feedback score, the feeding amount and feeding interval are adaptively updated, and the feeding willingness threshold is optimized to achieve adaptive closed-loop control. The fish aggregation stability is calculated based on the image data in the preprocessed data.

[0012] An intelligent feeding control system for aquaculture, comprising the following parts: Multimodal perception unit module, feeding intention monitoring module, intelligent feeding execution module and feedback optimization module; The multimodal sensing unit module collects raw monitoring data and performs standardized preprocessing to obtain preprocessed data; performs feature extraction on the preprocessed data to obtain multimodal behavioral features, and stores them in the feedback optimization module; the multimodal sensing unit module is connected to the feeding intention monitoring module, the intelligent feeding execution module, and the feedback optimization module; The feeding willingness monitoring module calculates the fish feeding willingness index based on multimodal behavioral characteristics and the dissolved oxygen concentration in the pre-processed data, and stores it in the feedback optimization module. When the fish feeding willingness index is updated, the intelligent feeding execution module is automatically notified; The intelligent feeding execution module compares the fish feeding willingness index with a preset feeding willingness threshold. When the fish feeding willingness index is greater than the feeding willingness threshold, the feeding process is initiated and the feeding interval is set in combination with an exponential decay function. Target detection and motion trajectory analysis are performed on the image data in the preprocessed data, and a feeding strategy is formulated. Based on the fish feeding willingness index and the feeding willingness threshold, the feeding amount is calculated and feeding is carried out in accordance with the feeding strategy. The feeding willingness threshold, feeding interval, and feeding amount are stored in the feedback optimization module. The feedback optimization module obtains feeding response data based on the multimodal behavioral characteristics after feeding, and calculates a comprehensive feedback score based on the stability of fish aggregation. Based on the comprehensive feedback score, it adaptively adjusts the feeding amount and feeding interval, and optimizes the feeding willingness threshold to achieve adaptive closed-loop control.

[0013] The beneficial effects of the technical solution of the present invention are: 1. Multimodal precision monitoring: Through the multimodal fusion of underwater cameras, water quality sensors, sound detection modules and auxiliary sensing equipment, fish behavior and environmental parameters are comprehensively collected. After feature extraction and normalization processing, behavioral characteristics of comparable scales are formed, which significantly improves the accuracy and real-time performance of the fish feeding willingness index calculation.

[0014] 2. Intelligent triggering and dynamic control: When the fish feeding willingness index exceeds the feeding willingness threshold, the feeding process is automatically triggered. The target detection and trajectory analysis algorithm is used to determine the gathering center and swimming direction of the fish school, so as to formulate a feeding strategy and select a single feeding port or multiple points for dispersed feeding, thereby reducing fish competition and feed waste.

[0015] 3. Feedback-driven continuous optimization: After feeding, a comprehensive feedback score is calculated based on fish aggregation stability and feeding response data. Based on this comprehensive feedback score, the feeding amount and feeding interval are optimized and updated through machine learning. The feeding willingness threshold is iteratively optimized through the policy gradient method combined with multiple feeding backtesting, thus achieving the unity of immediate adjustment and long-term optimization.

[0016] 4. Closed-loop intelligent control: The present invention forms a closed-loop control mechanism of "monitoring-execution-feedback-re-monitoring", which can automatically adjust feeding parameters when feeding status changes or environmental conditions fluctuate, thereby improving the stability, robustness and adaptability of the intelligent feeding control system.

[0017] 5. Reduce costs and improve efficiency: Through precise feeding, the waste of leftover bait is reduced, the water quality environment is improved, the feed cost is reduced, and at the same time, the uniform feeding and healthy growth of fish are promoted, thereby improving the breeding efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a structural diagram of an intelligent feeding control system for aquaculture according to the present invention; Figure 2 This is a flow chart of the intelligent feeding control method for aquaculture described in the present invention. DETAILED DESCRIPTION

[0019] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0020] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0021] The intelligent feeding control method and system for aquaculture provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0022] Refer to the attached Figure 1 , which shows a structural diagram of an aquaculture intelligent feeding control system provided by one embodiment of the present invention. The system includes the following parts: Multimodal perception unit module, feeding intention monitoring module, intelligent feeding execution module and feedback optimization module; The multimodal sensing unit module automatically collects raw monitoring data through multimodal sensing equipment deployed in the aquaculture water body, including underwater cameras, water quality sensor groups, sound detection modules, etc., and performs standardized preprocessing to obtain preprocessed data; the preprocessed data is subjected to feature extraction to obtain multimodal behavior features, which are stored in the feedback optimization module; the multimodal sensing unit module is connected to the feeding intention monitoring module, the intelligent feeding execution module, and the feedback optimization module; The feeding willingness monitoring module calculates the fish feeding willingness index based on multimodal behavioral characteristics and the dissolved oxygen concentration in the pre-processed data, and stores it in the feedback optimization module. When the fish feeding willingness index is updated, the intelligent feeding execution module is automatically notified; The intelligent feeding execution module compares the fish feeding willingness index with a preset feeding willingness threshold. When the fish feeding willingness index is greater than the feeding willingness threshold, the feeding process is initiated and the feeding interval is set in combination with an exponential decay function. Target detection and motion trajectory analysis are performed on the image data in the preprocessed data, and a feeding strategy is formulated. Based on the fish feeding willingness index and the feeding willingness threshold, the feeding amount is calculated and feeding is carried out in accordance with the feeding strategy. The feeding willingness threshold, feeding interval, and feeding amount are stored in the feedback optimization module. The feedback optimization module acquires the multimodal behavioral characteristics after feeding in real time, obtains feeding response data (such as fish aggregation, feeding duration, residual bait ratio, etc.), and calculates the comprehensive feedback score based on the stability of fish aggregation; based on the comprehensive feedback score, the feeding amount and feeding interval are adaptively updated, and the feeding willingness threshold is optimized through back-verification using the policy gradient method; the updated results are written into the feedback optimization module and synchronized to the intelligent feeding execution module, thereby realizing the adaptive closed-loop control of monitoring-execution-feedback-re-monitoring.

[0023] Refer to the attached Figure 2 , which shows a flow chart of an intelligent feeding control method for aquaculture provided by one embodiment of the present invention, the method comprising the following steps: S1. Collecting raw monitoring data and performing standardized preprocessing to obtain preprocessed data; extracting features from the preprocessed data to obtain multimodal behavioral features; and calculating the fish feeding willingness index based on the multimodal behavioral features and the preprocessed data. Multimodal sensing equipment is deployed in the aquaculture water body to collect raw monitoring data. The multimodal sensing equipment includes: an underwater camera (responsible for collecting image data, including fish phase, fish density and behavior), a water quality sensor group (responsible for collecting environmental physical and chemical data, including dissolved oxygen (DO) concentration, temperature, pH, and, if necessary, conductivity and ammonia nitrogen concentration), a sound detection module (for capturing feeding sounds and group activity sounds), and other auxiliary sensors (buoy-type position sensors, current meters, and light sensors). The raw monitoring data is subjected to standardized preprocessing to obtain preprocessed data. The standardized preprocessing process includes cleaning, denoising, missing information completion, time synchronization, and normalization and dedimensionalization to ensure that data of different modalities are comparable at a unified scale. The methods used are technical means well known to those skilled in the art and are not detailed here. Use existing feature engineering technology to extract features from pre-processed data and output multimodal behavioral features; Furthermore, based on the multimodal behavioral characteristics, combined with the dissolved oxygen concentration in the preprocessed data and introducing the dynamic amplification factor of group movement, the fish feeding willingness index was calculated to comprehensively represent the feeding needs of the fish in the short term. The linear weighting of the multimodal behavioral characteristics was combined with the environmental sensitivity correction to define the specific formula of the fish feeding willingness index as follows: , in, Indicates the current time The output of the fish feeding willingness index is a continuous value between 0 and 1, which is used to indicate the strength of the fish feeding willingness. The larger the value of the fish feeding willingness index, the more likely the fish are to need to feed, thus guiding the execution delay and edge computing of the feeding strategy; The Sigmoid function is an S-shaped nonlinear function that smoothly controls the effect of dissolved oxygen concentration on fish feeding willingness by mapping the input value to the range of 0–1. is the dissolved oxygen sensitivity coefficient, obtained through experiments, which is a positive value and is usually between 0.1 and 10. The larger the value, the more sensitive the fish's feeding willingness is to changes in dissolved oxygen; It is behavioral characteristics; It is The feature weight of each behavioral feature is set according to expert experience and the value range is [-1,1]; is the total number of behavioral features; is the weighted sum of multimodal behavioral features; It is the current moment The dissolved oxygen concentration after pretreatment was taken from the preprocessed data; The dissolved oxygen concentration threshold is set according to the expert experience method to characterize the critical point between fish feeding activity and inhibition. Its value range is 3 to 7. When the dissolved oxygen concentration after pretreatment is Above the dissolved oxygen concentration threshold When the dissolved oxygen concentration after pretreatment is lower than the dissolved oxygen concentration threshold, the feeding willingness of fish is weakened. It is the current moment The rate of change of the kinetic energy of a school of fish per unit time is used to quantify the dynamic mutation of the group behavior. This is obtained by performing optical flow analysis on the multimodal behavior characteristics, calculating the regional mean of the square of the optical flow velocity amplitude, and performing time difference. This is a technical means well known to those skilled in the art and will not be described in detail here. is the dynamic amplification factor, by comparing The actual feeding intensity is fitted using a regression method to amplify (or attenuate) the changes in feeding willingness caused by sudden changes in group dynamics, and the regression fitting results are limited to 0–2 using a Sigmoid scaling function. , then the desire to eat is amplified when exercise intensifies. The smaller the value, the weaker the effect of the movement change; the actual feeding intensity is obtained by the existing image recognition and acoustic detection methods, which is used to characterize the feeding activity of the fish during the feeding process, and will not be described in detail here; the regression method is a technical means well known to those skilled in the art, and will not be described in detail here; the first item , introducing the dissolved oxygen concentration after pretreatment The difference from the dissolved oxygen concentration threshold is the dissolved oxygen concentration after pretreatment. When the index is low, the feeding willingness index of fish is suppressed; the second is the weighted sum of multimodal behavioral features, representing the comprehensive feature results extracted from direct monitoring data such as vision, acoustics, and fish density; the third It is the dynamic amplification factor of group movement, emphasizing the enhancement or weakening of the fish feeding hospital index when the group changes suddenly (such as sudden accelerated aggregation). It is used to make a secondary correction based on the weighted sum of behavioral characteristics, reflecting the real-time effect of group dynamics.

[0024] S2. Determine whether to start the feeding process based on the fish feeding willingness index. When feeding is started, calculate the feeding amount based on the fish feeding willingness index and feed the fish based on the feeding interval. After feeding, obtain feeding response data based on multimodal behavioral characteristics and calculate a comprehensive feedback score. Based on the comprehensive feedback score, adaptively update the feeding amount and feeding interval to achieve closed-loop control.

[0025] The fish feeding willingness index is compared with the feeding willingness threshold to determine whether the trigger condition is met: when the fish feeding willingness index is greater than the feeding willingness threshold (i.e., the trigger condition is met), the aquaculture intelligent feeding control system starts the feeding process; when the fish feeding willingness index is less than or equal to the feeding willingness threshold, the feeding process is stopped; the feeding willingness threshold According to the expert experience method, the value range is 0.3 to 0.7; Set the feeding interval when the trigger conditions are met , combined with the exponential decay function, the feeding interval Fish feeding willingness index Establish a mapping relationship so that when the fish feeding willingness index When the feeding interval increases According to the exponential law, when the fish feeding willingness index When the feeding interval is reduced The extension is in an exponential law, so as to realize dynamic setting; Further, the fish aggregation center and swimming direction are obtained by processing and analyzing the preprocessed image data through a target detection and motion trajectory analysis algorithm, and a feeding strategy is formulated to determine whether to feed at a single port or multiple ports: if the fish are highly concentrated near a certain feeding port and swim in the same direction, single-port feeding is adopted; if the fish are dispersed or have high density in multiple regions (i.e., the spatial distribution of the fish is multi-centered or covers a large area), multiple-port feeding is adopted; Further, the feeding amount is calculated based on the fish feeding willingness index, combined with the feeding willingness threshold and the maximum single-port feeding amount for a single time; when feeding at a single port, if the feeding amount is greater than the maximum single-port feeding amount for a single time, the excess part is divided into batches and fed at a feeding interval; when feeding at multiple ports, the fish density in the coverage area of each feeding port, the distance between each feeding port and the fish aggregation center, and the residual feed are used to calculate the weight by the product method, and then the feeding amount is distributed to each feeding port in proportion to the weight; the preprocessed image data is obtained from the preprocessed data, and the target detection and motion trajectory analysis algorithm and the product method are well-known technical means to those skilled in the art, which will not be described here; the residual feed condition is obtained by identifying the number or proportion of un-eaten feed particles from the preprocessed image data through a target detection and segmentation algorithm, and can also be corrected by acoustic detection or water quality parameter changes, which will not be described here; The specific calculation formula of the feeding amount is: , wherein, is the feeding amount calculated in the current period, representing the final feeding output result of the intelligent aquaculture feeding control system; is the maximum single-port feeding amount for a single time, which is the upper limit value of feeding specified by the equipment or the breeding strategy, representing the maximum feeding amount that cannot be exceeded in the most ideal feeding state for a single-port feeding for a single time; represents the sampling time index of the intelligent aquaculture feeding control system, used to identify the calculation period of the fish feeding willingness index and the corresponding feeding amount; is the fish feeding willingness index at time ; is the feeding willingness threshold, with a value range of 0.5-0.9, which is the minimum willingness requirement for starting feeding set according to expert experience; is the current detected residual feed amount, i.e., the amount of feed remaining after the last or previous feeding; is the maximum allowed residual feed amount, which is set to 5%-20% of the maximum single-port feeding amount for a single time according to the expert experience method, and if the residual feed amount exceeds the maximum allowed residual feed amount, it is considered that the feeding is excessive, and the subsequent feeding needs to be reduced or stopped; represents the intensity ratio of the fish feeding willingness index at the current time to the feeding willingness threshold. is the residual bait correction factor. If the residual bait amount is close to the maximum allowed residual bait amount (i.e. ≈ ), the residual bait correction factor approaches 0, indicating that no more feeding is needed. If the residual bait amount is small ( ≈0), the residual bait correction factor is close to 1, indicating that normal feeding is possible; After feeding, based on the multimodal behavioral characteristics, the feeding response data (including feeding duration, residual bait ratio and feeding efficiency, etc.) and the stability of fish aggregation are obtained, and normalized and weighted fusion calculations are performed to obtain a comprehensive feedback score. , whose value range is 0 to 1. The larger the value of the comprehensive feedback score, the better the feeding effect. The fish aggregation stability is calculated by target detection and trajectory analysis of the pre-processed image data, for example, by the spatial distribution variance of the fish school, the consistency of the speed direction, or the amplitude of the center of gravity fluctuation of the group. These are technical means well known to those skilled in the art and will not be described in detail here. Based on feeding amount and comprehensive feedback rating , using existing machine learning optimization methods to optimize feeding amount and feeding interval Adaptively update and, based on multiple feeding results, back-verify using the policy gradient method to optimize the feeding willingness threshold ; The updated results are written into the feedback optimization module and synchronized to the intelligent feeding execution module, thereby realizing the adaptive closed-loop control of monitoring-execution-feedback-re-monitoring; the machine learning optimization method and the policy gradient method are both technical means well known to those skilled in the art and will not be elaborated here.

[0026] In summary, an intelligent feeding control method and system for aquaculture is completed.

[0027] The order in which the embodiments of the invention are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0028] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0029] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. An intelligent feeding control method for aquaculture, characterized in that: The following steps are involved: S1. Collect original monitoring data and perform standardized preprocessing to obtain preprocessed data; Perform feature extraction on the preprocessed data to obtain multimodal behavior features; Based on multimodal behavioral characteristics and combined with preprocessed data, the fish feeding willingness index is calculated; S2. Determine whether to start the feeding process based on the fish feeding willingness index; When feeding is started, the feeding amount is calculated based on the fish feeding willingness index and the feeding interval is combined with feeding. After feeding, the feeding response data is obtained based on the multimodal behavioral characteristics and the comprehensive feedback score is calculated. Based on the comprehensive feedback score, the feeding amount and feeding interval are adaptively updated to achieve closed-loop control.

2. The intelligent feeding control method for aquaculture according to claim 1, characterized in that: Said S1 specifically includes: The feeding willingness index of fish schools was calculated by taking a weighted sum of multimodal behavioral characteristics, combining the dissolved oxygen concentration in the preprocessed data, and introducing a group movement dynamic amplification factor.

3. The intelligent feeding control method for aquaculture according to claim 2, characterized in that: Said S1 specifically includes: By performing optical flow analysis on the multimodal behavioral characteristics, the rate of change of the kinetic energy of the fish school in unit time is calculated, and the dynamic amplification coefficient is introduced to obtain the dynamic amplification factor of the group movement.

4. The intelligent feeding control method for aquaculture according to claim 1, characterized in that: Said S2 specifically includes: The fish feeding willingness index is compared with the preset feeding willingness threshold. When the fish feeding willingness index is greater than the feeding willingness threshold, the feeding process is started. Combined with the exponential decay function, the feeding interval is set, and a mapping relationship is established between the feeding interval and the fish feeding willingness index.

5. The intelligent feeding control method for aquaculture according to claim 4, characterized in that: Said S2 specifically includes: Target detection and motion trajectory analysis are performed on the image data in the preprocessed data to obtain the fish gathering center and swimming direction, and formulate feeding strategies.

6. The intelligent feeding control method for aquaculture according to claim 5, characterized in that: Said S2 specifically includes: Based on the fish feeding willingness index, combined with the feeding willingness threshold and the preset single-mouth maximum feeding amount, the feeding amount is calculated and feeding is carried out in combination with the feeding strategy.

7. The intelligent feeding control method for aquaculture according to claim 6, characterized in that: Said S2 specifically includes: After feeding is executed, normalization and weighted fusion calculations are performed based on the feeding response data and combined with the stability of fish aggregation to obtain a comprehensive feedback score. Based on the comprehensive feedback score, the feeding amount and feeding interval are adaptively updated, and the feeding willingness threshold is optimized to achieve adaptive closed-loop control. The fish aggregation stability is calculated based on the image data in the preprocessed data.

8. An intelligent feeding control system for aquaculture, applied to the intelligent feeding control method for aquaculture according to claim 1, characterized in that: Includes the following sections: Multimodal perception unit module, feeding intention monitoring module, intelligent feeding execution module and feedback optimization module; The multimodal sensing unit module collects raw monitoring data and performs standardized preprocessing to obtain preprocessed data; Perform feature extraction on the preprocessed data to obtain multimodal behavior features and store them in the feedback optimization module; The multimodal sensing unit module is connected to the feeding intention monitoring module, the intelligent feeding execution module and the feedback optimization module; The feeding willingness monitoring module calculates the fish feeding willingness index based on multimodal behavioral characteristics and the dissolved oxygen concentration in the pre-processed data, and stores it in the feedback optimization module. When the fish feeding willingness index is updated, the intelligent feeding execution module is automatically notified; The intelligent feeding execution module compares the fish feeding willingness index with the preset feeding willingness threshold. When the fish feeding willingness index is greater than the feeding willingness threshold, the feeding process is started and the feeding interval is set in combination with the exponential decay function. Perform target detection and motion trajectory analysis on the image data in the preprocessed data, and formulate a feeding strategy; calculate the feeding amount based on the fish feeding willingness index and the feeding willingness threshold, and feed the fish according to the feeding strategy; store the feeding willingness threshold, feeding interval and feeding amount in the feedback optimization module; The feedback optimization module obtains feeding response data based on the multimodal behavioral characteristics after feeding, and calculates the comprehensive feedback score based on the aggregation stability of the fish; Based on the comprehensive feedback score, the feeding amount and feeding interval are adaptively adjusted, and the feeding willingness threshold is optimized to achieve adaptive closed-loop control.

Images