Automatic grading catching method and system for perches based on living body dynamic perception

Abstract

The invention relates to the technical field of intelligent automatic grading, in particular to an automatic grading fishing method and system for perches based on living body dynamic perception. The method comprises the following steps: acquiring multi-source perch state sensing data, and based on the multi-source perch state sensing data, performing cross-modal fusion analysis of living body behaviors and physiological appearances according to preset body shape reference parameters to obtain a body shape-stress fusion evaluation information set; based on this, analyzing the matching relationship between the perch body type state, the stress state and the grading fishing target pool, and generating a perch differentiation grading guide strategy set; on the basis, dynamic flow direction regulation and control are carried out through a bubble curtain, the stress perch individuals are drained to the redundant pool, after the stress states of the stress perch individuals are relieved through anti-stress treatment in the redundant pool, secondary flow guide distribution is carried out on the stress perch individuals according to body type evaluation, and a perch automatic grading catching log is output. And the automation level and the sorting accuracy of aquaculture grading fishing are improved.

Description

Technical Field

[0001] This application relates to the field of intelligent automated grading technology, and in particular to an automated grading and harvesting method and system for bass based on live dynamic perception. Background Technology

[0002] Current methods for grading and harvesting sea bass mainly rely on mechanical screening based on body length or static image recognition technology. These methods typically classify sea bass by body size through screen aperture or two-dimensional visual measurement, but they do not consider the dynamic behavior and appearance changes of sea bass during the harvesting process due to their high oxygen consumption and stress-prone physiological characteristics. Existing technologies lack multimodal real-time perception and fusion analysis of the live state of sea bass (such as fin movement, gill cover rhythm, body color, etc.), making it impossible to identify stressed individuals. This results in stressed sea bass being directly sorted into target ponds, exacerbating the risk of injury or even death.

[0003] Current methods mostly follow fixed patterns in sorting paths and strategies, making it impossible to dynamically guide and regulate based on the real-time stress state and body size differences of sea bass. In particular, it is difficult to achieve the refined operation of "first stress isolation and recovery, then grading according to size". This leads to the superposition of stress in sea bass during the grading process, reduced sorting accuracy and survival rate, and affects the farming efficiency and quality of sea bass. Summary of the Invention

[0004] This application provides an automated grading and harvesting method and system for bass based on live dynamic sensing to solve the above-mentioned problems.

[0005] In a first aspect, this application provides an automated grading and harvesting method for bass based on live dynamic sensing, the method comprising:

[0006] Acquire multi-source bass state perception data. Based on this data and preset body size baseline parameters, perform cross-modal fusion analysis of live behavior and physiological appearance to obtain a body size-stress fusion assessment information set. Based on this body size-stress fusion assessment information set, analyze the matching relationship between bass body size status, stress status, and graded harvesting target pools to generate a differentiated bass grading guidance strategy set. Based on this differentiated bass grading guidance strategy set, implement dynamic flow direction control through a bubble curtain to divert stressed bass individuals to a redundant pool. After stress relief treatment in the redundant pool, the stressed bass individuals are reassigned based on body size assessment, and an automated bass grading harvesting log is output.

[0007] The above technical solutions employ non-contact methods such as dynamic flow control via bubble curtains to gently implement strategies. Redundant pools are set up to provide specialized care and secondary diversion for stressed individuals, minimizing mechanical damage and stress responses caused by existing fishing methods and ensuring fish health and product quality. The fully automated recording of fishing logs provides robust data support for standardization, traceability, and continuous optimization of the production process. This effectively improves the automation level, sorting accuracy, and animal welfare of graded fishing in aquaculture, and is of great significance for promoting the development of smart fisheries.

[0008] Optionally, the multi-source bass state perception data includes fin dynamic frequency, gill cover movement rhythm, and body surface color change information; based on the fin dynamic frequency and combined with the gill cover movement rhythm, the activity level of the live bass behavior is analyzed using behavioral rhythm analysis technology to obtain behavioral stress characterization information; based on the body surface color change information, the reflective properties and regional color distribution of the bass body surface are analyzed using multi-region optical feature extraction technology to obtain appearance physiological state information; based on the behavioral stress characterization information and combined with the appearance physiological state information, a correlation coupling analysis between behavior and appearance is performed to determine the individual bass stress state and simultaneously assess the bass body size, resulting in the body size-stress fusion assessment information set that simultaneously includes stress state identification and body size classification results.

[0009] Optionally, based on the dynamic frequency of the fins, the oscillation differences between the hard spines and soft rays of the dorsal fin of the bass are analyzed using time-domain periodic signal analysis technology to obtain the rhythmic characteristics of fin movement; based on the gill cover movement rhythm, the physiological characteristics of the bass as a high-oxygen-consuming carnivorous fish are analyzed using rhythm and amplitude joint analysis technology to analyze the regular changes in the frequency and amplitude of gill cover opening and closing under static, swimming, and stress conditions to obtain the rhythmic characteristics of gill movement; based on the rhythmic characteristics of fin movement, combined with the rhythmic characteristics of gill movement, a comparative analysis of the movement rhythms of different body parts of the bass is performed to identify the synergistic relationship or abnormal disorder between fin oscillation and gill cover movement to obtain the behavioral stress characterization information.

[0010] Optionally, based on the surface color change information, and according to preset fishing light source information, the specular and diffuse reflection distribution areas of the sea bass's body surface under fishing reflected light are analyzed to obtain the reflective distribution characteristics of the sea bass's body surface; based on the reflective distribution characteristics, multi-region optical feature extraction technology is used to analyze the color depth and distribution uniformity of different areas of the sea bass's body surface to obtain the color distribution characteristics of the sea bass's body surface; based on the color distribution characteristics, a collaborative analysis of light signal pattern abnormalities and color distribution abnormalities is performed to determine the fading phenomenon, mottled phenomenon, or gloss disorder phenomenon on the sea bass's body surface caused by stress response, thereby obtaining the appearance physiological state information.

[0011] Optionally, based on the aforementioned abnormal imbalance relationship and combined with the aforementioned luster disorder phenomenon, a synergistic comparative analysis of behavioral rhythm disorder and optical signal disorder is performed to determine the stress state of the individual bass and obtain a preliminary stress state determination result; based on the preliminary stress state determination result, combined with the fin movement rhythm characteristics and the gill movement rhythm characteristics, the duration and trend of the bass's abnormal behavior are analyzed, and through stability assessment technology, transient physiological fluctuations and persistent stress responses are distinguished to obtain a stress state identifier verified by dynamic tracking; simultaneously, based on the color classification... The study analyzes the interference of the physical outline size and surface condition of the sea bass on optical assessment by examining the distribution characteristics. Through dual calibration of outline size and optical features, misjudgments of body shape caused by abnormal surface condition are eliminated, resulting in a calibrated body shape classification result. Based on the stress state identifier and the calibrated body shape classification result, the stress level is coupled with the specific body shape specification according to the correlation pattern between the behavioral rhythm of the fins and gills and the surface color and reflectivity of the sea bass under stress, forming a one-to-one corresponding individualized state profile, thus obtaining the body shape-stress fusion assessment information set.

[0012] Optionally, based on the stress state identifier and the body size classification result, according to the preset specifications and functional definitions of the graded fishing target pools, the target pool category that each bass individual conforms to in terms of body size and the suitability of stress state for entering the corresponding target pool are analyzed to obtain individual-level pool matching suitability information; based on the pool matching suitability information, bass are divided into two groups, stressed individuals and non-stressed individuals, according to the stress state identifier, and for each group, combined with the body size classification result, the final target pool to be guided is analyzed to obtain a preliminary guidance decision; based on the preliminary guidance decision, an instruction unit containing a unique identifier, current judgment status, specified guidance path and target pool is generated for each individual, and the instruction units of all individuals are aggregated and conflict resolved to resolve guidance conflicts caused by path intersection or instantaneous resource competition, resulting in a coordinated and executable set of differentiated graded guidance strategies for bass.

[0013] Optionally, based on the stress state identifier and the functional definition of the graded fishing target pool, a stress-oriented pool risk quantification assessment is conducted to analyze the expected recovery period of the stressed bass individual when entering the redundant pool and the probability of increased stress caused by entering the graded fishing target pool, thereby obtaining pool risk information corresponding to the stress state; based on the body size classification results and the specification requirements of the graded fishing target pool, the body size parameters of each bass individual are analyzed to determine the corresponding preset specification range of the target pool, thereby obtaining pool conformity information corresponding to the body size specification; the pool risk information and the pool conformity information are combined to construct the pool matching suitability information at the individual level.

[0014] Optionally, based on the pool matching suitability information, through temporal correlation and event backtracking analysis of the fishing process data, the entire process node information of each type of bass individual from initial perception, state determination, guidance execution to final pool allocation is analyzed to obtain the individual fishing trajectory chain; based on the individual fishing trajectory chain, according to the dynamic flow direction control process implemented by the bubble curtain, the linkage analysis of fishing operation and fish response status is performed, and the actual execution efficiency of the guidance strategy and the state recovery progress of the stressed bass individuals in the redundant pool are recorded to obtain the fishing process efficiency record; based on the individual fishing trajectory chain, combined with the fishing process efficiency record, the timestamp, individual identifier, state determination sequence, guidance action, pool change and key efficiency indicators are integrated and formatted for encapsulation to output the bass automated graded fishing log.

[0015] Optionally, based on the individual fishing trajectory chain, the temporal and spatial alignment technology of the regulatory action and fish school displacement is used to analyze the temporal changes of the bubble curtain operation and the correspondence between the displacement direction and velocity changes of individuals in the corresponding area of ​​the fish school, thereby obtaining bubble curtain regulation-fish school displacement response correlation information. Based on the bubble curtain regulation-fish school displacement response correlation information, combined with the stress state identifier, the behavioral differences between stressed and non-stressed bass individuals in terms of swimming path compliance, speed adjustment sensitivity, and group coordination when facing the same or different bubble curtain regulation parameters are analyzed through group response difference analysis, thereby obtaining regulation response characteristic information. Based on the regulation response characteristic information, the real-time adaptability analysis of the regulation strategy is used to analyze the deviation between the actual overall fish school flow pattern, individual diversion accuracy, and expected guidance path caused by different bubble curtain regulation actions triggered to execute the bass differentiated graded guidance strategy set. Based on the degree of deviation and the distribution of stressed bass individuals, the suitability of the bubble curtain parameter setting is dynamically retrospectively and quantitatively evaluated to obtain the linkage analysis results of fishing operation and fish school response state.

[0016] Secondly, this application provides an automated grading and harvesting system for bass based on live dynamic sensing, the system comprising:

[0017] The fusion assessment module is used to acquire multi-source bass state perception data. Based on the multi-source bass state perception data, and according to preset body size benchmark parameters, it performs cross-modal fusion analysis of live behavior and physiological appearance to obtain a body size-stress fusion assessment information set. The grading guidance module is used to analyze the matching relationship between bass body size status, stress status and grading target pools based on the body size-stress fusion assessment information set, and generate a differentiated grading guidance strategy set for bass. The grading fishing module is used to implement dynamic flow direction control through a bubble curtain based on the differentiated grading guidance strategy set for bass, to guide stressed bass individuals to a redundant pool, and after stress relief treatment in the redundant pool, to further allocate them according to body size assessment, and output an automated grading fishing log for bass. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram illustrating an application scenario provided in one embodiment of this application;

[0020] Figure 2 A flowchart of an automated graded harvesting method for bass based on live dynamic sensing provided in an embodiment of this application;

[0021] Figure 3 This is a schematic diagram of an automated grading and harvesting system for bass based on live dynamic sensing, provided as an embodiment of this application. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0023] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article, unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0024] The embodiments of this application will now be described in further detail with reference to the accompanying drawings.

[0025] Current methods for grading and harvesting sea bass mainly employ mechanical sieving based on body length or static image recognition, classifying them according to screen aperture or two-dimensional visual measurement. However, these methods do not fully consider the dynamic behavior and appearance changes exhibited by sea bass during the harvesting process due to their high oxygen consumption and stress-prone physiological characteristics. They also lack the ability to perceive and fuse the multimodal real-time state of live fish, resulting in the inability to effectively identify stressed individuals. Consequently, such sea bass are still sorted into the target pond, increasing the risk of injury and mortality.

[0026] Based on this, this application provides an automated grading and harvesting method and system for bass based on live dynamic perception. Through non-contact, gentle harvesting methods with dynamic control of bubble curtains, combined with the setting of redundant pools for special care and diversion of sensitive individuals, the mechanical damage and stress response caused by existing harvesting methods are significantly reduced, ensuring fish health and product quality. The entire process automatically generates harvesting logs, realizing standardized production, full traceability, and continuous optimization. This effectively improves the automation level, sorting accuracy, and animal welfare level of grading and harvesting, and has positive significance for the development of smart fisheries.

[0027] Figure 1 This application provides an illustration of an application scenario. In the current process of grading and harvesting sea bass, the method provided in this application significantly reduces fish injury and stress. Through fully automated recording of the entire process, it achieves efficient, accurate, and traceable grading and harvesting, thereby improving production efficiency and animal welfare, and powerfully promoting the development of smart fisheries.

[0028] Specifically, the method provided in this application can be applied to any server. The server interacts with the water quality sensor to obtain multi-source bass status perception data provided by the water quality sensor. The strategy is gently executed through non-contact means such as dynamic flow direction control via bubble curtain. Redundant pools are set up to provide specialized care and secondary diversion for stressed individuals. The automated grading and harvesting log of bass is output to aquaculture harvesters, providing solid data support for the standardization, traceability and continuous optimization of the production process.

[0029] For specific implementation details, please refer to the following examples.

[0030] Figure 2 This is a flowchart illustrating an embodiment of an automated grading and harvesting method for bass based on live dynamic sensing. The method of this embodiment can be applied to the server in the above scenario. Figure 2 As shown, the method includes:

[0031] S201. Obtain multi-source bass state perception data. Based on the multi-source bass state perception data and according to the preset body size benchmark parameters, perform cross-modal fusion analysis of live behavior and physiological appearance to obtain a body size-stress fusion assessment information set.

[0032] Multi-source bass status perception data can be a collection of various data reflecting bass body size, behavior, physiological characteristics, and environmental conditions, sourced from water quality sensors deployed within the aquaculture area. Preset body size baseline parameters can be standard parameters used to determine bass body size grades. Cross-modal fusion analysis of live behavior and physiological appearance involves cross-domain integration and correlation analysis of bass behavioral and physiological appearance data from different dimensions, using deep learning algorithms and feature association modeling methods. The body size-stress fusion assessment information set can be a collection of information including the body size grade determination result for each bass, a quantitative score of stress status, and related characteristic data.

[0033] Specifically, in large-scale, high-density aquaculture, existing grading and harvesting methods mainly rely on manual observation or simple mechanical screening. This method is inefficient, labor-intensive, and easily causes mechanical damage and strong stress reactions to fish during the harvesting process, leading to increased disease and mortality rates in subsequent farming and seriously affecting economic benefits. On the other hand, some existing automated technologies often focus only on a single indicator, such as analyzing body shape through images or judging environmental anomalies solely through water quality. They lack a comprehensive and real-time understanding of the complex state of living organisms, making it difficult to effectively distinguish and handle individuals in a state of stress while grading.

[0034] S202. Based on the body size-stress fusion assessment information set, analyze the matching relationship between the body size status and stress status of bass and the target pond for graded fishing, and generate a set of differentiated graded guidance strategies for bass.

[0035] The target ponds for graded harvesting can be dedicated aquaculture ponds or harvesting receiving ponds divided according to the size of the bass. The differentiated grading and guidance strategy set for bass can be a set of personalized guidance programs developed for bass of different sizes and under different stress conditions.

[0036] Specifically, even after obtaining the fusion assessment information set, effective execution is still impossible without a decision-making mechanism that can intelligently link the assessment results with specific operational actions. By constructing a key bridge from "state perception" to "action decision-making" and establishing a set of matching rules and optimization algorithms, the matching relationship between the state of each individual and the functional requirements of each target pool can be dynamically analyzed. For example, for individuals that meet the size requirements and have a "normal" stress assessment, the optimal strategy is to guide them into the final fishing channel; for individuals that meet the size requirements but have a "high" stress assessment, direct fishing and transportation would pose extremely high risks.

[0037] S203. Based on the differentiated grading and guidance strategy set for bass, dynamic flow direction control is implemented through a bubble curtain to guide stressed bass individuals to a redundant pool. After stress relief treatment is carried out in the redundant pool, the stressed bass individuals are reassigned based on body size assessment, and an automated grading and harvesting log for bass is output.

[0038] A bubble curtain can be a virtual barrier formed by releasing bubbles into the water through a bubble curtain generator. Dynamic flow direction control can be a method of regulating water flow by adjusting the intensity, range, and direction of bubble release in real time according to a graded guidance strategy, thereby achieving precise guidance of bass movement. Stressed bass individuals can be bass exhibiting stress responses (such as erratic swimming, surface congestion, and rapid breathing) due to environmental changes, fishing disturbances, etc. Redundant ponds can be dedicated ponds for temporarily storing and treating stressed bass. Anti-stress treatment can be a series of technical means to alleviate the stress response of bass. Secondary flow allocation can be a process of re-allocating flow based on body size assessment results for bass after the stress state has been relieved. An automated bass grading and harvesting log can be a document recording key data from the entire grading and harvesting process.

[0039] Specifically, strategies without effective and gentle implementation methods will not be effective. The strong contact nature of existing net fishing is the main source of stress. Providing a humanized physical implementation solution that matches the intelligent strategy is crucial. After the hierarchical guidance strategy set is generated, the bubble curtain generator that matches the guidance path specified in the strategy will be activated first. By precisely controlling the air pressure and opening and closing sequence of different sections of the pipeline, a dynamically moving or directional closed bubble curtain wall is generated in the water. This gently separates the bass individuals or groups identified as "high stress" from the main fish group and drives them into the redundant pool without contacting the fish.

[0040] The method provided in this embodiment gently implements strategies through non-contact means such as dynamic flow control via bubble curtains, and sets up redundant pools to provide specialized care and secondary diversion for stressed individuals. This minimizes the mechanical damage and stress response caused by existing fishing methods, ensuring fish health and product quality. The fully automated recording of the fishing log provides solid data support for the standardization, traceability, and continuous optimization of the production process, effectively improving the automation level, sorting accuracy, and animal welfare level of graded fishing in aquaculture, which is of great significance for promoting the development of smart fisheries.

[0041] In some embodiments, the multi-source bass state perception data includes fin dynamic frequency, gill cover movement rhythm, and body surface color change information. Based on the fin dynamic frequency and combined with the gill cover movement rhythm, the activity level of the live bass behavior is analyzed using behavioral rhythm analysis technology to obtain behavioral stress characterization information. Based on the body surface color change information, the reflective properties and regional color distribution of the bass body surface are analyzed using multi-region optical feature extraction technology to obtain appearance physiological state information. Based on the behavioral stress characterization information and combined with the appearance physiological state information, a correlation coupling analysis between behavior and appearance is performed to determine the stress state of the individual bass and simultaneously assess the bass body size, resulting in a body size-stress fusion assessment information set that simultaneously includes stress state identification and body size classification results.

[0042] Fin dynamic frequency refers to the number of swings and amplitude changes of the dorsal and pectoral fins of a bass within a unit of time. Gill cover movement rhythm refers to the frequency, amplitude, and periodic changes of the opening and closing of the bass's gill covers. Body surface color change information refers to the dynamic changes in color depth, distribution uniformity, and reflectivity in different areas of the bass's body surface. Behavioral rhythm analysis technology is a specialized technique used to uncover the intrinsic patterns, synergistic relationships, and abnormal characteristics of biological movement behavior. Multi-region optical feature extraction technology is a technique that, through optical signal filtering, feature point recognition, region segmentation, and information fusion, accurately extracts key optical information such as reflectivity distribution range, color gradient changes, and texture uniformity from different functional areas of the organism's body surface. Behavioral stress characterization information refers to a set of behavioral features that reflect whether a bass is in a stress state. Appearance and physiological state information refers to physiological state-related information obtained based on the extraction of optical features from the bass's body surface. Correlation coupling analysis is a method for in-depth correlation mining, collaborative verification, and complementary analysis of feature information from two or more dimensions.

[0043] Specifically, in automated grading and harvesting scenarios, grading based solely on a single dimension of information (such as a rough estimate of body length or weight) is insufficient to handle the complex dynamic changes during live harvesting. Sea bass, as a high-value aquaculture species, are highly susceptible to stress when disturbed during harvesting operations, exhibiting abnormal behavior (such as rapid, uncoordinated swimming) and physiological changes (such as temporary fading of body color and loss of luster). If individuals already under stress are directly sorted into target ponds based on body size, it will not only exacerbate their stress, leading to slow growth, weakened immunity, or even death, causing economic losses, but may also trigger group fights due to their panicked collisions in the pond. Stress disrupts the stability of the entire grading system. To address this issue, the following approach is employed: First, raw data is collected using underwater high-speed camera units and multispectral imaging units deployed in the grading channels. For fin dynamic frequencies, a deep learning-based attitude estimation model (such as HRNet) is used to continuously track key points of the dorsal and caudal fins of the bass in the video stream. By calculating the periodic changes in the displacement of these key points, the oscillation frequency and rhythmic waveform per unit time (e.g., per second) are extracted. For gill cover movement rhythm, semantic segmentation and motion amplification are performed on the gill cover region in the side-view image to accurately quantify its opening and closing frequency (e.g., number of openings per minute) and amplitude timing signals. These two timing signals are then input into the behavioral rhythm analysis module. Signal processing techniques (such as short-time Fourier transform to extract the dominant frequency, or cross-correlation analysis to determine the phase synchronization of fin and gill movements) are used to construct behavioral stress characterization information. For example, a typical mismatch pattern is identified: "high-frequency fin vibration (e.g., 10Hz) but sluggish gill cover movement (e.g., frequency drops to half of normal)." On the other hand, for information on changes in body color, the multispectral imaging unit acquires images under specific wavelength light sources (such as visible light and near-infrared light), and uses multi-region optical feature extraction technology to first divide the fish image into multiple regions of interest such as head, trunk, and tail. Then, it extracts the HSV color space statistical features (such as hue mean and variance), LBP texture features, and the intensity and distribution features of specular reflection spots in each region to obtain information on the appearance and physiological state, such as quantifying the abnormal phenomenon of "color saturation in a specific area on the side of the body decreasing by more than 30% and accompanied by irregular mottled texture".Finally, through correlation and coupling analysis, the quantitative feature vectors of the above-mentioned behaviors and appearances are fused and input into a pre-trained lightweight classification model (e.g., gradient boosting decision tree). This model, by learning from a large number of labeled samples, can establish a mapping relationship between "fin-gill rhythm disorder > threshold A" and "body lateral gloss uniformity < threshold B" and "high stress" state. Simultaneously, through a calibration sub-model, stable contour information in appearance features (e.g., near-infrared edges unaffected by color) is used to compensate for the body size measurement error caused by stress posture distortion (e.g., correcting the apparent body length by 3-5%). Finally, a body size-stress fusion evaluation information set is output, which simultaneously contains "stress state identifier (e.g., level II)" and "calibrated body size classification result (e.g., large size L)," completing the accurate perception and integrated diagnosis of the multi-dimensional state of live bass.

[0044] The method provided in this embodiment integrates cross-modal analysis of behavior and appearance, and simultaneously achieves accurate determination of the stress state of sea bass and calibration assessment of body size. This effectively solves the problem of misjudgment caused by neglecting stress response when grading based solely on body size. It can dynamically identify and separate highly stressed individuals, thereby significantly reducing the risk of subsequent mortality and growth inhibition caused by fishing operations. At the same time, it ensures the accuracy of grading results and improves the economic benefits and animal welfare of aquaculture production.

[0045] In some embodiments, based on the dynamic frequency of the fins, the oscillation differences between the hard spines and soft rays of the dorsal fin of the sea bass are analyzed using time-domain periodic signal analysis technology to obtain the rhythmic characteristics of fin movement; based on the rhythm of gill cover movement, the physiological characteristics of the sea bass as a high-oxygen-consuming carnivorous fish are analyzed using rhythm and amplitude joint analysis technology, and the regular changes in the frequency and amplitude of gill cover opening and closing under static, swimming, and stress conditions are analyzed to obtain the rhythmic characteristics of gill movement; based on the rhythmic characteristics of fin movement, combined with the rhythmic characteristics of gill movement, a comparative analysis of the movement rhythms of different body parts of the sea bass is conducted to identify the synergistic relationship or abnormal disorder between fin oscillation and gill cover movement, and to obtain behavioral stress characterization information.

[0046] Time-domain periodic signal analysis is a technique used to extract the periodic characteristics of signals over time. The dorsal fin spines refer to the hard, supportive, spiny parts of the bass's dorsal fin. The soft rays refer to the soft, flexible, ray-like parts of the bass's dorsal fin that move in tandem with the spines, exhibiting different oscillation patterns, which are important for distinguishing fin movement rhythms. Oscillation differences refer to the varying frequencies, amplitudes, and rhythms of the oscillations between the dorsal fin spines and soft rays. Fin movement rhythm characteristics can be obtained by analyzing the dynamic frequencies of the fins, reflecting characteristic information that reflects the regularity of fin movement. The gill cover movement rhythm refers to the temporal pattern of the opening and closing of the bass's gill covers. Combined rhythm and amplitude analysis is a technique that comprehensively analyzes both the rhythm and amplitude dimensions of movement simultaneously. High-oxygen-consuming carnivorous fish refers to the physiological attribute of the bass, characterized by a high metabolic rate and a strong demand for oxygen. Physiological characteristics specifically refer to the inherent metabolic and respiratory characteristics of the bass as a high-oxygen-consuming carnivorous fish. The frequency of gill opening and closing can be defined as the number of times the gill covers of a bass open and close per unit time. The amplitude can be defined as the maximum distance the gill covers can open and close. A static condition can be a state where the bass is not moving significantly and its breathing is stable. A swimming condition can be a state where the bass is actively swimming, at which point its metabolism and respiratory intensity will change, and the gill cover movement will adjust accordingly. A stress condition can be a state where the bass exhibits a stress response after being stimulated by the external environment. Regular changes can be defined as predictable and distinguishable trends in the frequency and amplitude of gill opening and closing under different conditions such as static, swimming, and stress. The rhythmic characteristics of gill movement can be obtained by analyzing the rhythm of gill cover movement, reflecting the regularity of gill respiratory movement. The rhythmic characteristics of different body parts can be defined as the individual temporal patterns of movement of the bass's fins and gill covers. Comparative analysis can be an analytical method that compares the characteristics of two or more objects to identify their similarities and differences. Synergistic relationships can refer to the consistent frequency and rhythm of fin movements and gill cover movements. An abnormal imbalance can be a mismatch in the frequency and rhythm of fin movements and gill cover movements.

[0047] Specifically, in the automated grading and harvesting of bass, accurately and early identification of an individual's stress state is a crucial prerequisite for the success of subsequent grading guidance. Stress responses lead to changes in bass behavior patterns, but these changes are often multi-site and coordinated abnormalities, rather than drastic fluctuations in a single indicator. Relying solely on a single, isolated behavioral indicator (such as observing only swimming speed or only fin frequency) for judgment is highly prone to misjudgment. To address this issue: First, for the dynamic frequency of the fins, time-domain periodic signal analysis technology is employed: a pre-trained convolutional neural network model (such as U-Net) is used to perform pixel-level segmentation of the bass dorsal fin region in consecutive frame images, extracting its contour centerline; subsequently, the time-series data of the angle change of this centerline in the image sequence is filtered and standardized. The system analyzes and estimates the power spectral density of the fins. For example, it finds that during normal cruising, the main frequency of the fin oscillation signal is stable (between 1.5 and 2.5 times per second), and the phase difference between the oscillations of the spines and soft rays remains constant. When the system detects an abnormal high-frequency component in the signal that is higher than 4.5 times per second, and this component is out of sync with the main propulsion rhythm, this feature (high-frequency tremor, rhythm instability) is recorded as part of the fin movement rhythm feature. Secondly, for the gill cover movement rhythm, a combined rhythm and amplitude analysis technique is applied: using optical flow combined with a region tracking algorithm, the movement trajectory of the posterior edge of the gill cover is accurately located and tracked to generate its opening and closing displacement-time curve. Peak detection is performed on the curve to calculate the duration (rhythm) and displacement peak (amplitude) of the continuous opening and closing cycle. For example, a baseline model is established where the gill cover opening and closing rhythm is stable during normal resting (e.g., a cycle of approximately once per second) with sufficient amplitude. Once a pattern is identified where "the rhythm significantly accelerates (e.g., the cycle shortens to more than twice per second) while the amplitude simultaneously shrinks (e.g., drops to less than 60% of the baseline amplitude)" across multiple consecutive cycles, this "high-frequency, low-amplitude" feature is marked as a key gill movement rhythm feature. Finally, the core comparative analysis is performed: the system synchronizes the two types of features on the time axis and quantifies the synchronicity of the fin rhythm signal and the gill rhythm signal within a specific time window by calculating indicators such as dynamic time warping distance or phase lock value. For example, within an observation window of a certain duration (e.g., 5 seconds), if the algorithm calculates that the rhythm synchronization index of the two is consistently below a set threshold (e.g., 0.3), it indicates that the fin oscillation and gill opening and closing have lost their normal synergistic relationship and exhibit an abnormal disordered relationship; conversely, if the synchronization index is consistently high, it is determined to be a synergistic relationship. The qualitative or quantitative determination of this relationship is encapsulated together with the aforementioned extracted fine-grained behavioral characteristics to generate the final behavioral stress representation information rich in deep-seated behavioral correlation information.

[0048] The method provided in this embodiment deeply analyzes and correlates the intrinsic rhythms of key locomotor organs in sea bass, achieving a leap from surface observation to mechanistic correlation of behavioral patterns. It can accurately isolate normal behavioral fluctuations caused by individual vitality differences and specifically capture deep-seated inter-system coordination disorder signals triggered by stress. This lays a solid and interference-resistant data foundation for building a highly reliable behavioral stress judgment model and significantly improves the accuracy of subsequent overall state assessment.

[0049] In some embodiments, based on information about changes in body color, and according to preset information about the fishing light source, the specular and diffuse reflection distribution areas of the sea bass's body surface under the reflected light are analyzed to obtain the reflective distribution characteristics of the sea bass's body surface. Based on the reflective distribution characteristics, multi-region optical feature extraction technology is used to analyze the color depth and distribution uniformity of different areas of the sea bass's body surface to obtain the color distribution characteristics of the sea bass's body surface. Based on the color distribution characteristics, a synergistic analysis of abnormal light signal patterns and abnormal color distribution is performed to determine the fading, mottled, or glossy disorder phenomena on the sea bass's body surface caused by stress response, thereby obtaining information on the physiological state of the appearance.

[0050] Preset fishing light source information can be artificially set fishing scene light source parameters to ensure the accuracy of body color detection. Specular reflection distribution area can be a strong reflection area produced by smooth parts of the sea bass's body surface (such as the intact scale area on the back) under fishing light illumination. Diffuse reflection distribution area can be a weak reflection area produced by rough or irregularly structured parts of the body surface (such as gaps between scales on the abdomen, fin edges). Reflection distribution characteristics can be a comprehensive representation of the location, range, and reflection intensity of specular and diffuse reflection areas on the sea bass's body surface. Multi-region optical feature extraction technology can be a technique that can separately analyze optical information of different divided areas of the sea bass's body surface. Color depth can be the intensity of color in a specific area of ​​the sea bass's body surface (such as dark bluish-gray in a normal state and light grayish-white after stress). Distribution uniformity can be the consistency of color on the entire body surface and in different areas (such as uniform color without difference under normal conditions, and local color differences after stress). Color distribution characteristics can be a comprehensive description of the distribution and uniformity of color depth on the sea bass's body surface in different areas. Abnormal light signal patterns can manifest as deviations in the intensity and wavelength distribution of light reflected from the sea bass's body surface from the standard pattern under normal physiological conditions. Abnormal color distribution can result in deviations from normal color depth and uniformity on the sea bass's body surface. Fading can be a significant lightening of the overall color of the sea bass's body surface compared to its normal state. Mottled appearance can be the appearance of irregular, randomly distributed patches of light and dark color on the sea bass's body surface. Disordered luster can be the appearance of unstable intensity and irregular distribution of reflected light on the sea bass's body surface.

[0051] Specifically, in automated grading and harvesting of bass, the physiological state of appearance is a key basis for determining their stress state. Relying solely on behavioral data is prone to misjudgment due to short-term fluctuations. However, changes in body surface coloration and mottled appearance caused by stress can accurately reflect the physiological state, compensating for the limitations of behavioral analysis. Furthermore, appearance information is an essential dimension for constructing a body shape-stress fusion assessment set; its absence leads to a one-sided assessment and affects the accuracy of grading. Simultaneously, it provides a basis for subsequent anti-stress treatment and secondary diversion, making it a core and indispensable element for achieving automated and precise harvesting. To address these issues: Based on pre-designed preset harvesting light source information (e.g., using an LED array with a color temperature of 6000K and an illuminance of 300 lux to simulate a specific natural light spectrum), lights are deployed above and to the sides of the harvesting area to provide stable and consistent basic illumination for the entire analysis. Under these conditions, a high-frame-rate underwater industrial camera is used to synchronously acquire video streams of bass swimming, from which high-quality image sequences of body surface color changes are extracted. For a single frame image, a deep learning-based instance segmentation model (such as Mask R-CNN) is first used to accurately separate the bass target, eliminating interference from the water background. Then, computer vision algorithms are used to analyze its specular and diffuse reflection distribution areas: by calculating the image's brightness gradient and local contrast, and based on a preset brightness threshold (for example, initially identifying the top 5% of pixel brightness values ​​as specular highlights), bright specular reflection spots and relatively uniform diffuse reflection areas are segmented. Then, parameters such as the area ratio and spatial clustering of these areas are statistically calculated to quantify the reflective distribution characteristics. Next, multi-region optical feature extraction technology was applied, specifically using a localization method based on key points of the sea bass body (such as the snout, the origin of the dorsal fin, and the caudal peduncle). The body surface image was adaptively divided into several physiologically significant sub-regions (for example, divided into four main regions: head, anterior trunk, posterior trunk, and tail). Within each sub-region, the image was converted from the RGB color space to the HSV color space, which is more favorable for separating brightness and color information. The histogram distribution of hue (H) and saturation (S) was calculated, and their mean, standard deviation, and entropy were statistically analyzed. This was used to depict the central tendency, dispersion, and disorder of color in each region, and finally, a color distribution feature describing the spatial non-uniformity of color was constructed.Finally, the specific phenomenon is determined through the collaborative analysis of abnormal light signal patterns and abnormal color distribution. This relies on a pre-built rule base and pattern recognition model. For example, when the algorithm detects that the proportion of specular reflection area in the trunk area of ​​a bass is abnormally reduced (e.g., below 30% of the average of healthy samples), and at the same time the average color saturation of the area is significantly reduced (e.g., the reduction rate exceeds 20%) and the color distribution entropy value is increased (indicating color mottledness), the system collaboratively determines that the individual has gloss disorder and mottledness. The entire analysis process integrates physical reflection models and regional statistical features to achieve objective and refined diagnosis of fading, mottledness, or gloss disorder, and outputs structured appearance and physiological state information.

[0052] The method provided in this embodiment enables precise and structured analysis of the complex and subtle optical changes on the body surface of sea bass caused by stress. It transforms non-contact visual observation into a quantifiable description of the physiological state of appearance that integrates multiple indicators, providing reliable and detailed input for subsequent cross-modal fusion with behavioral information. This fundamentally improves the accuracy and robustness of automatic identification of stress state based on appearance.

[0053] In some embodiments, based on the abnormal imbalance relationship and combined with the gloss disorder phenomenon, a synergistic comparative analysis of behavioral rhythm disorder and optical signal disorder is performed to determine the stress state of individual bass and obtain a preliminary stress state determination result. Based on the preliminary stress state determination result, combined with the fin movement rhythm characteristics and gill movement rhythm characteristics, the duration and trend of abnormal behavior of bass are analyzed. Through stability assessment technology, transient physiological fluctuations and persistent stress responses are distinguished to obtain a stress state identifier verified by dynamic tracking. Simultaneously, based on color distribution characteristics, the degree of interference of the physical outline size and body surface condition of the bass's body surface on optical assessment is analyzed. Through dual calibration of outline size and optical characteristics, misjudgment of body shape caused by abnormal body surface condition is eliminated to obtain a calibrated body shape classification result. Based on the stress state identifier and combined with the calibrated body shape classification result, according to the correlation pattern between the behavioral rhythm of the bass's fins and gills and the body surface color and reflectivity under stress, the stress level is coupled with the specific body shape specification to form a one-to-one corresponding individualized state profile and obtain a body shape-stress fusion assessment information set.

[0054] Body size assessment involves measuring and classifying the physical dimensions of sea bass. Stability assessment techniques are specialized methods used to analyze the duration and trends of behavioral abnormalities, distinguishing between transient physiological fluctuations and persistent stress responses. Dual calibration of outline dimensions and optical features is a method that simultaneously combines data from the sea bass's physical outline dimensions and optical characteristics to correct body size assessment results. Individualized state profiling involves coupling the sea bass's stress level with its specific body size, resulting in a comprehensive description containing unique state information for each individual.

[0055] Specifically, in automated grading and harvesting of bass, relying solely on single behavioral or appearance information for assessment can easily misjudge short-term physiological fluctuations as continuous stress, or misjudge body size due to abnormalities on the body surface, affecting the accuracy of grading. Furthermore, automated harvesting requires establishing precise status files for each bass, but the lack of behavioral and appearance correlation analysis makes it difficult to form a comprehensive understanding and develop targeted guidance strategies. At the same time, it can integrate previously scattered data and transform it into a practical fusion assessment information set to support subsequent processes. It is also key to ensuring the survival of bass and improving industry efficiency, so this solution is indispensable. To address the aforementioned issues: First, a synergistic comparative analysis of behavioral rhythm disorder and optical signal disturbance is employed. This involves aligning and matching the spatiotemporal patterns of "abnormal dysregulation" in fin and gill movements from the behavioral analysis module (e.g., the dorsal fin swing frequency is abnormally increased to several times the normal value, while the gill cover opening and closing rhythm is significantly slowed to less than half the normal frequency, indicating severe asynchrony) with "glossy disorder" from the appearance analysis module (e.g., through multi-region optical feature extraction, the specular reflection spot area above the lateral line of the fish body is abnormally enlarged and discretely distributed, replacing the original uniform diffuse reflection pattern). This allows for a preliminary determination of whether the bass is under stress. Next, to distinguish whether the abnormality is a transient physiological fluctuation or a persistent stress, stability assessment technology is introduced to track and analyze the duration and trend of the initial abnormal behavioral signals. For example, the system continuously monitors the fin movement rhythm of an individual initially identified as abnormal. If the high-frequency shaking occurs only within a few seconds and quickly returns to the baseline, it is classified as a transient fluctuation (such as being startled by a momentary light or shadow). If the gill movement rhythm shows that the gill cover opening and closing amplitude continuously decreases and the frequency becomes disordered within a few minutes, with an increasing trend, it is confirmed as a persistent stress response (such as due to a slow decrease in dissolved oxygen). Based on this, a dynamically verified "stress status label" is assigned. In parallel operation along the body size assessment line, a dual calibration method of contour size and optical features is employed to eliminate interference from temporary surface conditions on optical measurements. The system first extracts the initial physical contour based on high-resolution images, but simultaneously analyzes the "color distribution characteristics." For example, when a change in reflectivity is observed in a localized area of ​​the fish's abdomen due to the attachment of trace amounts of suspended matter, resulting in a decrease in the optical contrast of the contour boundary, the calibration algorithm intelligently corrects and interpolates the contour boundary based on the normal color and reflectivity distribution characteristics of the surrounding area (such as the uniform scale reflectivity pattern in adjacent areas), thereby eliminating interference and outputting the "calibrated body size classification result." Finally, through correlation and coupling analysis, the validated stress markers (such as "moderate persistent stress") are deeply bound to the calibrated body size results (such as "body length within the large size range"), generating an individualized profile for each fish containing multi-dimensional status labels, thus converging into an accurate "body size-stress fusion assessment information set."

[0056] The method provided in this embodiment significantly improves the accuracy and reliability of state determination, effectively avoiding misjudgments caused by single signal anomalies or transient interference. Its high-fidelity individual state profile provides a precise decision-making basis for subsequent differentiated graded guidance (such as accurately diverting stressed fish to redundant ponds), thereby achieving a comprehensive improvement in the accuracy, efficiency, and welfare protection of automated graded fishing.

[0057] In some embodiments, based on stress status identifiers and body size classification results, and according to the preset specifications and functional definitions of graded fishing target pools, the target pool category that each individual bass meets in terms of body size and the suitability of stress status for entering the corresponding target pool are analyzed to obtain individual-level pool matching suitability information. Based on the pool matching suitability information, bass are divided into two groups: stressed individuals and non-stressed individuals, according to stress status identifiers. For each group, combined with body size classification results, the final target pool to be guided is analyzed to obtain preliminary guidance decisions. Based on the preliminary guidance decisions, an instruction unit containing a unique identifier, current judgment status, specified guidance path, and target pool is generated for each individual. The instruction units of all individuals are aggregated and conflict resolved to resolve guidance conflicts caused by path intersections or instantaneous resource competition, resulting in a coordinated and executable set of differentiated graded guidance strategies for bass.

[0058] A graded fishing target pond can be a pre-designed pond specifically for accommodating bass of different sizes and physiological states. Specification requirements can be specific standards defining the body size of the bass that the graded fishing target pond can accommodate. Functional definition can be the pre-defined purpose and positioning of the graded fishing target pond. Pond matching suitability information can be based on the individual bass's stress status and body size classification results, combined with the specification requirements and functional definition of the graded fishing target pond, to analyze the degree of compatibility between the individual and the target pond. Preliminary guidance decision-making can be a decision-making scheme that preliminarily determines the final target pond to which each individual should be guided, based on their body size classification results, for both stressed and non-stressed individuals. Command units can be guidance command carriers generated individually for each bass individual. Aggregation and conflict resolution can be the integration of command units from all individuals. Guidance conflict can refer to situations where the designated guidance paths of multiple bass individuals overlap or too many individuals simultaneously flood into the same target pond entrance, affecting guidance efficiency.

[0059] Specifically, after accurately assessing the body size and stress state of individual bass based on dynamic perception, directly guiding the catch still faces significant challenges. The problem lies in the gap between the assessment results and the final action decision. Without a systematic strategy, the fishing process will become blind and inefficient. To address this issue: First, a combination of rule-based reasoning and risk quantification assessment is used to analyze the suitability of individuals with each target pool. The system includes built-in "specification requirements" (e.g., sales pool #1 only accepts fish with a body length of 35 cm or more) and "functional definitions" (e.g., redundant pool #2 is dedicated to stress recovery) for each graded target pool. By comparing the fish's body size data with the pool specifications, "pool type compliance information corresponding to body size specifications" is obtained. Simultaneously, based on stress indicators combined with historical data models, the potential risks of individuals in this state entering various functional pools are assessed (e.g., the probability of increased damage to highly stressed fish entering the sales pool), generating "pool type risk information corresponding to the stress state." After merging the two, accurate "pond-matching suitability information" is obtained (for example, a fish may fit into sales pond #1 in terms of size, but due to high stress, its risk assessment value for entering that pond is too high, so it is judged as "currently unsuitable"). Subsequently, the system performs classification decisions based on this information. For example, all individuals marked as "highly stressed," regardless of their size, are initially uniformly guided to redundant ponds, while "non-stressed" individuals are directly assigned to their corresponding sales or breeding ponds according to their size, forming an "initial guidance decision." To execute this decision, the system generates a structured "instruction unit" for each individual, containing a unique ID, status, planned path (e.g., via diversion gate 3), and target pool. Finally, a centralized scheduler performs "aggregation and conflict resolution" on all instruction units: the scheduler uses simulation and timing planning algorithms to detect and resolve potential resource competition. For example, when the system finds that multiple instruction units require to occupy the same physical channel within the same millisecond time window, it dynamically adjusts the passage timing of these individuals (e.g., introducing a small delay or changing the order) or enables alternative paths. All coordination calculations are completed instantaneously, ultimately outputting a "bass differentiated graded guidance strategy set" that is conflict-free in both time and space and can be directly executed by downstream actuators such as bubble curtains.

[0060] The method provided in this embodiment transforms the abstract individual state assessment into a specific, orderly, and executable group guidance action plan, effectively avoiding secondary damage or resource waste caused by the mismatch between individual state and target environment. At the same time, it solves the physical path and resource competition conflicts that may occur when guiding multiple individuals concurrently, ensuring the smoothness, efficiency, and reliability of the entire automated grading and fishing process, and significantly improving the grading accuracy and fish welfare level.

[0061] In some embodiments, based on stress state identifiers and combined with the functional definition of graded fishing target pools, stress-oriented pool risk quantification assessment is used to analyze the expected recovery period of stressed bass individuals entering redundant pools and the probability of increased stress caused by entering graded fishing target pools, thereby obtaining pool risk information corresponding to the stress state; based on body size classification results and combined with the specification requirements of graded fishing target pools, the body size parameters of each bass individual are analyzed to determine the corresponding preset specification range of the target pool, thereby obtaining pool conformity information corresponding to the body size specification; pool risk information and pool conformity information are combined to construct individual-level pool matching suitability information.

[0062] Stress-oriented pool-based risk quantification assessment can be a method that focuses on the stress state of sea bass and quantifies the impact of different pool types on their stress levels. The estimated recovery period can be estimated based on factors such as the degree of sea bass stress and the environmental conditions of redundant pools, predicting the time required for an individual stressed sea bass to recover from stress in a redundant pool. The probability of stress exacerbation can be the likelihood that a stressed sea bass will experience a further aggravation of its stress response after entering a non-redundant graded fishing target pool due to poor environmental adaptation. The preset size range can be a range of body size parameters defined according to the size requirements of the graded fishing target pools. Pool conformity information can be information reflecting whether the body size parameters of an individual sea bass fall within the preset size range of a specific graded fishing target pool.

[0063] Specifically, existing bass grading and harvesting techniques typically rely solely on physical dimensions (such as weight and length) for mechanical sorting, completely ignoring the crucial biological state of live fish inevitably experiencing stress during harvesting and transportation. This "one-size-fits-all" purely standardized grading method has serious flaws: placing bass individuals in a state of stress (such as exhibiting rapid swimming and dull coloration) alongside healthy individuals into target breeding or sales ponds not only disrupts the order of the fish population due to their abnormal behavior, but may also exacerbate their stress response due to new stimuli such as sudden environmental changes and density pressure, leading to anorexia, decreased immunity, or even secondary infections and death, causing direct economic losses and posing risks to subsequent breeding management or sales. To address these issues: First, for stressed individuals, a "stress-oriented pond-specific risk quantification assessment model" is activated. This model incorporates data based on a large number of... The stress-response knowledge base, trained from historical aquaculture data, allows the model to calculate the "expected recovery period" (e.g., approximately 3 hours) of an individual after entering the redundant pool, based on the individual's current stress level and specific characteristics (such as the percentage of gill opening and closing frequency exceeding the baseline value), combined with standard treatment process parameters of the "redundant pool" (defined as a functional pool with a still, shaded environment and the ability to administer anti-stress agents). Simultaneously, the model simulates the scenario where the individual is directly assigned to a "graded harvesting target pool" (e.g., a high-density cement pool for temporarily holding medium-sized marketable fish), comprehensively assessing the matching relationship between the potential stress intensity of the new environment (such as water flow velocity and similar density) and its current stress vulnerability, quantifying and predicting the "probability of increased stress" (e.g., a calculated probability of 40%). These two quantitative results—recovery period and risk probability—together constitute the individual's "pool-specific risk information." In parallel, the system performs "size range matching analysis" on all individuals (regardless of whether they are stressed or not): it quickly compares the "body size classification result" with the pre-defined "preset size range" of each target pool (for example, the system defines target pool 1 as receiving fish >450 grams, pool 2 as receiving fish 300-450 grams, and pool 3 as receiving fish <300 grams). For an individual weighing 380 grams, the analysis results clearly indicate that their "pool matching information" is "meets the specifications of Pool 2". Ultimately, the decision fusion engine integrates these two pieces of information to generate structured "individual-level pool matching suitability information" for each individual. A typical description would be: "ID-00157, medium size (380g), meets the requirements of Target Pool 2. However, due to the current state of moderate stress, the estimated risk of increased stress from direct entry into Pool 2 is 40%. Therefore, the decision priority is: first guide them to the redundant pool for recovery (estimated to take 3 hours), and after the stress is relieved, reassign them to Pool 2 based on their size." This information provides the unique and sufficient basis for subsequently generating precise and differentiated physical guidance instructions.

[0064] The method provided in this embodiment can effectively avoid secondary damage to stressed individuals during the grading process, significantly reduce subsequent morbidity and mortality caused by improper operation, and ensure the overall health of the fish population and the economic benefits of aquaculture. At the same time, this refined decision-making basis enables the entire process to implement the optimized process of "prioritizing isolation of stressed individuals and accurately grading based on health," which not only improves the scientific nature and accuracy of grading but also provides core technical support for achieving more efficient and humane modern aquaculture management.

[0065] In some embodiments, based on pool matching suitability information, the entire process node information of each type of bass individual is analyzed from initial perception, state determination, guidance execution to final pool allocation through temporal correlation and event backtracking analysis of the fishing process data, to obtain the individual fishing trajectory chain; based on the individual fishing trajectory chain, the dynamic flow direction control process is implemented according to the bubble curtain, and the linkage analysis between fishing operation and fish response status is carried out, recording the actual execution efficiency of the guidance strategy and the state recovery progress of stressed bass individuals in redundant pools, to obtain the fishing process efficiency record; based on the individual fishing trajectory chain, combined with the fishing process efficiency record, the timestamp, individual identifier, state determination sequence, guidance action, pool change and key efficiency indicators are integrated and formatted for encapsulation, and the bass automated graded fishing log is output.

[0066] The individual harvesting trajectory chain can be the entire process node information for each type of bass individual from initial perception, state determination, guidance execution to final pool allocation. The dynamic flow direction control process implemented by the bubble curtain can be the actual execution process of releasing bubbles to form a specific flow direction and guide the movement of individual bass. The linkage analysis between harvesting operations and fish school response can be the analysis process of correlating and decomposing equipment operations and fish school behavioral responses during the harvesting process. The actual execution efficiency of the guidance strategy can be a quantitative indicator of the effectiveness of the differentiated and graded guidance strategy in successfully guiding bass individuals to the target area in practical applications. The state recovery progress can be an indicator of the process by which stressed bass individuals transition from a stressed state to a normal state after receiving anti-stress treatment in a redundant pool. The harvesting process efficiency record can be an information set recording key efficiency data such as guidance efficiency and the recovery status of stressed individuals. The state determination sequence can be the determination results of each stress state and body size classification during the initial to final allocation process of bass individuals. The guidance action can be the specific guidance operation implemented through bubble curtain control. Key efficiency indicators can be key data reflecting the operational effectiveness of the harvesting system.

[0067] Specifically, in automated grading and harvesting of bass, constructing an automated grading and harvesting log can completely record the entire process data of bass from initial perception to final pond allocation. This provides a basis for tracing the source of problems such as grading errors and poor recovery of stressed individuals, ensuring quality control. The efficiency indicators in the log, such as guidance efficiency and recovery progress of stressed individuals, are key data support for optimizing bubble curtain parameters and guidance strategies, facilitating system iteration. The log also meets the compliance requirements of aquaculture, providing reliable records for production control, market supervision, and quality traceability, ensuring the stable and efficient operation of the harvesting system. To address the above issues: First, time-series correlation and event backtracking analysis technology is used to construct an "individual harvesting trajectory chain" for each bass assigned a unique ID. Specifically, this technology concatenates and reassembles discrete events scattered across different time points (e.g., abnormal skin color detected by a visual sensor at time t1, moderate stress identified by a fusion analysis module and matched to a redundant pool at time t2, and a guidance instruction to the northwest corner received at time t3) based on strict timestamps and individual IDs, thus forming a complete digital trajectory from entry perception to final pool placement. Next, to generate a "fishing process efficiency record," a linkage analysis of the dynamic control process of the bubble curtain is initiated. First, spatiotemporal alignment technology is used to precisely match and compare the opening and closing sequence and intensity changes of the bubble curtain (e.g., adjusting the bubble generator power in the southeast area to 70% of its rated value for 5 seconds during a certain guidance phase) with changes in fish school displacement captured by high-definition cameras (e.g., the average swimming speed of fish in this area increases from 0.1 m / s to 0.3 m / s, and the overall flow direction deflects by 15 degrees). This generates a "bubble curtain control - fish school displacement response" record. Based on the "linked information," the system further employs a group response difference analysis method. Using known stress state markers, it compares and analyzes the behavioral differences of different groups in response to the same control command. For example, when faced with the same moderate-intensity bubble screen, non-stressed fish groups may exhibit highly compliant, orderly group turning, while stressed individuals may exhibit abnormal responses such as path disorder, accelerated escape, or leaving the group. These characteristics are extracted as "control response characteristic information." Ultimately, the real-time adaptability analysis of the control strategy is used for effectiveness evaluation: the system calculates the degree of consistency between the actual diversion path and the preset guidance path (e.g., 20 fish were expected to be introduced into channel 3, but 18 were actually detected, resulting in a diversion accuracy of 90%), and combines this with monitoring data on the recovery of stressed fish in the redundant pool (e.g., subsequent image analysis showed that their gill cover movement rhythm recovered from disorder to normal within 30 minutes). This comprehensive approach quantifies the execution efficiency and recovery effect of the guidance strategy.All the aforementioned multidimensional data—including structured individual trajectory chains, performance records containing key quantitative indicators (such as diversion accuracy and state recovery time), and associated original control parameters—are automatically integrated by the system and encapsulated in a preset JSON format, ultimately outputting a detailed, structured "bass automated grading and harvesting log" that can be read by machines and viewed by humans.

[0068] The method provided in this embodiment enables full traceability of the fishing process, provides a basis for tracing the source of problems, ensures graded quality, and the recorded performance data can support the optimization of bubble curtain parameters, guidance strategies, etc., help the system iterative upgrade, and the standardized logs meet the compliance requirements of aquaculture, which facilitates production control, market supervision and quality traceability.

[0069] In some embodiments, based on individual fishing trajectory chains, the temporal and spatial alignment technology of the regulatory action and fish school displacement is used to analyze the temporal changes of bubble curtain operation and the correspondence between the displacement direction and velocity changes of individuals in the corresponding area of ​​the fish school, thereby obtaining bubble curtain regulation-fish school displacement response correlation information. Based on the bubble curtain regulation-fish school displacement response correlation information, combined with stress state indicators, the behavioral differences between stressed and non-stressed bass individuals in terms of swimming path compliance, speed adjustment sensitivity, and group coordination when facing the same or different bubble curtain regulation parameters are analyzed through group response difference analysis, thereby obtaining regulation response characteristic information. Based on the regulation response characteristic information, the real-time adaptability analysis of regulation strategies is used to analyze the deviation between the actual overall fish school flow pattern, individual diversion accuracy, and expected guidance path caused by different bubble curtain regulation actions triggered to execute the bass differentiated graded guidance strategy set. Based on the degree of deviation and the distribution of stressed bass individuals, the suitability of bubble curtain parameter settings is dynamically retrospectively and quantitatively evaluated, thereby obtaining the linkage analysis results of fishing operation and fish school response state.

[0070] Spatiotemporal alignment technology between regulatory effects and fish school displacement can be used to match and correlate changes in bubble curtain operation with the displacement response of individuals in the corresponding area of ​​the fish school in terms of time and space. Bubble curtain regulation-fish school displacement response correlation information can reflect the correspondence between changes in the timing of bubble curtain operation and changes in the direction and speed of fish school displacement. Group response difference analysis can be an analytical method for analyzing behavioral differences between stressed and non-stressed bass individuals under the same or different bubble curtain regulation parameters. Regulation response characteristic information can be a set of information characterizing the differences in swimming path compliance, speed adjustment sensitivity, and group coordination among bass in different states. Real-time adaptability analysis of regulation strategies can be an analytical method used to evaluate the deviation between the actual fish school flow pattern and individual diversion accuracy caused by bubble curtain regulation actions and the expected guidance path, and to retrospectively analyze the suitability of parameters. The suitability of bubble curtain parameter settings can be the degree of matching between the various operating parameters of the bubble curtain and the needs of fish school regulation.

[0071] Specifically, in automated grading and harvesting of bass, dynamic control of the bubble curtain is the core method for grading guidance. However, the difference between the stressed and non-stressed states of bass leads to different responses to control. Without the linkage analysis between harvesting operations and fish responses, problems such as a disconnect between control and response (e.g., bubble curtain adjustment failing to achieve the expected guidance effect), exacerbated stress in stressed individuals due to inappropriate parameters, and inability to detect parameter deviations can easily occur, resulting in inaccurate grading, low efficiency, and reduced bass survival. To address these issues, a high-precision timing module and underwater positioning network (such as an acoustic or computer vision-based tracking system) are used to generate control commands for each bubble curtain (e.g., "at A..."). (A bubble curtain with an intensity of 70% is activated in area A for 3 seconds.) The data is stamped with a millisecond-level timestamp, and the continuous spatial coordinates and velocity vectors of each bass in area A during this period are recorded simultaneously. Through a specialized spatiotemporal alignment algorithm, the time series of "bubble curtain activation period and intensity curve" is superimposed and correlated with the spatial series of "fish displacement direction and velocity changes". This allows for the quantification of bubble curtain regulation-fish displacement response correlation information, such as "when the bubble curtain in area A is activated at 70% intensity for 2 seconds, the average swimming direction of the fish in the area will deviate by about 30 degrees towards the predetermined guidance path, and the average speed will decrease to 40% of the original level". Next, the system invokes the pre-assigned stress status tags (from front-end perception fusion analysis) for each fish to perform a group response difference analysis: it groups the same batch of data into "stressed" and "non-stressed" categories. The comparison reveals that, faced with the same bubble curtain stimulus, the path deflection angle of the stressed individuals shows extremely high dispersion (some may deflect by 60 degrees, while others show almost no reaction), and their speed decrease is significantly delayed (a significant speed change may only occur 1.5 seconds after the stimulus begins). In contrast, the non-stressed individuals react more consistently and promptly. Based on these extracted regulatory response characteristics, the system enters a real-time adaptability analysis loop: for example, if the current instruction is "to guide the three stressed large fish in area A to the redundant pool entrance within 5 seconds," the system predicts and executes a set of regulatory combinations based on the aforementioned characteristics (such as first activating the guidance direction with a 50% intensity pulse, and then fine-tuning according to the real-time position). Subsequently, the system continuously compares the actual movement trajectories of the three target fish with the expected paths, calculates the "individual diversion accuracy" (e.g., only two fish arrive at the entrance after 5 seconds, accuracy is 66.7%), and analyzes the reasons for deviations (e.g., the unsuccessful fish attempted to escape in the opposite direction under pulse stimulation). At this point, the system dynamically backtracks and evaluates, determining that the initial pulse intensity is still too high for the stressed individual. Therefore, it immediately generates parameter optimization suggestions (e.g., using a gradient-increasing intensity for similar stressed individuals next time, starting from 30%), and updates this empirical knowledge to the regulatory strategy library, thereby achieving closed-loop optimization based on real-time biofeedback, and finally outputting linkage analysis results to guide subsequent precise regulation.

[0072] The method provided in this embodiment accurately captures the correlation between bubble curtain regulation and fish school response, clarifies the differentiated behavioral characteristics of bass under different stress states, breaks the limitation of "only regulating without evaluation", provides a basis for judging whether the regulation is up to standard, and can optimize bubble curtain parameters by analyzing deviations, improve the accuracy and stability of fish school diversion, and also avoid exacerbating stress in stressed individuals due to unsuitable parameters, ensuring their survival quality, forming a closed-loop feedback that allows the regulation strategy to be dynamically adapted, and simultaneously improves fishing efficiency, grading accuracy and bass survival quality.

[0073] Figure 3 This is a schematic diagram of the structure of an automated grading and harvesting system for bass based on live dynamic sensing, provided in an embodiment of this application. Figure 3 As shown, the automated grading and fishing system 300 for bass based on live dynamic perception in this embodiment includes: a fusion evaluation module 301, a grading guidance module 302, and a grading and fishing module 303.

[0074] The fusion assessment module 301 is used to acquire multi-source bass state perception data. Based on the multi-source bass state perception data, and according to preset body size benchmark parameters, it performs cross-modal fusion analysis of live behavior and physiological appearance to obtain a body size-stress fusion assessment information set. The grading guidance module 302 is used to analyze the matching relationship between bass body size status, stress status and grading target pool based on the body size-stress fusion assessment information set, and generate a differentiated grading guidance strategy set for bass. The grading fishing module 303 is used to implement dynamic flow direction control through a bubble curtain based on the differentiated grading guidance strategy set for bass, guide stressed bass individuals to a redundant pool, and after the stressed bass individuals are relieved of stress through anti-stress treatment in the redundant pool, they are re-guided and allocated according to body size assessment, and the automated grading fishing log for bass is output.

[0075] Optionally, the fusion evaluation module 301, when performing cross-modal fusion analysis of live behavior and physiological appearance based on the multi-source bass state perception data and according to preset body size benchmark parameters to obtain a body size-stress fusion evaluation information set, specifically performs the following: the multi-source bass state perception data includes fin dynamic frequency, gill cover movement rhythm, and body surface color change information; based on the fin dynamic frequency and combined with the gill cover movement rhythm, the live behavior activity of the bass is analyzed using behavioral rhythm analysis technology to obtain behavioral stress characterization information; based on the body surface color change information, the reflective properties and regional color distribution of the bass body surface are analyzed using multi-region optical feature extraction technology to obtain appearance physiological state information; based on the behavioral stress characterization information and combined with the appearance physiological state information, a correlation coupling analysis of behavior and appearance is performed to determine the individual stress state of the bass and simultaneously evaluate the bass body size, obtaining the body size-stress fusion evaluation information set that simultaneously includes stress state identification and body size classification results.

[0076] Optionally, the fusion evaluation module 301, during the construction of the behavioral stress representation information, is specifically used for: based on the fin dynamic frequency, analyzing the swing differences between the hard spines and soft rays of the dorsal fin of the bass using time-domain periodic signal analysis technology to obtain fin movement rhythm characteristics; based on the gill cover movement rhythm, analyzing the physiological characteristics of the bass as a high-oxygen-consuming carnivorous fish using rhythm and amplitude joint analysis technology, analyzing the regular changes in the frequency and amplitude of gill cover opening and closing under static, swimming, and stress conditions to obtain gill movement rhythm characteristics; based on the fin movement rhythm characteristics, combined with the gill movement rhythm characteristics, conducting a comparative analysis of the movement rhythms of different body parts of the bass, identifying the synergistic relationship or abnormal disorder between fin swing and gill cover movement, and obtaining the behavioral stress representation information.

[0077] Optionally, the fusion evaluation module 301, during the construction of the appearance physiological state information, is specifically used for: based on the surface color change information, analyzing the specular and diffuse reflection distribution areas of the sea bass's body surface under the fishing reflected light according to the preset fishing light source information, to obtain the reflective distribution characteristics of the sea bass's body surface; based on the reflective distribution characteristics, analyzing the color depth and distribution uniformity of different areas of the sea bass's body surface through multi-region optical feature extraction technology, to obtain the color distribution characteristics of the sea bass's body surface; based on the color distribution characteristics, performing a collaborative analysis of light signal mode abnormalities and color distribution abnormalities, determining the fading phenomenon, mottled phenomenon, or gloss disorder phenomenon of the sea bass's body surface caused by stress response, to obtain the appearance physiological state information.

[0078] Optionally, the fusion evaluation module 301, when performing the correlation coupling analysis between behavior and appearance to determine the stress state of an individual bass and simultaneously evaluating the bass's body size to obtain the body size-stress fusion evaluation information set containing both stress state identifiers and body size classification results, is specifically used for: based on the abnormal imbalance relationship and combined with the luster disorder phenomenon, performing a synergistic comparative analysis of behavioral rhythm disorder and optical signal disorder to determine the stress state of an individual bass and obtain a preliminary stress state determination result; based on the preliminary stress state determination result, combined with the fin movement rhythm characteristics and the gill movement rhythm characteristics, analyzing the duration and trend of abnormal bass behavior, and using stability evaluation technology. The system distinguishes between transient physiological fluctuations and persistent stress responses, obtaining stress state identifiers verified through dynamic tracking. Simultaneously, based on the color distribution characteristics, it analyzes the interference of the physical outline size and surface condition of the sea bass's body on optical assessment. Through dual calibration of outline size and optical characteristics, it eliminates misjudgments of body shape caused by abnormal surface condition, obtaining calibrated body shape classification results. Based on the stress state identifiers and the calibrated body shape classification results, and according to the correlation pattern between the behavioral rhythm of the sea bass's fins and gills and the color and reflectivity of its body surface under stress, it couples the stress level with specific body shape specifications to form a one-to-one corresponding individualized state profile, obtaining the body shape-stress fusion assessment information set.

[0079] Optionally, the graded guidance module 302, during the construction of the bass differentiated graded guidance strategy set, is specifically used for: based on the stress state identifier and combined with the body size classification result, according to the preset specification requirements and functional definitions of the graded fishing target pool, analyzing the target pool category that each bass individual conforms to in terms of body size and the suitability impact of stress state on entering the corresponding target pool, to obtain individual-level pool matching suitability information; based on the pool matching suitability information, dividing the bass into two groups, stressed individuals and non-stressed individuals, according to the stress state identifier, and for each group, combined with the body size classification result, analyzing the final target pool to be guided to, to obtain a preliminary guidance decision; based on the preliminary guidance decision, generating an instruction unit for each individual containing a unique identifier, current judgment state, specified guidance path and target pool, and aggregating and resolving conflicts of all individual instruction units to resolve guidance conflicts caused by path intersection or instantaneous resource competition, to obtain a coordinated and executable bass differentiated graded guidance strategy set.

[0080] Optionally, the grading guidance module 302, when analyzing the target pond category that each bass individual conforms to in terms of body size and the impact of stress state on its suitability for entering the corresponding target pond, and obtaining individual-level pond matching suitability information, specifically performs the following: Based on the stress state identifier and combined with the functional definition of the grading fishing target pond, it analyzes the expected recovery period of the stressed bass individual when entering the redundant pond and the probability of stress aggravation caused by entering the grading fishing target pond through stress-guided pond risk quantification assessment, to obtain pond risk information corresponding to the stress state; Based on the body size classification result and combined with the specification requirements of the grading fishing target pond, it analyzes the preset specification range of the target pond corresponding to the body size parameters of each bass individual, to obtain pond conformity information corresponding to the body size specification; and constructs the individual-level pond matching suitability information by combining the pond risk information and the pond conformity information.

[0081] Optionally, the graded fishing module 303, during the construction of the automated graded fishing log for bass, is specifically used for: based on the pool matching suitability information, analyzing the entire process node information of each type of bass individual from initial perception, state determination, guidance execution to final pool allocation through temporal correlation and event backtracking analysis of the fishing process data, to obtain the individual fishing trajectory chain; based on the individual fishing trajectory chain, performing linkage analysis of fishing operation and fish response state according to the dynamic flow direction control process implemented by the bubble curtain, recording the actual execution efficiency of the guidance strategy and the state recovery progress of the stressed bass individual in the redundant pool, to obtain the fishing process efficiency record; based on the individual fishing trajectory chain, combined with the fishing process efficiency record, integrating and formatting the timestamp, individual identifier, state determination sequence, guidance action, pool change, and key efficiency indicators, and outputting the automated graded fishing log for bass.

[0082] Optionally, the graded fishing module 303, when performing the linkage analysis of fishing operations and fish school response states during the dynamic flow direction control process based on the bubble curtain, is specifically used for: analyzing the temporal changes of the bubble curtain operation and the correspondence between the changes in displacement direction and velocity of individuals in the corresponding area of ​​the fish school, based on the individual fishing trajectory chain and the spatiotemporal alignment technology of the control effect and fish school displacement, to obtain bubble curtain control-fish school displacement response correlation information; and, based on the bubble curtain control-fish school displacement response correlation information and the stress state identifier, analyzing the stressed bass individuals and non-stressed bass individuals through group response difference analysis. When individual fish face the same or different bubble curtain control parameters, their behavioral differences in swimming path compliance, speed adjustment sensitivity, and group coordination are analyzed to obtain control response characteristic information. Based on the control response characteristic information, the deviation between the actual overall fish flow pattern, individual diversion accuracy, and expected guidance path caused by different bubble curtain control actions triggered by the execution of the bass differentiated graded guidance strategy set is analyzed through real-time adaptability analysis. According to the degree of deviation and the distribution of stressed bass individuals, the suitability of bubble curtain parameter settings is dynamically retrospectively and quantitatively evaluated to obtain the linkage analysis results between fishing operations and fish response status.

[0083] The system in this embodiment can be used to execute the methods of any of the above embodiments, and its implementation principle and technical effect are similar, so they will not be described again here.

Claims

1. An automated grading and harvesting method for bass based on live dynamic sensing, characterized in that, include: Acquire multi-source bass state perception data, and based on the multi-source bass state perception data, perform cross-modal fusion analysis of live behavior and physiological appearance according to preset body size benchmark parameters to obtain a body size-stress fusion assessment information set; Based on the body size-stress fusion assessment information set, the matching relationship between the body size status, stress status and graded fishing target pond of bass is analyzed, and a set of differentiated graded guidance strategies for bass is generated. Based on the aforementioned differentiated grading and guidance strategy set for bass, dynamic flow direction control is implemented through a bubble curtain to guide stressed bass individuals to a redundant pool. After stress relief treatment is carried out in the redundant pool, the stressed bass individuals are reassigned based on body size assessment, and an automated grading and harvesting log for bass is output.

2. The method according to claim 1, characterized in that, Based on the multi-source bass state perception data, and according to preset body size benchmark parameters, a cross-modal fusion analysis of live behavior and physiological appearance is performed to obtain a body size-stress fusion assessment information set, including: The multi-source bass state perception data includes fin dynamic frequency, gill cover movement rhythm and body surface color change information. Based on the dynamic frequency of the fins and the rhythm of the gill cover movement, the activity level of the live behavior of the bass is analyzed by behavioral rhythm analysis technology to obtain behavioral stress characterization information. Based on the information on changes in body color, the reflective properties and regional color distribution of the sea bass's body surface are analyzed using multi-region optical feature extraction technology to obtain information on its appearance and physiological state. Based on the behavioral stress representation information and combined with the appearance physiological state information, a correlation coupling analysis between behavior and appearance is performed to determine the stress state of individual bass and simultaneously assess the body size of the bass, resulting in the body size-stress fusion assessment information set that simultaneously includes stress state identifiers and body size classification results.

3. The method according to claim 2, characterized in that, The process of constructing the behavioral stress representation information includes: Based on the dynamic frequency of the fins, the swing differences between the hard spines and soft rays of the dorsal fin of the sea bass are analyzed using time-domain periodic signal analysis technology to obtain the rhythmic characteristics of fin movement. Based on the gill cover movement rhythm, the physiological characteristics of the sea bass as a high-oxygen-consuming carnivorous fish are analyzed by combining rhythm and amplitude analysis technology. The regular changes in the frequency and amplitude of gill cover opening and closing under static, swimming and stress conditions are analyzed to obtain the gill movement rhythm characteristics. Based on the fin movement rhythm characteristics and the gill movement rhythm characteristics, a comparative analysis of the movement rhythms of different body parts of the bass is conducted to identify the synergistic relationship or abnormal disorder between fin swaying and gill cover movement, thereby obtaining the behavioral stress characterization information.

4. The method according to claim 3, characterized in that, The process of constructing the appearance and physiological state information includes: Based on the information on the color change of the body surface, and according to the preset information on the fishing light source, the distribution area of ​​specular reflection and diffuse reflection of the sea bass body surface under the fishing reflected light is analyzed to obtain the reflective distribution characteristics of the sea bass body surface. Based on the aforementioned reflective distribution characteristics, the color depth and distribution uniformity of different regions on the body surface of the sea bass are analyzed using multi-region optical feature extraction technology to obtain the color distribution characteristics of the sea bass body surface. Based on the aforementioned color distribution characteristics, a synergistic analysis of abnormal light signal patterns and abnormal color distribution is performed to determine the fading, mottled, or glossy disturbances on the sea bass's body surface caused by stress response, thereby obtaining the aforementioned physiological state information.

5. The method according to claim 4, characterized in that, The process involves performing a correlation coupling analysis between behavior and appearance to determine the stress state of individual bass and simultaneously assess their body size, resulting in a body size-stress fusion assessment information set that includes both stress state identifiers and body size classification results. This set includes: Based on the aforementioned abnormal imbalance, and combined with the aforementioned luster disorder, a synergistic comparative analysis of behavioral rhythm disorder and optical signal disorder was conducted to determine the stress state of individual bass and obtain preliminary stress state determination results. Based on the preliminary stress state determination results, combined with the fin movement rhythm characteristics and the gill movement rhythm characteristics, the duration and trend of abnormal behavior in bass were analyzed. Through stability assessment technology, transient physiological fluctuations and persistent stress responses were distinguished, and stress state indicators verified by dynamic tracking were obtained. Simultaneously, based on the color distribution characteristics, the degree of interference of the physical outline size and body surface condition of the sea bass on the optical evaluation is analyzed. Through dual calibration of outline size and optical characteristics, misjudgment of body shape caused by abnormal body surface condition is eliminated, and the calibrated body shape classification result is obtained. Based on the stress state identifier and the calibrated body size classification results, the stress level is coupled with the specific body size specification according to the correlation pattern between the fin and gill behavior rhythm and the body color and reflectivity of the bass under stress, forming a one-to-one corresponding individualized state profile, and thus obtaining the body size-stress fusion assessment information set.

6. The method according to claim 5, characterized in that, The process of constructing the differentiated grading guidance strategy set for sea bass includes: Based on the stress state identifier and the body size classification results, according to the preset specifications and functional definitions of the graded fishing target ponds, the target pond category that each bass individual meets in terms of body size and the impact of stress state on the suitability of entering the corresponding target pond are analyzed to obtain individual-level pond matching suitability information. Based on the pool matching suitability information, the bass are divided into two groups: stressed individuals and non-stressed individuals according to the stress status identifier. For each group, the final target pool to be guided is analyzed in combination with the body size classification results to obtain a preliminary guidance decision. Based on the initial guidance decision, an instruction unit containing a unique identifier, current judgment status, specified guidance path, and target pool is generated for each individual. The instruction units of all individuals are aggregated and conflict-resolved to resolve guidance conflicts caused by path intersections or instantaneous resource competition, resulting in a coordinated and executable set of differentiated guidance strategies for sea bass.

7. The method according to claim 6, characterized in that, The analysis examines the target pond category that each individual bass meets in terms of body size and the impact of stress state on its suitability for entering the corresponding target pond, yielding individual-level pond matching suitability information, including: Based on the stress state identifier and the functional definition of the graded fishing target pool, the expected recovery period of the stressed bass when entering the redundant pool and the probability of stress aggravation caused by entering the graded fishing target pool are analyzed through stress-oriented pool risk quantification assessment to obtain the pool risk information corresponding to the stress state. Based on the body size classification results and the specification requirements of the graded fishing target ponds, the body size parameters of each individual bass are analyzed to determine the corresponding preset specification range of the target pond, thereby obtaining the pond type compliance information corresponding to the body size specification. The pool risk information and the pool suitability information are used to construct the individual-level pool matching suitability information.

8. The method according to claim 7, characterized in that, The process of constructing the automated grading and harvesting log for bass includes: Based on the pool matching suitability information, through the temporal correlation and event backtracking analysis of the fishing process data, the entire process node information of each type of bass individual from initial perception, state judgment, guidance execution to final pool allocation is analyzed to obtain the individual fishing trajectory chain; Based on the individual fishing trajectory chain, and according to the dynamic flow direction control process implemented by the bubble curtain, the linkage analysis between fishing operation and fish response status is carried out, and the actual execution efficiency of the guidance strategy and the recovery progress of the stressed bass individuals in the redundant pool are recorded to obtain the fishing process efficiency record. Based on the individual fishing trajectory chain and combined with the fishing process efficiency record, the timestamp, individual identifier, status judgment sequence, guiding action, pool type change and key efficiency indicators are integrated and formatted into a log for automated graded fishing of bass.

9. The method according to claim 8, characterized in that, The process of dynamically controlling the flow direction based on the bubble curtain, and performing a linkage analysis between the fishing operation and the fish school response status, includes: Based on the individual fishing trajectory chain, by using the spatiotemporal alignment technology of the control effect and the fish school displacement, the temporal changes of the bubble curtain operation are analyzed, and the correspondence between the displacement direction and velocity changes of the corresponding individuals in the fish school is obtained, thus obtaining the bubble curtain control-fish school displacement response correlation information. Based on the bubble curtain regulation-fish school displacement response correlation information, combined with the stress state identifier, through group response difference analysis, the behavioral differences between stressed and non-stressed bass individuals in swimming path compliance, speed adjustment sensitivity and group coordination when facing the same or different bubble curtain regulation parameters are analyzed, and regulation response characteristic information is obtained. Based on the aforementioned regulatory response characteristic information, the deviation between the actual overall fish flow pattern, individual diversion accuracy, and expected guidance path caused by different bubble curtain regulatory actions triggered by the execution of the differentiated graded guidance strategy set for bass is analyzed through real-time adaptability analysis. Based on the degree of deviation and the distribution of stressed bass individuals, the suitability of the bubble curtain parameter settings is dynamically retrospectively and quantitatively evaluated to obtain the linkage analysis results between fishing operations and fish response status.

10. An automated grading and harvesting system for bass based on live dynamic sensing, characterized in that, The method applied to any one of claims 1-9 includes: The fusion assessment module is used to acquire multi-source bass state perception data. Based on the multi-source bass state perception data, and according to the preset body size benchmark parameters, it performs cross-modal fusion analysis of live behavior and physiological appearance to obtain a body size-stress fusion assessment information set. The graded guidance module is used to analyze the matching relationship between the body size status, stress status and graded fishing target pond of the bass based on the body size-stress fusion assessment information set, and generate a set of differentiated graded guidance strategies for bass. The graded harvesting module is used to dynamically control the flow direction of sea bass based on the differentiated graded guidance strategy set of sea bass, through a bubble curtain, to guide stressed sea bass individuals to a redundant pool, and after the stressed sea bass individuals are relieved of stress through anti-stress treatment in the redundant pool, they are reassigned according to body size assessment, and the automated graded harvesting log of sea bass is output.

Images