Search

CN-122018315-A - Bionic fish perception data driven cultivation running state modeling system and method

CN122018315ACN 122018315 ACN122018315 ACN 122018315ACN-122018315-A

Abstract

The invention relates to the technical field of aquaculture running state sensing and management, and discloses a bionic fish sensing data driven aquaculture running state modeling system and method, wherein the system comprises the steps of continuously sensing and collecting water disturbance caused by fish swarm activities, fish space migration track and local environment response influenced by behavior state change to form a sensing data set; constructing a time recursive behavior change description sequence along the direction of a fish body movement path to generate a multi-scale behavior evolution characteristic expression, extracting candidate state fragments by identifying structural inflection points of a behavior mode from stable transition to stress, mapping to corresponding cultivation operation environment units to generate a local operation state association structure, and carrying out self-adaptive adjustment on acquisition density and characteristic weights of bionic fish perception data in different water areas and time periods to form a state representation result of a current cultivation operation situation. The invention has the advantages of improving the timeliness and accuracy of the identification of the cultivation running state.

Inventors

  • WANG DI
  • MENG QINGYU
  • XIAO YANG

Assignees

  • 深圳大深传感科技有限公司

Dates

Publication Date
20260512
Application Date
20260203

Claims (10)

  1. 1. A modeling method for a cultivation running state driven by bionic fish perception data is characterized by comprising the following steps: Continuously sensing and collecting water disturbance, fish space migration track and local environment response influenced by behavior state change caused by fish swarm activities, and performing space attribution reforming on the collected original sensing data set based on the fish activity range and the water space continuity characteristic to form a sensing data set which can be unfolded along the fish movement path; Based on a perception data set, constructing a time recursive behavior change description sequence along the direction of a fish body movement path, introducing behavior offset between a fish shoal historical activity baseline and a current acquisition period, and generating a multi-scale behavior evolution characteristic expression; According to the multi-scale behavior evolution characteristic expression, extracting candidate state fragments reflecting abnormal germination of the culture running state by identifying structural inflection points of the behavior mode from stable transition to stress; Mapping the candidate state fragments to corresponding cultivation operation environment units, and establishing a corresponding relation between the candidate state fragments and the cultivation operation environment by combining a water body circulation mode, cultivation density distribution and recent operation records to generate a local operation state association structure; Based on the local running state association structure, the acquisition density and the characteristic weight of the bionic fish perception data in different water areas and time periods are adaptively adjusted, and a state representation result of the current cultivation running situation is formed.
  2. 2. The method for modeling a culture running state driven by bionic fish perception data according to claim 1, wherein the continuous sensing and collecting process of water disturbance, fish space migration track and local environment response influenced by behavior state change caused by fish shoal activities is as follows: A bionic fish sensing unit with autonomous swimming capability is arranged, and a water flow velocity perturbation signal, relative displacement information of the fish body and environmental response parameters of corresponding areas are synchronously obtained in the cruising movement of different water depths and space areas; Performing time stamp marking and local caching on the data acquired by each bionic fish sensing unit, and completing time consistency correction of multi-unit sensing data based on a wireless synchronization mechanism; Uploading the time-corrected perception data to a culture operation management node to form an original perception data set containing water disturbance, fish migration and environmental response.
  3. 3. The method for modeling a simulated fish-aware data-driven farming operation state of claim 2, wherein the step of forming a set of sensory data deployable along a path of fish motion comprises: according to the space positioning information of the bionic fish sensing unit, determining the water space position and the associated movable range corresponding to each sensing data; based on the spatial communication relation of the culture water body, carrying out path association mapping and region merging on data collected by the original perception data at different moments and different spatial positions; and sequentially aligning and expanding the merged sensing data set according to the actual motion path of the bionic fish to generate a sensing data set which can be expanded along the motion path of the fish body.
  4. 4. The modeling method for a bionic fish perception data driven cultivation running state according to claim 3, wherein the process of constructing a time recursive behavior change description sequence along the direction of a fish body movement path is as follows: using continuous space nodes on the fish body motion path corresponding to the perception data set as indexes, and arranging corresponding data according to time sequence; calculating the variation of the fish body motion state, the water disturbance intensity and the environmental response parameters based on the data difference of adjacent nodes in the sensing data set between adjacent time slices; The variable quantity is serially connected in a time recursion mode, and spatially continuous data in the data set are converted into time evolution description, so that a behavior change description sequence representing the evolution of fish shoal behavior along with time is formed.
  5. 5. The method for modeling a culture operation state driven by bionic fish perception data according to claim 4, wherein the process of generating the multi-scale behavioral evolution characteristic expression is as follows: Extracting normal behavior characteristics of the fish shoal under the water area and season conditions corresponding to the behavior change description sequence from the historical culture operation data, and constructing a fish shoal historical activity baseline; Comparing the current behavior change description sequence with a historical activity baseline time-period by time-period, and calculating the offset of the behavior change description sequence relative to a normal state in each time period to form a behavior offset sequence arranged according to time indexes; and based on the behavior offset sequence, performing aggregation and unfolding processing on different time scales to generate multi-scale behavior evolution characteristic expression reflecting the behavior evolution characteristics of the fish shoal.
  6. 6. The method for modeling a culture operation state driven by bionic fish perception data according to claim 5, wherein the process of extracting candidate state fragments reflecting abnormal germination of the culture operation state is as follows: Performing continuous time window analysis on the multi-scale behavioral evolution characteristic expression to obtain stability indexes of the behavioral characteristics changing along with time under different time scales; Based on the stability index, identifying a mutation point of the behavior feature in the time dimension, and judging whether the mutation point accords with a feature mode of transition from a stable state to a stress state or not by combining the behavior offset in a corresponding time period; And intercepting the mutation point corresponding time period which is judged to be in accordance with the transition characteristic mode from the multi-scale behavior evolution characteristic expression to form a candidate state segment reflecting abnormal germination of the culture running state.
  7. 7. The method for modeling a simulated fish-aware data driven farming operation state of claim 6, wherein mapping candidate state segments to corresponding farming operation environment units comprises: based on path expansion type sensing data corresponding to the candidate state fragments in the sensing data set, extracting a space motion path of the bionic fish during abnormal behavior and continuous distribution characteristics in a time dimension, and determining a main space coverage range of the abnormal behavior in the culture water body; Matching the main space coverage with a functional partition, an equipment layout area or a water management unit in the aquaculture water according to the concentration degree of the main space coverage in the water space and the space adjacent relation between the aquaculture operation structure; And screening the culture operation environment units consistent with the space characteristics of the abnormal behaviors from the matching result, and establishing a space corresponding relation between the candidate state fragments and the corresponding culture operation environment units.
  8. 8. The method for modeling a simulated fish-aware data driven farming operating state of claim 7, wherein the process of generating a local operating state correlation structure is: Acquiring water body circulation parameters, culture density information and recent operation records corresponding to the culture operation environment units, and forming an operation information set representing the current operation conditions; Performing association matching on candidate state fragments in the corresponding relation between the running information set and the space in the time dimension and the space dimension, and analyzing the influence relation of different running conditions on the generation and evolution of the abnormal behavior fragments; And constructing a local running state association structure for describing the influence relationship between the candidate state fragments and the cultivation running environment unit based on the association matching result.
  9. 9. The method for modeling a state of a farming operation driven by bionic fish perception data according to claim 8, wherein the process of forming the state characterization result of the current farming operation situation is: Identifying a water area range and a corresponding risk degree of concentrated distribution of abnormal behaviors based on an influence relation between the candidate state fragments represented in the local running state association structure and the culture running environment unit; according to the abnormal concentrated water area and risk degree obtained by recognition, adjusting the tour frequency and the sensing data acquisition density of the bionic fish sensing units in the corresponding water area; According to the risk weight of the candidate state segment in the local running state association structure, differential characteristic weights are distributed to the perception data acquired in different time periods; And integrating the acquired sensing data subjected to density and characteristic weight adjustment to generate a state characterization result reflecting the current cultivation running situation.
  10. 10. A bionic fish perception data driven farming operation state modeling system for use in a method according to any one of claims 1-9, comprising: The continuous sensing module is used for continuously collecting water disturbance, fish migration track and local environment response caused by fish swarm activities, and performing space attribution reconstruction on the original sensing data to form a sensing data set which can be unfolded along a fish motion path; The behavior modeling module is used for constructing a behavior change description sequence of time recursion along the fish body motion path based on the perception data set and generating a multi-scale behavior evolution characteristic expression; The abnormality extraction module is used for identifying structural inflection points of the behavior from stable transition to stress from the multi-scale behavior evolution characteristic expression and extracting corresponding candidate state fragments; the association construction module is used for mapping the candidate state fragments with the cultivation operation environment unit and generating a local operation state association structure describing the corresponding relation of the candidate state fragments and the cultivation operation environment unit; And the situation characterization module is used for adaptively adjusting the acquisition density and the characteristic weight of the perception data based on the local running state association structure to form a state characterization result of the current cultivation running situation.

Description

Bionic fish perception data driven cultivation running state modeling system and method Technical Field The invention relates to the technical field of aquaculture operation state sensing and management, in particular to a bionic fish sensing data driven aquaculture operation state modeling system and method. Background At present, in the aquaculture operation management process, usually, by arranging a fixed water quality sensor in an aquaculture water body, continuously monitoring environmental parameters such as water temperature, dissolved oxygen, pH value and the like, and judging an aquaculture operation state based on a preset threshold, however, in an actual aquaculture scene, especially in an outdoor pond or industrial circulating water aquaculture environment, due to factors such as uneven water body flow, centralized feeding, local biological oxygen consumption enhancement and the like, short-time and local microscale environmental anomalies often occur in the water body, the anomalies often do not lead to the whole water quality parameters exceeding an alarm threshold, but obvious physiological stress and behavior changes such as abnormal swimming speed, gathering and floating or avoiding of a specific water area are generated on the fish shoal, the state judgment is mainly carried out based on the environmental parameters of fixed points, the operation anomalies which take the fish body response as a guide are difficult to be perceived in time, the identification lag of the aquaculture operation state is easy to be caused, the optimal intervention opportunity is missed, and the fish stress risk and the aquaculture loss are increased. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a bionic fish perception data driven cultivation running state modeling system and method, which have the advantages of improving the timeliness and accuracy of cultivation running state identification and solve the problems in the background art. In order to achieve the purpose of improving the timeliness and accuracy of the identification of the culture running state, the invention provides the following technical scheme that the method for modeling the culture running state driven by the bionic fish perception data comprises the following steps: Continuously sensing and collecting water disturbance, fish space migration track and local environment response influenced by behavior state change caused by fish swarm activities, and performing space attribution reforming on the collected original sensing data set based on the fish activity range and the water space continuity characteristic to form a sensing data set which can be unfolded along the fish movement path; Based on a perception data set, constructing a time recursive behavior change description sequence along the direction of a fish body movement path, introducing behavior offset between a fish shoal historical activity baseline and a current acquisition period, and generating a multi-scale behavior evolution characteristic expression; According to the multi-scale behavior evolution characteristic expression, extracting candidate state fragments reflecting abnormal germination of the culture running state by identifying structural inflection points of the behavior mode from stable transition to stress; Mapping the candidate state fragments to corresponding cultivation operation environment units, and establishing a corresponding relation between the candidate state fragments and the cultivation operation environment by combining a water body circulation mode, cultivation density distribution and recent operation records to generate a local operation state association structure; Based on the local running state association structure, the acquisition density and the characteristic weight of the bionic fish perception data in different water areas and time periods are adaptively adjusted, and a state representation result of the current cultivation running situation is formed. Preferably, the continuous sensing and collecting process for the water disturbance, the fish space migration track and the local environment response influenced by the behavior state change caused by the fish swarm activity comprises the following steps: A bionic fish sensing unit with autonomous swimming capability is arranged, and a water flow velocity perturbation signal, relative displacement information of the fish body and environmental response parameters of corresponding areas are synchronously obtained in the cruising movement of different water depths and space areas; Performing time stamp marking and local caching on the data acquired by each bionic fish sensing unit, and completing time consistency correction of multi-unit sensing data based on a wireless synchronization mechanism; Uploading the time-corrected perception data to a culture operation management node to form an original perception data set containing water disturbance, fish