An aquaculture supervision system based on multimodal aquaculture data analysis

Abstract

The present invention relates to the technical field of aquaculture supervision. The present invention relates to an aquaculture supervision system based on multimodal aquaculture data analysis. It includes a multimodal data acquisition unit, an aquaculture data classification unit, an aquaculture growth analysis unit, a healthy growth analysis unit, and a feeding analysis unit. The multimodal data acquisition unit is used to obtain water quality data of an aquaculture area, and at the same time collect video data of fish in the aquaculture area, as well as feeding records of the aquaculture area; by performing quantitative statistics on the fish video data corresponding to the healthy growth rate, combining the feeding records to construct a "growth stage-fish quantity-feeding amount" historical database, forming a reusable feeding strategy model, matching the historical database based on real-time fish quantity and body shape data, and calculating the recommended feeding amount through arithmetic average to avoid overfeeding or underfeeding, thereby optimizing feed costs and reducing the risk of water pollution.

Description

Technical Field

[0001] The present invention relates to the technical field of aquatic product supervision, and in particular to an aquatic product supervision system based on multimodal aquaculture data analysis. Background Art

[0002] In aquaculture, feeding management is a key factor influencing aquaculture profitability. Existing technologies rely primarily on the experience of aquaculture personnel or on programmed feeding patterns with fixed times and dosages. The core purpose of these models is to meet fish growth needs and maintain basic aquaculture operations.

[0003] However, in large-scale aquaculture scenarios, the number of fish is huge and the growth cycle is long, which requires a high level of accuracy in feeding strategies. The existing feeding methods cannot obtain multi-dimensional data such as fish number, body shape changes, and water quality environment in real time. Feeding only relies on experience or fixed procedures, which can easily lead to overfeeding, resulting in feed waste and water pollution, or insufficient feeding, affecting fish growth rate, which directly affects breeding cost control and fish health status. In order to reduce this situation, an aquaculture supervision system based on multimodal breeding data analysis is proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide an aquaculture supervision system based on multimodal aquaculture data analysis to solve the problems raised in the above background technology.

[0005] To achieve the above objectives, an aquaculture supervision system based on multimodal aquaculture data analysis is provided, which includes a multimodal data acquisition unit, an aquaculture data classification unit, an aquaculture growth analysis unit, a health growth analysis unit, and a feeding analysis unit;

[0006] The multimodal data acquisition unit is used to obtain water quality data of the aquaculture area, and simultaneously collect video data of fish in the aquaculture area, as well as record feeding in the aquaculture area;

[0007] The breeding data classification unit is used to analyze the body size data of fish in the aquaculture area based on historical fish video data, and save the breeding period according to the development trend of the body size data, and analyze the growth period by combining the breeding period with the feeding record of the same period;

[0008] The aquaculture growth analysis unit is used to perform health analysis based on water quality data, screen and retain growth periods based on the health analysis results, and then combine the retained growth periods with body size data to perform initial growth rate analysis at different stages to obtain the initial growth rate range of different growth stages of fish;

[0009] The healthy growth analysis unit is used to set a growth rate threshold, perform a comparison analysis based on the growth rate threshold and the initial growth rate range, determine the healthy growth rate range based on the analysis results, and then extract the feeding records and fish video data corresponding to the healthy growth rate;

[0010] The feeding analysis unit is used to analyze the number of fish based on the fish video data extracted by the health growth analysis unit, obtain the feeding records corresponding to different numbers of fish at different fish growth stages based on the analysis results, and then analyze the feeding quantity in combination with the real-time fish video data, and send the analyzed feeding quantity to the aquatic supervision end.

[0011] As a further improvement of the present technical solution, the multimodal data acquisition unit establishes a data transmission connection with the aquaculture supervision end, thereby obtaining water quality data collected by water quality sensors in the aquaculture area, fish video data collected by cameras in the aquaculture area, and feeding records corresponding to each feeding in the aquaculture area from the aquaculture management end.

[0012] As a further improvement of the present technical solution, in the multimodal data acquisition unit, the real-time data is the data of the process of placing fry in the aquaculture area before the fish are salvaged, and the historical data is the data other than the real-time data.

[0013] As a further improvement of this technical solution, the breeding data classification unit includes a body shape analysis module and a growth period analysis module;

[0014] The body shape analysis module is used to analyze the body shape data of fish in the aquaculture area based on the historical fish video data, and determine the fish body shape data corresponding to the timestamp of the historical fish video data based on the analysis results;

[0015] The growth period analysis module is used to perform fry breeding analysis in the aquaculture area based on fish video data. When it is detected that fry are re-placed into the aquaculture area, the breeding period is terminated and a new breeding period is started. At the same time, the feeding records within the corresponding storage time range are extracted according to the storage time of the breeding period. The growth period is analyzed based on the feeding records, and the breeding period is divided into multiple growth periods.

[0016] As a further improvement of the present technical solution, the body shape analysis module collects images of fish in the aquaculture area, calculates the body shape of the collected fish images, and then aggregates all the fish images to calculate an average value, and uses the calculated average value as the fish body shape data of the aquaculture area;

[0017] In the growth period analysis module, the breeding period is from the entry of fry to the entry of new fry, and the time between the two fry entries is regarded as the breeding period;

[0018] At the same time, the growth period is from the start of feeding to the next feeding, and the time between two feedings is regarded as the growth period.

[0019] As a further improvement of this technical solution, the breeding growth analysis unit includes a health analysis module and an initial analysis module;

[0020] The health analysis module is used to set a health threshold for water quality data, and then compare the historical water quality data corresponding to the growth period with the health threshold. If the historical water quality data is not within the health threshold, it is judged to be unhealthy and will not be retained. Conversely, if the historical water quality data is within the health threshold, it is judged to be healthy and will be retained.

[0021] The initial analysis module is used to combine the healthy growth periods retained by the health analysis module with the body size data to perform body size growth analysis in adjacent periods, obtain the body size growth rate between the healthy growth periods, and then summarize the body size growth rates corresponding to all growth periods to obtain the initial growth rate range of different growth stages of fish.

[0022] As a further improvement of this technical solution, the healthy growth analysis unit includes a threshold setting module and a growth comparison module;

[0023] The threshold setting module is used to set growth rate thresholds for different growth stages of fish;

[0024] The growth comparison module is used to compare and analyze the growth rate thresholds of the same growth stage with the initial growth rate range, delete the growth rates that do not meet the growth rate thresholds in the initial growth rate range, and retain only the growth rates that meet the growth rate thresholds, thereby forming a healthy growth rate range, and extract feeding records and fish video data for the time period corresponding to the healthy growth rate range.

[0025] As a further improvement of the present technical solution, the feeding analysis unit includes a feeding analysis module and a feeding recommendation module;

[0026] The feeding analysis module is used to analyze the number of fish in the video data of fish within the healthy growth rate range, and then match the analysis results with the feeding records to obtain the feeding records corresponding to different numbers of fish at different fish growth stages;

[0027] The feeding recommendation module is used to analyze the fish quantity and body size data based on real-time fish video data, match the real-time fish quantity and real-time body size data obtained from the analysis with the feeding records obtained by the feeding analysis module, obtain the feeding records corresponding to the real-time fish video data, and then analyze the average feeding quantity of the feeding records, and send the analyzed feeding quantity to the aquatic supervision end, so that the aquatic supervision end feeds the fish according to the analyzed feeding quantity.

[0028] As a further improvement of this technical solution, the steps of performing the average feeding quantity analysis in the feeding recommendation module are as follows:

[0029] ;

[0030] Among them, F match To match the historical feeding record set of real-time data, f ak is the kth feeding amount corresponding to the ath number of fish, q a is the number of fish in the historical data, N target is the historical fish population set corresponding to the target growth stage, N real is the number of fish detected in real time, β is the number matching tolerance, F target A collection of all historical feeding records corresponding to the target growth stage;

[0031] ;

[0032] in, is the recommended average feeding amount, |F match | is the set F match The number of elements in .

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] 1. This aquaculture monitoring system based on multimodal aquaculture data analysis counts fish video data corresponding to healthy growth rates, and combines feeding records to build a "growth stage-fish number-feeding amount" historical database to form a reusable feeding strategy model. Based on the real-time fish number and body shape data matching the historical database, the recommended feeding amount is calculated through arithmetic average to avoid overfeeding or underfeeding, achieve feed cost optimization and reduce water pollution risks. At the same time, the recommended feeding amount is sent to the aquaculture monitoring end, supporting automatic feeding or manual confirmation, improving the flexibility and efficiency of aquaculture operations, and is particularly suitable for standardized management in large-scale aquaculture scenarios.

[0035] 2. In this aquaculture monitoring system based on multimodal aquaculture data analysis, fish body shape is quantitatively analyzed through image recognition technology, and the aquaculture process is divided into multiple growth periods in combination with feeding records to achieve segmented and refined management of the growth process. At the same time, water quality health thresholds are set based on industry standards or historical experience, and water quality data within the growth period is automatically compared. Unhealthy periods are eliminated, and only stages that meet growth conditions are retained for growth rate analysis to ensure data validity. The initial growth rate is then calculated and compared with the healthy growth rate threshold to establish a growth rate range for different growth stages, providing a quantitative basis for early warning of growth abnormalities during the aquaculture process. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is the overall structural principle diagram of the present invention. DETAILED DESCRIPTION

[0037] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0038] See also Figure 1 As shown, the purpose of this embodiment is to provide an aquaculture supervision system based on multimodal aquaculture data analysis, including a multimodal data acquisition unit, an aquaculture data classification unit, an aquaculture growth analysis unit, a healthy growth analysis unit and a feeding analysis unit.

[0039] The multimodal data acquisition unit is used to obtain water quality data in the aquaculture area, collect video data of fish in the aquaculture area, and record feeding in the aquaculture area;

[0040] The multimodal data acquisition unit establishes a data transmission connection with the aquaculture supervision end, thereby obtaining water quality data collected by water quality sensors in the aquaculture area from the aquaculture management end (based on the established transmission connection, the water quality data of the aquaculture area collected by the water quality sensors are extracted from the corresponding module of the aquaculture management end according to the set transmission protocol, including multi-dimensional indicators such as temperature, pH value, and dissolved oxygen content), fish video data collected by the camera device in the aquaculture area (access the camera device data storage module in the aquaculture management end to obtain fish video data collected by the camera device in the aquaculture area, covering video content of different time periods and different aquaculture areas for subsequent analysis), and feeding records corresponding to each feeding in the aquaculture area (extract the feeding records corresponding to each feeding from the feeding management-related module of the aquaculture management end, including feeding time, feeding type, feeding quantity and other information, and synchronize these records to the aquaculture supervision system through the data transmission connection).

[0041] Clarify the connection between the aquatic supervision end and the aquatic management end, determine the permissions and channels for obtaining data from the aquatic management end, and clearly define the three types of data to be obtained: water quality data collected by water quality sensors, fish video data collected by cameras, and feeding records;

[0042] In the multimodal data acquisition unit, real-time data refers to the data collected when fry are placed in the aquaculture area but the fish are not yet harvested, and historical data refers to data other than real-time data.

[0043] The breeding data classification unit is used to analyze the body size data of fish in the aquaculture area based on historical fish video data. At the same time, it saves the breeding period according to the development trend of the body size data, and combines the breeding period with the feeding records of the same period to analyze the growth period.

[0044] The breeding data classification unit includes body shape analysis module and growth period analysis module;

[0045] The body shape analysis module is used to analyze the body shape data of fish in the aquaculture area based on the historical fish video data, and determine the timestamp of the historical fish video data corresponding to the fish body shape data based on the analysis results;

[0046] The body shape analysis module collects images of fish in the aquaculture area, calculates the body shape of the collected fish images, and then summarizes all the fish images to calculate the average value. The calculated average value is used as the fish body shape data of the aquaculture area. The specific steps are as follows:

[0047] Historical video data screening: Filter historical fish video data for analysis from the multimodal data acquisition unit, and clarify the video time range required for analysis and the video content corresponding to the breeding area;

[0048] Fish image acquisition: For selected historical fish video data, image capture is performed at set time intervals (such as every second, every several frames) to obtain image samples of fish in the breeding area, covering images of different fish individuals at different time periods;

[0049] Image body size calculation: For each collected fish image, image recognition and processing technology (such as deep learning-based target detection and image segmentation algorithms) is used to identify the fish outline and key parts (such as the head, tail, and body), and calculate the body parameters of each fish, such as body length, body width, and estimated weight. Then, the body parameters of the fish calculated in all images are collected, and the same type of parameters (such as body length and weight) are summarized separately. The average value of the parameter in all images is calculated, and the average value of each parameter is integrated as the body size data of the fish in the aquaculture area and associated with the corresponding video timestamp. The formula is as follows:

[0050] ;

[0051] in, is the average value of the body parameters of fish in the breeding area, n is the total number of fish images collected for analysis, l i is the body parameter (body length) value of a single fish identified and calculated in the i-th image.

[0052] The growth period analysis module is used to analyze the fry breeding in the aquaculture area based on the fish video data. When it is detected that the fry are re-placed into the aquaculture area, the breeding period is terminated and a new breeding period is started. At the same time, the feeding records within the corresponding storage time range are extracted according to the storage time of the breeding period. The growth period is analyzed based on the feeding records and the breeding period is divided into multiple growth periods.

[0053] In the growth period analysis module, the breeding period is from the entry of fry to the entry of new fry, and the time between the two fry entries is regarded as the breeding period;

[0054] At the same time, the growth period is from the start of feeding to the next feeding, and the time between two feedings is regarded as the growth period. The specific steps are as follows:

[0055] Fish fry release monitoring: Continuously analyze fish video data and use image recognition technology (such as target detection algorithms to identify fish fry characteristics) to monitor whether new fish fry are released in the aquaculture area;

[0056] Breeding period division: When new fry are detected entering, the time period from the last fry entry to the current new fry entry is marked as a breeding period, the old breeding period ends, and the timing of the new breeding period is started, recording the start (time of the previous fry entry) and end (time of the current new fry entry) of each breeding period;

[0057] Feeding record extraction: For the divided breeding period, according to the start-end time range of the saved period, all feeding records within the time period are filtered out from the feeding record data to clarify the time and amount of each feeding;

[0058] Growth period division: Taking the feeding time in the feeding record as the node, the time period between two adjacent feedings is divided into the growth period, that is, the time interval from the start of the first feeding (feeding time point) to the start of the next new feeding (new feeding time point). In this way, the breeding period is subdivided into multiple growth periods, and the corresponding relationship between the breeding period and the growth period is established.

[0059] The aquaculture growth analysis unit is used to perform health analysis based on water quality data, screen and retain growth periods based on the health analysis results, and then combine the retained growth periods with body size data to perform initial growth rate analysis at different stages to obtain the initial growth rate range of fish at different growth stages;

[0060] The breeding growth analysis unit includes a health analysis module and an initial analysis module;

[0061] The health analysis module is used to set health thresholds for water quality data, and then compare the historical water quality data corresponding to the growth period with the health thresholds. If the historical water quality data is not within the health threshold, it is judged to be unhealthy and will not be retained. Conversely, if the historical water quality data is within the health threshold, it is judged to be healthy and will be retained.

[0062] Based on the growth needs of aquaculture species, industry standards or historical experience, health threshold ranges are set for key water quality indicators (such as temperature, pH value, dissolved oxygen content, ammonia nitrogen content, etc.). For example, if the water temperature suitable for fish growth is 20-30℃, the water temperature health threshold is set at 20℃-30℃.

[0063] The initial analysis module is used to combine the healthy growth periods retained by the health analysis module with the body data to perform body growth analysis in adjacent periods, obtain the body growth rate between healthy growth periods, and then summarize the body growth rates corresponding to all growth periods to obtain the initial growth rate range of fish at different growth stages.

[0064] Select adjacent healthy growth periods in chronological order, and based on the body shape data of the previous period, calculate the body shape change of the next period relative to the previous period. For example, first take period A and the period B that follows it, and subtract the body shape data of period A from the body shape data of period B to obtain the body shape change from period A to B;

[0065] Calculate the growth rate based on the changes in body size during adjacent time periods. Collect the growth rates calculated for each pair of adjacent healthy growth periods and then categorize and summarize them according to the fish's growth stage (such as the larval stage, growth stage, etc., which can be divided based on the breeding time or body shape characteristics). Organize the growth rate data within each growth stage and determine its minimum, maximum, and average values ​​to obtain the initial growth rate range for different fish growth stages. The formula is as follows:

[0066] ;

[0067] Among them, R j is the body growth rate, S j+1 is the fish size data corresponding to the j+1th healthy growth period, S j is the fish body size data corresponding to the j-th healthy growth period.

[0068] The healthy growth analysis unit is used to set the growth rate threshold, perform comparative analysis based on the growth rate threshold and the initial growth rate range, determine the healthy growth rate range based on the analysis results, and then extract the feeding records and fish video data corresponding to the healthy growth rate;

[0069] The healthy growth analysis unit includes a threshold setting module and a growth comparison module;

[0070] The threshold setting module is used to set the growth rate threshold for different growth stages of fish;

[0071] The breeding process is divided into different growth stages based on the fish growth cycle (such as seedling stage, juvenile stage, and adult stage) or body shape characteristics. For each stage, the corresponding growth rate threshold range is set in combination with industry standards or historical breeding data (for example, the weekly growth rate of body length in the seedling stage must be ≥5%, and in the adult stage it must be ≥2%).

[0072] The growth comparison module is used to compare and analyze the growth rate thresholds of the same growth stage with the initial growth rate range. The growth rates that do not meet the growth rate thresholds in the initial growth rate range are deleted, and only the growth rates that meet the growth rate thresholds are retained, thereby forming a healthy growth rate range. The feeding records and fish video data for the period corresponding to the healthy growth rate range are then extracted.

[0073] The feeding analysis unit is used to analyze the number of fish based on the fish video data extracted by the health growth analysis unit, obtain the feeding records corresponding to different numbers of fish at different fish growth stages based on the analysis results, and then analyze the feeding quantity in combination with the real-time fish video data, and send the analyzed feeding quantity to the aquatic supervision end.

[0074] The feeding analysis unit includes a feeding analysis module and a feeding recommendation module;

[0075] The feeding analysis module is used to analyze the number of fish in the video data of fish within the healthy growth rate range, and then match the analysis results with the feeding records to obtain the feeding records corresponding to different fish numbers at different fish growth stages;

[0076] For fish video data corresponding to the healthy growth rate range, the target detection algorithm is used to identify individual fish in the video and count the number of fish in each time period;

[0077] Classify the fish count by growth stage, match the feeding records within the same stage (including feeding time, quantity, and type), and establish a historical database of "growth stage-fish number-feeding amount";

[0078] The feeding recommendation module is used to analyze the fish quantity and body size data based on real-time fish video data, match the real-time fish quantity and real-time body size data obtained from the analysis with the feeding records obtained by the feeding analysis module, obtain the feeding records corresponding to the real-time fish video data, and then analyze the average feeding quantity of the feeding records. The analyzed feeding quantity is sent to the aquaculture supervision end, so that the aquaculture supervision end can feed the fish according to the analyzed feeding quantity.

[0079] The system analyzes the real-time video stream of the current aquaculture area, identifies the number of fish in real time, and estimates the size of individual fish using image measurement technology. Based on the real-time number and size of fish, the system locates their growth stage (e.g., judging them as seedlings by their size). The system then searches the historical database for feeding records of fish at the same growth stage and with similar numbers of fish, and selects matching feeding data sets.

[0080] The arithmetic average of the feeding amounts in the matching feeding records is calculated to obtain the recommended real-time feeding amount, and the result is sent to the aquaculture supervision end to trigger automatic feeding or manual confirmation of feeding operations. The formula is as follows:

[0081] ;

[0082] Among them, F match To match the historical feeding record set of real-time data, f ak is the kth feeding amount corresponding to the ath number of fish, q a is the number of fish in the historical data, N target is the historical fish population set corresponding to the target growth stage, N real is the number of fish detected in real time, β is the number matching tolerance, F target It is a collection of all historical feeding records corresponding to the target growth stage, including the feeding amount data corresponding to different fish numbers in that stage.

[0083] ;

[0084] in, is the recommended average feeding amount, calculated from the matching feeding records, |F match | is the set F match The number of elements in (i.e., the number of matching feeding records).

[0085] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and improvements may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and improvements fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. An aquaculture monitoring system based on multimodal aquaculture data analysis, characterized by: It includes multimodal data acquisition unit, breeding data classification unit, breeding growth analysis unit, health growth analysis unit and feeding analysis unit; The multimodal data acquisition unit is used to obtain water quality data of the aquaculture area, and simultaneously collect video data of fish in the aquaculture area, as well as record feeding in the aquaculture area; The breeding data classification unit is used to analyze the body size data of fish in the aquaculture area based on historical fish video data, and save the breeding period according to the development trend of the body size data, and analyze the growth period by combining the breeding period with the feeding record of the same period; The breeding data classification unit includes a body shape analysis module and a growth period analysis module; The body shape analysis module is used to analyze the body shape data of fish in the aquaculture area based on the historical fish video data, and determine the fish body shape data corresponding to the timestamp of the historical fish video data based on the analysis results; The growth period analysis module is used to perform fry breeding analysis in the aquaculture area based on the fish video data. When it is detected that fry are re-placed into the aquaculture area, the breeding period is terminated and a new breeding period is started. At the same time, the feeding records within the corresponding storage time range are extracted according to the storage time of the breeding period. The growth period is analyzed based on the feeding records and the breeding period is divided into multiple growth periods. The aquaculture growth analysis unit is used to perform health analysis based on water quality data, screen and retain growth periods based on the health analysis results, and then combine the retained growth periods with body size data to perform initial growth rate analysis at different stages to obtain the initial growth rate range of different growth stages of fish; The breeding growth analysis unit includes a health analysis module and an initial analysis module; The health analysis module is used to set a health threshold for water quality data, and then compare the historical water quality data corresponding to the growth period with the health threshold. If the historical water quality data is not within the health threshold, it is judged to be unhealthy and will not be retained. Conversely, if the historical water quality data is within the health threshold, it is judged to be healthy and will be retained. The initial analysis module is used to combine the healthy growth periods retained by the health analysis module with the body size data to perform body size growth analysis in adjacent periods, obtain the body size growth rate between the healthy growth periods, and then summarize the body size growth rates corresponding to all growth periods to obtain the initial growth rate range of different growth stages of fish; The healthy growth analysis unit is used to set a growth rate threshold, perform a comparison analysis based on the growth rate threshold and the initial growth rate range, determine the healthy growth rate range based on the analysis results, and then extract the feeding records and fish video data corresponding to the healthy growth rate; The healthy growth analysis unit includes a threshold setting module and a growth comparison module; The threshold setting module is used to set growth rate thresholds for different growth stages of fish; The growth comparison module is used to compare and analyze the growth rate threshold of the same growth stage with the initial growth rate range, delete the growth rates that do not meet the growth rate threshold in the initial growth rate range, and only retain the growth rates that meet the growth rate threshold, thereby forming a healthy growth rate range, and extract feeding records and fish video data for the period corresponding to the healthy growth rate range; The feeding analysis unit is used to analyze the number of fish according to the fish video data extracted by the health growth analysis unit, obtain the feeding records corresponding to different numbers of fish at different fish growth stages according to the analysis results, and then analyze the feeding quantity in combination with the real-time fish video data, and send the analyzed feeding quantity to the aquatic supervision end; The feeding analysis unit includes a feeding analysis module and a feeding recommendation module; The feeding analysis module is used to analyze the number of fish in the video data of fish within the healthy growth rate range, and then match the analysis results with the feeding records to obtain the feeding records corresponding to different numbers of fish at different fish growth stages; The feeding recommendation module is used to analyze the fish quantity and body size data based on real-time fish video data, match the real-time fish quantity and real-time body size data obtained from the analysis with the feeding records obtained by the feeding analysis module, obtain the feeding records corresponding to the real-time fish video data, and then analyze the average feeding quantity of the feeding records, and send the analyzed feeding quantity to the aquatic supervision end, so that the aquatic supervision end feeds the fish according to the analyzed feeding quantity.

2. The aquaculture monitoring system based on multimodal aquaculture data analysis according to claim 1, characterized in that: The multimodal data acquisition unit establishes a data transmission connection with the aquaculture supervision end, thereby obtaining water quality data collected by water quality sensors in the aquaculture area, fish video data collected by cameras in the aquaculture area, and feeding records corresponding to each feeding in the aquaculture area from the aquaculture management end.

3. The aquaculture monitoring system based on multimodal aquaculture data analysis according to claim 1, characterized in that: In the multimodal data acquisition unit, the real-time data refers to the data collected during the process of placing fry in the aquaculture area but not fishing the fish, and the historical data refers to the data other than the real-time data.

4. The aquaculture monitoring system based on multimodal aquaculture data analysis according to claim 1, characterized in that: The body shape analysis module collects images of fish in the aquaculture area, calculates the body shape of the collected fish images, and then aggregates all the fish images to calculate the average value, and uses the calculated average value as the fish body shape data of the aquaculture area; In the growth period analysis module, the breeding period is from the entry of fry to the entry of new fry, and the time between the two fry entries is regarded as the breeding period; At the same time, the growth period is from the start of feeding to the next feeding, and the time between two feedings is regarded as the growth period.

5. The aquatic product monitoring system based on multimodal aquaculture data analysis according to claim 1, characterized in that: The steps of the feeding recommendation module for analyzing the average feeding quantity are as follows: ； Among them, F match To match the historical feeding record set of real-time data, f ak is the kth feeding amount corresponding to the ath number of fish, q a is the number of fish in the historical data, N target is the historical fish population set corresponding to the target growth stage, N real is the number of fish detected in real time, β is the number matching tolerance, F target A collection of all historical feeding records corresponding to the target growth stage; ； in, is the recommended average feeding amount, |F match | is the set F match The number of elements in .

Images