Intelligent control system and method for aquaculture

Abstract

The invention provides an intelligent control system and method for aquaculture. The intelligent control system comprises a dissolved oxygen detector, a pH detector, an ammonia nitrogen detector, a redox potentiometer, a structured data front-end processor, an underwater binocular vision observation device, a camera, an unstructured data front-end processor, a cloud processor and front-end equipment. The cloud processor makes a decision according to the information of the structured data front-end processor and the unstructured data front-end processor, and controls the front-end equipment to adjust the operation of the culture pond. According to the aquaculture monitoring system, aquaculture water quality key parameters such as dissolved oxygen, pH value, ammonia nitrogen and oxidation-reduction potential are monitored, aquatic animals are monitored in real time in the aquaculture process, the aquaculture environment is dynamically regulated and controlled, and the intelligence of aquaculture is improved. The intelligent control system for aquaculture can be used as a minimum unit template, is efficient and stable in operation, and facilitates the realization of whole-course real-time monitoring, autonomous regulation and control, multi-link coordinated linkage control and intelligent decision-making of aquaculture through copying and popularization.

Description

Technical Field

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

[0002] As an emerging aquaculture model, factory aquaculture has transformed from a labor-intensive industry to an automated mechanized industry. While developing rapidly, it has also brought many problems due to the lack of technical capabilities and supervision mechanisms, such as overfeeding, blind feeding of drugs and feeding, and unreasonable oxygenation operations. These problems have led to the deterioration of water quality conditions, thereby breeding pathogens and affecting the growth and survival of fish. This has led to prominent problems in factory aquaculture, such as high energy consumption, high labor intensity, serious feed waste, and low intelligence. Therefore, the creation of green, efficient, and intelligent aquaculture factories has become a research hotspot and development trend.

[0003] However, there are prominent problems in smart aquaculture factories, such as difficulty in perceiving information about the breeding group of objects, difficulty in dynamically regulating the breeding environment, and lack of effective intelligent control equipment. In addition, smart aquaculture factories lack modular and standardized replication and promotion models, which to some extent restricts their further development. Summary of the invention

[0004] In response to the above technical problems, the present invention provides an intelligent control system and method for aquaculture, which can monitor aquatic animals in real time during the breeding process, realize dynamic regulation of the breeding environment, and green and efficient control of aquaculture, thereby improving the intelligence of aquaculture.

[0005] The present invention achieves the above technical objectives through the following technical means.

[0006] An intelligent control system for aquaculture, comprising a dissolved oxygen detector, a pH detector, an ammonia nitrogen detector, an oxidation-reduction potential meter, a structured data front-end processor, an underwater binocular vision observation device, a camera, an unstructured data front-end processor, a cloud processor and front-end equipment;

[0007] The dissolved oxygen detector is used to detect the dissolved oxygen in the breeding pond; the pH detector is used to detect the pH value of the breeding pond; the ammonia nitrogen detector is used to detect the ammonia nitrogen content in the breeding pond; the redox potential meter is used to detect the redox potential of the breeding pond; the dissolved oxygen detector, pH detector, ammonia nitrogen detector, and redox potential meter are respectively connected to the structured data front-end processor, and the detected data are transmitted to the structured data front-end processor;

[0008] The underwater binocular vision observation device is used to collect images of aquatic animals underwater in the breeding pond, and the camera is used to collect images of aquatic animals on the surface of the breeding pond; the underwater binocular vision observation device and the camera are respectively connected to the unstructured data front-end processor, and the collected images are transmitted to the unstructured data front-end processor;

[0009] The structured data front-end processor and the unstructured data front-end processor are respectively connected to the cloud processor, and the cloud processor makes decisions based on the information of the structured data front-end processor and the unstructured data front-end processor to control the front-end equipment to adjust the operation of the breeding pond.

[0010] In the above scheme, the front-end equipment includes an inspection robot, a handling robot, a feeding robot, an underwater cleaning robot, an aerator, a water exchange pump and an automatic pH adjustment device;

[0011] The dissolved oxygen detector, pH detector, ammonia nitrogen detector, redox potential meter underwater binocular visual observation device and camera are mounted on the inspection robot;

[0012] The inspection robot, the handling robot, the feeding robot, the underwater cleaning robot, the aerator and the water exchange pump are respectively connected to the cloud processor;

[0013] The inspection robot is used to inspect the corresponding breeding pond according to the instructions of the cloud processor, and transmit the pH value, ammonia nitrogen content, and redox potential data collected in the breeding pond to the structured data front-end processor, and transmit the collected images of underwater aquatic animals in the breeding pond and the images of aquatic animals on the water surface to the unstructured data front-end processor;

[0014] The transport robot is used to transport the feed or underwater cleaning robot to a designated breeding pond according to the instructions of the cloud processor;

[0015] The feeding robot is used to feed aquatic animals in the breeding pond water according to the instructions of the cloud processor;

[0016] The underwater cleaning robot is used to clean the designated breeding pond according to the instructions of the cloud processor;

[0017] The aerator is used to increase oxygen to the designated breeding pond according to the instructions of the cloud processor;

[0018] The water exchange pump is used to exchange water in a designated breeding pond according to instructions from a cloud processor;

[0019] The automatic pH adjustment device is used to adjust the pH of a designated breeding pond by adding alkaline materials or acidic materials according to the instructions of the cloud processor.

[0020] In the above scheme, the structured data front-end processor processes the received pH value, ammonia nitrogen content, and redox potential data of the breeding pond to obtain the environmental change data of the breeding pond and transmit it to the cloud processor;

[0021] The unstructured data front-end processor processes the received images of underwater aquatic animals in the breeding pond and the images of aquatic animals on the water surface, obtains water surface fluctuations, the behavior, quantity, phenotype, and disease conditions of aquatic animals, and transmits them to the cloud processor.

[0022] A control method according to the intelligent control system for aquaculture comprises the following steps:

[0023] The dissolved oxygen detector detects the dissolved oxygen in the breeding pond; the pH detector detects the pH value of the breeding pond; the ammonia nitrogen detector detects the ammonia nitrogen content in the breeding pond; the redox potential meter detects the redox potential of the breeding pond; the dissolved oxygen detector, pH detector, ammonia nitrogen detector, and redox potential meter transmit the detected data to the structured data front-end processor respectively; the structured data front-end processor processes the received pH value, ammonia nitrogen content, and redox potential data of the breeding pond to obtain the environmental change data of the breeding pond and transmit it to the cloud processor;

[0024] The underwater binocular vision observation device collects images of aquatic animals underwater in the breeding pond, and the camera collects images of aquatic animals on the surface of the breeding pond; the underwater binocular vision observation device and the camera respectively transmit the collected images to the unstructured data front-end processor; the unstructured data front-end processor processes the received images of aquatic animals underwater in the breeding pond and the images of aquatic animals on the surface of the water, obtains water surface fluctuations, the behavior, quantity, phenotype, and disease conditions of aquatic animals, and transmits them to the cloud processor;

[0025] The cloud processor issues instructions based on the information from the structured data front-end processor and the unstructured data front-end processor to control the front-end equipment to adjust the operation of the breeding pond.

[0026] In the above scheme, when the cloud processor determines that the oxygen content detected by the dissolved oxygen detector is lower than the preset value, the aerator is controlled to increase the running time for oxygenation; when the oxygen content is higher than the preset value, the aerator is controlled to reduce the running time for oxygen reduction; the cloud processor predicts the dynamic change of dissolved oxygen in the water body through the dissolved oxygen dynamic balance equation to control the running time of the aerator, and the dissolved oxygen dynamic balance equation is:

[0027]

[0028] Where: O(T) is the dissolved oxygen concentration in the aquaculture pond water at time T; O intis the dissolved oxygen at the initial moment; O0(t) is the amount of dissolved oxygen delivered to the aquaculture pond by the aerator per unit time, which is related to the oxygenation efficiency and the circulating water flow rate; O1(t) is the oxygen consumption of the aquaculture object per unit time; O2(t) is the oxygen consumption of the residual bait, feces and water per unit time during the feeding period; b is the oxygen escape from the surface of the aquaculture pond per unit time, which is set as a constant.

[0029] In the above scheme, when the cloud processor determines that the pH value detected by the pH detector is lower than the preset value, the pH automatic adjustment device is controlled to add alkaline materials such as lime; when the cloud processor determines that the pH value detected by the pH detector is too high than the preset value, the pH automatic adjustment device is controlled to add acidic materials such as acetic acid.

[0030] In the above scheme, when the cloud processor determines that the ammonia nitrogen concentration detected by the ammonia nitrogen detector exceeds the threshold, the ammonia nitrogen accumulation equation is used to predict the dynamic changes of ammonia nitrogen in the water body, thereby controlling the water exchange pump to exchange water;

[0031] The ammonia nitrogen accumulation equation is: N(t)=N0-N1, where N0=f(x) is the amount of ammonia nitrogen produced from feed sources per unit time, and it gradually increases linearly with the passage of time N0∝x; N1 is the water purification capacity of the facility per unit time.

[0032] In the above scheme, when the cloud processor determines that the redox potential detected by the redox potential meter is lower than the preset value, the aerator is controlled to increase the redox potential; when the redox potential is higher than the preset value, the feeding robot is controlled to reduce the amount of feed.

[0033] In the above scheme, the cloud-based processor determines the satiety of the fish school by the degree of water surface fluctuation based on the image of aquatic animals on the water surface during feeding captured by the camera processed by the unstructured data front-end processor.

[0034] In the above scheme, the cloud processor collects images of underwater aquatic animals in the breeding pond according to the underwater binocular vision observation device processed by the unstructured data front-end processor, obtains the behavior, quantity, phenotype, and disease status of the aquatic animals, and compares them with the established growth law model library and growth discrimination knowledge base to perform quantitative analysis of aquatic animal behavior, growth status estimation, and disease identification.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] The present invention establishes monitoring of key parameters of aquaculture water quality such as dissolved oxygen, pH value, ammonia nitrogen, and redox potential, monitors aquatic animals in real time during the aquaculture process, realizes dynamic regulation of the aquaculture environment, green and efficient control of aquaculture, and improves the intelligence of aquaculture. The intelligent control system for aquaculture described in the present invention can be used as the minimum unit template, which operates efficiently and stably. Through replication and promotion, it is conducive to realizing real-time monitoring, autonomous regulation, multi-link coordinated linkage control and intelligent decision-making of aquaculture throughout the process. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 is a schematic diagram of an intelligent control system for aquaculture according to one embodiment of the present invention;

[0038] Figure 2 It is a schematic diagram of the smallest unit of an intelligent control system for aquaculture according to an embodiment of the present invention applied in an aquaculture factory.

[0039] In the figure, 1. Dissolved oxygen detector; 2. pH detector; 3. Ammonia nitrogen detector; 4. Redox potential meter; 5. Structured data front-end processor; 6. Underwater binocular vision observation device; 7. Camera; 8. Unstructured data front-end processor; 9. Cloud processor. DETAILED DESCRIPTION

[0040] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention. In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances.

[0041] like Figure 1 and 2 As shown, an intelligent control system for aquaculture includes a dissolved oxygen detector 1, a pH detector 2, an ammonia nitrogen detector 3, an oxidation-reduction potential meter 3, a structured data front-end processor 4, an underwater binocular vision observation device 5, a camera 7, an unstructured data front-end processor 8, a cloud processor 9 and front-end equipment;

[0042] The dissolved oxygen detector 1 is used to detect the dissolved oxygen in the breeding pond; the pH detector 2 is used to detect the pH value of the breeding pond; the ammonia nitrogen detector 3 is used to detect the ammonia nitrogen content in the breeding pond; the redox potential meter 3 is used to detect the redox potential of the breeding pond; the dissolved oxygen detector 1, the pH detector 2, the ammonia nitrogen detector 3, and the redox potential meter 3 are respectively connected to the structured data front-end processor 4, and the detected data are transmitted to the structured data front-end processor 4;

[0043] The underwater binocular vision observation device 5 is used to collect images of aquatic animals underwater in the breeding pond, and the camera 7 is used to collect images of aquatic animals on the surface of the breeding pond; the underwater binocular vision observation device 5 and the camera 7 are respectively connected to the unstructured data front-end processor 8, and the collected images are transmitted to the unstructured data front-end processor 8;

[0044] The structured data front-end processor 4 and the unstructured data front-end processor 8 are respectively connected to the cloud processor 9. The cloud processor 9 makes decisions based on the information of the structured data front-end processor 4 and the unstructured data front-end processor 8, determines the priority of each process, and controls the front-end equipment to adjust the operation of the breeding pond.

[0045] In the above scheme, the front-end equipment includes an inspection robot, a handling robot, a feeding robot, an underwater cleaning robot, an aerator, a water exchange pump and an automatic pH adjustment device;

[0046] The dissolved oxygen detector 1, pH detector 2, ammonia nitrogen detector 3, redox potential meter 3, underwater binocular visual observation device 5 and camera 7 are mounted on the inspection robot;

[0047] The inspection robot, the handling robot, the feeding robot, the underwater cleaning robot, the aerator and the water exchange pump are respectively connected to the cloud processor 9;

[0048] The inspection robot is used to inspect the corresponding breeding pond according to the instructions of the cloud processor 9, and transmit the pH value, ammonia nitrogen content, and redox potential data of the breeding pond to the structured data front-end processor 4, and transmit the collected images of underwater aquatic animals in the breeding pond and the images of aquatic animals on the water surface to the unstructured data front-end processor 8;

[0049] The transport robot is used to transport the feed or underwater cleaning robot to the designated breeding pond according to the instructions of the cloud processor 9;

[0050] The feeding robot is used to feed aquatic animals in the breeding pond water according to the instructions of the cloud processor 9;

[0051] The underwater cleaning robot is used to clean the designated breeding pond according to the instructions of the cloud processor 9;

[0052] The aerator is used to increase oxygen to the designated breeding pond according to the instructions of the cloud processor 9;

[0053] The water exchange pump is used to exchange water in the designated breeding pond according to the instruction of the cloud processor 9;

[0054] The automatic pH adjustment device is used to adjust the pH of a designated breeding pond by adding alkaline materials or acidic materials according to the instructions of the cloud processor 9.

[0055] In the above scheme, the structured data front-end processor 4 processes the received data of pH value, ammonia nitrogen content, and redox potential of the breeding pond to obtain the environmental change data of the breeding pond and transmits it to the cloud processor 9;

[0056] The unstructured data front-end processor 8 processes the received images of underwater aquatic animals in the breeding pond and the images of aquatic animals on the water surface, obtains water surface fluctuations, the behavior, quantity, phenotype, and disease conditions of aquatic animals, and transmits them to the cloud processor 9.

[0057] The data collected by the dissolved oxygen detector 1, pH detector 2, ammonia nitrogen detector 3, redox potential meter 3, underwater binocular visual observation device 5 and camera 7 carried by the inspection robot are conducive to monitoring the key factors of important aquaculture water quality in the breeding pond and the behavior, quantity, phenotype, disease, etc. of typical aquatic animals.

[0058] A control method according to the intelligent control system for aquaculture comprises the following steps:

[0059] The dissolved oxygen detector 1 detects the dissolved oxygen in the breeding pond; the pH detector 2 detects the pH value of the breeding pond; the ammonia nitrogen detector 3 detects the ammonia nitrogen content in the breeding pond; the redox potential meter 3 detects the redox potential of the breeding pond; the dissolved oxygen detector 1, the pH detector 2, the ammonia nitrogen detector 3, and the redox potential meter 3 respectively transmit the detected data to the structured data front-end processor 4; the structured data front-end processor 4 processes the received pH value, ammonia nitrogen content, and redox potential data of the breeding pond to obtain the environmental change data of the breeding pond and transmit it to the cloud processor 9;

[0060] The underwater binocular vision observation device 5 collects images of aquatic animals underwater in the breeding pond, and the camera 7 collects images of aquatic animals on the surface of the breeding pond; the underwater binocular vision observation device 5 and the camera 7 respectively transmit the collected images to the unstructured data front-end processor 8; the unstructured data front-end processor 8 processes the received images of aquatic animals underwater in the breeding pond and the images of aquatic animals on the surface of the water, obtains water surface fluctuations, the behavior, quantity, phenotype, and disease conditions of aquatic animals, and transmits them to the cloud processor 9;

[0061] The cloud processor 9 issues instructions based on the information of the structured data front-end processor 4 and the unstructured data front-end processor 8 to control the front-end equipment to adjust the operation of the breeding pond.

[0062] In the above scheme, when the cloud processor 9 determines that the oxygen content detected by the dissolved oxygen detector 1 is lower than the preset value, the aerator is controlled to increase the running time for oxygenation; when the oxygen content is higher than the preset value, the aerator is controlled to reduce the running time for oxygen reduction; the cloud processing predicts the dynamic change of dissolved oxygen in the water body through the dissolved oxygen dynamic balance equation to control the running time of the aerator, and the dissolved oxygen dynamic balance equation is:

[0063]

[0064] Where: O(T) is the dissolved oxygen concentration in the aquaculture pond water at time T; O int is the dissolved oxygen at the initial moment; O0(t) is the amount of dissolved oxygen delivered to the aquaculture pond by the aerator per unit time, which is related to the oxygenation efficiency and the circulating water flow rate; O1(t) is the oxygen consumption of the aquaculture object per unit time; O2(t) is the oxygen consumption of the residual bait, feces and water per unit time during the feeding period; b is the oxygen escape from the surface of the aquaculture pond per unit time, which is set as a constant.

[0065] Preferably, the dissolved oxygen content in the aquaculture water is controlled at 5-8 mg / L, and should be maintained at at least above 4 mg / L.

[0066] In the above scheme, when the cloud processor 9 determines that the pH value detected by the pH detector 2 is lower than the preset value, the pH automatic adjustment device is controlled to add alkaline materials such as lime; when the cloud processor 9 determines that the pH value detected by the pH detector 2 is too high than the preset value, the pH automatic adjustment device is controlled to add acidic materials such as acetic acid.

[0067] Preferably, in aquaculture, the pH value is set at 6.5 to 8.5.

[0068] In the above scheme, when the cloud processor 9 determines that the ammonia nitrogen concentration detected by the ammonia nitrogen detector 3 exceeds the threshold value, the ammonia nitrogen accumulation equation is used to predict the dynamic change of ammonia nitrogen in the water body to control the water change pump to change water;

[0069] The ammonia nitrogen accumulation equation is: N(t)=N0-N1, where N0=f(x) is the amount of ammonia nitrogen produced from feed sources per unit time, and it gradually increases linearly with the passage of time N0∝x; N1 is the water purification capacity of the facility per unit time.

[0070] Preferably, the ammonia nitrogen concentration during the breeding process should be less than 0.2 mg / L. For water bodies for shrimp and crab farming, the ammonia nitrogen level should be ≤0.1 mg / L; the level for seedling water bodies is more stringent and should be ≤0.035 mg / L. In the above scheme, when the cloud processor 9 determines that the redox potential detected by the redox potential meter 4 is lower than the preset value, the aerator is controlled to increase the redox potential; when the redox potential is higher than the preset value, the feeding robot is controlled to reduce the amount of bait.

[0071] Preferably, in the aquaculture environment, the redox potential ranges from 360 to 390 mV, which is conducive to the metabolic process of microorganisms and is more suitable for aquaculture organisms. In the above scheme, the cloud processor 9 determines the satiety of the fish school by the degree of water surface fluctuation based on the image of aquatic animals on the water surface captured by the camera 7 processed by the unstructured data front-end processor 8 during feeding;

[0072] Water surface fluctuation degree = α(L / H)C

[0073] Among them, H is the wave height; L is the wavelength; C is the wave speed; α is a dimensionless proportional factor, which is used to adjust the influence weight of different parameters on the fluctuation degree index. Its value needs to be calibrated or determined experimentally according to the actual situation;

[0074] If the image of aquatic animals on the water surface during feeding, i.e., the water surface fluctuation caused by fierce competition for food in the image of the school of fish feeding, is higher than the maximum threshold value A0 of the water surface fluctuation, it is determined that the school of fish is highly hungry and needs to be fed more;

[0075] If the water surface fluctuation degree caused by the fish feeding in the fish school feeding image is within the water surface fluctuation degree threshold A0-A1, it is determined that the hunger level of the fish school is moderate and it is necessary to continue feeding;

[0076] If the fish in the fish feeding image have no obvious reaction to the feeding, and the water surface fluctuation is lower than the set minimum threshold value A1 of the water surface fluctuation, it is determined that the hunger level of the fish is low and there is no need to continue feeding.

[0077] In the above scheme, the cloud processor 9 collects images of underwater aquatic animals in the breeding pond according to the underwater binocular vision observation device 5 processed by the unstructured data front-end processor 8, obtains the behavior, quantity, phenotype, and disease status of the aquatic animals, and compares them with the established growth law model library and growth discrimination knowledge base to perform quantitative analysis of aquatic animal behavior, growth status estimation, and disease identification;

[0078] The number of fish is calculated by the following steps:

[0079] Collect sample data of farmed objects during feeding, calculate biomass with density estimation network, collect samples of farmed ponds at different times, lighting conditions, fish density and fish species combinations, annotate images, mark the position of each fish with a bounding box, crop, scale and normalize images, convert them into point annotations suitable for density estimation, i.e. the center point of each fish, detect fish using YOLO target detection network, and then calculate the number based on the detection results. If the density of fish is very high and it is difficult to distinguish individuals, use U-Net based segmentation method to segment each fish body, and then calculate the number of segmented areas; obtain sample density map of farmed objects, build sample data of farmed objects and overall biomass estimation model. At the same time, use robot inspection method to obtain the number of organisms on the inspection route, compare the number of organisms on the inspection route with the overall biomass, and get the overall biomass.

[0080] The growth law model is established by the following method:

[0081] The growth law was estimated by using the Von Bertalanffy nonlinear growth equation model based on the measured weight, length and height, and a growth law model was established.

[0082] W=A / (1-Be-*t)

[0083] Among them, W is body weight; A is the maximum growth value; k is the instantaneous growth rate; B is a constant; and t is age in months.

[0084] The videos and pictures observed by the underwater binocular vision observation device 5 and the camera 7 are transmitted to the unstructured data front-end processor 8. In a specific embodiment of the present invention, preferably, the unstructured data front-end processor 8 is combined with complex background image restoration and enhancement, target recognition and segmentation, and underwater three-dimensional precision measurement technology for processing and then transmitted to the cloud processor 9. The cloud processor 9 establishes a growth law model library and a growth discrimination knowledge base for quantitative analysis of aquatic animal behavior, growth status estimation, and disease identification.

[0085] The present invention utilizes front-end equipment such as inspection robots, feeding robots, handling robots, aerators, water exchange pumps and other related equipment to complete the processes of inspection, feeding, handling, oxygenation, water exchange and the like, and uploads them to the cloud processor 9. The cloud processor 9 senses and regulates the operating status of the factory-based intelligent breeding system, breeding environment change data, and the growth and biosafety status tracking of the breeding objects.

[0086] The present invention tracks underwater moving targets, measures underwater fish appearance characteristics, estimates growth status, and identifies and predicts diseases. It uses the data from the underwater binocular vision observation device 5 to compare with the database and model library. At the same time, it makes the most appropriate decision based on the laws of aquatic production, completes daily management, operation scheduling, and abnormal disputes, and realizes real-time monitoring and processing during the breeding process.

[0087] The present invention establishes monitoring of key parameters of aquaculture water quality such as dissolved oxygen, pH value, ammonia nitrogen, redox potential, and abnormal processing, which is conducive to solving the problems of insufficient coordinated control capabilities of aquaculture water quality in large-scale aquaculture factories, poor automatic control accuracy, and low control timeliness. It is conducive to achieving real-time monitoring, autonomous control, coordinated linkage control of multiple links, and intelligent decision-making throughout the aquaculture process.

[0088] The present invention uses the underwater binocular visual observation device 5 and the camera 7 for real-time observation, which is conducive to establishing a breeding model and a growth law model library for breeding objects to explore the behavior, quantity, phenotype, and diseases of typical aquatic animals.

[0089] In a specific embodiment of the present invention, a knowledge base for judging the growth of farmed animals is established based on expert experience and professional field data, including the normal range of data such as body length, body width and weight of farmed animals in different growth periods. A basic database of farmed animals is constructed, including farmed species, release time, release quantity, etc.

[0090] In a specific embodiment of the present invention, complex background image restoration and enhancement, target recognition and segmentation, and underwater three-dimensional precise measurement are combined, group behavior and morphological characteristics are acquired through online sampling by machine vision, and then image segmentation and local reconstruction are used to acquire the morphological and behavioral characteristics of some independent individuals in the group, thereby solving the problems of group behavior feature extraction and analysis, and solving the difficulties of precise matching of sparse features of underwater targets, which is beneficial to the study of quantitative analysis of aquatic animal behavior, growth status estimation, and disease identification.

[0091] In a specific embodiment of the present invention, an underwater binocular visual observation device 5 suitable for underwater environment is selected, and it can be moved to a position that can fully cover the breeding area during inspection, so as to capture the real-time status of the breeding object. In addition, the camera 7 has a wireless transmission function, and can instantly transmit the captured image to the unstructured data front-end processor 8, and use the standard network transmission protocol IP for data transmission to ensure the reliability and integrity of the data, and use the transmission encryption method to ensure the security of the data. A variety of imaging technologies such as binocular stereo vision, illumination and white balance, image enhancement, noise suppression, color correction, and image segmentation and recognition are integrated to improve the image quality, and at the same time obtain the real weight data of the breeding object. The convex hull and contour algorithm are used to perform instance segmentation of the breeding object image, and the biological characteristic parameters of the breeding object are extracted, including parameters such as body length, body width, and area. Then, according to the experimental observation data of different breeding objects, the least squares method or the random gradient descent method is used to fit the relationship model between the characteristic parameters of the breeding object and the weight. Combined with the growth discrimination knowledge base and the breeding object basic database, the breeding object is subjected to rule-based growth discrimination, the discrimination result is obtained, and finally the growth discrimination knowledge base is established.

[0092] The intelligent control system for aquaculture described in the present invention can be used as a minimum unit template, which can operate efficiently and stably. Through replication and promotion, it is conducive to achieving real-time monitoring, autonomous regulation, multi-link coordinated linkage control and intelligent decision-making throughout the aquaculture process.

[0093] It should be understood that although this specification is described according to various embodiments, not every embodiment contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

[0094] The series of detailed descriptions listed above are only specific descriptions of feasible embodiments of the present invention. They are not intended to limit the scope of protection of the present invention. All equivalent embodiments or changes that do not deviate from the technical spirit of the present invention should be included in the scope of protection of the present invention.

Claims

1. An intelligent control system for aquaculture, characterized in that: It includes a dissolved oxygen detector (1), a pH detector (2), an ammonia nitrogen detector (3), an oxidation-reduction potential meter (3), a structured data front-end processor (4), an underwater binocular vision observation device (5), a camera (7), an unstructured data front-end processor (8), a cloud processor (9) and front-end equipment; The dissolved oxygen detector (1) is used to detect the dissolved oxygen in the breeding pond; the pH detector (2) is used to detect the pH value of the breeding pond; the ammonia nitrogen detector (3) is used to detect the ammonia nitrogen content in the breeding pond; the redox potential meter (3) is used to detect the redox potential of the breeding pond; the dissolved oxygen detector (1), the pH detector (2), the ammonia nitrogen detector (3), and the redox potential meter (3) are respectively connected to the structured data front-end processor (4), and transmit the detected data to the structured data front-end processor (4); The underwater binocular visual observation device (5) is used to collect images of aquatic animals underwater in the breeding pond, and the camera (7) is used to collect images of aquatic animals on the surface of the breeding pond; the underwater binocular visual observation device (5) and the camera (7) are respectively connected to the unstructured data front-end processor (8), and the collected images are transmitted to the unstructured data front-end processor (8); The structured data front-end processor (4) and the unstructured data front-end processor (8) are respectively connected to the cloud processor (9), and the cloud processor (9) makes decisions based on the information of the structured data front-end processor (4) and the unstructured data front-end processor (8), and controls the front-end equipment to adjust the operation of the breeding pond.

2. The intelligent control system for aquaculture according to claim 1, characterized in that: The front-end equipment includes an inspection robot, a handling robot, a feeding robot, an underwater cleaning robot, an aerator, a water exchange pump and an automatic pH adjustment device; The dissolved oxygen detector (1), pH detector (2), ammonia nitrogen detector (3), oxidation-reduction potential meter (3), underwater binocular visual observation device (5) and camera (7) are mounted on the inspection robot; The inspection robot, the transport robot, the feeding robot, the underwater cleaning robot, the aerator and the water exchange pump are respectively connected to the cloud processor (9); The inspection robot is used to inspect the corresponding breeding pond according to the instructions of the cloud processor (9), and transmit the pH value, ammonia nitrogen content, and redox potential data collected in the breeding pond to the structured data front-end processor (4), and transmit the collected images of underwater aquatic animals in the breeding pond and the images of aquatic animals on the water surface to the unstructured data front-end processor (8); The transport robot is used to transport the feed or the underwater cleaning robot to a designated breeding pond according to the instructions of the cloud processor (9); The feeding robot is used to feed aquatic animals in the breeding pond water according to the instructions of the cloud processor (9); The underwater cleaning robot is used to clean the designated breeding pond according to the instructions of the cloud processor (9); The aerator is used to increase oxygen in the designated breeding pond according to the instruction of the cloud processor (9); The water exchange pump is used to exchange water in a designated breeding pond according to instructions from a cloud processor (9); The automatic pH adjustment device is used to adjust the pH of a designated breeding pond by adding alkaline materials or acidic materials according to the instructions of the cloud processor (9).

3. The intelligent control system for aquaculture according to claim 1, characterized in that: The structured data front-end processor (4) processes the received data on pH value, ammonia nitrogen content, and redox potential of the breeding pond to obtain environmental change data of the breeding pond and transmits the data to the cloud processor (9); The unstructured data front-end processor (8) processes the received images of underwater aquatic animals in the breeding pond and the images of aquatic animals on the water surface, obtains water surface fluctuations, the behavior, quantity, phenotype, and disease conditions of the aquatic animals, and transmits them to the cloud processor (9).

4. A control method for an intelligent control system for aquaculture according to any one of claims 1 to 3, characterized in that: The following steps are involved: The dissolved oxygen detector (1) detects the dissolved oxygen in the breeding pond; the pH detector (2) detects the pH value of the breeding pond; the ammonia nitrogen detector (3) detects the ammonia nitrogen content in the breeding pond; the redox potential meter (3) detects the redox potential of the breeding pond; the dissolved oxygen detector (1), the pH detector (2), the ammonia nitrogen detector (3), and the redox potential meter (3) respectively transmit the detected data to the structured data front-end processor (4); the structured data front-end processor (4) processes the received pH value, ammonia nitrogen content, and redox potential data of the breeding pond to obtain the environmental change data of the breeding pond and transmits it to the cloud processor (9); The underwater binocular visual observation device (5) collects images of aquatic animals underwater in the breeding pond, and the camera (7) collects images of aquatic animals on the surface of the breeding pond; the underwater binocular visual observation device (5) and the camera (7) respectively transmit the collected images to the unstructured data front-end processor (8); the unstructured data front-end processor (8) processes the received images of aquatic animals underwater in the breeding pond and the images of aquatic animals on the surface of the water, obtains water surface fluctuations, the behavior, quantity, phenotype, and disease status of aquatic animals, and transmits them to the cloud processor (9); The cloud processor (9) issues instructions based on the information from the structured data front-end processor (4) and the unstructured data front-end processor (8) to control the front-end equipment to adjust the operation of the breeding pond.

5. The control method of the intelligent control system for aquaculture according to claim 4, characterized in that: When the cloud processor (9) determines that the oxygen content detected by the dissolved oxygen detector (1) is lower than a preset value, the aerator is controlled to increase the operating time to increase oxygen; when the oxygen content is higher than the preset value, the aerator is controlled to reduce the operating time to reduce oxygen; the cloud processor predicts the dynamic change of dissolved oxygen in the water body through the dissolved oxygen dynamic balance equation to control the operating time of the aerator, and the dissolved oxygen dynamic balance equation is: Where: O(T) is the dissolved oxygen concentration in the aquaculture pond water at time T; O int is the dissolved oxygen at the initial moment; O0(t) is the amount of dissolved oxygen delivered to the aquaculture pond by the aerator per unit time, which is related to the oxygenation efficiency and the circulating water flow rate; O1(t) is the oxygen consumption of the aquaculture object per unit time; O2(t) is the oxygen consumption of the residual bait, feces and water per unit time during the feeding period; b is the oxygen escape from the surface of the aquaculture pond per unit time, which is set as a constant.

6. The control method of the intelligent control system for aquaculture according to claim 4, characterized in that: When the cloud processor (9) determines that the pH value detected by the pH detector (2) is lower than a preset value, the pH automatic adjustment device is controlled to add an alkaline material such as lime; when the cloud processor (9) determines that the pH value detected by the pH detector (2) is too high than the preset value, the pH automatic adjustment device is controlled to add an acidic material such as acetic acid.

7. The control method of the intelligent control system for aquaculture according to claim 4, characterized in that: When the cloud processor (9) determines that the ammonia nitrogen concentration detected by the ammonia nitrogen detector (3) exceeds a threshold value, the ammonia nitrogen accumulation equation is used to predict the dynamic change of ammonia nitrogen in the water body, thereby controlling the water exchange pump to exchange water; The ammonia nitrogen accumulation equation is: N(t)=N0-N1, where N0=f(x) is the amount of ammonia nitrogen produced from feed sources per unit time, and it gradually increases linearly with the passage of time N0∝x; N1 is the water purification capacity of the facility per unit time.

8. The control method of the intelligent control system for aquaculture according to claim 4, characterized in that: When the cloud processor (9) determines that the redox potential detected by the redox potential meter (4) is lower than a preset value, it controls the aerator to increase the redox potential; when the redox potential is higher than the preset value, it controls the feeding robot to reduce the amount of feeding.

9. The control method of the intelligent control system for aquaculture according to claim 4, characterized in that: The cloud processor (9) determines the satiety of the fish school by the degree of water surface fluctuation based on the image of aquatic animals on the water surface captured by the camera (7) during feeding and processed by the unstructured data front-end processor (8).

10. The control method of the intelligent control system for aquaculture according to claim 4, characterized in that: The cloud processor (9) collects images of underwater aquatic animals in the breeding pond based on the underwater binocular visual observation device (5) processed by the unstructured data front-end processor (8), obtains the behavior, quantity, phenotype, and disease status of the aquatic animals, and compares them with the established growth law model library and growth discrimination knowledge library to perform quantitative analysis of aquatic animal behavior, growth status estimation, and disease identification.

Images