A computer vision-based grouper stimulation cryptomyxidosis damage prevention intelligent breeding method and device

Abstract

This invention belongs to the field of aquaculture technology and provides a computer vision-based intelligent aquaculture method and device for the prevention and control of Cryptocaryon fasciculata disease in grouper. The method includes: upper layer image acquisition, grouper swimming and distribution data recognition, abnormal behavior monitoring, bottom image acquisition, cyst data detection, green algae coverage calculation, ultraviolet lamp irradiation, feed supplementation, and overbreeding monitoring. This invention provides data support for automated management through grouper swimming and distribution data, upper layer monitoring results, cyst data, and green algae coverage, reducing reliance on manual labor. Through ultraviolet lamp irradiation strategies, feed supplementation strategies, and overbreeding judgment strategies, automated grouper aquaculture is achieved, reducing damage to fry and improving the efficiency of Cryptocaryon fasciculata disease prevention and control.

Description

Technical Field

[0001] This invention relates to the field of aquaculture technology, and in particular to an intelligent aquaculture method and device for the prevention and control of Cryptocaryon irritans disease in grouper based on computer vision. Background Technology

[0002] Cryptocrystal irritans, a major parasitic ciliate of marine bony fish, has a life cycle consisting of four stages: trophozoite, cyst precursor, cyst, and larva. The cyst stage is the core link in the spread of disease. When the water temperature is between 25°C and 27°C, a single cyst can hatch in 3 to 7 days, releasing hundreds of ciliate larvae. These larvae invade the fish through wounds on the body surface or thin mucus layers, leading to large-scale mortality of fish fry and causing billions of dollars in losses to the global marine aquaculture industry annually.

[0003] Existing prevention and control technologies have significant shortcomings: 1) Outdated monitoring methods: Traditional manual observation relies on the experience of fish farmers, resulting in a missed detection rate of over 30% for cysts, and it is impossible to track the dynamic distribution of cysts in real time; 2) Limited prevention and control methods: Currently, the prevention and control of Cryptocaryoniasis mainly involves chemical control, physical control, and immunotherapy. Chemical control mainly uses chemical drugs such as formalin, methylene blue, and potassium permanganate, alone or in combination, to kill Cryptocaryonia. These drugs can cause varying degrees of harm to fish, users, and the environment, and are prone to drug residues, endangering food safety and the environment. Physical control mainly includes freshwater soaking, temperature control, and ultraviolet light irradiation. Although these methods are safe, they are time-consuming, labor-intensive, and can easily damage fish. Summary of the Invention

[0004] In order to overcome the shortcomings of the existing technology, the purpose of this invention is to provide a smart aquaculture method and device for the prevention and control of Cryptocaryon irritans disease in grouper based on computer vision, which solves the problems of poor monitoring effect and large damage of prevention and control methods in existing aquaculture methods.

[0005] To achieve the above objectives, the present invention provides the following solution:

[0006] A computer vision-based intelligent aquaculture method for the prevention and control of Cryptocaryon irritans disease in grouper includes:

[0007] Images of grouper in the target area were acquired using a top-view camera and a side-view camera inside the first breeding area.

[0008] Based on the top-view and side-view images of the fish, the YOLO11 model is used to detect and track the grouper in the target area, and the swimming and distribution data of the grouper are obtained.

[0009] The average swimming speed and the density of gas stone accumulation are monitored on the swimming and distribution data of the grouper to obtain the upper layer monitoring results. When the upper layer monitoring results exceed the preset normal range, an abnormal behavior alarm is triggered.

[0010] The bottom of the target area is captured by a camera on the inner wall of the second breeding area to obtain an image of the bottom of the bucket;

[0011] Based on the water temperature of the target area, a detection threshold is set, and the YOLO11 model is used to detect the capsules in the target area based on the bottom image of the bucket and the detection threshold to obtain capsule data;

[0012] The fish body top view image is filtered using the RGB color space to obtain a green algae region mask. The proportion of green algae pixels in the target area is calculated based on the green algae region mask to obtain the green algae coverage rate.

[0013] When the upper-layer monitoring results exceed the normal range, if the capsule density in the capsule data is greater than the preset density threshold, then the ultraviolet lamp is activated to irradiate the target area in a pulsed manner.

[0014] When the green algae coverage is less than the preset feeding threshold, the algae-eating fish feed is increased by 30% in the target area within 10 minutes.

[0015] When the water temperature in the target area exceeds the preset temperature range, if the cyst density is greater than the density threshold or the green algae coverage is greater than the over-proliferation threshold, an early warning message is sent to the management personnel.

[0016] Preferably, the average swimming speed and airstone aggregation density of the grouper's swimming and distribution data are monitored to obtain upper-layer monitoring results. When the upper-layer monitoring results exceed a preset normal range, an abnormal behavior alarm is triggered, including:

[0017] The total distance and average swimming speed of a single grouper within a unit time window are calculated based on the grouper swimming and distribution data. If the average swimming speed is greater than a preset speed threshold for 10 consecutive minutes, the grouper swimming and distribution data are judged as abnormal behavior.

[0018] Mark the location of the air stone in the top view image and the side view image of the fish body, and delineate the ROI area with a radius of 0.5m at the location of the air stone. Based on the swimming and distribution data of the grouper, count the number of fish in the ROI area to obtain the fish density data. If the fish density data is greater than the preset fish density for 5 consecutive minutes, the swimming and distribution data of the grouper is judged as abnormal clustering.

[0019] If ≥30% of the fish in a single frame of the fish top-view image and the fish side-view image trigger either the abnormal behavior or the abnormal clustering, and this is repeated ≥3 times within 10 consecutive minutes, an abnormal behavior alarm will be issued.

[0020] Preferably, a detection threshold is set based on the water temperature of the target area, and the YOLO11 model is used to detect capsules in the target area based on the bucket bottom image and the detection threshold to obtain capsule data, including:

[0021] A baseline threshold of 25°C is set, and the difference between the water temperature in the target area and 25°C is calculated to obtain the temperature difference. The threshold is then adjusted downwards by 1 unit per meter for every 2°C increase in temperature. 2 The strategy adjusts the baseline threshold to obtain the detection threshold;

[0022] Based on the detection threshold, the YOLO11 model is used to identify the cysts in the barrel bottom image to obtain the cyst pixel coordinates.

[0023] The capsule pixel coordinates are transformed to a 1m×1m grid with the bottom center of the target area as the origin, and the capsule density is calculated based on the capsule pixel coordinates to obtain the capsule data.

[0024] Preferably, the top-view image of the fish is filtered using the RGB color space to obtain a green algae region mask. The percentage of green algae pixels within the target area is calculated based on the green algae region mask to obtain the green algae coverage rate, including:

[0025] The fish top view image is converted to RGB color space, and the areas with green channels greater than 150 and red channels greater than 80 in the converted fish top view image are extracted to obtain the green algae region mask;

[0026] The green algae coverage rate is obtained by calculating the ratio of the total number of green algae pixels in the green algae region mask to the total number of pixels in the target region.

[0027] Preferably, when the upper-layer monitoring results exceed the normal range, if the capsule density in the capsule data is greater than a preset density threshold, then the ultraviolet lamp is activated to pulse-irradiate the target area, further comprising:

[0028] When the capsule density is detected to be greater than the density threshold, the ultraviolet lamp is turned on and irradiated on the target area for 2 hours with a cycle of 30 seconds on and 10 seconds off.

[0029] If the capsule density is greater than the density threshold after a 1-hour interval following the irradiation, the irradiation duration of the target area will be extended to 4 hours.

[0030] Preferably, when the green algae coverage is less than a preset feeding threshold, the algae-eating fish feed is increased by 30% within 10 minutes in the target area, including:

[0031] When the green algae coverage is less than 15%, increase the amount of algae-eating fish feed in the target area by 30% within ten minutes.

[0032] When the green algae coverage is in the range of 15% to 80%, maintain the feeding amount of algae-eating fish feed;

[0033] When the green algae coverage exceeds 80%, a mesh blockage information message is sent to the management personnel.

[0034] Preferably, a computer vision-based intelligent aquaculture device for the prevention and control of Cryptocaryon irritans disease in grouper includes: a top support, a circulating water inlet, a water quality sensor, a top camera, a feeder, an intelligent aquaculture tank shell, a mesh isolation layer, an upper circulating water outlet, a lower circulating water outlet, a first side wall camera, a second side wall camera, an inner ring of ultraviolet lamps, an outer ring of ultraviolet lamps, and an ultraviolet lamp base.

[0035] The top bracket and the feeder are respectively fixed to the top edge of the intelligent aquaculture tank shell; the water quality sensor and the top camera are fixed to the lower side of the top bracket; the circulating water inlet is located on the top side of the intelligent aquaculture tank shell; the upper circulating water outlet and the lower circulating water outlet are respectively located in the middle and bottom of the intelligent aquaculture tank shell; the first side wall camera and the second side wall camera are respectively located on the inner side of the upper and lower parts of the intelligent aquaculture tank shell; the mesh isolation layer is located in the middle of the interior of the intelligent aquaculture tank shell; the mesh isolation layer is above the upper circulating water outlet; the ultraviolet lamp base is located at the bottom of the interior of the intelligent aquaculture tank shell; the inner ultraviolet lamp ring and the outer ultraviolet lamp ring are located on the ultraviolet lamp base; the outer ultraviolet lamp ring is outside the inner ultraviolet lamp ring.

[0036] The present invention discloses the following technical effects:

[0037] This invention provides a computer vision-based intelligent aquaculture method and device for the prevention and control of Cryptocaryon stimuli disease in grouper. By using grouper swimming and distribution data, upper layer monitoring results, cyst data, and green algae coverage, it solves the problem of poor monitoring effects of existing methods and realizes automated monitoring of the grouper aquaculture environment. Through ultraviolet lamp irradiation strategy, feed supplementation strategy, and over-breeding judgment strategy, it overcomes the shortcomings of existing prevention and control methods that cause significant damage and improves the efficiency of Cryptocaryon stimuli disease prevention and control. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1 A schematic diagram of a computer vision-based intelligent aquaculture process for the prevention and control of Cryptocaryon irritans disease in grouper, provided in an embodiment of the present invention.

[0040] Figure 2 This is a three-dimensional structural diagram of the intelligent breeding tank provided in an embodiment of the present invention;

[0041] Figure 3 This is a three-dimensional structural cross-sectional view of the intelligent breeding tank provided in an embodiment of the present invention;

[0042] Figure 4 This is a schematic diagram of the ultraviolet lamp chassis provided in an embodiment of the present invention;

[0043] Figure 5 The control system logic flowchart provided for an embodiment of the present invention.

[0044] Explanation of reference numerals in attached figures:

[0045] 1-Top bracket, 2-Circulating water inlet, 3-Water quality sensor, 4-Top camera, 5-Feeder, 6-Smart aquaculture tank shell, 7-Mesh isolation layer, 8-Upper circulating water outlet, 9-Lower circulating water outlet, 10-First side wall camera, 11-Second side wall camera, 12-Inner ring of ultraviolet lamp, 13-Outer ring of ultraviolet lamp, 14-Base of ultraviolet lamp panel. Detailed Implementation

[0046] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] The purpose of this invention is to provide a computer vision-based intelligent aquaculture method and device for the prevention and control of Cryptocaryon irritans disease in grouper, which solves the problems of poor monitoring effectiveness and significant damage caused by existing aquaculture methods.

[0048] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0049] Figure 1 This is a schematic diagram of an intelligent aquaculture process for the prevention and control of Cryptocaryon irritans disease in grouper based on computer vision, as provided in an embodiment of the present invention. Figure 1 As shown, this invention provides a computer vision-based intelligent aquaculture method for the prevention and control of Cryptocaryon irritans disease in grouper, comprising:

[0050] Step 100: Use the top camera 4 and the inner side wall camera of the first breeding area to collect images of the grouper in the target area, and obtain a top view image and a side view image of the fish.

[0051] Step 200: Based on the top view image and the side view image of the fish, use the YOLO11 model to detect and track the grouper in the target area to obtain the grouper swimming and distribution data;

[0052] Step 300: Monitor the average swimming speed and the density of gas stones in the swimming and distribution data of the grouper to obtain the upper layer monitoring results. When the upper layer monitoring results exceed the preset normal range, an abnormal behavior alarm is triggered.

[0053] Step 400: Use the camera on the inner wall of the second breeding area to capture images of the bottom of the target area to obtain images of the bottom of the bucket;

[0054] Step 500: Set a detection threshold based on the water temperature of the target area, and use the YOLO11 model to detect the capsules in the target area based on the bottom image of the bucket and the detection threshold to obtain capsule data;

[0055] Step 600: Filter the fish body top view image using RGB color space to obtain a green algae region mask. Calculate the percentage of green algae pixels in the target area based on the green algae region mask to obtain the green algae coverage rate.

[0056] Step 700: When the upper layer monitoring result exceeds the normal range, if the capsule density in the capsule data is greater than the preset density threshold, then the ultraviolet lamp is activated to irradiate the target area in a pulsed manner.

[0057] Step 800: When the green algae coverage is less than the preset feeding threshold, increase the amount of algae-eating fish feed in the target area by 30% within 10 minutes.

[0058] Step 900: When the water temperature in the target area exceeds the preset temperature range, if the cyst density is greater than the density threshold or the green algae coverage is greater than the over-proliferation threshold, an early warning message is sent to the management personnel.

[0059] Specifically, the average swimming speed and airstone aggregation density of the grouper's swimming and distribution data are monitored to obtain upper-layer monitoring results. When the upper-layer monitoring results exceed a preset normal range, an abnormal behavior alarm is triggered, including:

[0060] The total distance and average swimming speed of a single grouper within a unit time window are calculated based on the grouper swimming and distribution data. If the average swimming speed is greater than a preset speed threshold for 10 consecutive minutes, the grouper swimming and distribution data are judged as abnormal behavior.

[0061] Mark the location of the air stone in the top view image and the side view image of the fish body, and delineate the ROI area with a radius of 0.5m at the location of the air stone. Based on the swimming and distribution data of the grouper, count the number of fish in the ROI area to obtain the fish density data. If the fish density data is greater than the preset fish density for 5 consecutive minutes, the swimming and distribution data of the grouper is judged as abnormal clustering.

[0062] If ≥30% of the fish in a single frame of the fish top-view image and the fish side-view image trigger either the abnormal behavior or the abnormal clustering, and this is repeated ≥3 times within 10 consecutive minutes, an abnormal behavior alarm will be issued.

[0063] Preferably, a detection threshold is set based on the water temperature of the target area, and the YOLO11 model is used to detect capsules in the target area based on the bucket bottom image and the detection threshold to obtain capsule data, including:

[0064] A baseline threshold of 25°C is set, and the difference between the water temperature in the target area and 25°C is calculated to obtain the temperature difference. The threshold is then adjusted downwards by 1 unit per meter for every 2°C increase in temperature. 2 The strategy adjusts the baseline threshold to obtain the detection threshold;

[0065] Based on the detection threshold, the YOLO11 model is used to identify the cysts in the barrel bottom image to obtain the cyst pixel coordinates.

[0066] The capsule pixel coordinates are transformed to a 1m×1m grid with the bottom center of the target area as the origin, and the capsule density is calculated based on the capsule pixel coordinates to obtain the capsule data.

[0067] Optionally, the top-view image of the fish is filtered using the RGB color space to obtain a green algae region mask. The percentage of green algae pixels within the target area is calculated based on the green algae region mask to obtain the green algae coverage rate, including:

[0068] The fish top view image is converted to RGB color space, and the areas with green channels greater than 150 and red channels greater than 80 in the converted fish top view image are extracted to obtain the green algae region mask;

[0069] The green algae coverage rate is obtained by calculating the ratio of the total number of green algae pixels in the green algae region mask to the total number of pixels in the target region.

[0070] Optionally, when the upper-layer monitoring results exceed the normal range, if the capsule density in the capsule data is greater than a preset density threshold, then the ultraviolet lamp is activated to pulse-irradiate the target area, further comprising:

[0071] When the capsule density is detected to be greater than the density threshold, the ultraviolet lamp is turned on and irradiated on the target area for 2 hours with a cycle of 30 seconds on and 10 seconds off.

[0072] If the capsule density is greater than the density threshold after a 1-hour interval following the irradiation, the irradiation duration of the target area will be extended to 4 hours.

[0073] Optionally, when the green algae coverage is less than a preset feeding threshold, the algae-eating fish feed is increased by 30% within 10 minutes in the target area, including:

[0074] When the green algae coverage is less than 15%, increase the amount of algae-eating fish feed in the target area by 30% within ten minutes.

[0075] When the green algae coverage is in the range of 15% to 80%, maintain the feeding amount of algae-eating fish feed;

[0076] When the green algae coverage exceeds 80%, a mesh blockage information message is sent to the management personnel.

[0077] refer to Figures 2 to 4 A computer vision-based intelligent aquaculture device for the prevention and control of Cryptocaryon irritans disease in grouper includes: a top support 1, a circulating water inlet 2, a water quality sensor 3, a top camera 4, a feeder 5, an intelligent aquaculture tank shell 6, a mesh isolation layer 7, an upper circulating water outlet 8, a lower circulating water outlet 9, a first side wall camera 10, a second side wall camera 11, an inner ultraviolet lamp ring 12, an outer ultraviolet lamp ring 13, and an ultraviolet lamp plate base 14.

[0078] The top bracket 1 and the feeder 5 are respectively fixed to the top edge of the intelligent breeding tank shell 6; the water quality sensor 3 and the top camera 4 are fixed to the lower side of the top bracket 1; the circulating water inlet 2 is located on the top side of the intelligent breeding tank shell 6; the upper circulating water outlet 8 and the lower circulating water outlet 9 are respectively located in the middle and bottom of the intelligent breeding tank shell 6; the first side wall camera 10 and the second side wall camera 11 are respectively located on the inner side of the upper and lower parts of the intelligent breeding tank shell 6; the mesh isolation layer 7 is located in the middle of the interior of the intelligent breeding tank shell 6; the mesh isolation layer 7 is above the upper circulating water outlet 8; the ultraviolet lamp base 14 is located at the bottom of the interior of the intelligent breeding tank shell 6; the inner ultraviolet lamp ring 12 and the outer ultraviolet lamp ring 13 are located on the ultraviolet lamp base 14; the outer ultraviolet lamp ring 13 is outside the inner ultraviolet lamp ring 12.

[0079] Specifically, the 3D monitoring module includes: camera equipment: consisting of a top underwater camera (resolution ≥1080P, waterproof rating IP68), a camera on the inner wall of the first aquaculture area (180° wide-angle lens), and a camera on the inner wall of the second aquaculture area (waterproof rating IP68), which are fixed to the tank body by corrosion-resistant brackets to achieve 360° image acquisition without blind spots inside the aquaculture tank; sensor unit: integrating a water temperature sensor (accuracy ±0.5℃) to collect water environment parameters in real time.

[0080] Furthermore, the intelligent disinfection module includes: an ultraviolet lamp assembly: employing a 254nm low-pressure mercury lamp (50-100W power), equipped with a parabolic reflector (reflectivity ≥90%) to achieve full coverage irradiation of the cysts at the bottom of the tank; its control logic: when the camera in the second breeding area detects a cyst density >5 cysts / m²... 2 At that time, the ultraviolet lamp is triggered to provide pulsed irradiation (on for 30 seconds, off for 10 seconds), and the cumulative irradiation time is dynamically adjusted according to the capsule density (density > 10 capsules / m³). 2 (At that time, it was extended to 4 hours).

[0081] Preferably, the adaptive feeding system includes: feeding equipment: including an algae monitoring camera (6 megapixels resolution) on the first aquaculture area grid and a quantitative feeder 5 (accuracy ±2%); its control strategy: setting the green algae coverage threshold T1=15%, when the monitoring value C<T1, the feeder 5 automatically increases the feeding amount by 30%.

[0082] Optionally, the edge computing unit incorporates the YOLO11 target detection algorithm, trained based on capsule morphology features, achieving an accuracy of ≥95% and a false detection rate of <5%; the capsule detection threshold is dynamically adjusted based on water temperature (the threshold is lowered by 1 capsule / m for every 2°C increase in water temperature). 2 ), to improve adaptability to complex environments.

[0083] Specifically, feature extraction and analysis: This step utilizes a multimodal data fusion algorithm and a hierarchical feature extraction framework to accurately quantify abnormal grouper behavior, cyst count, and algae coverage. Abnormal grouper behavior recognition (based on an improved YOLO11 algorithm): Image acquisition and preprocessing: Top camera 4 (1080P, 5 frames / second) and the inner wall camera of the first aquaculture area (180° wide-angle, 3 frames / second) acquire real-time top-down and side-view images of the fish. After Gaussian filtering to remove water ripple noise, the images are input into the YOLO11 model for fish detection and tracking.

[0084] Furthermore, the core metrics are calculated as follows:

[0085] 1) Average swimming speed monitoring: Track a single grouper by video stream and calculate its total distance traveled over 300 frames (1 minute) to obtain its average swimming speed V. If V ≥ 10 cm / s is detected for 10 consecutive minutes (normal state V < 5 cm / s), it is judged as "abnormal behavior".

[0086] 2) Airstone Density Detection: Mark the locations of airstones in the image and delineate a ROI region with a radius of 0.5m. Count the number of fish within this region using YOLOv11. When a fish density > 5 fish / m² is detected... 2 If the duration is ≥5 minutes, it is judged as "cluster clustering anomaly".

[0087] 3) Abnormal triggering rules:

[0088] When any of the above-mentioned abnormalities occur in ≥30% of the fish in a single frame, and are triggered repeatedly ≥3 times within 10 consecutive minutes, the edge computing unit generates an abnormal behavior alarm and triggers the capsule detection module to start an expedited scan.

[0089] Preferably, the detection of the number of cysts at the bottom of the tank (precise identification of small targets) involves image acquisition and target recognition: a camera (IP68 waterproof rating) on ​​the inner wall of the second aquaculture area acquires images of the tank bottom at a frequency of 2 frames / second. After contrast enhancement preprocessing, the images are input into the YOLO11 model to detect cysts (training data labels cyst diameter 50-300μm, surface ridge features). Dynamic threshold adjustment: the edge computing unit automatically corrects the detection threshold based on water temperature sensor data (accuracy ±0.5℃): baseline threshold: 5 cysts / m. 2 (At a water temperature of 25℃); for every 2℃ increase in temperature, the threshold is lowered by 1 unit / m. 2 (For example, at 28℃, the threshold is adjusted to 4 per m) 2 This improves the sensitivity of capsule recognition under high-temperature conditions.

[0090] Furthermore, the density quantization method is as follows: the detected capsule pixel coordinates are converted into physical positions through camera calibration parameters, a 1m×1m grid is divided with the center of the bucket bottom as the origin, the number of capsules in each grid is calculated, and the density per unit area is output.

[0091] Specifically, the algae coverage calculation for the grid platform (lightweight morphological analysis) involves image acquisition and segmentation: The algae monitoring camera (6 megapixels) on the grid platform acquires top-down images in real time. The algae regions are separated by RGB color space filtering (extracting green channel values ​​>150 and red channel values ​​>80), and then binarized to generate an algae mask. Coverage calculation: The percentage of algae pixels is calculated using the contour area statistical method: C = (total number of algae pixels / total number of pixels on the grid platform) × 100%. When C < 15%, it is determined that there is insufficient algae, triggering the logic for supplementing algae-eating fish feed.

[0092] Preferably, in the multimodal data fusion and decision-making process, the edge computing unit integrates three types of data through timestamp alignment technology (error < 100ms) and executes hierarchical control logic: Abnormal behavior linkage: if the abnormal behavior alarm and the capsule density > 5 / m 2 Simultaneously triggered, the UV lamp pulse irradiation is immediately started (30s on / 10s off), with an initial duration of 2 hours, which is then dynamically extended according to the capsule density.

[0093] Furthermore, the algae-feeding closed loop is implemented: when the green algae coverage C < 10%, the quantitative feeder 5 automatically increases the feeding amount by 30%, and checks again after 30 minutes. If the standard is still not met, the feed is added again.

[0094] Preferably, the environmental parameter warning is as follows: if the water temperature is >27℃ and the capsule density is >4 capsules / m³ 2 If the green algae coverage rate is >40% (excessive algae growth), a text message alert will be sent to the administrator via the 4G module, prompting manual intervention.

[0095] refer to Figure 5 Intelligent decision-making and execution:

[0096] 1) Intelligent control for cyst elimination, trigger condition: when the camera in the second breeding area detects a cyst density at the bottom of the tank > 5 cysts / m². 2 Immediately trigger the UV lamp pulse irradiation mode: 30 seconds on, 10 seconds off, initial cumulative irradiation time is 2 hours. Dynamic control rule: if capsule density > 10 capsules / m² 2 The system automatically extends the cumulative irradiation time to 4 hours. After irradiation, the system automatically rechecks the capsule density; if it is still >5 capsules / m², the system will detect it. 2 After a one-hour interval, a second round of irradiation is initiated until the density is less than 5 particles / m². 2 Environmental parameter linkage: When the water temperature is 27℃, the capsule detection threshold is automatically lowered by 1 capsule / m³. 2(That is, the trigger density is adjusted to 4 per m) 2 Meanwhile, the duration of each UV lamp activation was extended to 40 seconds to improve disinfection efficiency during high-temperature periods.

[0097] 2) Adaptive feeding of algae-eating fish food, triggering condition:

[0098] When the green algae coverage is less than 15%, increase the amount of algae-eating fish feed by 30% in the target area within ten minutes;

[0099] When the green algae coverage is between 15% and 80%, maintain the feeding amount of algae-eating fish feed;

[0100] When the green algae coverage exceeds 80%, a mesh blockage information message is sent to the management personnel.

[0101] 3) Multi-dimensional abnormal data alarm, warning trigger conditions: abnormal biological indicators: abnormal behavior of grouper lasting ≥10 minutes (e.g., clustering >5 fish / m²) 2 Or average swimming speed ≥10cm / s); green algae coverage >40% (risk of excessive algae growth); algae-eating fish survival rate <60% (need to check for clogged mesh or insufficient feed). Abnormal environmental indicators: water temperature >30℃ or <20℃. Early warning execution logic: Abnormal data is sent to the administrator terminal in real time via the 4G communication module. The content includes: abnormal type (e.g., "cyst density exceeds standard" or "excessive algae growth"); specific parameter values ​​(e.g., "cyst density: 6 / m³"). 2 "Water temperature: 31℃"); recommended operations (such as "Please check the effect of UV lamp disinfection" and "Algae need to be removed manually").

[0102] Preferably, multimodal linkage prevention and control: when abnormal behavior alarms and capsule density > 5 / m 2 When triggered simultaneously, the system automatically upgrades to "emergency prevention and control mode": the ultraviolet lamp pulse interval is shortened to 20 seconds on / 10 seconds off.

[0103] Specifically, the main body of the intelligent aquaculture tank is a cylindrical polyethylene container (1.5m in diameter and 2m in height) with a smooth, non-adhesive inner wall. The first and second aquaculture zones are separated by a polypropylene mesh platform with a 3mm aperture. A camera is installed on the inner wall of the second aquaculture zone, 30cm below the mesh platform, to monitor the cysts at the bottom of the tank. Inner and outer ultraviolet lamp rings are fixed to the bottom of the tank to ensure no blind spots in illumination.

[0104] Furthermore, once the system is started, the top camera 4 captures images of grouper swimming at a frequency of 5 frames per second, identifying abnormal behavior using the YOLO11 algorithm. If more than 30% of the grouper show abnormalities for 10 consecutive minutes, the edge computing unit simultaneously analyzes the images from the second aquaculture area camera. If the cyst density reaches 5, the ultraviolet lamp is triggered to activate pulse irradiation (30s on / 10s off), which automatically stops after 2 hours. During this period, when the water temperature sensor detects that the water temperature has risen to 28℃, the cyst detection threshold is automatically adjusted to 4, improving the sensitivity of prevention and control.

[0105] Specifically, the technical terminology defines the stimulation of Cryptocaryon cysts as: spherical dormant bodies formed after the trophozoite detaches, with a diameter of 50-300 μm and ridge-like protrusions on the surface, which is a key stage for disease transmission; and the green algae coverage rate as: the proportion of green algae area on the grid platform surface, calculated by an image segmentation algorithm, with a threshold of 15% as the trigger condition for supplementing algae-eating fish feed.

[0106] The beneficial effects of this invention are as follows:

[0107] This invention provides data support for automated management by using grouper swimming and distribution data, upper layer monitoring results, cyst data, and green algae coverage, reducing reliance on manual labor. Through ultraviolet lamp irradiation strategies, feed supplementation strategies, and overbreeding judgment strategies, it realizes automated grouper farming, reduces damage to fry, and improves the prevention and control efficiency of cryptocaryon disease.

[0108] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0109] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A computer vision-based intelligent aquaculture method for the prevention and control of Cryptocaryon irritans disease in grouper, characterized in that, include: An apparatus for implementing the method includes: an outer shell of an intelligent aquaculture tank, a circulating water inlet, an upper circulating water outlet, a lower circulating water outlet, and a mesh isolation layer; The upper circulating water outlet and the lower circulating water outlet are respectively located in the middle and bottom of the outer shell of the intelligent breeding tank; the interior of the intelligent breeding tank is divided into a first breeding area and a second breeding area by a mesh isolation layer; The methods include: Images of grouper in the target area were acquired using a top-view camera and a side-view camera inside the first breeding area. Based on the top-view and side-view images of the fish, the YOLO11 model is used to detect and track the grouper in the target area, and the swimming and distribution data of the grouper are obtained. The average swimming speed and airstone aggregation density of the grouper swimming and distribution data are monitored to obtain upper-layer monitoring results. When the upper-layer monitoring results exceed the preset normal range, an abnormal behavior alarm is issued. The airstone aggregation density monitoring is specifically carried out by marking the airstone positions in the top view image and the side view image of the fish body, delineating a ROI area with a radius of 0.5m for the airstone positions, and counting the number of fish in the ROI area based on the grouper swimming and distribution data to obtain fish density data. The bottom of the target area is captured by a camera on the inner wall of the second breeding area to obtain an image of the bottom of the bucket; Based on the water temperature of the target area, a detection threshold is set, and the YOLO 11 model is used to detect cysts in the target area based on the bottom image of the bucket and the detection threshold to obtain cyst data; a baseline threshold at 25℃ is set, and the difference between the water temperature of the target area and 25℃ is calculated to obtain the temperature difference value. The threshold is then adjusted down by 1 cyst / m for every 2℃ increase in temperature. 2 The strategy adjusts the baseline threshold to obtain the detection threshold; The fish body top view image is filtered using the RGB color space to obtain a green algae region mask. The proportion of green algae pixels in the target area is calculated based on the green algae region mask to obtain the green algae coverage rate. When the upper-layer monitoring results exceed the normal range, if the capsule density in the capsule data is greater than the preset density threshold, then the ultraviolet lamp is activated to irradiate the target area in a pulsed manner. When the green algae coverage is less than the preset feeding threshold, the algae-eating fish feed is increased by 30% in the target area within 10 minutes. When the water temperature in the target area exceeds the preset temperature range, if the cyst density is greater than the density threshold or the green algae coverage is greater than the over-proliferation threshold, an early warning message is sent to the management personnel.

2. The intelligent aquaculture method for the prevention and control of Cryptocaryon irritans disease in grouper based on computer vision according to claim 1, characterized in that, The average swimming speed and airstone aggregation density of the grouper's swimming and distribution data are monitored to obtain upper-layer monitoring results. When the upper-layer monitoring results exceed the preset normal range, an abnormal behavior alarm is triggered, including: The total distance and average swimming speed of a single grouper within a unit time window are calculated based on the grouper swimming and distribution data. If the average swimming speed is greater than a preset speed threshold for 10 consecutive minutes, the grouper swimming and distribution data are judged as abnormal behavior. If the fish density data is greater than the preset fish density for 5 consecutive minutes, the grouper swimming and distribution data will be judged as abnormal clustering. If ≥30% of the fish in a single frame of the top-view image or the side-view image of the fish trigger either abnormal behavior or abnormal clustering, and this is repeated ≥3 times within 10 consecutive minutes, an abnormal behavior alarm will be issued.

3. The intelligent aquaculture method for the prevention and control of Cryptocaryon irritans disease in grouper based on computer vision according to claim 1, characterized in that, Based on the water temperature of the target area, a detection threshold is set, and the YOLO11 model is used to detect capsules in the target area based on the bucket bottom image and the detection threshold to obtain capsule data, including: Based on the detection threshold, the YOLO11 model is used to identify the cysts in the barrel bottom image to obtain the cyst pixel coordinates. The capsule pixel coordinates are transformed to a 1m×1m grid with the bottom center of the target area as the origin, and the capsule density is calculated based on the capsule pixel coordinates to obtain the capsule data.

4. The intelligent aquaculture method for the prevention and control of Cryptocaryon irritans disease in grouper based on computer vision according to claim 1, characterized in that, The fish's top-view image is filtered using the RGB color space to obtain a green algae region mask. Based on this mask, the percentage of green algae pixels within the target area is calculated to obtain the green algae coverage rate, including: The fish top view image is converted to RGB color space, and the areas with green channels greater than 150 and red channels greater than 80 in the converted fish top view image are extracted to obtain the green algae region mask; The green algae coverage rate is obtained by calculating the ratio of the total number of green algae pixels in the green algae region mask to the total number of pixels in the target region.

5. The intelligent aquaculture method for the prevention and control of Cryptocaryon irritans disease in grouper based on computer vision according to claim 1, characterized in that, When the upper-layer monitoring results exceed the normal range, if the capsule density in the capsule data is greater than a preset density threshold, then the ultraviolet lamp is activated to pulse-irradiate the target area, further comprising: When the capsule density is detected to be greater than the density threshold, the ultraviolet lamp is turned on and irradiated on the target area for 2 hours with a cycle of 30 seconds on and 10 seconds off. If the capsule density is greater than the density threshold after a 1-hour interval following the irradiation, the irradiation duration of the target area will be extended to 4 hours.

6. The intelligent aquaculture method for the prevention and control of Cryptocaryon irritans disease in grouper based on computer vision according to claim 1, characterized in that, When the green algae coverage is less than a preset feeding threshold, increase the amount of algae-eating fish feed by 30% within 10 minutes in the target area, including: When the green algae coverage is less than 15%, increase the amount of algae-eating fish feed in the target area by 30% within ten minutes. When the green algae coverage is in the range of 15% to 80%, maintain the feeding amount of algae-eating fish feed; When the green algae coverage exceeds 80%, a mesh blockage information message is sent to the management personnel.

7. A computer vision-based intelligent aquaculture device for the prevention and control of Cryptocaryon irritans disease in grouper, characterized in that, The intelligent aquaculture method for preventing and controlling Cryptocaryon irritans disease in grouper based on computer vision, as described in claim 1, comprises: a top support, a circulating water inlet, a water quality sensor, a top camera, a feeder, an intelligent aquaculture tank shell, a mesh isolation layer, an upper circulating water outlet, a lower circulating water outlet, a first side wall camera, a second side wall camera, an inner ring of ultraviolet lamps, an outer ring of ultraviolet lamps, and an ultraviolet lamp base. The top bracket and the feeder are respectively fixed to the top edge of the intelligent aquaculture tank shell; the water quality sensor and the top camera are fixed to the lower side of the top bracket; the circulating water inlet is located on the top side of the intelligent aquaculture tank shell; the upper circulating water outlet and the lower circulating water outlet are respectively located in the middle and bottom of the intelligent aquaculture tank shell; the first side wall camera and the second side wall camera are respectively located on the inner side of the upper and lower parts of the intelligent aquaculture tank shell; the mesh isolation layer is located in the middle of the interior of the intelligent aquaculture tank shell; the mesh isolation layer is above the upper circulating water outlet; the ultraviolet lamp base is located at the bottom of the interior of the intelligent aquaculture tank shell; the inner ultraviolet lamp ring and the outer ultraviolet lamp ring are located on the ultraviolet lamp base; the outer ultraviolet lamp ring is outside the inner ultraviolet lamp ring.

Images