Intelligent feeding control method and intelligent feeding device based on fish behavior analysis

Abstract

The invention discloses an intelligent feeding device based on fish behavior analysis, comprising a feeding device, a camera device, a communication device and a computer. The feeding device comprises a feeding device, a sliding device and a control device. The feeding device comprises a large storage bin, multiple small storage bins and a spreading bin. Feed ports are respectively arranged between the large storage bin and the small storage bin, between the small storage bin and the spreading bin and between the multiple small storage bins. The large storage bin, the small storage bin and the spreading bin are separated by partitions. The sliding device comprises a first slide rail, a second slide rail, a first screw rod, a second screw rod, a transmission sleeve, a fixed frame, a slider and a transmission shaft. The first screw rod passes through the slider under the fixed frame, and the parallel second screw rod passes through the slider under the first slide rail respectively, and the transmission shaft is connected with the slider under the first slide rail. The control device comprises a PLC, a stepping motor and a transmission belt. The stepping motor is used to receive control instructions from the PLC and control the transmission belt and the transmission sleeve to drive the feeding device to move.

Description

Technical Field

[0001] The present invention relates to the technical field of aquaculture, and in particular to an intelligent feeding control method and an intelligent feeding device based on fish behavior analysis. Background Art

[0002] In aquaculture, feed management is crucial to fish growth and farming efficiency. Traditional feeding methods often rely on manual judgment and experience, which makes it difficult to achieve accurate feeding, resulting in feed waste and water pollution. With the development of the Internet of Things, artificial intelligence, and computer vision technology, intelligent feeding equipment has gradually been applied to aquaculture. Through real-time monitoring and analysis of fish behavior, accurate feeding can be achieved, improving feed utilization and farming efficiency.

[0003] K-means algorithm is a widely used clustering algorithm that can realize unsupervised classification of data. It is very suitable for analyzing the activity patterns of fish and providing a scientific basis for determining the feeding location. In traditional aquaculture, the determination of feeding amount mainly depends on the experience and observation of farmers, which lacks scientificity and accuracy and easily leads to feed waste. SVM classifier is a powerful classification tool that can realize accurate classification of complex data and is very suitable for evaluating fish activity and feeding intensity.

[0004] Traditional feeding methods mostly rely on manual feeding, which is not only time-consuming and laborious, but also difficult to accurately control the feeding position and feeding amount, which easily leads to feed waste or malnutrition. With the development of intelligent technology, some automatic feeding equipment has appeared on the market, but these automatic feeding equipment often have low feeding accuracy and cannot meet the precise feed quantity requirements of specific animals. Summary of the invention

[0005] The purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art and to provide an intelligent feeding control method and an intelligent feeding device based on fish behavior analysis, which monitors fish behavior in real time through a camera, controls the discharge amount and discharge position of the intelligent feeding device according to the visual analysis results, and realizes precise feeding.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] An intelligent feeding device based on fish behavior analysis includes a feeding device, a camera device, a communication device and a computer. The camera device is connected to the computer through the communication device. The camera device includes an above-water camera and an underwater camera. The above-water camera and the underwater camera are fixed at both ends of a bracket. The bracket is fixed to the side wall of a breeding pond, and one end of the bracket is located above the breeding pond, and the other end is located in the breeding pond.

[0008] The feeding device is located above the breeding pond, and comprises a feeding device, a sliding device and a control device. The feeding device is arranged on the sliding device, and the sliding device is fixed on the bracket;

[0009] The feeding device includes a large storage bin, multiple small storage bins and a spreading bin, the multiple small storage bins are fitted with a rotating shaft, and the rotating shaft drives the multiple small storage bins to rotate, the multiple small storage bins are fixed to the spreading bin by a fixed shaft, and feeding ports are respectively provided between the large storage bin and the small storage bin, the small storage bin and the spreading bin, and the multiple small storage bins, and the large storage bin and the small storage bin, the small storage bin and the spreading bin, and the multiple small storage bins are separated by partitions;

[0010] The sliding device includes a first slide rail, a second slide rail, a first screw rod, a second screw rod, a screw rod fixing seat, a bearing seat, a transmission sleeve, a fixing frame, a sliding block and a transmission shaft.

[0011] The first screw rod is connected to the first slide rail, and the second screw rod is connected to the second slide rail through a screw rod fixing seat, and the bearing seat and the transmission sleeve are connected to the first screw rod and the second screw rod respectively;

[0012] The first screw rod is perpendicular to the parallel second screw rod, and the first screw rod is located above the second screw rod. The feeding device is connected to the fixed frame, the fixed frame is fixed above the slider, the first screw rod passes through the slider below the fixed frame, a slider is provided below the first slide rail, the parallel second screw rods pass through the sliders below the first slide rail respectively, and the transmission shaft is connected to the slider below the first slide rail;

[0013] The control device includes a PLC, a stepper motor and a transmission belt. The PLC is connected to the computer through a communication device. The stepper motor is respectively connected to the PLC and the transmission sleeve and the transmission belt. The stepper motor is used to receive control instructions from the PLC and control the transmission belt and the transmission sleeve to drive the feeding device to move.

[0014] Furthermore, the partition includes a first partition, a second partition and a third partition. The large storage bin is separated from the small storage bin by the first partition, the small storage bin is separated from the spreading bin by the second partition, and multiple small storage bins are separated by the third partition.

[0015] Furthermore, the feeding device also includes a silo fixing seat, which is sleeved on the outer walls of the plurality of small storage silos and connected to the fixing frame.

[0016] Furthermore, the stepper motor includes a first stepper motor, a second stepper motor and a third stepper motor. The first stepper motor is located under the fixed frame and is fixed to the slider through a motor fixing seat. The second stepper motor is connected to the transmission sleeve connected to the first screw rod, and the third stepper motor is connected to the transmission sleeve connected to the second screw rod.

[0017] Furthermore, the control device also includes a temperature sensor and a light sensor, and the temperature sensor and the light sensor are respectively connected to the PLC.

[0018] An intelligent feeding control method based on fish behavior analysis, applied to any of the above-mentioned intelligent feeding devices based on fish behavior analysis, comprises the following steps:

[0019] Obtain the behavior images, temperature and light information of farmed fish, pre-process the fish behavior images, and extract the fish behavior characteristics based on the fish behavior images through image processing algorithms;

[0020] Based on the behavioral characteristics of fish, the clustering algorithm is used to determine the degree and location of fish aggregation, and the SVM algorithm is used to identify the feeding intensity of fish in combination with temperature and light information.

[0021] The feeding area is determined according to the degree and location of fish aggregation, the feeding amount is determined according to the feeding intensity of the fish, and the feeding device is controlled to accurately deliver the feed to the feeding area according to the feeding amount.

[0022] Furthermore, the clustering algorithm adopts the K-means algorithm, which is as follows:

[0023] The number of clusters is set to K, that is, the number of preset feeding locations. Through iterative calculation, the fish activity area is divided into k clusters. Each cluster represents a potential feeding location. The goal is to minimize the sum of the squares of the distances from each fish to the center, expressed as:

[0024]

[0025] In the formula, K is the number of clusters, C i is the point set of the i-th cluster, x is the point belonging to C i Data points, μ i is the center of the ith cluster, ‖x-μ i ‖ 2 is the data point x and the cluster center μ i The square of the Euclidean distance between them;

[0026] According to the clustering results, the center point of each cluster is selected as the feeding position. These feeding positions can cover the activity area of ​​the fish to the greatest extent, ensure that the feed can be evenly distributed, and improve feed utilization. According to the growth of the fish, water quality changes and seasonal changes, the number of clusters K and feeding positions are dynamically adjusted to meet the needs of different breeding stages.

[0027] Furthermore, combined with temperature and light information, the SVM algorithm is used to identify the feeding intensity of fish, specifically:

[0028] Based on the fish behavior images, the dynamic characteristics of fish speed, acceleration, turning angle, and activity area are identified through image processing algorithms; combined with the temperature and light characteristics related to fish appetite, because when the water temperature and light are suitable, the fish appetite is strong, while at high temperature, low temperature, and high light intensity, the fish appetite will decrease;

[0029] Obtain multidimensional features of speed, acceleration, turning angle, activity area, temperature, and light parameters, and input these multidimensional features into a pre-trained SVM classification model to identify different levels of fish feeding intensity, including low-intensity feeding, medium-intensity feeding, and high-intensity feeding;

[0030] For fish with low feeding intensity, adopt the strategy of reducing feeding amount or adjusting feeding time to avoid food waste and water quality deterioration; for fish with moderate feeding intensity, maintain the regular feeding plan; for fish with high feeding intensity, increase feeding frequency or improve food quality to meet their strong appetite and growth needs.

[0031] Furthermore, for each parameter of speed, acceleration, number of turns, activity area, temperature, and light, the weight value of the parameter's influence on feeding intensity is calculated through SVM, and the speed weight ω1, acceleration weight ω2, turn weight ω3, activity area weight ω4, temperature weight ω5, and light weight ω6 are obtained. Each standardized parameter value is multiplied by its corresponding weight, and these weighted values ​​are added to obtain the feeding intensity index.

[0032] Furthermore, the formula for calculating the feeding intensity index is:

[0033] FI = ω1*speed + ω2*acceleration + ω3*number of turns + ω4*activity area + ω5*temperature + ω6*light;

[0034] When the feeding intensity index FI is between 0.7 and 1, it is high intensity, and the feeding amount needs to be increased; when the feeding intensity index FI is between 0.4 and 0.7, it is medium intensity, and there is no need to adjust the feeding amount; when the feeding intensity index FI is between 0 and 0.4, it is low intensity, and the feeding amount needs to be reduced;

[0035] The specific feeding amount needs to be calculated using the feeding intensity index FI combined with the number of breeding days and the number of fish farmed. As the weight and volume of the fish gradually increase, the amount of feed required will also increase accordingly. The initial value of the feeding amount is determined based on the experience of the staff and is dynamically adjusted through an algorithm later.

[0036] Compared with the prior art, the intelligent feeding device based on fish behavior analysis of the present invention feeds the feed to the fish gathering area to prevent the feed from sinking to the bottom. Through the cooperation of the slide rail and the screw rod, a slider that can move on the slide rail is designed to adjust the position of the feeding device. Multiple small storage bins are used to fix the feeding amount of each small storage bin, so as to ensure that when the large storage bin is full, the feeding amount is fixed, so as to achieve accurate and automatic feeding.

[0037] The computer uses a clustering algorithm to determine the degree and location of fish aggregation and sends it to the PLC. The PLC controls the feeding device to move to the feeding position through a stepper motor to avoid feeding feed to areas with too few fish, which causes feed to sink to the bottom and be wasted. The computer uses an SVM classifier to analyze the characteristics of fish feeding intensity and determines the amount of feed to be fed based on the calculated feeding intensity, thereby saving breeding costs while meeting the feeding needs of farmed fish at different growth stages. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 This is a schematic diagram of the structure of an intelligent feeding device based on fish behavior analysis.

[0039] Figure 2 It is a schematic diagram of the overall structure of the feeding device and the sliding device.

[0040] Figure 3 It is a schematic diagram of the feeding device and connecting parts structure.

[0041] Figure 4 This is a schematic diagram of the overall structure of the silo.

[0042] Figure 5 This is a schematic diagram of the silo structure from a bird's-eye view.

[0043] Figure 6 This is a schematic diagram of the structure of a small storage silo from a bird's-eye view.

[0044] Figure 7 This is a schematic diagram of the material spreading silo structure.

[0045] Figure 8 This is a flow chart of the intelligent feeding control method based on fish behavior analysis.

[0046] Fig. 9 Schematic diagram of the K-Means clustering algorithm.

[0047] Fig.10 Schematic diagram of the SVM classifier process.

[0048] Description of Figure Numbers:

[0049] Feeding equipment - 10; Camera equipment - 20; Communication equipment - 30; Computer - 40; Bracket - 50; Breeding pond - 60;

[0050] Feeding device: large storage bin-11; small storage bin-12; spreading bin-13; rotating shaft-14; fixed shaft-15; feeding port-16; first partition-171; second partition-172; third partition-173; bin fixing seat-18;

[0051] Sliding device: first slide rail-211; second slide rail-212; first screw rod-221; second screw rod-222; screw rod fixing seat-23; bearing seat-24; transmission sleeve-25; limiting gasket-26; fixing frame-27; slider-28; transmission shaft-29;

[0052] Control device: PLC-31; first stepper motor-32; second stepper motor-33; third stepper motor-34; transmission belt-35; motor fixing seat-36; temperature sensor-37; light sensor-38. DETAILED DESCRIPTION

[0053] The intelligent feeding control method based on fish behavior analysis and the intelligent feeding device of the present invention are further described below in conjunction with the accompanying drawings and specific embodiments.

[0054] See also Figure 1 and Figure 2 The present invention discloses an intelligent feeding device based on fish behavior analysis, including a feeding device 10, a camera device 20, a communication device 30 and a computer 40. The camera device 20 is connected to the computer 40 through the communication device 30. The camera device 20 includes an above-water camera and an underwater camera. The above-water camera and the underwater camera are fixed at both ends of a bracket 50. The bracket 50 is fixed to the side wall of a breeding pond 60, and one end of the bracket 50 is located above the breeding pond 60, and the other end is located in the breeding pond 60. The feeding device 10 is located above the breeding pond 60. The feeding device 10 includes a feeding device, a sliding device and a control device. The feeding device is arranged on the sliding device, and the sliding device is fixed on the bracket 50.

[0055] See also Figure 3 , Figure 4 and Figure 7 The feeding device includes a large storage bin 11, multiple small storage bins 12 and a spreading bin 13. The multiple small storage bins 12 are fitted with a rotating shaft 14, and the rotating shaft 14 drives the multiple small storage bins 12 to rotate. The multiple small storage bins 12 and the spreading bin 13 are fixed by a fixed shaft 15. Feed ports 16 are respectively provided between the large storage bin 11 and the small storage bin 12, the small storage bin 12 and the spreading bin 13, and the multiple small storage bins 12. The large storage bin 11 and the small storage bin 12, the small storage bin 12 and the spreading bin 13, and the multiple small storage bins 12 are separated by partitions.

[0056] See also Figure 5 and Figure 6In a specific embodiment, the partition includes a first partition 171, a second partition 172 and a third partition 173. The large storage bin 11 and the small storage bin 12 are separated by the first partition 171, the small storage bin 12 and the spreading bin 13 are separated by the second partition 172, and the plurality of small storage bins 12 are separated by the third partition 173. The feeding device also includes a bin fixing seat 18, which is sleeved on the outer walls of the plurality of small storage bins 12 and connected to the fixing frame 27.

[0057] See also Figure 2 The sliding device includes a slide rail, a screw rod, a screw rod fixing seat 23, a bearing seat 24, a transmission sleeve 25, a limit washer 26, a fixing frame 27, a slider 28 and a transmission shaft 29. The slide rail includes a first slide rail 211 and a second slide rail 212, and the screw rod includes a first screw rod 221 and a second screw rod 222. The first screw rod 221 is fixedly connected to the first slide rail 211, and the second screw rod 222 is fixedly connected to the second slide rail 212 through a screw rod fixing seat 23.

[0058] The bearing seat 24 and the transmission sleeve 25 are respectively connected to the first screw rod 221 and one of the second screw rods 222, and the bearing seat 24 and the other screw rod fixing seat 23 are respectively connected to the other second screw rod 222. The limiting gasket 26 is respectively sleeved on the first screw rod 221 and the second second screw rod 222.

[0059] The first screw rod 221 is perpendicular to the two parallel second screw rods 222, and the first screw rod 221 is located above the two second screw rods 222. The feeding device is connected to the fixed frame 27, the fixed frame 27 is fixed above the slider 28, the first screw rod 221 passes through the slider 28 below the fixed frame 27, the slider 28 is provided below the first slide rail 211, the two parallel second screw rods 222 respectively pass through the slider 28 below the first slide rail 211, and the transmission shaft 29 is connected to the slider 28 below the first slide rail 211.

[0060] The control device includes a PLC 31, a first stepper motor 32, a second stepper motor 33, a third stepper motor 34, a transmission belt 35, a motor fixing seat 36, a temperature sensor 37 and a light sensor 38. The PLC 31 is connected to a computer 40 via a communication device 30, and the first stepper motor 32, the second stepper motor 33, the third stepper motor 34, the temperature sensor 37 and the light sensor 38 are connected to the PLC 31 respectively.

[0061] The first stepper motor 32 is located below the fixing frame 27 and fixed to the slider 28 through the motor fixing seat 36, the second stepper motor 33 is connected to the transmission sleeve 25 connected to the first screw rod 221, and the third stepper motor 34 is connected to the transmission sleeve 25 connected to the second screw rod 222. The first stepper motor 32, the second stepper motor 33, and the third stepper motor 34 are used to receive control instructions from the PLC 31, and control the transmission belt 35 or the transmission sleeve 25 to drive the feeding device or the sliding device to move.

[0062] The computer 40 establishes a connection with the camera device 20 via the communication device 30 to obtain the behavior characteristics of the fish in the breeding pond 60 in real time, thereby dynamically adjusting the feeding device 10.

[0063] PLC31 is responsible for controlling the second stepper motor 33, thereby changing the first relative position of the feeding device through the transmission shaft 29 and the first screw rod 221; PLC31 controls the third stepper motor 34 to change the second relative position of the feeding device through the second screw rod 222, and accurately controls the feeding position of the feeding device through the first and first movements. The screw rod fixing seat 23 is responsible for connecting the transmission sleeve 25 to fix the first screw rod 222 and the second screw rod 222, and the first screw rod 222 and the second screw rod 222 are responsible for limiting the movement of the slider 28.

[0064] PLC31 controls the temperature sensor 37 and the light sensor 38 to collect relevant data, and sends the data to the communication device 30 and then transmits it to the computer 40. After receiving the data transmitted by the communication device 30, PLC31 controls the first stepper motor 32 to drive the rotating shaft 14 on the small storage bin 12 and the small storage bin 12 to rotate through the transmission belt 35. At this time, the feed in the large storage bin 11 enters the small storage bin 12, and the feed in the small storage bin 12 enters the spreading bin 13. The bin fixing seat 18 is responsible for fixing the feeding device, the motor fixing seat 36 is responsible for fixing the first stepper motor 32, and the slider 28 is responsible for driving the entire feeding device to move along the first direction and the second direction.

[0065] When the middle rotating shaft 14 of the small storage bin 12 drives the small storage bin 12 to rotate, the feed in the large storage bin 11 enters the small storage bin 12 through the feed inlet 16 between the large storage bin 11 and the multiple small storage bins 12. At this time, the feed in the small storage bin 12 on the opposite side of the rotating shaft 14 passes through the feed inlet 16 between the multiple small storage bins 12, along the feed inlet 16 between the multiple small storage bins 12 and the spreading bin 13, and is sent to the bottom spreading bin 13 for spreading. The first partition 171 and the second partition 172 are responsible for isolating the large storage bin 11, the small storage bin 12, and the spreading bin 13. The small storage bins 12 are separated by the third partition 173, and the bottom spreading bin 13 is connected to the top by the fixed shaft 15.

[0066] The intelligent feeding device based on fish behavior analysis of the present invention feeds the feed to the fish gathering area to prevent the feed from sinking to the bottom. A movable slide rail is designed to adjust the position of the feeding bin. Multiple silos are used to fix the feeding amount of each silo, so as to ensure that when the silo is full, the feeding amount is fixed, so as to achieve accurate and automatic feeding.

[0067] See also Figure 8 The present invention also discloses an intelligent feeding control method based on fish behavior analysis, which is applied to any of the above-mentioned intelligent feeding devices based on fish behavior analysis, and comprises the following steps:

[0068] Obtain the behavior images, temperature and light information of farmed fish, pre-process the fish behavior images, and extract the fish behavior characteristics based on the fish behavior images through image processing algorithms;

[0069] Based on the behavioral characteristics of fish, the clustering algorithm is used to determine the degree and location of fish aggregation, and the SVM algorithm is used to identify the feeding intensity of fish in combination with temperature and light information.

[0070] The feeding area is determined according to the degree and location of fish aggregation, the feeding amount is determined according to the feeding intensity of the fish, and the feeding device is controlled to accurately deliver the feed to the feeding area according to the feeding amount.

[0071] In a specific embodiment, the above-water camera and underwater camera of the camera device 20 are used to collect fish behavior images in the breeding pond 60 in real time and send them to the computer 40 for analysis. The collected images are subjected to pre-processing operations such as filtering, binarization, and target detection, and then the extracted behavioral characteristic parameters such as the position and speed of the fish school are input into the subsequent clustering algorithm and SVM classifier to analyze the feeding area and feeding amount.

[0072] The computer uses a clustering algorithm to determine the degree and location of fish aggregation and sends it to the PLC. The PLC controls the feeding device to move to the feeding position through a stepper motor to avoid feeding feed to areas with too few fish, which causes feed to sink to the bottom and be wasted. The computer uses an SVM classifier to analyze the characteristics of fish feeding intensity and determines the amount of feed to be fed based on the calculated feeding intensity, thereby saving breeding costs while meeting the feeding needs of farmed fish at different growth stages.

[0073] See also Fig. 9 In a specific implementation, the clustering algorithm adopts the K-means algorithm model, sets the number of clusters to K, that is, the number of preset feeding locations, and divides the fish activity area into k clusters through iterative calculation. Each cluster represents a potential feeding location. The goal is to minimize the sum of the squares of the distances from each fish to the center, expressed as:

[0074]

[0075] In the formula, K is the number of clusters, C i is the point set of the i-th cluster, x is the point belonging to C i Data points, μ i is the center of the ith cluster, ‖x-μ i ‖ 2 is the data point x and the cluster center μ i The square of the Euclidean distance between them.

[0076] According to the clustering results, the center point of each cluster is selected as the feeding position. These feeding positions can cover the activity area of ​​the fish to the greatest extent, ensure that the feed can be evenly distributed, and improve feed utilization. According to the growth of the fish, water quality changes and seasonal changes, the number of clusters K and feeding positions are dynamically adjusted to meet the needs of different breeding stages.

[0077] See also Fig.10 , the monitored data is input into the SVM classifier model for training and learning. The SVM classifier learns the complex relationship between fish activity, feeding intensity and feeding amount by analyzing a large amount of historical data, and establishes a classification model. During the training process, a large amount of sample data needs to be prepared, including fish images under different activity states and feeding intensities, as well as corresponding feeding amount records. Through continuous iteration and optimization, the SVM classifier can accurately classify and evaluate new monitoring data.

[0078] In a specific implementation, the fish feeding intensity is identified by combining temperature and light information through the SVM algorithm. Based on the fish behavior image, the dynamic characteristics of the fish's speed, acceleration, turning angle, and activity area are identified through the image processing algorithm; combined with the temperature and light characteristics related to the fish's appetite, because when the water temperature and light are suitable, the fish's appetite is strong, while at high temperature, low temperature, and high light intensity, the fish's appetite will decrease.

[0079] The multidimensional features of speed, acceleration, turning angle, activity area, temperature, and light parameters are obtained, and these multidimensional features are input into a pre-trained SVM classification model to identify different levels of fish feeding intensity, including low-intensity feeding, medium-intensity feeding, and high-intensity feeding.

[0080] For fish with low feeding intensity, adopt the strategy of reducing feeding amount or adjusting feeding time to avoid food waste and water quality deterioration; for fish with moderate feeding intensity, maintain the regular feeding plan; for fish with high feeding intensity, increase feeding frequency or improve food quality to meet their strong appetite and growth needs.

[0081] In a specific implementation, for each parameter of speed, acceleration, number of turns, activity area, temperature, and light, the weight value of the parameter's influence on the feeding intensity is calculated by SVM, and the speed weight ω1, acceleration weight ω2, turn weight ω3, activity area weight ω4, temperature weight ω5, and light weight ω6 are obtained. Each standardized parameter value is multiplied by its corresponding weight, and these weighted values ​​are added to obtain the feeding intensity index. The formula for calculating the feeding intensity index is:

[0082] FI = ω1*speed + ω2*acceleration + ω3*number of turns + ω5*activity area + ω5*temperature + ω6*light;

[0083] When the feeding intensity index FI is between 0.7 and 1, it is high intensity and the feeding amount needs to be increased; when the feeding intensity index FI is between 0.4 and 0.7, it is medium intensity and no adjustment of the feeding amount is required; when the feeding intensity index FI is between 0 and 0.4, it is low intensity and the feeding amount needs to be reduced, as shown in Table 1.

[0084] The specific feeding amount needs to be calculated using the feeding intensity index FI combined with the number of breeding days and the number of fish farmed. As the weight and volume of the fish gradually increase, the amount of feed required will also increase accordingly. The initial value of the feeding amount is determined based on the experience of the staff and is dynamically adjusted through an algorithm later.

[0085] Table 1 Feeding intensity and feeding amount percentage

[0086] Feeding intensity Feeding amount percentage High strength 70％～100％ Medium Strength 40％～70％ Low intensity 0～40％

[0087] A plurality of small storage bins 12 separated by third partitions 173 arranged at equal intervals are arranged in a circle to form the main body of the feeding device. Each small storage bin 12 has a fixed volume and can store a preset amount of feed, such as 100 grams. A second partition 172 that can be opened and closed is provided at the bottom of the small storage bin 12 to control the release of the feed. Equipped with a first stepper motor 32, the number of small storage bins 12 that need to be opened is calculated based on the result of calculating the amount of feed required for each feeding. The PLC 31 sends a command to the first stepper motor 32 to rotate a fixed angle to control the number of small storage bins 12 opened. The opening and closing of the second partition 172 is connected to the second partition 172 through the first stepper motor 32, the transmission belt 35, and the rotating shaft 14 to ensure whether the material can be discharged. By driving it to rotate to the specified position, the corresponding number of small storage bins 12 are opened in turn, thereby realizing accurate control of the feeding amount.

[0088] The above description is a detailed description of the preferred feasible embodiments of the present invention, but the embodiments are not intended to limit the scope of the patent application of the present invention. All equivalent changes or modified changes completed under the technical spirit disclosed by the present invention should fall within the patent scope covered by the present invention.

Claims

1. An intelligent feeding control method based on fish behavior analysis, characterized in that: The following steps are involved: Obtain the behavior images, temperature and light information of farmed fish, pre-process the fish behavior images, and extract the fish behavior characteristics based on the fish behavior images through image processing algorithms; Based on the behavioral characteristics of fish, the clustering algorithm is used to determine the degree and location of fish aggregation, and the SVM algorithm is used to identify the feeding intensity of fish in combination with temperature and light information. Determine the feeding area according to the degree and location of fish aggregation, determine the feeding amount according to the feeding intensity of fish, and control the feeding device to accurately deliver the feed to the feeding area according to the feeding amount; The clustering algorithm uses the K-means algorithm, which is as follows: The number of clusters is set to K, that is, the number of preset feeding locations. Through iterative calculation, the fish activity area is divided into k clusters. Each cluster represents a potential feeding location. The goal is to minimize the sum of the squares of the distances from each fish to the center, expressed as: In the formula, K is the number of clusters, C i is the point set of the i-th cluster, x is the point belonging to C i Data points, μ i is the center of the ith cluster, ‖x-μ i ‖ 2 is the data point x and the cluster center μ i The square of the Euclidean distance between them; According to the clustering results, the center point of each cluster is selected as the feeding position. These feeding positions can cover the activity area of ​​the fish to the greatest extent, ensure that the feed can be evenly distributed, and improve feed utilization. According to the growth of the fish, water quality changes and seasonal changes, the number of clusters K and feeding positions are dynamically adjusted to meet the needs of different breeding stages.

2. The intelligent feeding control method based on fish behavior analysis according to claim 1 is characterized in that: Combining temperature and light information, the SVM algorithm is used to identify the feeding intensity of fish, specifically: Based on the fish behavior images, the dynamic characteristics of fish speed, acceleration, turning angle, and activity area are identified through image processing algorithms; combined with the temperature and light characteristics related to fish appetite, because when the water temperature and light are suitable, the fish appetite is strong, while at high temperature, low temperature, and high light intensity, the fish appetite will decrease; Obtain multidimensional features of speed, acceleration, turning angle, activity area, temperature, and light parameters, and input these multidimensional features into a pre-trained SVM classification model to identify different levels of fish feeding intensity, including low-intensity feeding, medium-intensity feeding, and high-intensity feeding; For fish with low feeding intensity, adopt the strategy of reducing feeding amount or adjusting feeding time to avoid food waste and water quality deterioration; for fish with moderate feeding intensity, maintain the regular feeding plan; for fish with high feeding intensity, increase feeding frequency or improve food quality to meet their strong appetite and growth needs.

3. The intelligent feeding control method based on fish behavior analysis according to claim 2 is characterized in that: For each parameter of speed, acceleration, number of turns, activity area, temperature, and light, the weight value of the parameter's influence on feeding intensity is calculated through SVM, and the speed weight ω1, acceleration weight ω2, turn weight ω3, activity area weight ω4, temperature weight ω5, and light weight ω6 are obtained. Each standardized parameter value is multiplied by its corresponding weight, and these weighted values ​​are added together to obtain the feeding intensity index.

4. The intelligent feeding control method based on fish behavior analysis according to claim 3 is characterized in that: The formula for calculating the feeding intensity index is: FI = ω1*speed + ω2*acceleration + ω3*number of turns + ω4*activity area + ω5*temperature + ω6*light; When the feeding intensity index FI is between 0.7 and 1, it is high intensity, and the feeding amount needs to be increased; when the feeding intensity index FI is between 0.4 and 0.7, it is medium intensity, and there is no need to adjust the feeding amount; when the feeding intensity index FI is between 0 and 0.4, it is low intensity, and the feeding amount needs to be reduced; The specific feeding amount needs to be calculated using the feeding intensity index FI combined with the number of breeding days and the number of fish farmed. As the weight and volume of the fish gradually increase, the amount of feed required will also increase accordingly. The initial value of the feeding amount is determined based on the experience of the staff and is dynamically adjusted through an algorithm later.

5. An intelligent feeding device based on fish behavior analysis, using the intelligent feeding control method based on fish behavior analysis according to any one of claims 1 to 4, characterized in that: It includes feeding equipment, video equipment, communication equipment and a computer. The video equipment is connected to the computer through the communication equipment. The video equipment includes an above-water camera and an underwater camera. The above-water camera and the underwater camera are fixed at both ends of a bracket. The bracket is fixed to the side wall of the breeding pool, and one end of the bracket is located above the breeding pool, and the other end is located in the breeding pool; The feeding device is located above the breeding pond, and comprises a feeding device, a sliding device and a control device. The feeding device is arranged on the sliding device, and the sliding device is fixed on the bracket; The feeding device includes a large storage bin, multiple small storage bins and a spreading bin, the multiple small storage bins are fitted with a rotating shaft, and the rotating shaft drives the multiple small storage bins to rotate, the multiple small storage bins are fixed to the spreading bin by a fixed shaft, and feeding ports are respectively provided between the large storage bin and the small storage bin, the small storage bin and the spreading bin, and the multiple small storage bins, and the large storage bin and the small storage bin, the small storage bin and the spreading bin, and the multiple small storage bins are separated by partitions respectively; The sliding device includes a first slide rail, a second slide rail, a first screw rod, a second screw rod, a screw rod fixing seat, a bearing seat, a transmission sleeve, a fixing frame, a slider and a transmission shaft. The first screw rod is connected to the first slide rail, and the second screw rod is connected to the second slide rail through the screw rod fixing seat. The bearing seat and the transmission sleeve are connected to the first screw rod and the second screw rod respectively. The first screw rod is perpendicular to the parallel second screw rod, and the first screw rod is located above the second screw rod. The feeding device is connected to the fixed frame, the fixed frame is fixed above the slider, the first screw rod passes through the slider below the fixed frame, a slider is provided below the first slide rail, the parallel second screw rods pass through the sliders below the first slide rail respectively, and the transmission shaft is connected to the slider below the first slide rail; The control device includes a PLC, a stepper motor and a transmission belt. The PLC is connected to the computer through a communication device. The stepper motor is respectively connected to the PLC and the transmission sleeve and the transmission belt. The stepper motor is used to receive control instructions from the PLC and control the transmission belt and the transmission sleeve to drive the feeding device to move.

6. The intelligent feeding device based on fish behavior analysis according to claim 5, characterized in that: The partitions include a first partition, a second partition and a third partition. The large storage bin is separated from the small storage bin by the first partition, the small storage bin is separated from the spreading bin by the second partition, and multiple small storage bins are separated by the third partition.

7. The intelligent feeding device based on fish behavior analysis according to claim 5, characterized in that: The feeding device also includes a silo fixing seat, which is sleeved on the outer walls of a plurality of small storage silos and connected to the fixing frame.

8. The intelligent feeding device based on fish behavior analysis according to claim 5, characterized in that: The stepper motor includes a first stepper motor, a second stepper motor and a third stepper motor. The first stepper motor is located below the fixing frame and is fixed to the slider through a motor fixing seat. The second stepper motor is connected to a transmission sleeve connected to the first screw rod. The third stepper motor is connected to a transmission sleeve connected to the second screw rod.

9. The intelligent feeding device based on fish behavior analysis according to claim 5, characterized in that: The control device also includes a temperature sensor and a light sensor, and the temperature sensor and the light sensor are connected to the PLC respectively.

Images