A visual-based fish adaptive feeding method and device

Abstract

The application relates to the technical field of fish culture feeding, in particular to a fish self-adaptive feeding method and device based on vision, which comprises the following steps: sequentially performing attracting feeding operation and formal feeding operation on fish in a culture area; acquiring a feeding video of the fish after the formal feeding operation; acquiring an average feeding behavior quantification index of the fish according to the feeding video; acquiring a feeding judgment result according to the average feeding behavior quantification index; and performing self-adaptive feeding on the fish according to the feeding judgment result. The method can characterize the feeding desire of a fish group based on quantified feeding behavior of the fish group, and realizes accurate feeding of the fish in the culture area.

Description

Technical Field

[0001] This invention relates to the field of fish farming feeding technology, and more specifically, to a vision-based adaptive feeding method and apparatus for fish. Background Technology

[0002] Precise feeding of fish is a crucial aspect of aquaculture. It involves providing fish with suitable feed in a precise and scientific manner based on their growth needs, environmental conditions, and feeding behavior patterns. This improves feed utilization, promotes healthy fish growth, reduces aquaculture costs, and is an important guarantee for the sustainable development of aquaculture.

[0003] Currently, manual feeding relies heavily on the experience and judgment of fish farmers, who determine the amount, timing, and location of feed based on observation of the fish population. While flexible, it lacks precision and consistency. Machine feeding, on the other hand, uses programmed settings and sensors to control the feeding equipment, distributing feed according to preset parameters such as time and amount. Although it achieves a degree of automation, it is not sensitive enough to real-time feeding behavior. Quantitative behavioral feeding combines advanced monitoring technologies, such as video surveillance and image analysis, attempting to precisely adjust feeding strategies through quantitative analysis of fish feeding behavior. However, in practical applications, it is subject to numerous interferences from the outdoor environment of fish farming, directly affecting the accuracy of fish feeding control. Summary of the Invention

[0004] This invention provides a vision-based adaptive feeding method and apparatus for fish, which is used to characterize the feeding desire of fish based on quantitative fish feeding behavior, and to achieve precise feeding of fish in aquaculture areas.

[0005] According to a first aspect of this application, a vision-based adaptive feeding method for fish is provided, the method comprising:

[0006] The fish in the aquaculture area are subjected to an attraction and feeding operation followed by a formal feeding operation; the formal feeding operation includes a first formal feeding operation and a second formal feeding operation performed sequentially.

[0007] After the formal feeding operation, the feeding video of the fish is acquired; the feeding video includes a first feeding video and a second feeding video, the first feeding video is acquired after the first formal feeding operation, and the second feeding video is acquired after the second formal feeding operation;

[0008] The average feeding behavior quantification index of the fish is obtained based on the feeding video; the average feeding behavior quantification index includes a first average feeding behavior quantification index corresponding to the first feeding video and a second average feeding behavior quantification index corresponding to the second feeding video;

[0009] The feeding judgment result is obtained based on the first average feeding behavior quantification index and the second average feeding behavior quantification index, and the fish are adaptively fed according to the feeding judgment result.

[0010] Understandably, the two-stage feeding design, employing attraction feeding and formal feeding, can attract and gather fish, with the feeding behavior quantification index calculated only during formal feeding. This reduces quantification errors caused by surface fluctuations resulting from fish gathering. Based on visual data, feeding behavior quantification indices are collected and extracted from feeding videos, transforming biological characteristics such as feeding intensity, speed, and group distribution into calculable parameters, improving the accuracy of calculations and judgments. By comparing and analyzing the first and second feeding videos, conditions for judging fish feeding desire and satiety are constructed, reflecting both the immediate feeding needs of fish and capturing trends in their feeding capacity. This adaptive feeding method allows the feeding amount to dynamically match the actual needs of fish, ensuring nutrient supply while reducing feed waste, lowering the risk of water pollution, and improving aquaculture efficiency and ecological benefits.

[0011] Optionally, the attraction and feeding operation continues for a preset attraction and feeding time to attract and gather the fish.

[0012] The first formal feeding operation lasts for a preset first feeding time, and the second formal feeding operation lasts for a preset second feeding time;

[0013] The attraction and feeding operation is separated from the first formal feeding operation by a preset first interval time;

[0014] After the first formal feeding operation is completed, a preset second interval is set before the second formal feeding operation begins, and the duration of the first feeding video is the second interval.

[0015] The duration of the second feeding video is the preset third interval time.

[0016] Understandably, a scientific feeding rhythm for fish was constructed through the design of refined time parameters: First, a preset feeding time is used to attract fish and gather them, ensuring the effectiveness of subsequent feeding; then, a first interval is used to eliminate initial feeding interference, allowing the fish to enter a natural feeding state; a second interval is set after the first formal feeding, and the first feeding video is collected simultaneously, ensuring the smooth acquisition of the first feeding video and avoiding excessive waiting by the fish after the first formal feeding; the third interval also ensures the smooth acquisition of the second feeding video, which can be compared with the first feeding video to monitor the fish's feeding behavior. By comparing the quantitative index of feeding behavior at different time points and combining feeding data at different stages, multi-dimensional conditions for judging feeding behavior and satiety are constructed, making adaptive feeding both responsive and physiologically adaptable to the fish, ultimately achieving the dual goals of precise feeding and resource optimization.

[0017] Optionally, obtaining the average feeding behavior quantification index of the fish based on the feeding video includes:

[0018] Extract several frames of feeding images corresponding to each of the feeding videos;

[0019] Each frame of the feeding image is preprocessed to obtain a preprocessed feeding image;

[0020] The feeding behavior quantification index of each frame of the preprocessed feeding image is obtained by using optical flow method.

[0021] The average feeding behavior quantification index corresponding to each feeding video is obtained based on the feeding behavior quantification index of all the preprocessed feeding images corresponding to each feeding video.

[0022] Understandably, the preprocessing stage effectively eliminates environmental interference such as water reflection and shadow noise, enhancing image features and improving the reliability of feature extraction. By tracking the motion vector of each pixel in the feeding image using optical flow, complex behaviors such as fish feeding intensity, swimming rhythm, and group cooperation are transformed into quantifiable dynamic indicators, offering greater spatiotemporal continuity compared to traditional manual observation or static feature recognition. Finally, by summing and averaging each frame of the feeding video, errors caused by instantaneous behavioral fluctuations are suppressed while preserving the overall trend of fish group feeding characteristics. This multi-level quantification system enables feeding decisions to be based on the actual feeding needs and behavioral feedback of the fish, ensuring nutrient supply while avoiding overfeeding, thereby improving feed utilization, reducing the risk of aquaculture pollution, and achieving refined aquaculture management.

[0023] Optionally, the step of processing each frame of the preprocessed feeding image using optical flow to obtain a quantitative index of feeding behavior for each frame of the preprocessed feeding image includes:

[0024] For each of the aforementioned feeding videos:

[0025] Obtain information about each pixel in each preprocessed frame of the feeding image;

[0026] The second preprocessed feeding image and the subsequent preprocessed feeding images are used as the processed images;

[0027] The preprocessed feeding image of the preceding frame adjacent to each of the processed images is obtained as the corresponding comparison image;

[0028] Based on the optical flow method, the corresponding pixel displacement amplitude is obtained according to the pixel information of each pixel in the processed image and the pixel information of the corresponding comparison image;

[0029] Preset the weights corresponding to the information of each pixel;

[0030] The feeding behavior quantification index corresponding to each pixel is obtained based on the pixel displacement amplitude, weight, and total number of pixels in the corresponding processed image.

[0031] The feeding behavior quantization index of the processed image is obtained based on the feeding behavior quantization index of all pixels.

[0032] Understandably, by capturing the instantaneous motion characteristics of fish during feeding through optical flow vector analysis, behaviors such as fish body swaying frequency and swimming trajectory are transformed into calculable physical parameters. A pixel weighting mechanism is introduced, setting differentiated contribution levels for regions at different distances from the feeding device. This highlights the dynamic characteristics of areas closer to the feeding area while suppressing interference from redundant data in areas further away from the feeding area. Through temporal analysis comparing adjacent frames, the continuous changes in feeding behavior are fully recorded, reflecting the dynamic growth or decline of fish feeding intentions more effectively than single-frame static analysis. This quantitative method overcomes the subjective limitations of traditional visual observation in assessing feeding behavior, accurately identifying the hunger / satisfaction state of fish, providing millisecond-level behavioral data support for adaptive feeding strategies, significantly improving the timeliness and accuracy of feeding decisions, while reducing the risk of misjudgment caused by environmental interference, ultimately achieving a dual optimization of aquaculture resource utilization and ecological benefits.

[0033] Optionally, the weights are represented as:

[0034]

[0035] in The x and y axes represent the horizontal and vertical coordinates of the feeding and throwing equipment, respectively. , These are the major and minor axes of the pre-defined elliptical feeding area with the feeding and throwing device as the origin; , These are the x and y coordinates of the pixel, respectively. For the pixel point The corresponding weights.

[0036] Understandably, this achieves spatial differentiation characterization of fish feeding areas. An elliptical weighted field is formed centered on the feeding device, and parameters are used to... , The system flexibly adapts to the feeding area morphology of fish in actual aquaculture areas, enhancing the representation of feeding behavior in key areas around the equipment. By constructing elliptical gradient weights through an improved Gaussian decay function, the weights decay smoothly from the center to the periphery, highlighting the pixel displacement contribution of fish in the feeding area while suppressing edge noise interference, thus improving the signal-to-noise ratio of feeding behavior quantification. The spatial heteroweighting mechanism deeply integrates biological spatial distribution characteristics with computer vision features, making the feeding behavior quantification index more consistent with the actual feeding patterns of fish, providing more accurate behavioral basis for subsequent feeding decisions, and ultimately achieving dynamic matching between feeding strategies and the physiological needs of fish.

[0037] Optionally, the feeding behavior quantification index of each processed image frame Represented as:

[0038]

[0039] in, The total number of pixels in the processed image; This represents the pixel displacement amplitude corresponding to the pixel information. The weights corresponding to the pixel information;

[0040] And / or, obtaining the average feeding behavior quantification index corresponding to each feeding video based on the feeding behavior quantification index of all the preprocessed feeding images corresponding to each feeding video includes:

[0041] The average feeding behavior quantification index is obtained by summing and averaging the feeding behavior quantification indices of all the preprocessed feeding images corresponding to each feeding video.

[0042] Understandably, the quantization formula combines pixel displacement amplitude with spatial weights. The weights are used to increase the contribution of pixels closer to the feeding area, highlighting the feeding intensity characteristics of the fish's feeding region. The average feeding behavior quantization index is obtained by summing and averaging each frame of the feeding video. This smooths out noise from instantaneous swimming interference in single-frame feeding images while preserving the continuity of the fish's feeding trend. This multi-layered computational architecture makes feeding behavior quantification both spatially sensitive and temporally robust, providing high-precision, low-interference behavioral characteristic indicators for subsequent feeding decisions. Ultimately, it achieves dynamic matching between feeding strategies and the actual needs of fish, effectively improving feed utilization and reducing environmental impact.

[0043] Optionally, the step of obtaining a feeding judgment result based on the first average feeding behavior quantification index and the second average feeding behavior quantification index, and adaptively feeding the fish based on the feeding judgment result, includes:

[0044] Calculate the first feeding behavior quantitative index, which is a preset first multiple, as the first comparison threshold;

[0045] If the second feeding behavior quantification index is greater than or equal to the first comparison threshold, then the fish will continue with the next formal feeding operation after the third interval.

[0046] If the second feeding behavior quantification index is less than the first comparison threshold, then the current total amount of feed fed in the aquaculture area and the current total mass of the fish are obtained;

[0047] The current total mass is calculated as a preset second multiple as a second comparison threshold;

[0048] If the current total amount of feed fed is greater than the second comparison threshold, then feeding the fish should be stopped.

[0049] Understandably, precise adaptive feeding control was achieved by quantifying feeding behavior and dynamically judging thresholds in the first and second feeding videos. Using a first average feeding index as a benchmark, a first comparison threshold was used to determine the feeding status of the fish after the second formal feeding. If the second average feeding behavior quantification index met the standard, the feeding rhythm continued, aligning with the fish's continuous feeding needs. When the fish's feeding intention was insufficient, a second comparison threshold was used to impose long-term total quantity constraints, combining the current total feed amount with the total fish mass. This prevented both short-term misjudgments leading to insufficient feeding and long-term overfeeding causing resource waste. The dual threshold conditions combined short-term behavioral responses with long-term growth needs, significantly improving feed utilization and reducing uneaten feed pollution while ensuring the fish's nutritional intake, ultimately achieving synergistic optimization of aquaculture efficiency and ecological benefits.

[0050] Optionally, the step of continuing the next formal feeding operation on the fish after the third interval includes:

[0051] The next formal feeding operation will be taken as the current formal feeding operation;

[0052] The current formal feeding operation lasts for a certain feeding time, and a current feeding video of a certain duration is obtained after the current formal feeding operation. The average feeding behavior quantitative index corresponding to the current formal feeding operation is obtained based on the current feeding video.

[0053] Whether to continue feeding the fish is determined based on the average feeding behavior quantification index corresponding to the current formal feeding operation and the average feeding behavior quantification index corresponding to the previous formal feeding operation.

[0054] Understandably, incorporating each formal feeding operation into an iterative loop, and through real-time analysis of the comparative analysis of the quantitative index of the "current" and "previous" feeding behavior, forms a dynamic judgment basis based on historical behavior, effectively avoiding misjudgments caused by fluctuations in single feeding events. By continuously acquiring current feeding videos and calculating the current quantitative index of feeding behavior, the feeding strategy and changes in fish behavior are tracked synchronously, responding promptly to changes in fish feeding needs over time and in different states. This recursive decision-making architecture can automatically optimize the feeding rhythm based on the dynamic evolution of fish feeding patterns, ultimately achieving a precise match between feed quantity and the actual needs of fish, reducing resource waste while ensuring a continuous supply for healthy fish growth.

[0055] Optionally, the method further includes:

[0056] Multiple feeding time periods are preset, and adaptive feeding is performed on the fish at the beginning of each feeding time period;

[0057] The cessation of feeding the fish includes:

[0058] If feeding the fish is stopped during the current feeding period, wait for the start of the next feeding period, reset all average feeding behavior quantification indices and feeding frequency to zero, and then resume the adaptive feeding of the fish.

[0059] Understandably, pre-setting multiple feeding time periods aligns with the biological rhythms of fish, and segmented management can accurately match the differences in feeding needs at different times, such as dawn and dusk, and day and night. The data clearing mechanism between different feeding time periods effectively isolates historical data interference, avoids misjudgments caused by the accumulation of behavior across time periods, and ensures independent optimization for each feeding cycle. The combination of periodic reset and adaptive algorithms ensures the rationality of current feeding while enabling continuous monitoring of aquaculture strategies through implicit correlation of cross-cycle data, ultimately achieving a spiral improvement in resource utilization and aquaculture efficiency.

[0060] According to a second aspect of this application, a vision-based adaptive fish feeding device is provided, the device comprising:

[0061] The feeding module is used to sequentially perform attraction and feeding operations and formal feeding operations on the fish in the aquaculture area; the formal feeding operation includes a first formal feeding operation and a second formal feeding operation performed sequentially.

[0062] The camera module is used to acquire feeding videos of the fish after the formal feeding operation; the feeding videos include a first feeding video and a second feeding video, the first feeding video being acquired after the first formal feeding operation and the second feeding video being acquired after the second formal feeding operation;

[0063] The processing module is used to obtain the average feeding behavior quantification index of the fish based on the feeding video; the average feeding behavior quantification index includes a first average feeding behavior quantification index corresponding to the first feeding video and a second average feeding behavior quantification index corresponding to the second feeding video;

[0064] The judgment module is used to obtain a feeding judgment result based on the first average feeding behavior quantification index and the second average feeding behavior quantification index, and to adaptively feed the fish according to the feeding judgment result.

[0065] Based on any of the above aspects, this application provides a vision-based adaptive feeding method and apparatus for fish, which sequentially performs attraction feeding operations and formal feeding operations on fish in a breeding area; the formal feeding operation includes a first formal feeding operation and a second formal feeding operation performed sequentially; after the formal feeding operation, feeding videos of the fish are acquired; the feeding videos include a first feeding video and a second feeding video, the first feeding video being acquired after the first formal feeding operation, and the second feeding video being acquired after the second formal feeding operation; an average feeding behavior quantification index of the fish is obtained based on the feeding videos; the average feeding behavior quantification index includes a first average feeding behavior quantification index corresponding to the first feeding video and a second average feeding behavior quantification index corresponding to the second feeding video; a feeding judgment result is obtained based on the first average feeding behavior quantification index and the second average feeding behavior quantification index, and adaptive feeding is performed on the fish based on the feeding judgment result. The method can achieve the following benefits:

[0066] • High accuracy in quantifying fish feeding behavior: Pre-feeding attracts and gathers fish, reducing errors caused by surface fluctuations during feeding behavior quantification; by calculating the feeding behavior quantification index for each feeding video, and precisely measuring the displacement fluctuation of each pixel in each frame of each feeding video, the feeding desire of fish at each interval can be accurately captured and represented with accurate quantification index data, improving the efficiency and accuracy of judgment, better enabling adaptive feeding of fish, and ensuring good environmental conditions for fish growth.

[0067] • Simple structure and easy implementation: The method for quantifying fish feeding behavior in this application only requires a camera for recording, a processor for the feeding video, and other conventional devices to ensure feeding in the aquaculture area. The structure and installation are very simple. In the operation and calculation process, only the processor is needed to process the feeding video and control the feeding operation. There is no need for too many complicated operation procedures, which ensures the simplicity and practicality of the implementation. Attached Figure Description

[0068] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0069] Figure 1 This is a schematic diagram of a vision-based adaptive feeding structure for fish provided in this embodiment.

[0070] Figure 2 This is a flowchart of a vision-based adaptive feeding method for fish provided in this embodiment.

[0071] Figure 3 This embodiment provides a process for obtaining a quantitative index of average feeding behavior in fish. Figure 1 .

[0072] Figure 4 This embodiment provides a process for obtaining a quantitative index of average feeding behavior in fish. Figure 2 .

[0073] Figure 5 This is a flowchart of the next formal feeding operation provided in this embodiment.

[0074] Figure 6 This is a schematic diagram of the functional modules of a vision-based adaptive fish feeding device provided in this embodiment.

[0075] Icons: 1-PLC controller, 2-camera equipment, 3-material throwing equipment, 4-hopper, 5-oxygenation equipment, 6-processor. Detailed Implementation

[0076] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this application. To better illustrate the following embodiments, some components in the drawings may be omitted, enlarged, or reduced, and do not represent the actual dimensions of the product; it is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0077] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0078] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0079] In today's fish farming industry, precise control of feeding strategies is crucial for ensuring fish nutrition and controlling water pollution. Traditional techniques typically involve manual monitoring and feeding of fish. However, this manual method relies on experience and judgment, leading to significant errors in feeding amount and timing, and a lack of objective standards, making scientific management difficult. While current quantitative behavioral feeding technologies based on video monitoring and image analysis can optimize feeding strategies using fish feeding behavior data, they face significant challenges in large-scale outdoor pond applications (typically over four acres): wind disturbances, wave fluctuations caused by aerator operation, and dynamic interference factors such as vegetation movement and human activity severely restrict the accurate extraction of feeding behavior characteristics. To address these complex scenarios, minimizing background noise interference in non-feeding areas has become a key breakthrough direction for improving the accuracy of fish feeding behavior recognition in dynamic environments.

[0080] This embodiment provides a technical solution that can solve the above problems. The specific implementation of this application will be described in detail below with reference to the accompanying drawings.

[0081] like Figure 1 As shown in the diagram, this embodiment provides a vision-based adaptive feeding structure for fish. The diagram illustrates the relevant structure of the aquaculture area. Preferably, the aquaculture area is a rectangular pond of more than four acres. A camera device 2 is installed in the middle of the bank of the aquaculture area to ensure real-time capture of the entire aquaculture area's water. The camera device 2 is connected to the input of the processor 6, enabling the processor 6 to process the feeding videos captured by the camera device 2. Preferably, the camera device 2 can be a high-definition waterproof camera, and the processor 6 can be a digital signal processor capable of processing the feeding videos represented as digital signals.

[0082] The feeding device 3 is fixed at the center of the rectangular aquaculture area, and the line connecting the positions of the feeding device 3 and the camera device 2 is parallel to one side of the rectangular aquaculture area, so that the camera device 2 can capture feeding videos centered on the feeding device 3.

[0083] The feed hopper 4 is installed on the outside right side of the breeding area and is used to replenish feed for the feeding device 3;

[0084] The oxygenation device 5 is installed in a corner of the aquaculture area. The PVC pipe (Polyvinyl Chloride Pipe) used for oxygen circulation is connected to the oxygenation device 5 and fixed to one side of the oxygenation device 5 to reduce water surface fluctuations caused by the operation of the oxygenation device 5 during feeding, thereby interfering with the quantification of water surface information in the feeding area.

[0085] The output of processor 6 is connected to the input of PLC controller 1 (Programmable Logic Controller), enabling processor 6 to process the input information received from camera device 2, analyze the real-time feeding desire of the fish through image processing technology, and then transmit the processing result to PLC controller 1. PLC controller 1 controls the working state of the feeding device 3 according to the processing result.

[0086] Understandably, in practical applications, this installation structure can be adapted to aquaculture areas of other shapes and sizes, and the equipment included in the installation structure can be appropriately adjusted according to the shape and size of the aquaculture area.

[0087] like Figure 2 As shown, this embodiment provides a vision-based adaptive feeding method for fish, which can be further divided into the following steps:

[0088] S100, Perform the attraction and feeding operation and the formal feeding operation on the fish in the breeding area in sequence; the formal feeding operation includes the first formal feeding operation and the second formal feeding operation performed in sequence.

[0089] In this embodiment, the attraction feeding operation releases feeding signals to the fish, causing them to gather in the feeding area. This reduces interference from subsequent water surface fluctuations caused by fish gathering on the quantification of feeding video information. The formal feeding operation is the actual feeding of the fish, through which they obtain feed and nutrients. By setting up a first and a second formal feeding operation performed sequentially, the feeding strategy can be adjusted by analyzing and comparing the changes in the fish's feeding desire after the two formal feeding operations, thus achieving adaptive feeding for the fish.

[0090] S200. After the formal feeding operation, acquire the feeding video of the fish; the feeding video includes a first feeding video and a second feeding video, the first feeding video is acquired after the first formal feeding operation, and the second feeding video is acquired after the second formal feeding operation;

[0091] In this embodiment, after the first formal feeding operation, it is necessary to acquire the corresponding first feeding video; similarly, after the second formal feeding operation, it is necessary to acquire the corresponding second feeding video. The first and second feeding videos can serve as the data basis for extracting the fish's feeding desire, providing data support for subsequent comparison of fish feeding desires.

[0092] Specifically, the attraction and feeding operation is performed for a preset attraction and feeding time to attract and gather the fish.

[0093] The first formal feeding operation lasts for a preset first feeding time, and the second formal feeding operation lasts for a preset second feeding time;

[0094] The attraction and feeding operation is separated from the first formal feeding operation by a preset first interval time;

[0095] After the first formal feeding operation is completed, a preset second interval is set before the second formal feeding operation begins, and the duration of the first feeding video is the second interval.

[0096] The duration of the second feeding video is the preset third interval time.

[0097] In this embodiment, it is necessary to preset the corresponding feeding time and interval time to make the entire adaptive feeding process more organized and scientific, and to make comparisons of data in the same dimension, thereby improving the accuracy of the comparison.

[0098] For example, in actual operation, the PLC controller 1 can first control the throwing device 3 to perform attraction feeding for a certain period of time. Preferably, the attraction feeding time can be set to 10 seconds.

[0099] After the attraction and feeding operation, a first interval is required before the first formal feeding operation is performed; preferably, the first interval can be set to 40 seconds.

[0100] After the first interval, the first formal feeding operation begins and continues for the first feeding time; preferably, the first feeding time can be set to 4 seconds.

[0101] After the first formal feeding operation, a second formal feeding operation is required after a second interval; preferably, the second interval can be set to 40 seconds; at the same time, after the first formal feeding operation, the camera device 6 starts to record the first feeding video for the second interval, which is used to monitor the feeding desire of the fish after the first formal feeding operation.

[0102] After the second interval, the second formal feeding operation begins and continues for the second feeding time; preferably, the first feeding time can be set to 4 seconds;

[0103] After the second formal feeding operation, a third interval is required before a subsequent judgment step is taken to determine whether to conduct the next formal feeding. Preferably, the third interval can be set to 40 seconds. At the same time, after the second formal feeding operation, a second feeding video is recorded using camera device 6 for a continuous third interval to monitor the fish's feeding desire after the second formal feeding operation, ensuring that the duration of the video recording is the same as that of the first feeding video, with a focus on comparing the fish's feeding desire.

[0104] S300. Obtain the average feeding behavior quantification index of the fish based on the feeding video; the average feeding behavior quantification index includes a first average feeding behavior quantification index corresponding to the first feeding video and a second average feeding behavior quantification index corresponding to the second feeding video;

[0105] In this embodiment, it is necessary to obtain the corresponding quantitative index of average feeding behavior based on each feeding video, so as to convert the feeding desire of fish into an objective and accurate value. This can avoid the error of the main board judgment under certain circumstances and improve the accuracy of extracting the feeding desire of fish.

[0106] Specifically, such as Figure 2 As shown, obtaining the average feeding behavior quantification index of the fish based on the feeding video may include the following steps:

[0107] S310. Extract several frames of feeding images corresponding to each feeding video;

[0108] In this embodiment, it is necessary to quantify and extract the feeding desire of fish in each frame of the feeding image, therefore it is necessary to obtain the sampling rate of the shooting device 3. That is, in a 1-second video of a food being eaten, it is possible to extract... Since it's a frame-by-frame video, the total duration of the feeding video needs to be extracted. A frame of images showing the act of feeding.

[0109] S320. Preprocess each frame of the feeding image to obtain a preprocessed feeding image;

[0110] In this embodiment, since optical flow is subsequently used to extract the displacement amplitude of each pixel in the feeding image, and the principle of optical flow is to track the displacement according to the brightness of each pixel, some brightness noise in the feeding image will seriously affect the displacement extraction. Therefore, it is necessary to preprocess each frame of the feeding image to repair the image quality affected by water surface reflection, which affects the extraction of fish feeding behavior, as well as the blurring features of water surface ripples caused by feeding, and enhance the ripple information features in each frame of the feeding image.

[0111] For example, in actual preprocessing, it is necessary to first convert the RGB (Red, Green, Blue) color space of each frame of the food image into the HSV (Hue, Saturation, Value) color space to separate the brightness (Value) from the color information (Hue and Saturation), so as to process the brightness channel independently without interfering with the color information.

[0112] In this embodiment, block histogram equalization is used to process the food image after color space conversion. By stretching the gray level distribution of each block image, the overall brightness of the sub-block is made more uniform, which facilitates subsequent optical flow processing.

[0113] First, each frame of the feeding image is divided into a preset number of blocks, and histogram equalization is performed on each sub-block using the following formula:

[0114]

[0115]

[0116] in, This represents the total number of pixels in each sub-block. The number of gray levels for each sub-block. The cutoff factor is , The average grayscale value for each sub-block. The mean squared error for each sub-block For the maximum permissible slope, This represents the upper limit of the shearing for each sub-histogram.

[0117] S330. Use optical flow to process each frame of the preprocessed feeding image to obtain the feeding behavior quantification index of each frame of the preprocessed feeding image;

[0118] Specifically, such as Figure 3 As shown, the step of processing each frame of the preprocessed feeding image using optical flow to obtain a quantitative index of feeding behavior for each frame of the preprocessed feeding image may include the following steps:

[0119] For each of the aforementioned feeding videos:

[0120] S331. Obtain information about each pixel in each frame of the preprocessed feeding image;

[0121] In this embodiment, the optical flow method requires processing each pixel of each frame of the feeding image, so it is necessary to first obtain the pixel information of each preprocessed frame of the feeding image. Preferably, the pixel information is the position information corresponding to the feeding image frame, including the horizontal coordinate and the vertical coordinate.

[0122] S332. Use the preprocessed feeding image of the second frame and the preprocessed feeding images of other frames thereafter as the processed images;

[0123] As can be understood, optical flow refers to the instantaneous velocity vector of pixels in an image caused by the movement of objects or the camera, which is expressed as the displacement of pixels in consecutive frames. The three prerequisites that optical flow must follow are: the brightness of pixels in consecutive frames is constant; each pixel has small, continuous motion in adjacent frames; and the motion of each pixel in adjacent frames has spatial consistency.

[0124] Therefore, whether the above-mentioned preprocessing and restoration of the food image meets the premise of the optical flow method, so that the displacement amplitude of each pixel can be extracted more accurately.

[0125] In this embodiment, since the optical flow method requires comparison and displacement extraction between two consecutive adjacent frames, the first frame of the feeding image is not processed by default. The preprocessed feeding images of the second frame and subsequent frames are used as the processed images, thereby extracting the displacement amplitude of each pixel in each processed image.

[0126] S333. Obtain the pre-processed feeding image of the previous frame adjacent to each of the processed images as the corresponding comparison image;

[0127] In this embodiment, since the optical flow method requires processing two consecutive frames of feeding images, it is necessary to obtain the preprocessed feeding image of the previous frame adjacent to each processed image as the corresponding comparison image to complete the displacement amplitude of each pixel in the optical flow method.

[0128] S334. Based on the optical flow method, obtain the corresponding pixel displacement amplitude according to the pixel information of each pixel in the processed image and the pixel information of the corresponding comparison image;

[0129] S335, Preset the weight corresponding to the information of each pixel;

[0130] In this embodiment, a larger Euclidean distance between a pixel in the feeding image and the origin is assigned a smaller weight value, indicating a lower importance of that pixel in quantifying the activity level of fish feeding behavior. The origin is defined as the location of the feeding device. Understandably, if fish have a strong feeding desire, they will continuously gather towards the feeding device, exhibiting active behavior and causing significant surface fluctuations near the device. Since the area near the feeding device is mostly inhabited by fish, pixels closer to the device have higher weights. Similarly, if fish have a low feeding desire, typically being satiated, they will not gather towards the feeding device, exhibiting stable behavior and resulting in minimal surface fluctuations in the area near the device and even the entire aquaculture area. Therefore, focusing on the area near the feeding device with high weights accurately reflects the fish's feeding desire.

[0131] In this embodiment, each pixel The original formula for the corresponding weights is:

[0132]

[0133] in, It is the distance between a pixel and the origin. It is a coefficient that adjusts the decay rate of the weighting coefficient;

[0134] However, due to the limitations of the natural environment in the breeding area, the camera was positioned on one side of the breeding area, resulting in the feeding area appearing to be approximately elliptical. (Any pixel in the feeding image...) and the origin The distance is calculated as follows:

[0135]

[0136] in and These are the lengths of the major and minor semi-axis of the elliptical feeding region, respectively; preferably, the feeding region is pre-determined and manually divided. and An appropriate length needs to be pre-set to divide the feeding area into suitable sizes.

[0137] Finally, the weights are adjusted using an improved Gaussian decay function. The original formula is transformed to obtain the optimized weights. .

[0138] Specifically, the weights are represented as follows:

[0139]

[0140] in The x and y axes represent the horizontal and vertical coordinates of the feeding and throwing equipment, respectively. , These are the major and minor axes of the pre-defined elliptical feeding area with the feeding and throwing device as the origin; , These are the x and y coordinates of the pixel, respectively. For the pixel point The corresponding weights.

[0141] S336. Obtain the feeding behavior quantification index corresponding to each pixel information based on the pixel displacement amplitude, weight, and the total number of pixels in the corresponding processed image.

[0142] S337. Obtain the feeding behavior quantization index of the processed image based on the feeding behavior quantization index of all pixels.

[0143] Specifically, the feeding behavior quantification index of each processed image frame Represented as:

[0144]

[0145] in, The total number of pixels in the processed image; This represents the pixel displacement amplitude corresponding to the pixel information. The weights are the information corresponding to each pixel.

[0146] S340. Obtain the average feeding behavior quantification index corresponding to each feeding video based on the feeding behavior quantification index of all the preprocessed feeding images corresponding to each feeding video.

[0147] Specifically, obtaining the average feeding behavior quantization index corresponding to each feeding video based on the feeding behavior quantization index of all preprocessed feeding images corresponding to each feeding video includes:

[0148] The average feeding behavior quantification index is obtained by summing and averaging the feeding behavior quantification indices of all the preprocessed feeding images corresponding to each feeding video.

[0149] In this embodiment, the above steps have already calculated the quantitative index of eating behavior corresponding to the eating video, and the average quantitative index of eating behavior for each eating video needs to be obtained by summing and averaging the quantitative indices of eating behavior for all the corresponding preprocessed eating images.

[0150] For example, as mentioned above, the duration of both the first and second feeding videos can be set to 40 seconds, and the timer can begin at the start of the first formal feeding operation. That is, the first formal feeding operation begins at 0 seconds, and the first formal feeding operation lasts for 4 seconds. Therefore, the first feeding video is started at 4 seconds and ends at 44 seconds.

[0151] Similarly, at 44 seconds, the second formal feeding operation begins, and the second formal feeding operation lasts for 4 seconds. Therefore, at 48 seconds, the second feeding video is started and ends at 88 seconds.

[0152] Therefore, the average quantitative index of eating behavior corresponding to the first eating video Average feeding behavior quantification index corresponding to the second feeding video for:

[0153]

[0154]

[0155] in, The sampling rate of the camera device. for -1 frame of feeding behavior quantification index, for -1 frame of feeding behavior quantification index.

[0156] S400: Obtain feeding judgment results based on the first average feeding behavior quantification index and the second average feeding behavior quantification index, and adaptively feed the fish according to the feeding judgment results.

[0157] Specifically, the step of obtaining a feeding judgment result based on the first average feeding behavior quantification index and the second average feeding behavior quantification index, and adaptively feeding the fish based on the feeding judgment result, includes:

[0158] Calculate the first feeding behavior quantitative index, which is a preset first multiple, as the first comparison threshold;

[0159] If the second feeding behavior quantification index is greater than or equal to the first comparison threshold, then the fish will continue with the next formal feeding operation after the third interval.

[0160] If the second feeding behavior quantification index is less than the first comparison threshold, then the current total amount of feed fed in the aquaculture area and the current total mass of the fish are obtained;

[0161] The current total mass is calculated as a preset second multiple as a second comparison threshold;

[0162] If the current total amount of feed fed is greater than the second comparison threshold, then feeding the fish should be stopped.

[0163] In this embodiment, a corresponding judgment and control strategy needs to be set based on the first average feeding behavior quantification index and the second average feeding behavior quantification index, and the fish are adaptively fed according to their feeding desire.

[0164] For example, the first multiple can be set to 0.9; therefore, if the second feeding behavior quantification index is greater than or equal to the first comparison threshold, specifically: Then, after the third interval, the fish will continue with the next formal feeding operation;

[0165] If the second feeding behavior quantification index is less than the first comparison threshold, specifically: Then, obtain the current total feed amount T in the aquaculture area and the current total mass of the fish. ;

[0166] For example, the second multiple can be set to 0.3; therefore, if the current total amount of feed fed is greater than the second comparison threshold, specifically: If this happens, stop feeding the fish.

[0167] By using the first and second average feeding behavior quantification indices, the fish's appetite after the first and second formal feedings is obtained, accurately determining whether the fish are eating. Through corresponding judgment strategies, the number and timing of feedings are controlled, enabling adaptive feeding of the fish, achieving precise nutrient supply to the fish and minimizing water pollution.

[0168] Specifically, such as Figure 4 As shown, the next formal feeding operation for the fish after the third interval includes:

[0169] S410. The next formal feeding operation is taken as the current formal feeding operation;

[0170] In this embodiment, if it is determined that the fish's appetite is still strong and the average feeding behavior quantification index after the second formal feeding is still high, then another formal feeding is required to ensure the fish's nutritional supply.

[0171] S420. The current formal feeding operation lasts for a certain feeding time, and a current feeding video of a certain duration is obtained after the current formal feeding operation. The average feeding behavior quantification index corresponding to the current formal feeding operation is obtained based on the current feeding video.

[0172] In this embodiment, the current formal feeding operation lasts for a certain feeding time. The certain feeding time can be the same as the feeding time of the first or second formal feeding, and can be adjusted appropriately according to the actual situation.

[0173] In acquiring the current feeding video of a certain duration, this duration needs to be the same as the recording duration of the previously recorded feeding video to ensure that the focus during comparison is on the average quantitative index of feeding behavior in the feeding videos. In this embodiment, the certain duration can be set to 40 seconds.

[0174] S430. Determine whether to continue feeding the fish based on the average feeding behavior quantification index corresponding to the current formal feeding operation and the average feeding behavior quantification index corresponding to the previous formal feeding operation.

[0175] In this embodiment, the same judgment method described above is used to determine whether to perform the next formal feeding, completing the iterative processing of judgment and feeding. It has the same scientific processing method to ensure that the processing judgment steps are carried out in an orderly manner.

[0176] Specifically, the method further includes:

[0177] Multiple feeding time periods are preset, and adaptive feeding is performed on the fish at the beginning of each feeding time period;

[0178] The cessation of feeding the fish includes:

[0179] If feeding the fish is stopped during the current feeding period, wait for the start of the next feeding period, reset all average feeding behavior quantification indices and feeding frequency to zero, and then resume the adaptive feeding of the fish.

[0180] Understandably, in general fish farming feeding operations, fish are usually fed in multiple time periods, such as in the morning, afternoon, and evening, which is in line with the biological feeding characteristics of fish.

[0181] Therefore, in this embodiment, the adaptive feeding method in this application is used to feed the fish in the aquaculture area during each feeding period. If it is determined that the fish have a low appetite during a feeding period, it means that the fish are almost satiated during that feeding period, so feeding the fish is stopped during that feeding period.

[0182] When the next feeding period begins, the fish start a new round of feeding behavior due to metabolic consumption. Therefore, it is necessary to reset all average feeding behavior quantification indices and feeding times to zero so as not to be affected by the feeding information of the previous round. The adaptive feeding method of this application can then be used to carry out a new round of adaptive feeding on the fish, completing multiple rounds of feeding and ensuring sufficient nutrition for the fish.

[0183] like Figure 5 As shown in the illustration, this application also provides a vision-based adaptive feeding device for fish. Optionally, the device includes:

[0184] The module includes a feeding module 511, a camera module 512, a processing module 513, and a judgment module 514, among which:

[0185] The feeding module 511 is used to sequentially perform attraction and feeding operations and formal feeding operations on the fish in the aquaculture area; the formal feeding operation includes a first formal feeding operation and a second formal feeding operation performed sequentially.

[0186] In this embodiment, the feeding module 511 can be used to perform... Figure 2 For a detailed description of the feeding module 511, please refer to the description of step S100 shown.

[0187] The camera module 512 is used to acquire feeding videos of the fish after the formal feeding operation; the feeding videos include a first feeding video and a second feeding video, wherein the first feeding video is acquired after the first formal feeding operation and the second feeding video is acquired after the second formal feeding operation;

[0188] In this embodiment, the camera module 512 can be used to perform... Figure 2 For a detailed description of the camera module 512, please refer to the description of step S200 shown.

[0189] Processing module 513 is used to obtain the average feeding behavior quantification index of the fish based on the feeding video; the average feeding behavior quantification index includes a first average feeding behavior quantification index corresponding to the first feeding video and a second average feeding behavior quantification index corresponding to the second feeding video;

[0190] In this embodiment, the processing module 513 can be used to execute... Figure 2 For a detailed description of the processing module 513, please refer to the description of step S300 shown.

[0191] The judgment module 514 is used to obtain a feeding judgment result based on the first average feeding behavior quantification index and the second average feeding behavior quantification index, and to adaptively feed the fish based on the feeding judgment result.

[0192] In this embodiment, the judgment module 514 can be used to perform... Figure 2 For a detailed description of the determination module 514, please refer to the description of step S400 shown.

[0193] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the technical solution of the present invention, and are not intended to limit the specific implementation of the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the claims of the present invention should be included within the protection scope of the claims of the present invention.

Claims

1. A vision-based adaptive fish feeding method, characterized in that, The method comprises: sequentially performing an attracting feeding operation and a formal feeding operation on fish in a culture area; the formal feeding operation comprises sequentially performed first and second formal feeding operations; after the formal feeding operation, obtaining feeding videos of the fish; the feeding videos comprise first and second feeding videos, the first feeding video is obtained after the first formal feeding operation, and the second feeding video is obtained after the second formal feeding operation; According to the feeding video, an average feeding behavior quantization index of the fish is obtained; the average feeding behavior quantization index includes a first average feeding behavior quantization index corresponding to the first feeding video and a second average feeding behavior quantization index corresponding to the second feeding video; wherein the average feeding behavior quantization index of the fish is obtained according to the feeding video, including: extracting a plurality of frame feeding images corresponding to each feeding video; respectively pre-processing each frame feeding image to obtain a pre-processed feeding image; using an optical flow method to process each frame pre-processed feeding image to obtain a feeding behavior quantization index of each frame pre-processed feeding image; obtaining the average feeding behavior quantization index corresponding to each feeding video according to the feeding behavior quantization index of all pre-processed feeding images corresponding to each feeding video; the feeding behavior quantization index of each frame pre-processed feeding image is obtained by using the optical flow method to process each frame pre-processed feeding image, including: for each feeding video: obtaining each pixel point information of each frame pre-processed feeding image; taking the second frame pre-processed feeding image and other frame pre-processed feeding images after the second frame pre-processed feeding image as a processing image; obtaining each adjacent previous frame pre-processed feeding image of each processing image as a corresponding comparison image; based on the optical flow method, the corresponding pixel point displacement amplitude is obtained according to each pixel point information of the processing image and the pixel point information of the corresponding comparison image; a weight corresponding to each pixel point information is pre-set; the weight is represented as: ; wherein X and Y are respectively a horizontal coordinate and a vertical coordinate of a feeding throwing device, , a and b are respectively a major semi-axis and a minor semi-axis of a pre-planned elliptical feeding area with the feeding throwing device as the origin; , X and Y are respectively a horizontal coordinate and a vertical coordinate of a pixel point, is a weight corresponding to the pixel point ; the feeding behavior quantization index corresponding to each pixel point information is obtained according to the pixel point displacement amplitude corresponding to each pixel point information, the weight and the total number of pixel points of the corresponding processing image; the feeding behavior quantization index of the processing image is obtained according to the feeding behavior quantization index of all pixel points. obtaining a feeding judgment result according to the first and second average feeding behavior quantification indexes, and performing adaptive feeding on the fish according to the feeding judgment result.

2. The method of claim 1, wherein, The attracting feeding operation lasts for a preset attracting feeding time, and is used for completing attracting and gathering of the fish. The first formal feeding operation lasts for a preset first feeding time, and the second formal feeding operation lasts for a preset second feeding time. The attracting feeding operation is separated from the first formal feeding operation by a preset first interval time. After the first formal feeding operation is completed, a second interval time is separated before the second formal feeding operation is started, and a time length of the first feeding video is the second interval time. A time length of the second feeding video is a preset third interval time.

3. The method of claim 1, wherein, The feeding behavior quantification index of the image of each frame is represented as: wherein, is the total number of pixel points in the processed image; is the pixel point displacement amplitude corresponding to the pixel point information, is the weight corresponding to the pixel point information; The average feeding behavior quantification index corresponding to each feeding video is obtained according to feeding behavior quantification indexes of all the preprocessed feeding images corresponding to each feeding video, and comprises: the average feeding behavior quantification index is obtained by summing and averaging feeding behavior quantification indexes of all the preprocessed feeding images corresponding to each feeding video.

4. The method of claim 1, wherein, The feeding judgment result is obtained according to the first and second average feeding behavior quantification indexes, and adaptive feeding is performed on the fish according to the feeding judgment result, and comprises: a first feeding behavior quantification index of a preset first multiple is calculated as a first comparison threshold value; if the second feeding behavior quantification index is greater than or equal to the first comparison threshold value, the next formal feeding operation is continued on the fish after the third interval time; if the second feeding behavior quantification index is less than the first comparison threshold value, a current total feeding amount of the culture area and a current total mass of the fish are obtained; a second multiple of the current total mass is calculated as a second comparison threshold value; if the current total feeding amount is greater than the second comparison threshold value, the fish is stopped from being fed.

5. The method of claim 4, wherein, The next formal feeding operation is continued on the fish after the third interval time, and comprises: the next formal feeding operation is taken as a current formal feeding operation; the current formal feeding operation lasts for a certain feeding time, a current feeding video of a certain time length is obtained after the current formal feeding operation, and an average feeding behavior quantification index corresponding to the current formal feeding operation is obtained according to the current feeding video; whether the fish is continued to be fed is judged according to the average feeding behavior quantification index corresponding to the current formal feeding operation and an average feeding behavior quantification index corresponding to a last formal feeding operation.

6. The method of claim 4, wherein, The method further comprises: presetting a plurality of feeding time periods, and performing the adaptive feeding on the fish at the beginning of the feeding time periods; the stopping of feeding the fish comprises: if the feeding of the fish is stopped in the current feeding time period, waiting for the beginning of the next feeding time period, clearing all the average feeding behavior quantification indexes and feeding times, and performing the adaptive feeding on the fish again.

7. A vision-based fish self-adaptive feeding device, characterized in that, The device comprises: a feeding module configured to sequentially perform an attracting feeding operation and a formal feeding operation on the fish in the breeding area; the formal feeding operation comprises sequentially performed first and second formal feeding operations; The camera module is used to obtain feeding videos of the fish after the formal feeding operation; the feeding videos include a first feeding video and a second feeding video, the first feeding video is obtained after the first formal feeding operation, and the second feeding video is obtained after the second formal feeding operation; wherein the average feeding behavior quantitative index of the fish is obtained according to the feeding videos, which includes: extracting a plurality of frames of feeding images corresponding to each feeding video; respectively pre-processing each frame of the feeding images to obtain pre-processed feeding images; using an optical flow method to process each frame of the pre-processed feeding images to obtain a feeding behavior quantitative index of each frame of the pre-processed feeding images; obtaining the average feeding behavior quantitative index corresponding to each feeding video according to the feeding behavior quantitative indexes of all the pre-processed feeding images corresponding to each feeding video; the optical flow method is used to process each frame of the pre-processed feeding images to obtain a feeding behavior quantitative index of each frame of the pre-processed feeding images, which includes: for each feeding video: obtaining each pixel point information of each frame of the pre-processed feeding images; taking the second frame of the pre-processed feeding images and other frames of the pre-processed feeding images after the second frame as processing images; taking each adjacent previous frame of the pre-processed feeding images as a corresponding comparison image; based on the optical flow method, obtaining a corresponding pixel point displacement amplitude according to each pixel point information of the processing image and the pixel point information of the corresponding comparison image; presetting a weight corresponding to each pixel point information; the weight is represented as: ; wherein X and Y are respectively a horizontal coordinate and a vertical coordinate of a feeding throwing device, , a and b are respectively a major semi-axis and a minor semi-axis of an elliptical feeding area with the feeding throwing device as an origin, , X and Y are respectively a horizontal coordinate and a vertical coordinate of a pixel point, is the weight corresponding to the pixel point ; a feeding behavior quantitative index corresponding to each pixel point information is obtained according to the pixel point displacement amplitude corresponding to each pixel point information, the weight and the total number of pixel points of the corresponding processing image; a feeding behavior quantitative index of the processing image is obtained according to the feeding behavior quantitative indexes of all the pixel points. a processing module configured to acquire the average feeding behavior quantification indexes of the fish according to the feeding videos; the average feeding behavior quantification indexes comprise first and second average feeding behavior quantification indexes corresponding to the first and second feeding videos, respectively; a judging module configured to acquire a feeding judgment result according to the first and second average feeding behavior quantification indexes, and perform adaptive feeding on the fish according to the feeding judgment result.

8. The apparatus of claim 7, wherein, The camera module further comprises: the attracting feeding operation lasts for a preset attracting feeding time, and is configured to complete the attracting and gathering of the fish; the first formal feeding operation lasts for a preset first feeding time, and the second formal feeding operation lasts for a preset second feeding time; the attracting feeding operation is separated from the first formal feeding operation by a preset first interval time; after the completion of the first formal feeding operation, a second interval time is provided before the start of the second formal feeding operation, and the time length of the first feeding video is the second interval time; the time length of the second feeding video is a preset third interval time.

9. The apparatus of claim 7, wherein, The processing module further comprises: The feeding behavior quantification index of the image of each frame is represented as: wherein, is the total number of pixel points in the processed image; is the pixel point displacement amplitude corresponding to the pixel point information, is the weight corresponding to the pixel point information. and / or, the acquisition of the average feeding behavior quantification index corresponding to each feeding video from the feeding behavior quantification indexes of all the preprocessed feeding images corresponding to each feeding video comprises: summing the feeding behavior quantification indexes of all the preprocessed feeding images corresponding to each feeding video and then averaging to obtain the average feeding behavior quantification index.

10. The apparatus of claim 7, wherein, The judging module further comprises: calculating a first feeding behavior quantification index of a preset first multiple as a first comparison threshold value; if the second feeding behavior quantification index is greater than or equal to the first comparison threshold value, continuing the next formal feeding operation on the fish after the third interval time; if the second feeding behavior quantification index is less than the first comparison threshold value, acquiring a current total feeding amount of feed in the breeding area and a current total mass of the fish; calculating the current total mass of the fish of a preset second multiple as a second comparison threshold value; if the current total feeding amount of feed is greater than the second comparison threshold value, stopping the feeding of the fish.

Images