Big-data aquaculture full-automatic intelligent feeding method and system
Through big data analysis, the collected water quality and environmental data and feeding images are dynamically adjusted, and the amount of feeding in aquaculture is solved, which has failed to consider water quality changes in traditional feeding methods, improved the scientificity and accuracy of feeding, and ensured the healthy growth of fish and the sustainability of water quality.
Patent Information
- Application Number
- CN202510606374.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-12
- Publication Date
- 2025-06-27
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional aquaculture feeding methods fail to effectively consider the impact of changes in water quality and environment on the feeding capacity of fish schools, resulting in too much or too little feeding, affecting the healthy growth of fish schools and aggravating the deterioration of water quality.
The fully automatic intelligent feeding method of aquaculture is adopted with big data. By collecting water quality and environmental data and feeding images, the degree of feeding in fish schools and the impact of water quality environment on fish schools are analyzed, and the feeding volume is dynamically adjusted to ensure that the appropriate feeding volume is selected under different water quality conditions.
It improves the scientificity and accuracy of feeding decisions, avoids the impact of the water quality environment on the feeding capacity of fish schools, and ensures the healthy growth of fish schools and the sustainability of water quality.
Smart Images
Figure CN120202978A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of data processing, and in particular, to a fully automatic intelligent feeding method and system for aquaculture of big data. Background Art
[0002] In the context of the continuous growth of the global population and the increasing demand for aquatic products, the aquaculture industry shoulders the important responsibilities of ensuring food supply security and promoting economic development. According to relevant statistics, in the past few decades, the global aquaculture production has shown a steady upward trend, providing a large amount of high-quality protein sources for humans. However, there are many drawbacks in the feeding management of traditional aquaculture models, which seriously restrict the sustainable development of the industry. For example, in traditional feeding methods (such as manual experience feeding, fixed-time and fixed-quantity feeding, etc.), farmers mainly rely on long-term accumulated experience to judge the feeding amount, but this method has significant defects. For example, when the water quality environment is poor, it is easy to cause overfeeding. When overfeeding, the feed is not only wasted but also causes water pollution, increasing the breeding cost; when underfeeding, the fish population will lack nutrition, grow slowly, and reduce the breeding efficiency. Therefore, the present invention proposes a fully automatic intelligent feeding method and system for aquaculture of big data.
[0003] It is known that traditional feeding methods (such as manual experience feeding, fixed-time and fixed-quantity feeding, etc.) do not consider the impact of changes in water quality environment on the feeding ability of aquatic products. If the water quality environment deteriorates and the feeding amount is not adjusted appropriately, it may overfeed the aquatic products when their feeding ability declines, affecting their healthy growth, and also causing the excessive feed to decompose in the water, leading to further deterioration of water quality. Summary of the Invention
[0004] In order to solve the technical problem that the traditional feeding method does not consider the impact of changes in water quality environment on the feeding ability of fish populations, resulting in overfeeding when the feeding amount is too much, affecting the growth of aquatic products and leading to further deterioration of water quality, the present invention provides a fully automatic intelligent feeding method and system for aquaculture of big data.
[0005] In the first aspect, the present invention provides a fully automatic intelligent feeding method for aquaculture of big data, adopting the following technical scheme: A fully automatic intelligent feeding method for aquaculture of big data includes the steps: Collect water quality environment data and feeding images during each feeding process; obtain the degree of influence of water quality environment data on the fish population during each feeding process; obtain the probability of each fish moving towards the feeding area in each frame of the feeding image during each feeding process; based on the probability, obtain the target image of each fish during each feeding process; use the ratio of the number of frames of the target image of each fish during each feeding process to the total number of frames of the feeding image during each feeding process as the probability that the movement trajectory of each fish during each feeding process tends to the feeding area; obtain the degree of fish feeding tendency during each feeding process , represents the degree of fish feeding tendency during the b-th feeding process; and respectively represent the movement speed of the i-th fish in the feeding video and the probability that the movement trajectory of the i-th fish tends to the feeding area during the b-th feeding process; I represents the number of fish during the b-th feeding process; Based on the degree of fish feeding tendency and the degree of influence of water quality environment data on the fish population, obtain the feeding requirement degree of each feeding process; cluster the water quality environment data in all feeding processes according to the feeding requirement degree to obtain each clustering cluster and the feeding weight of each clustering cluster; obtain the preference degree of each clustering cluster; use the product of the feeding weight of the clustering cluster corresponding to the maximum preference degree and the current feeding amount as the current corrected feeding amount.
[0006] The innovation of the present invention lies in that firstly, by analyzing the degree of influence of different water quality environment data on the fish population and the activity degree of the fish population under different water quality environment data in historical data, the feeding requirement degree under different water quality environment data is obtained, clarifying the relationship between different water quality environments and the feeding requirement degree, avoiding blind feeding, and improving the scientificity of feeding decision-making; then, clustering the water quality environment data in all feeding processes with the feeding requirement degree of each feeding process as the characteristic clustering to obtain each clustering cluster and the feeding weight of each clustering cluster; analyzing the similarity between the current water quality environment data and the historical water quality environment data to obtain the preference degree of each clustering cluster, and using the feeding weight of the clustering cluster corresponding to the maximum preference degree as the feeding weight of the current feeding, which can dynamically adjust the feeding weight of the current feeding according to the real-time water quality change, ensuring to select a more suitable feeding amount under different water quality conditions; finally, correcting the current feeding amount with the feeding weight of the current feeding, avoiding the influence of water quality environment data on the feeding ability of fish, and improving the accuracy of the current feeding.
[0007] Preferably, the obtaining the degree of influence of water quality environment data on the fish population during each feeding process includes: , represents the comfort factor of the temperature data during the b-th feeding process; represents the value of the c-th temperature data during the b-th feeding process; represents the suitable temperature for the growth of water; represents the number of temperature data during the b-th feeding process; , represents the influence degree of the water quality environment data on the fish population during the b-th feeding process; , and respectively represent the weights of the comfort factors of the temperature data, humidity data, and oxygen content data during the b-th feeding process; , and respectively represent the comfort factors of the temperature data, humidity data, and oxygen content data during the b-th feeding process.
[0008] It is convenient to subsequently obtain the feeding requirement degree of each feeding process.
[0009] Preferably, the obtaining of the possibility that each fish moves towards the feeding area in each feeding image during each feeding process includes: Denote the motion vector of the i-th fish in the j-th feeding image during the b-th feeding process as , where and represent the position coordinates of the i-th fish in the j-th feeding image and the (j + 1)-th feeding image during the b-th feeding process; Denote the vector from the i-th fish in the j-th feeding image to the feeding area during the b-th feeding process as ; where represents the position coordinates of the feeding area; The possibility that the i-th fish moves towards the feeding area in the j-th feeding image during the b-th feeding process is ; || represents the absolute value symbol.
[0010] Preferably, the obtaining of the position coordinates of the i-th fish in the j-th feeding image and the (j + 1)-th feeding image during the b-th feeding process includes: Preset a possibility threshold T, and obtain the possibility that the i-th fish moves towards the feeding area in the j-th feeding image during the b-th feeding process; If the possibility that the i-th fish moves towards the feeding area in the j-th feeding image during the b-th feeding process is greater than the possibility threshold T, denote the j-th feeding image as the target image of the i-th fish during the b-th feeding process.
[0011] Preferably, the obtaining of the position coordinates of the i-th fish in the j-th feeding image and the (j + 1)-th feeding image during the b-th feeding process includes: Perform frame segmentation on the video of the b-th feeding to obtain a number of feeding images; use the YOLO object detection algorithm to identify the position coordinates of each fish in each feeding image, and combine the DeepSORT algorithm with the Hungarian algorithm to match adjacent feeding images, so as to identify the position coordinates of the same fish in each feeding image of the video of the b-th feeding.
[0012] Preferably, the obtaining of the movement speed of the i-th fish in the feeding video during the b-th feeding process includes: ; In the formula, represents the movement speed of the i-th fish in the feeding video during the b-th feeding process; represents the position coordinates of the i-th fish in the j-th feeding image during the b-th feeding process; represents the position coordinates of the i-th fish in the (j + 1)-th feeding image during the b-th feeding process; represents the total number of frames of the feeding images during the b-th feeding process.
[0013] Preferably, the obtaining of the feeding demand degree of each feeding process includes: ; In the formula, represents the feeding demand degree of the b-th feeding process; represents the influence degree of the water quality environment data on the fish school during the b-th feeding process; represents the feeding tendency degree of the fish school during the b-th feeding process; exp() represents the exponential function with the natural constant as the base; exp() represents the exponential function with the natural constant as the base.
[0014] It is convenient to cluster the water quality environment data in each feeding process according to the feeding demand degree of each feeding process later.
[0015] Preferably, the obtaining of the feeding weight of each clustering cluster includes: ; In the formula, represents the feeding weight of the d-th clustering cluster; represents the average value of the feeding demand degrees in the d-th clustering cluster; represents the maximum value among the average values of the feeding demand degrees in all clustering clusters.
[0016] It is convenient to correct the feeding amount according to the feeding weight later.
[0017] Preferably, the obtaining of the preference degree of each clustering cluster includes: ; In the formula, represents the preference degree of the d-th cluster; represents the sum of comfort factors of water quality environment data for all dimensions during the current feeding; represents the sum of comfort factors of water quality environment data for all dimensions during the n-th feeding process in the d-th cluster; represents the number of feeding times in the d-th cluster; || represents the absolute value symbol; exp() represents the exponential function with the natural constant as the base.
[0018] It is convenient to use the feeding weight of the cluster corresponding to the maximum preference degree as the feeding weight of the current feeding, and the feeding weight of the current feeding can be dynamically adjusted according to the real-time water quality change, so as to ensure a more suitable feeding amount under different water quality conditions.
[0019] In the second aspect, the present invention provides a fully automatic intelligent feeding system for aquaculture using big data, adopting the following technical solutions: A fully automatic intelligent feeding system for aquaculture using big data includes: a processor and a memory, and the memory stores computer program instructions, which implement the above-mentioned fully automatic intelligent feeding method for aquaculture using big data when executed by the processor.
[0020] By adopting the above technical solutions, the above-mentioned fully automatic intelligent feeding method for aquaculture using big data is generated into a computer program and stored in the memory, so as to be loaded and executed by the processor, and thus a terminal device is made according to the memory and the processor, which is convenient to use.
[0021] The present invention has the following technical effects: The purpose of the present invention is to collect historical data, analyze the influence degree of different water quality environment data on the fish population and the activity degree of the fish population under different water quality environment data in the historical data, and then obtain the feeding demand degree under different water quality environment data, which is convenient for subsequent analysis of the similarity between the current water quality environment data and the historical water quality environment data to judge the feeding demand degree under the current water quality environment data and avoid blind feeding; then, clustering the water quality environment data in all feeding processes with the feeding demand degree of each feeding process as the characteristic clustering to obtain each cluster and the feeding weight of each cluster; analyzing the similarity between the current water quality environment data and the historical water quality environment data to obtain the preference degree of each cluster; using the feeding weight of the cluster corresponding to the maximum preference degree as the feeding weight of the current feeding; correcting the current feeding amount with the feeding weight of the current feeding, avoiding the influence of water quality environment data on the feeding ability of fish, and improving the accuracy of the current feeding. BRIEF DESCRIPTION OF THE DRAWINGS
[0022] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features, and advantages of the exemplary embodiments of the present invention will become readily understood. In the drawings, several embodiments of the present invention are shown in an exemplary rather than restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts.
[0023] Figure 1 It is a flowchart of a method for fully automatic intelligent feeding in aquaculture of big data in an embodiment of the present invention. Specific embodiments
[0024] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts fall within the protection scope of the present invention.
[0025] It should be understood that when terms such as "first" and "second" are used in the claims, specifications, and drawings of the present invention, they are only used to distinguish different objects and not to describe a specific order. The terms "including" and "comprising" used in the specifications and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their combinations.
[0026] An embodiment of the present invention discloses a method for fully automatic intelligent feeding in aquaculture of big data. Referring to Figure 1 , it includes steps S1 - S4: S1: Collect water quality environment data and feeding videos during each feeding process in the past year.
[0027] In an embodiment of the present invention, temperature data, humidity data, and oxygen content data during each feeding process in the past year are obtained from a database to obtain water quality environment data during each feeding process in the past year; and, synchronously, feeding videos during each feeding process in the past year are obtained from the database; The water quality environment data during each feeding process includes several dimensional features, namely temperature, humidity, and oxygen content.
[0028] S2: Obtain the comfort factor of the water quality environment data for each dimension during each feeding process, and based on the comfort factor, obtain the degree of influence of the water quality environment data on the fish population during each feeding process.
[0029] It should be noted that traditional feeding methods (such as manual experience feeding, timed and quantitative feeding, etc.) do not consider the impact of changes in water quality environment on the feeding ability of fish populations. If the water quality environment deteriorates and the feeding amount is not adjusted appropriately, it may lead to overfeeding of the fish population when their feeding ability declines, affecting their healthy growth. Moreover, excessive feed decomposes in the water, causing further deterioration of water quality. Therefore, in the present invention, water quality environment data during each historical feeding process is collected and analyzed. On the one hand, accurately obtain the degree of influence of water quality environment data during each feeding process on the fish population. On the other hand, analyze the feeding tendency degree of the fish population during feeding under the corresponding water quality environment data. Finally, based on these two obtained indicators, obtain the feeding requirement degree of each feeding process, which reflects the true demand degree of the fish population for feed under different water quality environment data.
[0030] It should be further noted that when the value of water quality environment data in any dimension during any feeding process has a small difference from the standard value of this dimension, it indicates that the water quality environment data in this dimension has not deteriorated, and at this time, the comfort factor of the water quality environment data in this dimension is relatively large. Through analysis, it can be known that the greater the comfort factor of water quality environment data in each dimension during each feeding process, the smaller the degree of influence of water quality environment data during each feeding process on the fish population. It is known that temperature directly affects the metabolic rate of fish, oxygen is a key substance for fish to carry out respiration and is also directly related to the survival of fish, while humidity mainly indirectly affects the fish population by influencing water evaporation and water vapor content in the air. Therefore, when obtaining the degree of influence of water quality environment data during each feeding process on the fish population, a relatively large weight is set for the comfort factors of temperature data and oxygen content data during each feeding process.
[0031] In the embodiments of the present invention, refer to aquaculture books according to the types of aquatic fish populations to obtain the suitable temperature for aquatic growth; obtain the suitable humidity for aquatic growth and the suitable oxygen content for aquatic growth. Obtain the comfort factor of temperature data during each feeding process: ; In the formula, represents the comfort factor of temperature data during the b-th feeding process; represents the value of the c-th temperature data during the b-th feeding process; represents the suitable temperature for aquatic growth; represents the number of temperature data during the b-th feeding process; The larger the value of
[0032] Similarly, obtain the comfort factor of the humidity data and the comfort factor of the oxygen content data during each feeding process.
[0033] Obtain the degree of influence of the water quality environment data on the fish population during each feeding process: ; In the formula, represents the degree of influence of the water quality environment data on the fish population during the b-th feeding process; represents the weight of the comfort factor of the temperature data during the b-th feeding process; represents the comfort factor of the temperature data during the b-th feeding process; represents the comfort factor of the humidity data during the b-th feeding process; represents the weight of the comfort factor of the humidity data during the b-th feeding process; represents the comfort factor of the oxygen content data during the b-th feeding process; represents the weight of the comfort factor of the oxygen content data during the b-th feeding process; In the embodiment of the present invention, it is preset ; ; , in other embodiments, the implementer can preset the value of the weight according to the specific implementation situation; the smaller the value of the comfort factor of any item of water quality environment data during the b-th feeding process, that is, the worse the water quality environment; it indicates that the degree of influence of this item of water quality environment data on the fish population is greater.
[0034] S3: Obtain the probability that the movement trajectory of each fish during each feeding process tends to the feeding area and the movement speed of each fish in the feeding video during each feeding process; based on the probability and the movement speed, obtain the degree of fish foraging during each feeding process; according to the degree of fish foraging during each feeding process and the degree of influence of the water quality environment data on the fish population during each feeding process, obtain the feeding requirement degree of each feeding process.
[0035] It should be noted that after obtaining the degree of influence of different water quality environment data on the fish population, analyze the degree of fish foraging when feeding under the corresponding water quality environment data. It is known that the feeding video during each feeding process is obtained. The present invention first obtains the movement trajectory of each fish during the feeding process according to the feeding video. If the movement trajectory of each fish in the feeding video tends more to the feeding area and the movement speed is faster, it indicates that the degree of fish foraging during the feeding process is higher.
[0036] In the embodiment of the present invention, the specific method for obtaining the degree of fish foraging during the b-th feeding process is as follows: Perform frame-by-frame processing on the b-th feeding video to obtain a number of feeding images; use the YOLO object detection algorithm to identify the position coordinates of each fish in each feeding image, and combine the DeepSORT algorithm with the Hungarian algorithm to match adjacent feeding images, thereby identifying the position coordinates of the same fish in each feeding image of the b-th feeding video; Obtain the coordinate position with the most occurrences of the fish's coordinate positions in all frames of feeding images, and record it as the position coordinates of the feeding area; Obtain the motion vector of the i-th fish in the j-th feeding image during the b-th feeding process, denoted as where, represents the position coordinates of the i-th fish in the j-th feeding image during the b-th feeding process; represents the position coordinates of the i-th fish in the (j + 1)-th feeding image during the b-th feeding process; Obtain the vector from the i-th fish in the j-th feeding image to the feeding area during the b-th feeding process, denoted as ; where, represents the position coordinates of the feeding area; Obtain the possibility that the i-th fish in the j-th feeding image during the b-th feeding process tends to move towards the feeding area: ; where, represents the motion vector of the i-th fish in the j-th feeding image during the b-th feeding process; represents the vector from the i-th fish in the j-th feeding image to the feeding area during the b-th feeding process; || represents the absolute value symbol.
[0037] Preset a possibility threshold T. In the embodiments of the present invention, the preset possibility threshold T = 0.5. In other embodiments, the implementer can preset the value of T according to the specific implementation situation. If the possibility that the i-th fish in the j-th feeding image during the b-th feeding process tends to move towards the feeding area is greater than the possibility threshold T, it means that the i-th fish in the j-th feeding image during the b-th feeding process is tending to move towards the feeding area, and the j-th feeding image is recorded as the target image of the i-th fish during the b-th feeding process.
[0038] Obtain the probability that the motion trajectory of each fish during the b-th feeding process tends to the feeding area: ; In the formula, represents the probability that the motion trajectory of the i-th fish during the b-th feeding process tends to the feeding area; represents the number of frames of the target image of the i-th fish during the b-th feeding process; represents the total number of frames of the feeding image during the b-th feeding process; among them, the proportion of the number of frames in which the fish move towards the feeding area in the total number of frames, the higher the proportion, the more the fish tend to move towards the feeding area.
[0039] Obtain the movement speed of each fish in the feeding video during the b-th feeding process: ; In the formula, represents the movement speed of the i-th fish in the feeding video during the b-th feeding process; represents the position coordinates of the i-th fish in the j-th feeding image during the b-th feeding process; represents the position coordinates of the i-th fish in the (j + 1)-th feeding image during the b-th feeding process; represents the total number of frames of the feeding image during the b-th feeding process.
[0040] Obtain the degree of fish flock's food-seeking during each feeding process: ; In the formula, represents the degree of fish flock's food-seeking during the b-th feeding process; represents the movement speed of the i-th fish in the feeding video during the b-th feeding process; represents the probability that the movement trajectory of the i-th fish during the b-th feeding process tends to the feeding area; I represents the number of fish during the b-th feeding process; when the movement speed of the fish in the feeding video is faster and the movement trajectory is more inclined to the feeding area during the feeding process, the degree of fish flock's food-seeking during the feeding process is greater.
[0041] So far, the degree of fish flock's food-seeking during the b-th feeding process has been obtained. According to the method for obtaining the degree of fish flock's food-seeking during the b-th feeding process, the degree of fish flock's food-seeking during each feeding process is obtained.
[0042] It should be noted that when the degree of fish flock's food-seeking is greater under any water quality environment data, it indicates that the feeding demand degree during the feeding process is relatively large at this time. However, since the quality of water environment data has a great impact on the feeding ability of fish. For example, within a suitable water temperature range, the metabolism of fish speeds up, the food intake increases, the degree of fish flock's food-seeking is higher and the demand for the feeding amount increases. While when the water temperature is too high or too low, the appetite and the degree of fish flock's food-seeking of fish will decline, so the demand for the feeding amount will also decrease. Therefore, considering comprehensively the degree of fish flock's food-seeking and the influence of water quality environment data during each feeding process, the feeding demand degree during each feeding process is obtained; if all water quality indicators are within the suitable range and the fish flock is active, the feeding demand degree during the feeding process is relatively large; if the water quality environment is not good, after reducing the degree of fish flock's food-seeking during the feeding process, the obtained feeding demand degree during the feeding process is relatively small, that is, when the water quality environment is not good, feeding should be carried out with caution.
[0043] Obtain the feeding demand degree of each feeding process: ; In the formula, represents the feeding demand degree of the b-th feeding process; represents the influence degree of water quality environment data on fish schools during the b-th feeding process; represents the degree of fish approaching food during the b-th feeding process; when the value of is larger, it indicates that the fish school in the b-th feeding process is more inclined to the feeding area, and at this time, the feeding demand degree of the b-th feeding process is larger. the larger the value of, it indicates that the water quality environment is relatively poor during the b-th feeding process, which affects the appetite of the fish school. Therefore, the larger the value of, the smaller the feeding demand degree of the b-th feeding process.
[0044] S4: Cluster the water quality environment data in each feeding process according to the feeding demand degree of each feeding process, and obtain the feeding weight of each cluster; obtain the preference degree of each cluster, and use the feeding weight of the cluster with the largest preference degree as the feeding weight of the current feeding; after correcting the current feeding amount according to the feeding weight of the current feeding, obtain the corrected feeding amount.
[0045] It should be noted that the feeding demand degrees under different water quality environment data have been obtained. Therefore, in the present invention, the water quality environment data in all feeding processes are clustered with the feeding demand degree of each feeding process as the feature clustering, and the water quality environment data with similar characteristics are grouped into one category to obtain several clusters. It is known that if the average value of the feeding demand degree in any cluster is larger, it indicates that the feeding weight of this cluster is larger. Also, since each cluster represents the feeding demand degree corresponding to the fish school under a specific water quality environment, the similarity between the water quality environment data in the current feeding process and the water quality environment data in each cluster is analyzed, and the feeding weight of the cluster with the water quality environment data in the current feeding process being approximated is used as the feeding weight of the current feeding, thereby correcting the current feeding amount and improving the accuracy of the feeding amount.
[0046] In the embodiment of the present invention, using the DBSCAN algorithm, the water quality environment data in all feeding processes are clustered with the feeding demand degree of each feeding process as the feature clustering to obtain each cluster, and each cluster contains the water quality environment data and feeding demand degree of several feeding processes. Obtain the feeding weight of each cluster: ; In the formula, Represents the feeding weight of the d-th clustering cluster; Represents the average value of the feeding demand degree in the d-th clustering cluster; Represents the maximum value among the average values of the feeding demand degrees in all clustering clusters; The closer the value of approaches 1, it indicates that the water quality environment data in each feeding process corresponding to this clustering cluster is better and the fish feeding degree is higher. At this time, the feeding weight of this clustering cluster is larger; Obtain the preference degree of each clustering cluster: ; In the formula, Represents the preference degree of the d-th clustering cluster; Represents the sum of the comfort factors of the water quality environment data in all dimensions during the current feeding; Represents the sum of the comfort factors of the water quality environment data in all dimensions during the n-th feeding process in the d-th clustering cluster; Represents the number of feedings in the d-th clustering cluster; || represents the absolute value symbol; exp() represents the exponential function with the natural constant as the base; The smaller the value of, it indicates that the water quality environment data during the current feeding is similar to the water quality environment data during the feeding process in the d-th clustering cluster. At this time, the preference degree of this clustering cluster is larger.
[0047] Take the feeding weight of the clustering cluster corresponding to the maximum value in the preference degree as the feeding weight of the current feeding.
[0048] Obtain the feeding amount of the current feeding according to the conventional empirical value, take the product of the feeding weight of the current feeding and the feeding amount of the current feeding as the corrected feeding amount of the current feeding, and feed the fish in batches with the corrected feeding amount of the current feeding.
[0049] The embodiment of the present invention also discloses a fully automatic intelligent feeding system for aquaculture of big data, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, it realizes a fully automatic intelligent feeding method for aquaculture of big data according to the present invention.
[0050] The above system also includes other components well known to those skilled in the art such as a communication bus and a communication interface. Their settings and functions are known in the art, so they will not be elaborated here.
[0051] In the present invention, the aforementioned memory may be any tangible medium that contains or stores a program, which can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium may be any suitable magnetic storage medium or magneto-optical storage medium, such as, a resistive random access memory, a dynamic random access memory, a static random access memory, an enhanced dynamic random access memory, a high bandwidth memory, a hybrid memory cube, etc., or any other medium that can be used to store the required information and can be accessed by an application program, a module, or both. Any such computer storage medium may be part of the device or accessible or connectable to the device.
[0052] Although this specification has shown and described several embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many changes, alterations, and alternative forms will occur to those skilled in the art without departing from the spirit and scope of the present invention. It should be understood that various alternatives to the embodiments of the present invention described herein may be employed in practicing the present invention.
[0053] The above are all the preferred embodiments of the present invention, and the protection scope of the present invention is not limited accordingly. Therefore, all equivalent changes made according to the structure, shape, and principle of the present invention shall be covered within the protection scope of the present invention.
Claims
1. A fully automatic intelligent feeding method for aquaculture based on big data, characterized in that: include: Collect water quality environmental data and feeding images during each feeding process; Obtain the impact of water quality data on fish populations during each feeding process; Obtain the possibility that each fish moves toward the feeding area in each feeding image frame during each feeding process; Based on the possibilities, a target image of each fish in each feeding process is obtained; The ratio of the number of frames of the target image of each fish in each feeding process to the total number of frames of the feeding image in each feeding process is used as the probability that the movement trajectory of each fish in each feeding process tends to the feeding area; Obtain the feeding degree of fish schools during each feeding process , Represents the feeding degree of fish school during the b-th feeding process; as well as They represent the movement speed of the ith fish in the feeding video during the bth feeding process and the probability that the movement trajectory of the ith fish tends to the feeding area; I represents the number of fish during the bth feeding process; Based on the feeding degree of the fish school and the influence of the water quality environment data on the fish school, the feeding demand degree of each feeding process is obtained; Clustering the water quality environment data in all feeding processes according to the feeding demand degree to obtain each cluster and the feeding weight of each cluster; The preferred degree of each cluster is obtained; the product of the feeding weight of the cluster corresponding to the maximum preferred degree and the current feeding amount is taken as the current corrected feeding amount.
2. According to the big data aquaculture fully automatic intelligent feeding method according to claim 1, it is characterized in that: The degree of influence of the water quality environment data on the fish school during each feeding process includes: , Represents the comfort factor of the temperature data during the b-th feeding process; Represents the value of the cth temperature data in the bth feeding process; Represents the suitable temperature for the growth of water plants; Represents the number of temperature data in the b-th feeding process; , Represents the degree of influence of water quality environmental data on fish schools during the b-th feeding process; , as well as Respectively represent the weights of the comfort factors of the temperature data, humidity data and oxygen content data during the b-th feeding process; , as well as They respectively represent the comfort factors of temperature data, humidity data and oxygen content data during the b-th feeding process.
3. According to the big data aquaculture fully automatic intelligent feeding method of claim 1, it is characterized in that: The method of obtaining the possibility that each fish in each feeding process tends to move toward the feeding area in each frame of the feeding image comprises: The motion vector of the i-th fish in the j-th feeding image during the b-th feeding process is recorded as ,in, as well as represents the position coordinates of the i-th fish in the j-th feeding image and the j+1-th feeding image during the b-th feeding process; the vector from the i-th fish in the j-th feeding image to the feeding area during the b-th feeding process is recorded as ;in, Represents the location coordinates of the feeding area; The probability that the i-th fish moves toward the feeding area in the j-th feeding image during the b-th feeding process is ;|| represents the absolute value symbol.
4. According to the big data aquaculture fully automatic intelligent feeding method of claim 1, it is characterized in that: The step of obtaining a target image of each fish during each feeding process includes: A possibility threshold T is preset to obtain the possibility that the ith fish moves toward the feeding area in the jth frame feeding image during the bth feeding process; if the possibility that the ith fish moves toward the feeding area in the jth frame feeding image during the bth feeding process is greater than the possibility threshold T, the jth frame feeding image is recorded as the target image of the ith fish during the bth feeding process.
5. According to the big data aquaculture fully automatic intelligent feeding method of claim 3, it is characterized in that: The acquisition of the position coordinates of the i-th fish in the j-th feeding image and the j+1-th feeding image during the b-th feeding process includes: The b-th feeding video is frame-processed to obtain several frames of feeding images. The YOLO target detection algorithm is used to identify the position coordinates of each fish in each frame of feeding image, and the DeepSORT algorithm is combined with the Hungarian algorithm to match adjacent frames of feeding images, and then the position coordinates of the same fish in each frame of feeding image of the b-th feeding video are identified.
6. According to the big data aquaculture fully automatic intelligent feeding method of claim 1, it is characterized in that: The acquisition of the movement speed of the i-th fish in the feeding video during the b-th feeding process includes: ； In the formula, represents the movement speed of the i-th fish in the feeding video during the b-th feeding process; Represents the position coordinates of the i-th fish in the j-th feeding image during the b-th feeding process; Represents the position coordinates of the ith fish in the j+1th feeding image during the bth feeding process; Represents the total number of frames of the feeding image during the b-th feeding process.
7. A fully automatic intelligent feeding method for aquaculture based on big data according to claim 1, characterized in that: The step of obtaining the feeding requirement degree of each feeding process includes: ； In the formula, Represents the feeding demand degree of the b-th feeding process; Represents the degree of influence of water quality environmental data on fish schools during the b-th feeding process; Represents the feeding degree of fish school during the b-th feeding process; exp() represents an exponential function with a natural constant as the base; exp() represents an exponential function with a natural constant as the base.
8. The fully automatic intelligent feeding method for aquaculture based on big data according to claim 1 is characterized in that: The acquisition of the feeding weight of each cluster includes: ; In the formula, Represents the feeding weight of the dth cluster; Represents the mean value of feeding demand in the dth cluster; Represents the maximum value of the mean value of feeding requirements in all clusters.
9. The fully automatic intelligent feeding method for aquaculture based on big data according to claim 1 is characterized in that: The obtaining of the preference level of each cluster includes: ； In the formula, Represents the preference degree of the dth cluster; Represents the sum of the comfort factors of water quality environmental data in all dimensions at the time of current feeding; Represents the sum of the comfort factors of water quality environmental data of all dimensions during the n-th feeding process in the d-th cluster; represents the number of times the material is fed in the dth cluster; || represents the absolute value symbol; exp() represents an exponential function with a natural constant as the base.
10. A big data aquaculture fully automatic intelligent feeding system, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a fully automatic intelligent feeding method for aquaculture based on big data according to any one of claims 1 to 9 is implemented.