Marine bait feeding metering method and system based on Internet of Things

Abstract

The invention relates to a marine bait feeding metering method and system based on the Internet of Things, and belongs to the technical field of fishery bait feeding, and the method comprises the steps: obtaining the characteristics of cultured organisms of a plurality of target regions, and determining the activity range and culture density of target aquatic organisms according to the characteristics of cultured organisms of the plurality of target regions; obtaining an evaluation result of the nutrient content of the required nutrients in the bait, and performing estimation in combination with the evaluation result, the activity range and the breeding density to obtain a preliminary feeding metering value; feeding according to the preliminary feeding metering value, and analyzing the survival rate of the target aquatic organisms to obtain a first metering correction scheme; feeding adjustment is conducted on the basis of the first metering correction scheme, the ammonia nitrogen decomposition rate in the current water body is calculated and analyzed, and a second metering correction scheme is obtained; and judging whether the bait feeding equipment has faults or not and maintaining the bait feeding equipment. According to the invention, accurate and efficient metering control can be carried out on marine bait feeding, the breeding quality is improved, and feed waste is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of fishery bait feeding, and in particular to a method and system for measuring bait feeding for a ship based on the Internet of Things. Background Art

[0002] In the current fishery farming industry, the introduction of advanced farming equipment can make fishery farming activities faster, more efficient and more convenient, and improve the overall quality of fishery farming. The most critical part of fishery farming is the feeding of target aquatic organisms. The accuracy of feed feeding determines the growth quality and survival rate of target aquatic organisms, and can also greatly control the output of feed costs. However, the existing feed feeding equipment is not equipped with an accurate feeding metering method, so that each specific feeding needs to be estimated and proportioned manually, which is time-consuming and labor-intensive, and the feeding amount of manual metering ratio still has a large error, resulting in a lot of waste and increasing the input of feeding costs. The feeding amount with inaccurate metering ratio will lead to nutritional deficiencies and imbalanced health management of target aquatic organisms, reduce the survival rate of farming, and too much feed will remain in the water body and decompose a large amount of ammonia nitrogen, pollute the farming water body, and is not conducive to high-quality farming activities. At the same time, the errors caused by equipment failures will also lead to inaccurate feeding amounts. Therefore, it is necessary to develop a method that can accurately and efficiently measure feed feeding to solve the above problems. Summary of the invention

[0003] The present invention overcomes the shortcomings of the prior art and provides a method and system for measuring bait feeding for ships based on the Internet of Things.

[0004] To achieve the above object, the technical solution adopted by the present invention is: The first aspect of the present invention provides a method for measuring bait feeding for a ship based on the Internet of Things, comprising the following steps: Acquire aquaculture image data of multiple target areas within a preset time period, and introduce a local binary pattern algorithm to perform feature extraction to obtain aquaculture biological features of the multiple target areas, and determine the activity range and aquaculture density of the target aquatic organisms based on the aquaculture biological features of the multiple target areas; Obtaining the required nutrients of the target aquatic organisms, analyzing and evaluating the nutrient content of the required nutrients in the current feed to obtain an evaluation result, and combining the evaluation result, the activity range and the breeding density to make an estimate to obtain a preliminary feeding measurement value; Feeding according to the preliminary feeding measurement value, obtaining the survival rate of the target aquatic organism, constructing a trained survival rate prediction model for prediction, obtaining a survival rate interval with insufficient nutrient content, analyzing whether the survival rate of the target aquatic organism is within the survival rate interval, and obtaining a first measurement correction plan; Based on the first metering correction scheme, feeding adjustment is performed to obtain the actual survival rate after the first correction. If the actual survival rate after the first correction is not within the survival rate range of water environment pollution, the ammonia nitrogen decomposition rate in the current water body is calculated and analyzed to obtain a second metering correction scheme; Based on the second metering correction scheme, feeding adjustments are made to obtain the actual survival rate after the second correction. If the actual survival rate after the second correction is not within the survival rate range of equipment failure, a curve chart is constructed to analyze whether there is a fault in the bait feeding equipment and repair and maintenance are performed.

[0005] Furthermore, in a preferred embodiment of the present invention, the aquaculture image data of multiple target areas within a preset time period are obtained, and a local binary pattern algorithm is introduced to perform feature extraction to obtain aquaculture biological features of multiple target areas, and the activity range and aquaculture density of target aquatic organisms are determined according to the aquaculture biological features of the multiple target areas, which specifically includes the following steps: Based on the Internet of Things technology, the remote sensing camera on the fishing boat is controlled to capture images of target aquatic organisms, and the aquaculture image data of multiple target areas within a preset time period are obtained; A local binary pattern algorithm is introduced to extract the features of the farmed organisms from the farmed image data of the plurality of target areas, each farmed image data is divided into a plurality of local blocks, and the pixel points in each of the local blocks are compared to generate a plurality of binary coded information; Converting the plurality of binary coded information into different local binary patterns, counting the frequencies of the different local binary patterns in each local block, drawing and outputting a local binary pattern histogram, and obtaining the aquaculture biological characteristics of the plurality of target areas; The Euclidean distances between the characteristics of the aquaculture organisms in the target areas are calculated, and a hot spot radius algorithm is introduced to plan the hot spot areas for the calculated multiple Euclidean distances, and a first activity hot spot map is constructed. The activity range and aquaculture density of the target aquatic organisms are determined according to the first activity hot spot map.

[0006] Furthermore, in a preferred embodiment of the present invention, the step of obtaining the required nutrients of the target aquatic organisms, analyzing and evaluating the nutrient content of the required nutrients in the current bait to obtain an evaluation result, and combining the evaluation result, the activity range and the breeding density to estimate and obtain a preliminary feeding measurement value specifically includes the following steps: Obtaining the species information of the target aquatic organism, searching the relevant nutritional information required for the normal growth of the target aquatic organism based on the big data network, obtaining the required nutritional components of the target aquatic organism, and obtaining the required intake of each nutritional component; Acquire several kinds of ingredient information of the current bait, construct a knowledge graph, search the nutritional components contained in the several kinds of ingredient information in the knowledge graph, obtain search results, screen and mark the ingredient information that is the same as the required nutritional components of the target aquatic organism in the search results, and obtain the target ingredient information; Based on the big data network, the standard nutritional content interval corresponding to the target ingredient is obtained, and the ratio quality of the target ingredient is obtained, and the fuzzy C-means algorithm is introduced to calculate the initial membership of the ratio quality and each standard nutritional content in the standard nutritional content interval, and a cluster center is constructed. The position of the cluster center is updated based on the membership until the maximum number of iterations is met, and multiple memberships are generated; Constructing an evaluation model, selecting the standard nutritional content corresponding to the maximum membership degree from the plurality of membership degrees as a result output, obtaining the nutritional content of the current bait, and importing the nutritional content of the current bait into the evaluation model for evaluation to obtain an evaluation result; The ratio between the required intake of each nutrient component and the nutrient content of the current bait is calculated to obtain an individual feeding ratio, and the individual feeding ratio, the evaluation result, the activity range and the breeding density are comprehensively estimated to obtain a preliminary feeding measurement value.

[0007] Further, in a preferred embodiment of the present invention, feeding is performed according to the preliminary feeding measurement value, the survival rate of the target aquatic organism is obtained, a trained survival rate prediction model is constructed for prediction, a survival rate interval with insufficient nutrient content is obtained, and whether the survival rate of the target aquatic organism is within the survival rate interval is analyzed to obtain a first measurement correction plan, which specifically includes the following steps: Inputting the preliminary feeding measurement value of the target area into the bait feeding device for metered feeding, and reconstructing the heat map after a preset time period to obtain a second activity heat map; Based on the second activity hotspot map, all non-hotspots and all feature overlap points in the target area are obtained, the non-hotspots and the feature overlap points are defined as low activity points of the target aquatic organisms, and the ratio of the low activity points of all target aquatic organisms to all hotspots is calculated in the second activity hotspot map to obtain the survival rate of the target aquatic organisms; The survival rates corresponding to the influence of different aquaculture conditions and factors are obtained through a big data network, and a survival rate prediction model is constructed based on a support vector machine algorithm. The survival rates corresponding to the influence of the different aquaculture conditions and factors are trained and verified in the survival probability prediction model to obtain a trained survival rate prediction model; wherein the different aquaculture conditions and factors include insufficient nutrient content, water environment pollution, and equipment failure; Extracting a survival rate interval with insufficient nutrient content based on the trained survival rate prediction model, and determining whether the survival rate of the target aquatic organism is within the survival rate interval with insufficient nutrient content; If not, a hash algorithm is introduced to calculate a hash value between the lower limit value in the survival rate interval and the survival rate of the target aquatic organism, and the required feeding amount of the target area is metered and regulated according to the hash value to obtain a first metering correction plan.

[0008] Further, in a preferred embodiment of the present invention, the feeding adjustment is performed based on the first metering correction scheme to obtain the actual survival rate after the first correction. If the actual survival rate after the first correction is not within the survival rate range of water environment pollution, the ammonia nitrogen decomposition rate in the current water body is calculated and analyzed to obtain the second metering correction scheme, which specifically includes the following steps: According to the first metering correction scheme, the feed feeding equipment is corrected in metering, and the survival rate of the corrected target aquatic organism is calculated to obtain the actual survival rate after a correction; The survival rate prediction model after training is used to extract the survival rate interval of water environment pollution. If the actual survival rate after the first correction is within the survival rate interval of water environment pollution, aquaculture water samples are extracted and tested by chemical testing methods to obtain the current ammonia nitrogen content index. Acquire multiple historical ammonia nitrogen content indexes within a preset time period, construct a series of ammonia nitrogen decomposition change models according to the multiple historical ammonia nitrogen content indexes and the current ammonia nitrogen content index, introduce a mean absolute error algorithm to calculate the model errors between the series of ammonia nitrogen decomposition change models, obtain multiple discrete error rates, screen out the minimum discrete error rate and define it as the current ammonia nitrogen decomposition rate; Obtaining a water body ammonia nitrogen pollution standard evaluation system based on a big data network, establishing a scoring rule according to the water body ammonia nitrogen pollution standard evaluation system, and scoring the current ammonia nitrogen decomposition rate according to the scoring rule to obtain a score value; Determine whether the score value is less than a preset score value. If so, calculate the error between the score value and the preset score value to obtain an error value. According to the error value, match the corresponding ammonia nitrogen decomposition rate in the water ammonia nitrogen pollution standard evaluation system and define it as the ammonia nitrogen decomposition error rate. Based on the ammonia nitrogen decomposition error rate, perform metering control on the first metering correction scheme to obtain a second metering correction scheme.

[0009] Further, in a preferred embodiment of the present invention, the feeding adjustment is performed based on the second metering correction scheme to obtain the actual survival rate after the second correction. If the actual survival rate after the second correction is not within the survival rate interval of the equipment failure, a curve chart is constructed to analyze whether there is a fault in the feed feeding equipment and repair and maintenance are performed, which specifically includes the following steps: The feed feeding equipment is corrected according to the second metering correction scheme, and the corrected survival rate of the target aquatic organism is recalculated to obtain the actual survival rate after the second correction, and the survival rate interval of the equipment failure is extracted based on the trained survival rate prediction model; If the actual survival rate after the secondary correction is within the survival rate range of the equipment failure, the feed feeding equipment is subjected to laser coverage scanning by laser point cloud scanning technology to obtain several groups of point cloud data of the feed feeding equipment, and the RANSAC algorithm is introduced to perform model fitting on the several groups of point cloud data to obtain a three-dimensional simulation model of the feed feeding equipment; The actual feeding rate of the bait feeding equipment within a preset time period is obtained by using the Internet of Things technology, the preset time period is divided into a plurality of uniform time nodes, a curve graph is constructed based on the plurality of uniform time nodes and the actual feeding rate corresponding to each time node, and an actual feeding rate-time change curve graph is obtained; The feeding amount is calculated by analyzing the actual feeding rate-time variation curve to obtain the actual feeding amount, and the curvature of the curve is calculated to obtain a first curvature; Based on the plurality of uniform time nodes, the actual feeding amount is simulated in the three-dimensional simulation model of the bait feeding equipment to obtain a simulated feeding rate corresponding to each time node, a simulated feeding rate-time variation curve diagram is constructed based on the simulated feeding rate corresponding to each time node, the feeding amount is calculated according to the simulated feeding rate-time variation curve diagram to obtain the simulated feeding amount, and the curvature of the curve is calculated to obtain a second curvature; Determine whether the actual feeding amount is less than the simulated feeding amount. If so, it indicates that there is a fault in the bait feeding equipment. Calculate the deviation between the first curvature and the second curvature to obtain a curvature deviation value. Repair and maintain the fault of the bait feeding equipment based on the curvature deviation value.

[0010] A second aspect of the present invention provides a ship bait feeding metering system based on the Internet of Things, wherein the ship bait feeding metering system based on the Internet of Things comprises a memory and a processor, wherein the memory stores a ship bait feeding metering method program based on the Internet of Things, and when the ship bait feeding metering method program based on the Internet of Things is executed by the processor, the following steps are implemented: Acquire aquaculture image data of multiple target areas within a preset time period, and introduce a local binary pattern algorithm to perform feature extraction to obtain aquaculture biological features of the multiple target areas, and determine the activity range and aquaculture density of the target aquatic organisms based on the aquaculture biological features of the multiple target areas; Obtaining the required nutrients of the target aquatic organisms, analyzing and evaluating the nutrient content of the required nutrients in the current feed to obtain an evaluation result, and combining the evaluation result, the activity range and the breeding density to make an estimate to obtain a preliminary feeding measurement value; Feeding according to the preliminary feeding measurement value, obtaining the survival rate of the target aquatic organism, constructing a trained survival rate prediction model for prediction, obtaining a survival rate interval with insufficient nutrient content, analyzing whether the survival rate of the target aquatic organism is within the survival rate interval, and obtaining a first measurement correction plan; Based on the first metering correction scheme, feeding adjustment is performed to obtain the actual survival rate after the first correction. If the actual survival rate after the first correction is not within the survival rate range of water environment pollution, the ammonia nitrogen decomposition rate in the current water body is calculated and analyzed to obtain a second metering correction scheme; Based on the second metering correction scheme, feeding adjustments are made to obtain the actual survival rate after the second correction. If the actual survival rate after the second correction is not within the survival rate range of equipment failure, a curve chart is constructed to analyze whether there is a fault in the bait feeding equipment and repair and maintenance are performed.

[0011] The present invention solves the technical defects existing in the background technology, and the beneficial technical effects of the present invention are: Acquire the characteristics of aquaculture organisms in multiple target areas, and determine the activity range and aquaculture density of the target aquatic organisms according to the characteristics of aquaculture organisms in the multiple target areas; analyze and evaluate the nutrient content of the required nutrients in the bait to obtain an evaluation result, and estimate in combination with the evaluation result, the activity range and the aquaculture density to obtain a preliminary feeding measurement value; feed according to the preliminary feeding measurement value to obtain the survival rate of the target aquatic organisms, analyze whether the survival rate of the target aquatic organisms is within the survival rate interval of insufficient nutrient content, and obtain a first measurement correction scheme; adjust the feeding based on the first measurement correction scheme to obtain the actual survival rate after the first correction, if the actual survival rate after the first correction is not within the survival rate interval of water environment pollution, calculate and analyze the ammonia nitrogen decomposition rate in the current water body to obtain a second measurement correction scheme; adjust the feeding based on the second measurement correction scheme to obtain the actual survival rate after the second correction, if the actual survival rate after the second correction is not within the survival rate interval of equipment failure, construct a curve chart to analyze whether there is a failure in the bait feeding equipment and perform repair and maintenance. The present invention can accurately and efficiently control the metering of ship bait feeding, thereby improving the breeding quality and reducing feed waste. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, drawings of other embodiments can be obtained based on these drawings without paying creative work.

[0013] Figure 1 A first method flow chart of a method for measuring bait feeding for ships based on the Internet of Things is shown; Figure 2 A second method flow chart of a method for measuring bait feeding for ships based on the Internet of Things is shown; Figure 3 A third method flow chart of a method for measuring bait feeding for ships based on the Internet of Things is shown; Figure 4 The system framework diagram of a ship bait feeding metering system based on the Internet of Things is shown. DETAILED DESCRIPTION

[0014] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0015] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0016] The first aspect of the present invention provides a method for feeding and metering bait for a ship based on the Internet of Things, such as Figure 1 As shown, the following steps are included: S102: Acquire aquaculture image data of multiple target areas within a preset time period, and introduce a local binary pattern algorithm to perform feature extraction to obtain aquaculture biological features of the multiple target areas, and determine the activity range and aquaculture density of the target aquatic organisms according to the aquaculture biological features of the multiple target areas; S104: Obtaining the required nutrients of the target aquatic organisms, analyzing and evaluating the nutrient content of the required nutrients in the current feed to obtain an evaluation result, and performing an estimation based on the evaluation result, the activity range and the breeding density to obtain a preliminary feeding measurement value; S106: feeding according to the preliminary feeding measurement value, obtaining the survival rate of the target aquatic organism, constructing a trained survival rate prediction model for prediction, obtaining a survival rate interval with insufficient nutrient content, analyzing whether the survival rate of the target aquatic organism is within the survival rate interval, and obtaining a first measurement correction plan; S108: adjusting feeding based on the first metering correction scheme to obtain the actual survival rate after the first correction; if the actual survival rate after the first correction is not within the survival rate range of water environment pollution, calculating and analyzing the ammonia nitrogen decomposition rate in the current water body to obtain a second metering correction scheme; S110: Adjust feeding based on the second metering correction scheme to obtain the actual survival rate after the second correction. If the actual survival rate after the second correction is not within the survival rate range of equipment failure, construct a curve chart to analyze whether there is a fault in the feed feeding equipment and perform repair and maintenance.

[0017] Furthermore, in a preferred embodiment of the present invention, the aquaculture image data of multiple target areas within a preset time period are obtained, and a local binary pattern algorithm is introduced to perform feature extraction to obtain aquaculture biological features of multiple target areas, and the activity range and aquaculture density of target aquatic organisms are determined according to the aquaculture biological features of the multiple target areas, which specifically includes the following steps: Based on the Internet of Things technology, the remote sensing camera on the fishing boat is controlled to capture images of target aquatic organisms, and the aquaculture image data of multiple target areas within a preset time period are obtained; A local binary pattern algorithm is introduced to extract the features of the farmed organisms from the farmed image data of the plurality of target areas, each farmed image data is divided into a plurality of local blocks, and the pixel points in each of the local blocks are compared to generate a plurality of binary coded information; Converting the plurality of binary coded information into different local binary patterns, counting the frequencies of the different local binary patterns in each local block, drawing and outputting a local binary pattern histogram, and obtaining the aquaculture biological characteristics of the plurality of target areas; The Euclidean distances between the characteristics of the aquaculture organisms in the target areas are calculated, and a hot spot radius algorithm is introduced to plan the hot spot areas for the calculated multiple Euclidean distances, and a first activity hot spot map is constructed. The activity range and aquaculture density of the target aquatic organisms are determined according to the first activity hot spot map.

[0018] It should be noted that for accurate measurement of feed feeding, it is first necessary to know the activity range of farmed organisms in the breeding area and the breeding density of farmed organisms within the activity range. The activity range and breeding density are key indicators for accurately estimating the required feeding amount for all farmed organisms in the breeding area; since the activity of the target aquatic organisms in the water is irregular, but the activity range and breeding density can remain unchanged, the activity range and breeding density of the target aquatic organisms are calculated and analyzed by obtaining the features of the breeding area image extraction; among them, the local binary pattern algorithm is used for image feature extraction. The algorithm represents the texture features of the image by locally binarizing the pixels in the image and constructing a local binary pattern. The advantage is that it has good robustness to the influence of light, noise, etc., and the feature calculation efficiency is high. In order to facilitate the observation of the activity range and breeding density of the target aquatic organisms, the hot spot radius algorithm is introduced to construct an activity hot spot map, and the distribution position and distribution density of the target aquatic organism features in the breeding area image within the preset time period are further analyzed, which can speed up the analysis rate, reduce redundant calculation steps, and improve efficiency. The present invention can analyze and determine the activity range and breeding density of target aquatic organisms by extracting image features and constructing an activity heat map, thereby improving data analysis and calculation rates and ensuring the accuracy of feeding measurement.

[0019] Furthermore, in a preferred embodiment of the present invention, the required nutrients of the target aquatic organisms are obtained, the nutrient content of the required nutrients in the current bait is analyzed and evaluated to obtain an evaluation result, and the evaluation result, the activity range and the breeding density are combined to estimate and obtain a preliminary feeding measurement value, such as Figure 2 As shown, the specific steps include: S202: Obtaining the species information of the target aquatic organism, searching the relevant nutritional information required for the normal growth of the species information of the target aquatic organism based on the big data network, obtaining the required nutritional components of the individual target aquatic organism, and obtaining the required intake of each nutritional component; S204: obtaining several kinds of ingredient information of the current bait, constructing a knowledge graph, searching the knowledge graph for the nutrients contained in the several kinds of ingredient information, obtaining a search result, screening and marking the ingredient information having the same nutrients as the required nutrients of the target aquatic organism in the search result, and obtaining the target ingredient information; S206: obtaining the standard nutritional content interval corresponding to the target ingredient based on the big data network, and obtaining the ratio quality of the target ingredient, introducing the fuzzy C-means algorithm to calculate the initial membership of the ratio quality and each standard nutritional content in the standard nutritional content interval, constructing a cluster center, and updating the position of the cluster center based on the membership until the maximum number of iterations is met, and generating multiple memberships; S208: constructing an evaluation model, selecting the standard nutritional content corresponding to the maximum membership degree from the plurality of membership degrees as a result output, obtaining the nutritional content of the current bait, and importing the nutritional content of the current bait into the evaluation model for evaluation to obtain an evaluation result; S210: Calculate the ratio between the required intake of each nutrient and the nutrient content of the current bait to obtain an individual feeding ratio, and make an estimate based on the individual feeding ratio, the evaluation result, the activity range and the breeding density to obtain a preliminary feeding measurement value.

[0020] It should be noted that the required nutrients for the target aquatic organisms include protein, carbohydrates, fat, vitamins and minerals; one of the key indicators for determining the precise measurement of feed feeding is to analyze the ratio of the bait ingredients and the content of the nutrients, so as to calculate the specific feed dosage that can meet the required nutrient content of the bait for the target aquatic organisms. If the bait is fed directly without a preliminary analysis and evaluation of the ratio and nutrient content of the bait ingredients, the nutrient content of the bait may be insufficient, resulting in nutritional deficiencies in the target aquatic organisms, which will simultaneously lead to a sharp drop in survival rate over time, which is not conducive to high-quality breeding; for the ratio of the bait ingredients and the nutrient content, the ratio of the bait ingredients and the nutrient content are analyzed and evaluated, so as to calculate the specific feed dosage that can meet the required nutrient content of the bait for the target aquatic organisms. The analysis of the ratio and nutrient content is to screen out the target ingredient information related to the required nutrients of the target aquatic organisms in the bait. Since the nutrient content of the target ingredient is not a fixed value, there is a certain range of values, so it is necessary to obtain the standard nutrient content interval corresponding to the target ingredient, and then introduce the fuzzy C mean to calculate the membership of the ratio quality of the target ingredient and the content of each nutrient in the standard nutrient content interval, and determine the nutrient content of the bait according to the membership, improve the accurate matching rate of the nutrient content, and reduce the occurrence of errors; the individual feeding ratio is the ratio of the nutrient feeding amount calculated based on the target individual's demand for nutrients and the nutrient content of the bait. The present invention can analyze the specific feeding dosage of the bait according to the required nutrient content of the bait of the target aquatic organism object, ensure the balanced feeding nutrition of the cultured organisms, and improve the quality of feeding health management.

[0021] It should be noted that the input data of the fuzzy C-means algorithm is the sample of the target batching quality coefficient. And the standard nutritional content characteristic coefficients of the target ingredients (such as protein and fat content). The calculation process of the initial membership degree is as follows: Set the coefficient sample set , where each sample represents a different ratio quality and a characteristic coefficient range of set standard nutritional content , each characteristic coefficient represents the central value of a standard nutrient content, and the number of standard nutrient content categories C (number of clusters) is set. Then, the number of standard nutrient content categories is used to perform matrix calculation on the membership of the coefficient sample set and the characteristic coefficient interval. The specific expression is:

[0022] in, It represents the membership degree of the coefficient sample point to the jth nutritional standard category.

[0023] For example, taking protein and fat content as an example, it is assumed that the ratio quality data contains 3 samples, each sample has 2 standard nutritional content characteristics, and it is desired to be divided into 2 categories (C=2), then the sample table of some input data is as follows:

[0024] Then the initial membership is:

[0025] Among them, U represents the initial membership of each ratio quality coefficient sample to the two categories.

[0026] It should be noted that the specific steps for constructing the evaluation model are as follows: first, the key indicators of bait nutritional evaluation and the historical evaluation coefficients of each key indicator for the current bait are obtained through the big data network, where the historical evaluation coefficients are used as input data. Then, these historical evaluation coefficients are split into 80% training sets and 20% test sets, and denoising operations such as missing value processing, deduplication, and outlier detection are performed to reduce the impact of redundant noise in data training on the evaluation results. Subsequently, the linear regression model architecture is used to fit the historical evaluation coefficients of the 80% test set, and the hyperparameters of the model architecture are continuously adjusted during the training process. The number ensures the stability and accuracy of training. After the training is completed, a preliminary evaluation model will be formed which can evaluate the standard nutritional content of bait according to key indicators and is in the testing stage. At this time, the remaining 20% ​​of the test set is used to perform a regularized evaluation test on the trained preliminary evaluation model to avoid overfitting or underfitting of the evaluation model. The learning rate of the model is adjusted according to the test results. After the test is completed, the evaluation model will output data in the form of a series of nutritional evaluation regression coefficients for key indicators of bait nutritional evaluation. This series of nutritional evaluation regression coefficients is the accurate linear expression of the final evaluation result of the current standard nutritional content of bait.

[0027] Further, in a preferred embodiment of the present invention, feeding is performed according to the preliminary feeding measurement value, the survival rate of the target aquatic organism is obtained, a trained survival rate prediction model is constructed for prediction, a survival rate interval with insufficient nutrient content is obtained, and whether the survival rate of the target aquatic organism is within the survival rate interval is analyzed to obtain a first measurement correction plan, which specifically includes the following steps: Inputting the preliminary feeding measurement value of the target area into the bait feeding device for metered feeding, and reconstructing the heat map after a preset time period to obtain a second activity heat map; Based on the second activity hotspot map, all non-hotspots and all feature overlap points in the target area are obtained, the non-hotspots and the feature overlap points are defined as low activity points of the target aquatic organisms, and the ratio of the low activity points of all target aquatic organisms to all hotspots is calculated in the second activity hotspot map to obtain the survival rate of the target aquatic organisms; The survival rates corresponding to the influence of different aquaculture conditions and factors are obtained through a big data network, and a survival rate prediction model is constructed based on a support vector machine algorithm. The survival rates corresponding to the influence of the different aquaculture conditions and factors are trained and verified in the survival probability prediction model to obtain a trained survival rate prediction model; wherein the different aquaculture conditions and factors include insufficient nutrient content, water environment pollution, and equipment failure; Extracting a survival rate interval with insufficient nutrient content based on the trained survival rate prediction model, and determining whether the survival rate of the target aquatic organism is within the survival rate interval with insufficient nutrient content; If not, a hash algorithm is introduced to calculate a hash value between the lower limit value in the survival rate interval and the survival rate of the target aquatic organism, and the required feeding amount of the target area is metered and regulated according to the hash value to obtain a first metering correction plan.

[0028] It should be noted that the calculated preliminary feeding measurement value is accurately measured and fed in the bait feeding equipment, but since the preliminary calculated feeding measurement value is a preliminary estimated value, there is a risk of error, which makes the measurement of bait feeding inaccurate, resulting in the survival rate of the target aquatic organisms in the breeding area not being improved or the survival rate increasing at a small rate. If such a phenomenon occurs, the preliminary feeding measurement value needs to be corrected in time to reduce the measurement error; after feeding the preliminary feeding dosage value for a preset time period, the second activity hotspot map of the target aquatic organisms is constructed again. When the target aquatic organisms die or have low activity in the breeding area, the features displayed in the image will not undergo a large change in activity. Therefore, non-hotspots are used to represent death or low activity in the activity hotspot map. Characteristics, the survival rate of the target aquatic organism after feeding can be calculated; since there is no comparative data, it is impossible to know what causes the survival rate, so the corresponding survival rate of the target aquatic organism under the influence of different breeding conditions can be predicted by constructing a prediction model; although the initial feeding measurement value has been evaluated for nutrient content, there may still be measurement errors that lead to malnutrition, so when the survival rate of the target aquatic organism after feeding belongs to the survival rate interval of malnutrition, it means that the main cause of the survival rate is malnutrition, so as to calculate an accurate correction value to correct and adjust the initial feeding measurement value, thereby improving the accuracy of metered feeding, and the lower limit value in the survival rate interval is the lowest survival rate value of the survival rate interval with insufficient nutrient content. The present invention can accurately analyze the cause of the survival rate of the target aquatic organism after the initial feeding measurement value, so as to correct the initial feeding measurement value and improve the feeding measurement quality.

[0029] It should be noted that for the construction of the survival rate prediction model, this method first collects and organizes the historical influencing factor data under different breeding conditions through the big data network and several aquatic biological feeding cases, such as water temperature parameters with insufficient nutrient content, dissolved oxygen parameters, pH value and feed type parameters, and obtains the corresponding historical survival rate labels of each historical influencing factor as input data; then the collected input data is subjected to denoising processing by removing missing values, outliers, data standardization or normalization to improve the convergence speed and stability of the model. Then, the input data set is divided into training set, test set and validation set according to the ratio of 7:2:1. At this time, the support vector machine algorithm is introduced to further select the appropriate support vector machine kernel function to adapt to different data feature distributions. According to the selected support vector machine kernel function, the penalty coefficient and kernel function parameters are adjusted in the support vector machine algorithm to form a survival rate prediction model of the support vector machine model architecture. Then, the historical influencing factor data and its historical survival rate labels in the training set are used to continuously train and learn the survival rate prediction model of the support vector machine model architecture, so that it can adaptively predict the relationship between breeding conditions and survival rate. Set the evaluation mean square error threshold, evaluate the trained survival rate prediction model on the validation set, and output the current evaluation mean square error. If the current evaluation mean square error exceeds the evaluation mean square error threshold, it means that there is an error in the prediction accuracy of the trained model. During the training process, adjust the feature extraction of the support vector machine model architecture and optimize the penalty coefficient and kernel function parameters to improve the generalization ability of the model prediction. Finally, use the test set data to evaluate the final trained survival rate prediction model to ensure that it performs well on unseen data. Applying the trained survival rate prediction model to the actual breeding environment can provide a prediction reference for the survival rate under different breeding conditions.

[0030] Furthermore, in a preferred embodiment of the present invention, the feeding adjustment is performed based on the first metering correction scheme to obtain the actual survival rate after the first correction. If the actual survival rate after the first correction is not within the survival rate range of the water environment pollution, the ammonia nitrogen decomposition rate in the current water body is calculated and analyzed to obtain a second metering correction scheme, such as Figure 3 As shown, the specific steps include: S302: performing metering correction on the feed feeding equipment according to the first metering correction scheme, and calculating the corrected survival rate of the target aquatic organism to obtain a corrected actual survival rate; S304: extracting the survival rate interval of water environment pollution from the trained survival rate prediction model, and if the actual survival rate after the first correction is within the survival rate interval of water environment pollution, extracting aquaculture water samples, and testing the aquaculture water samples by a chemical testing method to obtain a current ammonia nitrogen content index; S306: Obtain multiple historical ammonia nitrogen content indexes within a preset time period, construct a series of ammonia nitrogen decomposition change models according to the multiple historical ammonia nitrogen content indexes and the current ammonia nitrogen content index, introduce a mean absolute error algorithm to calculate the model errors between the series of ammonia nitrogen decomposition change models, obtain multiple discrete error rates, screen out the minimum discrete error rate and define it as the current ammonia nitrogen decomposition rate; S308: obtaining a water body ammonia nitrogen pollution standard evaluation system based on a big data network, establishing a scoring rule according to the water body ammonia nitrogen pollution standard evaluation system, and scoring the current ammonia nitrogen decomposition rate according to the scoring rule to obtain a score value; S310: Determine whether the score value is less than a preset score value. If so, calculate the error between the score value and the preset score value to obtain an error value. According to the error value, match the corresponding ammonia nitrogen decomposition rate in the water body ammonia nitrogen pollution standard evaluation system and define it as the ammonia nitrogen decomposition error rate. Based on the ammonia nitrogen decomposition error rate, perform metering control on the first metering correction scheme to obtain a second metering correction scheme.

[0031] It should be noted that after correcting the nutrient deficiency problem in the feeding metering, the second activity hotspot map is constructed again to calculate the survival rate of the target aquatic organisms in the breeding area. If the survival rate still does not meet the standard, it means that the first metering correction plan of the metering regulation still needs to be further revised. However, the problem this time is no longer nutrient deficiency. It may be that a large amount of ammonia nitrogen pollution in the water environment has caused the breeding quality of the target aquatic organisms to decline, resulting in a low survival rate. The ammonia nitrogen content in the water body is related to the decomposition rate. Therefore, the current ammonia nitrogen decomposition rate can be analyzed and calculated for further evaluation. In this way, it can be determined whether the current water environment quality is the problem of low survival rate of the target aquatic organisms caused by the regulation of the first metering correction plan. Regarding the method of obtaining the current ammonia nitrogen decomposition rate, it can be obtained through multiple historical data at different time nodes. The relevant ammonia nitrogen decomposition change model is constructed in the process of the historical ammonia nitrogen content index gradually reaching the current ammonia nitrogen content index. The discrete error of the model can be calculated to analyze the decomposition rate of ammonia nitrogen in the process. Compared with the traditional method of gradually calculating the ammonia nitrogen decomposition rate, the analysis rate is fast, the analysis result is accurate, and the result output with a low error rate can be achieved; finally, the ammonia nitrogen decomposition rate results are scored, and the score value can be used to know whether the low survival rate of the current target aquatic organisms is affected by the current ammonia nitrogen pollution in the water body, so as to make secondary corrections and adjustments to the first measurement correction plan according to the pollution situation of the ammonia nitrogen content. On the one hand, it ensures that the target aquatic organisms are further adequately supplemented with nutrients to avoid the nutrient deficiency problem caused by water environmental pollution to the target aquatic organisms; on the other hand, it reduces the increase in ammonia nitrogen content in the water body caused by bait residues, thereby improving the survival rate.

[0032] It should be noted that the specific steps for constructing the ammonia nitrogen decomposition change model are as follows: first, multiple historical ammonia nitrogen content indices of the aquaculture water body at different time nodes within a preset time period are extracted through the aquaculture log, and then multiple historical ammonia nitrogen content indices and the current ammonia nitrogen content index are normalized, and then it is determined whether the time series fluctuation index presented by the historical ammonia nitrogen content index and the current ammonia nitrogen content index is greater than the preset threshold. If it is greater, it means that the ammonia nitrogen decomposition of the historical ammonia nitrogen content index gradually reaching the current ammonia nitrogen content index is not stable, so the historical ammonia nitrogen content index and the current ammonia nitrogen content index are calculated by stable difference to generate a difference order; then the key is introduced. The decomposition spectrum of autocorrelation and partial autocorrelation of ammonia nitrogen in water bodies is used to jointly calculate the autocorrelation function and partial autophase function of ammonia nitrogen from various historical ammonia nitrogen content indexes to the current ammonia nitrogen content index through the two spectra. At this time, an autoregressive order and a moving average order are obtained to describe the variability of the time series decomposition of the input data. The output difference order, autoregressive order and moving average order are used to perform time series calculation fitting on the autoregressive integrated sliding average of the current ammonia nitrogen content index from various historical ammonia nitrogen content indexes. The series of models finally formed by the fitting are the concrete expression of the time series decomposition and change of ammonia nitrogen in aquaculture water bodies.

[0033] It should be noted that the water body ammonia nitrogen pollution standard evaluation system refers to an evaluation reference system for assessing the water quality of aquaculture based on national or international environmental quality standards, including but not limited to the "Fishery Water Quality Standard" (GB 11607-1989), OECD eutrophication evaluation standards and "Urban Wastewater Treatment Plant Pollutant Discharge Standard" (GB 18918-2002), etc., which can be obtained through big data network query. After querying the reference water body ammonia nitrogen pollution standard evaluation system, a water quality evaluation decision matrix is ​​constructed based on the ammonia nitrogen decomposition rate evaluation level and evaluation indicators recorded in the water body ammonia nitrogen pollution standard evaluation system. The specific form of the water quality evaluation decision matrix is:

[0034] in, It indicates the value of the ammonia nitrogen decomposition rate evaluation grade based on the evaluation index.

[0035] Since each evaluation indicator may have different importance, it is necessary to set an indicator weight vector , the indicator weight vector satisfies the normalization requirements of the following formula:

[0036] Next, a weighted normalized matrix is ​​obtained according to the weight vector of the indicator to obtain a weighted normalized water quality evaluation decision matrix. The weighted normalized water quality evaluation decision matrix is ​​used to calculate the positive ideal evaluation solution and negative ideal evaluation solution of the evaluation indicator for the ammonia nitrogen decomposition rate. The positive ideal evaluation solution and the negative ideal evaluation solution are the evaluation value range of the ammonia nitrogen decomposition rate planned by the water quality evaluation decision matrix. Since the scoring of ammonia nitrogen decomposition is based on the ammonia nitrogen decomposition rate evaluation level as a reference assignment, it is necessary to calculate the distance between each ammonia nitrogen decomposition rate evaluation level and the positive and negative ideal evaluation solutions, and this distance is the basis for measuring the strength of the ammonia nitrogen decomposition rate scoring. Finally, the above steps are repeated for different ammonia nitrogen decomposition rates, so as to form a complete scoring rule for different ammonia nitrogen decomposition rates, and then the current ammonia nitrogen decomposition rate is accurately scored under the constraints of the water body ammonia nitrogen pollution standard evaluation system.

[0037] For example, the scoring rules are as follows:

[0038] Among them, 100 from 100 to 80 is a positive ideal evaluation solution, and 80 is a negative ideal evaluation solution.

[0039] Further, in a preferred embodiment of the present invention, the feeding adjustment is performed based on the second metering correction scheme to obtain the actual survival rate after the second correction. If the actual survival rate after the second correction is not within the survival rate interval of the equipment failure, a curve chart is constructed to analyze whether there is a fault in the feed feeding equipment and repair and maintenance are performed, which specifically includes the following steps: The feed feeding equipment is corrected according to the second metering correction scheme, and the corrected survival rate of the target aquatic organism is recalculated to obtain the actual survival rate after the second correction, and the survival rate interval of the equipment failure is extracted based on the trained survival rate prediction model; If the actual survival rate after the secondary correction is within the survival rate range of the equipment failure, the feed feeding equipment is subjected to laser coverage scanning by laser point cloud scanning technology to obtain several groups of point cloud data of the feed feeding equipment, and the RANSAC algorithm is introduced to perform model fitting on the several groups of point cloud data to obtain a three-dimensional simulation model of the feed feeding equipment; The actual feeding rate of the bait feeding equipment within a preset time period is obtained by using the Internet of Things technology, the preset time period is divided into a plurality of uniform time nodes, a curve graph is constructed based on the plurality of uniform time nodes and the actual feeding rate corresponding to each time node, and an actual feeding rate-time change curve graph is obtained; The feeding amount is calculated by analyzing the actual feeding rate-time variation curve to obtain the actual feeding amount, and the curvature of the curve is calculated to obtain a first curvature; Based on the plurality of uniform time nodes, the actual feeding amount is simulated in the three-dimensional simulation model of the bait feeding equipment to obtain a simulated feeding rate corresponding to each time node, a simulated feeding rate-time variation curve diagram is constructed based on the simulated feeding rate corresponding to each time node, the feeding amount is calculated according to the simulated feeding rate-time variation curve diagram to obtain the simulated feeding amount, and the curvature of the curve is calculated to obtain a second curvature; Determine whether the actual feeding amount is less than the simulated feeding amount. If so, it indicates that there is a fault in the bait feeding equipment. Calculate the deviation between the first curvature and the second curvature to obtain a curvature deviation value. Repair and maintain the fault of the bait feeding equipment based on the curvature deviation value.

[0040] It should be noted that the failure of the feed feeding equipment is also one of the key factors affecting the survival rate of the target aquatic organisms. When the feed feeding equipment fails, it may cause large errors in the metering output rate, output volume and data monitoring of the feed, so that the feed feeding equipment cannot accurately meter and feed according to the second metering correction scheme, thereby causing the nutrition of the target aquatic organisms to decrease. Therefore, it is necessary to determine and repair the fault of the feed feeding equipment; for judging whether the feed feeding equipment has a fault, the error between the simulated feeding behavior and the historical actual feeding behavior can be determined by simulating feeding, so that the equipment can be judged to have a fault based on the error, which saves the steps of repeated manual field operations, improves the rate and accuracy of data acquisition and comparison, and reduces the occurrence of judgment errors; wherein, the method of simulating feeding is realized by constructing a three-dimensional simulation model of the feed feeding equipment, which is low in cost and easy to operate; the first curvature represents the changing trend of the actual feeding rate, and the second curvature represents the changing trend of the simulated feeding rate. By analyzing the changing trends of the two, the degree of the fault error of the equipment can be known, so that it is convenient for maintenance personnel to accurately and quickly repair the faulty equipment and improve the metering quality.

[0041] In addition, the method for measuring bait feeding for ships based on the Internet of Things further comprises the following steps: Obtain a fault time node of the bait feeding equipment, determine whether the fault time node is within a preset working time interval, and if so, obtain the remaining feeding metering task of the currently faulty bait feeding equipment; A distance threshold is preset, and the distance values ​​between the current ship and all other ships in the aquaculture area are obtained through satellite positioning technology. If the distance value is less than the distance threshold, the distance values ​​less than the distance threshold are extracted, and the ship corresponding to the minimum distance value is screened out from the extracted distance values, and the ship corresponding to the minimum distance value is defined as the closest point ship; Obtain the remaining bait metering task and the driving speed of the nearest ship, and calculate the remaining bait metering task and the driving speed of each nearest ship to obtain the time to complete the remaining bait metering task of the nearest ship; Based on the remaining required feeding metering task of the currently faulty bait feeding equipment and the travel speed, a calculation is performed to obtain the time required for the nearest ship to complete the remaining feeding metering task of the currently faulty bait feeding equipment; The time required for the nearest ship to complete the remaining feeding metering task of the currently faulty feeding equipment is added to the time for the nearest ship to complete the remaining feeding metering task to obtain the total feeding time, and whether the total feeding time is less than the preset total working time; If it is less than, a help signal will be sent to the nearest ship through the Internet of Things technology, and the remaining required feeding metering task of the current faulty bait feeding equipment will be uploaded to the control terminal of the bait feeding equipment on the nearest ship to continue completing the task.

[0042] It should be noted that when the ship feeds the target breeding area, if the feed feeding equipment fails in the middle of feeding, it will lead to the inability to continue to complete the remaining feeding tasks, so that the target aquatic organisms can take in the accurately measured feed nutrients, greatly affecting the quality of breeding growth, reducing the survival rate, and being unfavorable for the high-quality breeding of the target aquatic organisms; and when the ship is feeding the feed in the breeding areas such as the sea, rivers and lakes, there are usually multiple spare ships or multiple ships working together at the same time, on the one hand, the problem of low feeding efficiency of a single ship is reduced, and on the other hand, the problem of failure of the ship or the failure of the feed feeding equipment can be prevented. The problem of normal breeding work can be prevented. Therefore, when the feed feeding equipment on a certain ship fails, a group request signal can be sent to the ship closest to the current faulty ship, and the remaining unfinished feeding tasks of the current ship can be handed over to the nearest ship to continue to complete. The present invention can hand over the remaining feeding tasks to the nearest ship to continue to complete when the ship feed feeding equipment fails during the feeding process, thereby solving the problem of midway failure of the feed feeding equipment, improving feeding efficiency, ensuring the quality of breeding growth, and reducing the measurement error of breeding feeding.

[0043] In addition, the method for measuring bait feeding for ships based on the Internet of Things further comprises the following steps: Match bait varieties in the big data network based on the required nutrients of target aquatic organisms, obtain related bait varieties, and obtain the purchase cost corresponding to each bait variety; If the purchase cost is greater than the preset purchase cost, the bait varieties corresponding to the purchase cost greater than the preset purchase cost are extracted to obtain the bait varieties after a screening; Based on the big data network, the nutritional value of each bait variety after the first screening is obtained, and the bait varieties after the first screening whose nutritional value is lower than the preset nutritional value are eliminated to obtain the bait varieties after the second screening; Constructing a sorting table, and importing the purchase costs corresponding to the bait varieties after the secondary screening into the sorting table for sorting from large to small; After the sorting is completed, the secondary screened bait variety corresponding to the maximum purchase cost is extracted and output as the optimal feeding bait.

[0044] It should be noted that the purchase and selection of bait should fully consider multiple factors, among which the two most important factors for breeders are nutritional value and purchase cost. The selection of bait for target aquatic organisms should first consider whether the nutritional value of the bait variety can achieve the expected breeding effect, and the bait variety with higher nutritional value and suitable for the breeding of target aquatic organisms should be selected; secondly, there are multiple varieties of bait, and there are differences in the price of each. For the consumer perspective, the bait variety with lower purchase cost should be selected to be the most suitable, which can ensure that the expected breeding target is achieved, and the lowest purchase cost output can be guaranteed to be controlled, and the feeding experience of breeders is better improved. The present invention can be preliminarily screened according to the nutritional value of the bait variety, ensuring that the feeding and breeding performance meets the expected effect, and then the bait variety is screened twice according to the purchase cost, which can reduce the output of the bait purchase cost of the breeders, and has a high cost performance ratio, and has a further improved effect for the feeding and metering of the bait.

[0045] The second aspect of the present invention provides a bait feeding metering system for ships based on the Internet of Things, wherein the bait feeding metering system for ships based on the Internet of Things comprises a memory 41 and a processor 42, wherein the memory 41 stores a bait feeding metering method program for ships based on the Internet of Things, and when the bait feeding metering method program for ships based on the Internet of Things is executed by the processor 42, Figure 4 As shown, implement the following steps: Acquire aquaculture image data of multiple target areas within a preset time period, and introduce a local binary pattern algorithm to perform feature extraction to obtain aquaculture biological features of the multiple target areas, and determine the activity range and aquaculture density of the target aquatic organisms based on the aquaculture biological features of the multiple target areas; Obtaining the required nutrients of the target aquatic organisms, analyzing and evaluating the nutrient content of the required nutrients in the current feed to obtain an evaluation result, and combining the evaluation result, the activity range and the breeding density to make an estimate to obtain a preliminary feeding measurement value; Feeding according to the preliminary feeding measurement value, obtaining the survival rate of the target aquatic organism, constructing a trained survival rate prediction model for prediction, obtaining a survival rate interval with insufficient nutrient content, analyzing whether the survival rate of the target aquatic organism is within the survival rate interval, and obtaining a first measurement correction plan; Based on the first metering correction scheme, feeding adjustment is performed to obtain the actual survival rate after the first correction. If the actual survival rate after the first correction is not within the survival rate range of water environment pollution, the ammonia nitrogen decomposition rate in the current water body is calculated and analyzed to obtain a second metering correction scheme; Based on the second metering correction scheme, feeding adjustments are made to obtain the actual survival rate after the second correction. If the actual survival rate after the second correction is not within the survival rate range of equipment failure, a curve chart is constructed to analyze whether there is a fault in the bait feeding equipment and repair and maintenance are performed.

[0046] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for feeding and metering bait for ships based on the Internet of Things, characterized in that: The following steps are involved: Acquire aquaculture image data of multiple target areas within a preset time period, and introduce a local binary pattern algorithm to perform feature extraction to obtain aquaculture biological features of the multiple target areas, and determine the activity range and aquaculture density of the target aquatic organisms based on the aquaculture biological features of the multiple target areas; Obtaining the required nutrients of the target aquatic organisms, analyzing and evaluating the nutrient content of the required nutrients in the current feed to obtain an evaluation result, and combining the evaluation result, the activity range and the breeding density to make an estimate to obtain a preliminary feeding measurement value; Feeding according to the preliminary feeding measurement value, obtaining the survival rate of the target aquatic organism, constructing a trained survival rate prediction model for prediction, obtaining a survival rate interval with insufficient nutrient content, analyzing whether the survival rate of the target aquatic organism is within the survival rate interval, and obtaining a first measurement correction plan; Based on the first metering correction scheme, feeding adjustment is performed to obtain the actual survival rate after the first correction. If the actual survival rate after the first correction is not within the survival rate range of water environment pollution, the ammonia nitrogen decomposition rate in the current water body is calculated and analyzed to obtain a second metering correction scheme; Based on the second metering correction scheme, feeding adjustments are made to obtain the actual survival rate after the second correction. If the actual survival rate after the second correction is not within the survival rate range of equipment failure, a curve chart is constructed to analyze whether there is a fault in the bait feeding equipment and repair and maintenance are performed.

2. A method for feeding and metering bait for ships based on the Internet of Things according to claim 1, characterized in that: The method of acquiring aquaculture image data of multiple target areas within a preset time period, and introducing a local binary pattern algorithm to perform feature extraction to obtain aquaculture biological features of multiple target areas, and determining the activity range and aquaculture density of target aquatic organisms according to the aquaculture biological features of the multiple target areas specifically includes the following steps: Based on the Internet of Things technology, the remote sensing camera on the fishing boat is controlled to capture images of target aquatic organisms, and the aquaculture image data of multiple target areas within a preset time period are obtained; A local binary pattern algorithm is introduced to extract the features of the farmed organisms from the farmed image data of the plurality of target areas, each farmed image data is divided into a plurality of local blocks, and the pixel points in each of the local blocks are compared to generate a plurality of binary coded information; Converting the plurality of binary coded information into different local binary patterns, counting the frequencies of the different local binary patterns in each local block, drawing and outputting a local binary pattern histogram, and obtaining the aquaculture biological characteristics of the plurality of target areas; The Euclidean distances between the characteristics of the aquaculture organisms in the target areas are calculated, and a hot spot radius algorithm is introduced to plan the hot spot areas for the calculated multiple Euclidean distances, and a first activity hot spot map is constructed. The activity range and aquaculture density of the target aquatic organisms are determined according to the first activity hot spot map.

3. A method for feeding and metering bait for ships based on the Internet of Things according to claim 1, characterized in that: The step of obtaining the required nutrients for the target aquatic organisms, analyzing and evaluating the nutrient content of the required nutrients in the current feed to obtain an evaluation result, and estimating the preliminary feeding measurement value by combining the evaluation result, the activity range and the breeding density, specifically includes the following steps: Obtaining the species information of the target aquatic organism, searching the relevant nutritional information required for the normal growth of the target aquatic organism based on the big data network, obtaining the required nutritional components of the target aquatic organism, and obtaining the required intake of each nutritional component; Acquire several kinds of ingredient information of the current bait, construct a knowledge graph, search the nutritional components contained in the several kinds of ingredient information in the knowledge graph, obtain search results, screen and mark the ingredient information that is the same as the required nutritional components of the target aquatic organism in the search results, and obtain the target ingredient information; Based on the big data network, the standard nutritional content interval corresponding to the target ingredient is obtained, and the ratio quality of the target ingredient is obtained, and the fuzzy C-means algorithm is introduced to calculate the initial membership of the ratio quality and each standard nutritional content in the standard nutritional content interval, and a cluster center is constructed. The position of the cluster center is updated based on the membership until the maximum number of iterations is met, and multiple memberships are generated; Constructing an evaluation model, selecting the standard nutritional content corresponding to the maximum membership degree from the plurality of membership degrees as a result output, obtaining the nutritional content of the current bait, and importing the nutritional content of the current bait into the evaluation model for evaluation to obtain an evaluation result; The ratio between the required intake of each nutrient component and the nutrient content of the current bait is calculated to obtain an individual feeding ratio, and the individual feeding ratio, the evaluation result, the activity range and the breeding density are comprehensively estimated to obtain a preliminary feeding measurement value.

4. A method for feeding and metering bait for ships based on the Internet of Things according to claim 1, characterized in that: The method of feeding according to the preliminary feeding measurement value, obtaining the survival rate of the target aquatic organism, constructing a trained survival rate prediction model for prediction, obtaining a survival rate interval with insufficient nutrient content, analyzing whether the survival rate of the target aquatic organism is within the survival rate interval, and obtaining a first measurement correction plan specifically includes the following steps: Inputting the preliminary feeding measurement value of the target area into the bait feeding device for metered feeding, and reconstructing the heat map after a preset time period to obtain a second activity heat map; Based on the second activity hotspot map, all non-hotspots and all feature overlap points in the target area are obtained, the non-hotspots and the feature overlap points are defined as low activity points of the target aquatic organisms, and the ratio of the low activity points of all target aquatic organisms to all hotspots is calculated in the second activity hotspot map to obtain the survival rate of the target aquatic organisms; The survival rates corresponding to the influence of different aquaculture conditions and factors are obtained through a big data network, and a survival rate prediction model is constructed based on a support vector machine algorithm. The survival rates corresponding to the influence of the different aquaculture conditions and factors are trained and verified in the survival probability prediction model to obtain a trained survival rate prediction model; wherein the different aquaculture conditions and factors include insufficient nutrient content, water environment pollution, and equipment failure; Extracting a survival rate interval with insufficient nutrient content based on the trained survival rate prediction model, and determining whether the survival rate of the target aquatic organism is within the survival rate interval with insufficient nutrient content; If not, a hash algorithm is introduced to calculate a hash value between the lower limit value in the survival rate interval and the survival rate of the target aquatic organism, and the required feeding amount of the target area is metered and regulated according to the hash value to obtain a first metering correction plan.

5. The method for feeding and metering bait for a ship based on the Internet of Things according to claim 1, characterized in that: The method of adjusting feeding based on the first metering correction scheme to obtain the actual survival rate after the first correction, and if the actual survival rate after the first correction is not within the survival rate range of water environment pollution, calculating and analyzing the ammonia nitrogen decomposition rate in the current water body to obtain the second metering correction scheme, specifically includes the following steps: According to the first metering correction scheme, the feed feeding equipment is corrected in metering, and the survival rate of the corrected target aquatic organism is calculated to obtain the actual survival rate after a correction; The survival rate prediction model after training is used to extract the survival rate interval of water environment pollution. If the actual survival rate after the first correction is within the survival rate interval of water environment pollution, aquaculture water samples are extracted and tested by chemical testing methods to obtain the current ammonia nitrogen content index. Acquire multiple historical ammonia nitrogen content indexes within a preset time period, construct a series of ammonia nitrogen decomposition change models according to the multiple historical ammonia nitrogen content indexes and the current ammonia nitrogen content index, introduce a mean absolute error algorithm to calculate the model errors between the series of ammonia nitrogen decomposition change models, obtain multiple discrete error rates, screen out the minimum discrete error rate and define it as the current ammonia nitrogen decomposition rate; Obtaining a water body ammonia nitrogen pollution standard evaluation system based on a big data network, establishing a scoring rule according to the water body ammonia nitrogen pollution standard evaluation system, and scoring the current ammonia nitrogen decomposition rate according to the scoring rule to obtain a score value; Determine whether the score value is less than a preset score value. If so, calculate the error between the score value and the preset score value to obtain an error value. According to the error value, match the corresponding ammonia nitrogen decomposition rate in the water ammonia nitrogen pollution standard evaluation system and define it as the ammonia nitrogen decomposition error rate. Based on the ammonia nitrogen decomposition error rate, perform metering control on the first metering correction scheme to obtain a second metering correction scheme.

6. The method for feeding and metering bait for a ship based on the Internet of Things according to claim 1, characterized in that: The feeding adjustment is performed based on the second metering correction scheme to obtain the actual survival rate after the second correction. If the actual survival rate after the second correction is not within the survival rate interval of the equipment failure, a curve chart is constructed to analyze whether there is a failure in the feed feeding equipment and repair and maintain it, which specifically includes the following steps: The feed feeding equipment is corrected according to the second metering correction scheme, and the corrected survival rate of the target aquatic organism is recalculated to obtain the actual survival rate after the second correction, and the survival rate interval of the equipment failure is extracted based on the trained survival rate prediction model; If the actual survival rate after the secondary correction is within the survival rate range of the equipment failure, the feed feeding equipment is subjected to laser coverage scanning by laser point cloud scanning technology to obtain several groups of point cloud data of the feed feeding equipment, and the RANSAC algorithm is introduced to perform model fitting on the several groups of point cloud data to obtain a three-dimensional simulation model of the feed feeding equipment; The actual feeding rate of the bait feeding equipment within a preset time period is obtained by using the Internet of Things technology, the preset time period is divided into a plurality of uniform time nodes, a curve graph is constructed based on the plurality of uniform time nodes and the actual feeding rate corresponding to each time node, and an actual feeding rate-time change curve graph is obtained; The feeding amount is calculated by analyzing the actual feeding rate-time variation curve to obtain the actual feeding amount, and the curvature of the curve is calculated to obtain a first curvature; Based on the plurality of uniform time nodes, the actual feeding amount is simulated in the three-dimensional simulation model of the bait feeding equipment to obtain a simulated feeding rate corresponding to each time node, a simulated feeding rate-time variation curve diagram is constructed based on the simulated feeding rate corresponding to each time node, the feeding amount is calculated according to the simulated feeding rate-time variation curve diagram to obtain the simulated feeding amount, and the curvature of the curve is calculated to obtain a second curvature; Determine whether the actual feeding amount is less than the simulated feeding amount. If so, it indicates that there is a fault in the bait feeding equipment. Calculate the deviation between the first curvature and the second curvature to obtain a curvature deviation value. Repair and maintain the fault of the bait feeding equipment based on the curvature deviation value.

7. A bait feeding metering system for ships based on the Internet of Things, characterized in that: The system for measuring bait feeding for a ship based on the Internet of Things comprises a memory and a processor, wherein the memory stores a program for measuring bait feeding for a ship based on the Internet of Things, and when the program for measuring bait feeding for a ship based on the Internet of Things is executed by the processor, the following steps are implemented: Acquire aquaculture image data of multiple target areas within a preset time period, and introduce a local binary pattern algorithm to perform feature extraction to obtain aquaculture biological features of the multiple target areas, and determine the activity range and aquaculture density of the target aquatic organisms based on the aquaculture biological features of the multiple target areas; Obtaining the required nutrients of the target aquatic organisms, analyzing and evaluating the nutrient content of the required nutrients in the current feed to obtain an evaluation result, and combining the evaluation result, the activity range and the breeding density to make an estimate to obtain a preliminary feeding measurement value; Feeding according to the preliminary feeding measurement value, obtaining the survival rate of the target aquatic organism, constructing a trained survival rate prediction model for prediction, obtaining a survival rate interval with insufficient nutrient content, analyzing whether the survival rate of the target aquatic organism is within the survival rate interval, and obtaining a first measurement correction plan; Based on the first metering correction scheme, feeding adjustment is performed to obtain the actual survival rate after the first correction. If the actual survival rate after the first correction is not within the survival rate range of water environment pollution, the ammonia nitrogen decomposition rate in the current water body is calculated and analyzed to obtain a second metering correction scheme; Based on the second metering correction scheme, feeding adjustments are made to obtain the actual survival rate after the second correction. If the actual survival rate after the second correction is not within the survival rate range of equipment failure, a curve chart is constructed to analyze whether there is a fault in the bait feeding equipment and repair and maintenance are performed.

Images