A method of coordinated increase of fishery resources based on large water surface environment

Abstract

The invention belongs to the technical field of fishery resource management, and discloses a method for coordinated proliferation of fishery resources based on a large water surface environment; the method comprises: collecting water area characteristic data; the water area characteristic data comprises water quality data, bottom quality data, sound wave data and water area image; performing aquatic biological monitoring according to the sound wave data to obtain aquatic biological data; formulating a proliferation and recovery strategy in combination with the water quality data, bottom quality data and aquatic biological data; performing fish resource assessment according to the water area image to obtain fish resource data; optimizing proliferation objects in combination with the water quality data, bottom quality data, aquatic biological data and fish resource data; determining the proliferation and release capacity based on the proliferation objects; the invention not only improves the accuracy and intelligence level of fishery resource management, but also can dynamically adjust strategies to cope with environmental changes, balance population relationships in the ecosystem, and effectively improve the sustainable development capacity of fishery resources.

Description

Technical Field

[0001] The present invention relates to the technical field of fishery resource management, and more specifically, to a method for coordinated proliferation of fishery resources based on a large water surface environment. Background Art

[0002] Globally, fishery resources are facing severe challenges such as overfishing, habitat destruction and water pollution, resulting in a significant decline in fishery resources in many waters. Especially in large water environments, due to their ecological complexity and biodiversity, the management and protection of fishery resources is particularly important. In response to the above problems, the synergistic enhancement method came into being. Synergistic enhancement refers to promoting the common growth and reproduction of different fish and other aquatic organisms in waters through the combination of scientific management and multiple ecological strategies. It not only focuses on the proliferation of a single species, but also emphasizes the interaction between species in the ecosystem, striving to increase fishery production while maintaining ecological balance.

[0003] The patent application with publication number CN116663930A discloses a tropical marine ranch ecological management method, system, equipment and medium; including: obtaining ecological environment data and energy flow analysis data of the target marine ranch; conducting ecological assessment on the ecological environment data to determine the ecological comprehensive assessment result; the ecological comprehensive assessment result includes the health status assessment of the biological community and the stability assessment of the ecosystem; using the Ecopath energy flow model, conducting a biocarrying capacity assessment based on the energy flow analysis data and the ecological comprehensive assessment result to obtain the current biocarrying capacity; conducting ecological management based on the current biocarrying capacity and formulating an ecological management strategy; the ecological management strategy includes reproduction and release and fishing plans; this invention can unify the assessment standards and facilitate the ecological management of marine ranches.

[0004] However, when formulating ecological management strategies, the above methods only outline "conducting ecological management based on the current biological carrying capacity and formulating ecological management strategies", but do not provide detailed explanations of the strategy formulation process. They also lack the ability to quickly customize personalized strategies based on specific fishery characteristics when the large-scale water surface environment changes, resulting in insufficient versatility of the strategies, which affects the scientific nature and effectiveness of the strategies.

[0005] In view of this, the present invention proposes a method for coordinated proliferation of fishery resources based on a large water surface environment to solve the above problems. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art and achieve the above-mentioned purpose, the present invention provides the following technical solution: a method for coordinated proliferation of fishery resources based on a large water surface environment, comprising:

[0007] S1: Collect water area characteristic data; water area characteristic data includes water quality data, bottom quality data, sound wave data and water area image;

[0008] S2: Monitor aquatic organisms based on acoustic wave data and obtain aquatic organism data;

[0009] S3: Combine water quality data, bottom quality data and aquatic biological data to develop a restoration strategy;

[0010] S4: Evaluate fish resources based on water area images and obtain fish resource data;

[0011] S5: Combine water quality data, bottom quality data, aquatic organism data and fish resource data to select the target for proliferation;

[0012] S6: Determine the proliferation and release capacity based on the proliferation objects.

[0013] Furthermore, the aquatic biological data includes biological abundance and biological species; the biological species include aquatic plant species, plankton species and benthic species; the biological abundance is the number of each aquatic organism in the biological species;

[0014] The method for obtaining aquatic organism data comprises:

[0015] Process the acoustic wave data to obtain corresponding acoustic feature data; the acoustic feature data includes acoustic wave intensity, acoustic wave frequency, echo duration and acoustic wave scattering intensity; input the acoustic feature data into the trained biological prediction model to predict the corresponding set label, which is a digital label corresponding to the analysis set, and different analysis sets have different corresponding digital labels; obtain the corresponding analysis set according to the predicted set label, and use the analysis set as aquatic biological data; the analysis set is set by randomly combining different biological species and biological abundances, and one combination corresponds to one analysis set;

[0016] The training process of the biological prediction model includes:

[0017] A group of acoustic feature data is collected in advance, and corresponding set labels are set for the a group of acoustic feature data, where a is an integer greater than 1, and the acoustic feature data and the corresponding set labels are converted into a corresponding set of feature vectors; each set of feature vectors is used as the input of a biological prediction model, and the biological prediction model uses a set of predicted set labels corresponding to each group of acoustic feature data as output, and uses the actual set labels corresponding to each group of acoustic feature data as a prediction target, where the actual set labels are pre-set set labels corresponding to the acoustic feature data; minimizing the sum of prediction errors of all acoustic feature data is used as a training target; the biological prediction model is trained until the sum of prediction errors converges and the training is stopped; the biological prediction model is a deep neural network model.

[0018] Furthermore, the method for obtaining acoustic feature data includes:

[0019] According to the sound wave data, the sound pressure value within time T is obtained, and the root mean square sound pressure is calculated. The expression of the root mean square sound pressure is: ; In the formula, is the root mean square sound pressure, is the sound pressure value at time t, To integrate the square of the sound pressure value from time 0 to time t, ; Calculate the sound wave intensity based on the root mean square sound pressure; The expression of sound wave intensity is: ; In the formula is the sound wave intensity, is the water density, is the speed of sound waves in water;

[0020] Fast Fourier transform is used to convert the acoustic wave data from a time domain signal into a frequency signal to obtain the acoustic wave frequency;

[0021] Obtain the sound wave emission time and the sound wave reception time, subtract the sound wave emission time from the sound wave reception time, and obtain the echo duration signal;

[0022] Get the emission intensity, which is the intensity of the sound wave when the sonar sensor emits the sound wave; calculate the sound wave scattering intensity based on the emission intensity and the sound wave intensity; the expression of the sound wave scattering intensity is: ; In the formula, is the sound wave scattering intensity, is the emission intensity.

[0023] Further, the steps to develop a proliferation recovery strategy include:

[0024] Step S301: construct m restoration sets, where m is an integer greater than 1, and the restoration sets include the required introduction quantity of each aquatic plant and benthic organism, and the m restoration sets are all different;

[0025] Step S302: Randomly select from m recovery sets Different recovery sets are marked as selection sets. Each selection set is taken as an individual to generate a population. ;

[0026] Step S303: mutate each individual in the population to generate a mutant individual; the expression of the mutant individual is: ; In the formula is the jth mutant individual, is the jth individual, , are two individuals randomly selected from the population, is the variation factor, , ;

[0027] Step S304: using a crossover operation to combine each individual in the population with the corresponding variant individual to generate a new individual;

[0028] Step S305: respectively calculating the recovery degree of each individual in the population and the new individual, and updating the individuals in the population;

[0029] Step S306: Obtain the maximum recovery degree corresponding to the updated population, and subtract the maximum recovery degree corresponding to the updated population to obtain the recovery degree difference; preset a difference threshold, if the recovery degree difference is less than the difference threshold, generate a slow instruction, if the recovery degree difference is greater than or equal to the difference threshold, do not generate a slow instruction; preset a quantity threshold, count the number of slow instructions generated continuously, and mark it as the instruction quantity, if the instruction quantity is greater than or equal to the quantity threshold, enter step S307, if the instruction quantity is less than the quantity threshold, return to step S303;

[0030] Step S307: Obtain the recovery set corresponding to the maximum recovery degree.

[0031] Furthermore, in step S304, the method for generating a new individual is: preset crossover probability , ; For each variable in an individual, a The random number between the individual and the variable in the individual is the number of introductions required for each aquatic plant or benthic organism, and the variable corresponds to the introduction number one by one; the random number corresponding to each variable in the individual is compared with the crossover probability. If the random number is less than the crossover probability, the corresponding variable is replaced by the corresponding variable in the corresponding variant individual; if the random number is greater than or equal to the crossover probability, the corresponding variable is not replaced; a variable is randomly selected from all the variables in the individual, and the selected variable is replaced by the corresponding variable in the corresponding variant individual.

[0032] Furthermore, in step S305, the updating method is: if the recovery degree of an individual is less than the recovery degree of a corresponding new individual, the individual is updated to the corresponding new individual; if the recovery degree of an individual is greater than the recovery degree of a corresponding new individual, the corresponding individual is not updated; the calculation method of the recovery degree is: using water quality data, bottom sediment data and aquatic biological data as research data, using the research data and the recovery set corresponding to the individual as calculation data, inputting the calculation data into a trained recovery assessment model, and evaluating the corresponding recovery degree, where the recovery degree is the recovery effect of the ecological health status of the water area; the training process of the recovery assessment model is consistent with the training process of the biological prediction model, and both are neural network models.

[0033] Furthermore, the fish resource data includes fish abundance and fish species; the fish abundance is the number of each type of fish in the fish species;

[0034] The method for obtaining fish resource data comprises:

[0035] The target detection method is used to mark the fish in the water image with rectangular frames, and the rectangular frames correspond to the fish one by one; the water image is segmented according to the marked rectangular frames to segment the water sub-images, and the water sub-images correspond to the rectangular frames one by one; the trained fish recognition model is used to identify the fish in the water sub-images, and the recognition results are output, and the recognition results are the types of fish; according to the identified types of fish, the number of each type of fish is counted to obtain the abundance of fish;

[0036] The fish identification model training process includes:

[0037] Collect multiple water sub-images in advance, and annotate the fish in each water sub-image as fish species; convert different fish species into different digital annotations; divide the annotated water sub-images into a training set and a test set; use the training set to train a fish recognition model, and use the test set to test the fish recognition model; preset an error threshold, and when the mean of the prediction errors of all water sub-images in the test set is less than the error threshold, output a fish recognition model; the fish recognition model is a convolutional neural network model.

[0038] Furthermore, the objects of proliferation are fish species released into the waters;

[0039] The method of preferably proliferating an object comprises:

[0040] Step S501: establishing a fuzzy set, dividing each data in the preferred data into multiple fuzzy sets, the preferred data including research data and fish resource data;

[0041] Step S502: convert each preferred data into the membership degree of each corresponding fuzzy set through fuzzification technology; fuzzification is the process of converting an exact numerical value into the membership degree corresponding to a fuzzy set;

[0042] Step S503: define fuzzy rules;

[0043] Step S504: matching the fuzzified preferred data with the fuzzy rules, performing fuzzy reasoning, and obtaining the fuzzy reasoning result, which is the membership degree of the proliferation object;

[0044] Step S505: Defuzzify the fuzzy reasoning result to obtain the defuzzified result, which is the proliferation object; the defuzzification operation is: according to the fuzzy reasoning result, obtain the membership corresponding to each fish respectively; preset the membership threshold, compare the membership corresponding to each fish with the membership threshold respectively, and take the fish with a membership greater than the membership threshold as the proliferation object.

[0045] Furthermore, methods for determining the growth release capacity include:

[0046] The objects of enhancement are divided according to their feeding habits to obtain feeding types; feeding types include filter-feeding fish, herbivorous fish, benthic fish and detritus-feeding fish; enhancement and release capacity includes filter-feeding fish enhancement capacity, herbivorous fish enhancement capacity, benthic fish enhancement capacity and detritus-feeding fish enhancement capacity;

[0047] Obtain the number of aquatic plants, plankton, and benthic organisms based on biological abundance;

[0048] The expression for the proliferation capacity of filter-feeding fish is: ; In the formula, is the proliferation capacity of filter-feeding fish, is the number of plankton, is the first biological coefficient, Water volume;

[0049] The expression of the proliferation capacity of herbivorous fish is: ; is the proliferation capacity of herbivorous fish, The number of aquatic plants.

[0050] Furthermore, the expression for the proliferation capacity of benthic feeding fish is: ; The capacity for the proliferation of benthic feeding fish is is the number of benthic organisms, is the second biological coefficient, is the water area; obtain the region where the water area is located, and obtain the first biological coefficient and the second biological coefficient through a preset biological coefficient set, wherein the biological coefficient set includes the values ​​of the first biological coefficient and the second biological coefficient in different regions;

[0051] The expression for the proliferation capacity of detritus-feeding fish is:

[0052] ;

[0053] In the formula, is the proliferation capacity of detritus-feeding fish, is the organic carbon content, is the proportion of silver carp, is the proportion of bighead carp; the method for obtaining the proportion of silver carp and bighead carp is: according to the fish resource data, obtain the number of silver carp and bighead carp; add the number of silver carp to the number of bighead carp to obtain the total number of silver carp and bighead carp; divide the number of silver carp by the total number of silver carp and bighead carp to obtain the proportion of silver carp; divide the number of bighead carp by the total number of silver carp and bighead carp to obtain the proportion of bighead carp.

[0054] The technical effects and advantages of the method for coordinated proliferation of fishery resources based on a large water surface environment of the present invention are as follows:

[0055] Based on the large water surface environment, the ecological health status of the water body and the situation of fishery resources are comprehensively assessed by collecting and analyzing water quality, bottom quality, sound waves and water image data; deep learning technology is used to accurately obtain the species and abundance of aquatic organisms, and scientific reproduction and restoration strategies are formulated in combination with water quality data and bottom quality data; suitable fish species for reproduction are selected through fuzzy reasoning and defuzzification methods, and reasonable reproduction and release capacity is calculated based on their feeding habits and ecological needs to ensure the reasonable reproduction of fishery resources; it not only improves the accuracy and intelligence level of fishery resource management, but also can dynamically adjust strategies to cope with environmental changes, balance population relationships in the ecosystem, and effectively enhance the sustainable development capacity of fishery resources, thereby achieving a dual improvement in ecological and economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is a flow chart of a method for coordinated proliferation of fishery resources based on a large water surface environment according to Example 1 of the present invention;

[0057] Figure 2 This is a flow chart of the method for formulating a proliferation recovery strategy according to Example 1 of the present invention. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Example 1

[0059] See also Figure 1 As shown, the method for coordinated proliferation of fishery resources based on a large water surface environment described in this embodiment includes:

[0060] S1: Collect water area characteristic data; water area characteristic data includes water quality data, bottom quality data, sound wave data and water area images.

[0061] Water quality data refers to data that describes the quality and characteristics of water bodies and is used to assess the health status and degree of pollution of water bodies. Water quality data includes water temperature, pH value, turbidity, dissolved oxygen concentration, ammonia nitrogen concentration, etc. Water quality data is obtained by water quality monitoring sensors installed in water areas. Water quality monitoring sensors integrate multiple sensors, such as temperature sensors, dissolved oxygen sensors, ammonia nitrogen sensors, etc. Water areas refer to large water surface environments, such as lakes, rivers, wetlands, etc.

[0062] Bottom sediment data refers to data that describes the characteristics of sediments at the bottom of water bodies and is used to reflect the living environment of benthic organisms. Bottom sediment data includes particle size (the size of particles in sediments), organic matter content, nutrient concentration, etc. Bottom sediment data is obtained by obtaining bottom sediment samples from bottom sediment samplers (such as grab samplers or sediment collectors) distributed and installed in the water area, and is obtained through detection by sensors such as laser particle size analyzers, organic matter analyzers, and chemical analyzers.

[0063] Acoustic wave data is the data information obtained when the acoustic wave signal propagates in the water body. The acoustic wave data is obtained by sonar sensors installed in underwater monitoring stations.

[0064] The water area image is a visual image of the water area environment, and the water area image is obtained by underwater cameras distributed and installed in the water area.

[0065] S2: Monitor aquatic organisms based on acoustic wave data and obtain aquatic organism data.

[0066] Aquatic biological data include biological abundance and biological species; biological species include aquatic plant species (such as water plantain, duckweed, etc.), plankton species (such as water fleas, diatoms, etc.) and benthic species (such as snails, freshwater shrimps, dragonfly larvae, etc.); biological abundance is the number of each aquatic organism in the biological species, and aquatic organisms include aquatic plants, plankton and benthic organisms.

[0067] Methods for obtaining aquatic organism data include:

[0068] The acoustic wave data are processed to obtain corresponding acoustic feature data; the acoustic feature data include acoustic wave intensity (acoustic energy passing through a unit area), acoustic wave frequency (the number of times the acoustic wave vibrates per unit time), echo duration (the length of time required for the reflected signal to return to the sensor after the acoustic wave is emitted) and acoustic wave scattering intensity (the energy of the acoustic wave signal scattered back after the acoustic wave interacts with an object in the water); the acoustic feature data are input into a trained biological prediction model to predict the corresponding set label, which is a digital label corresponding to the analysis set, and different analysis sets have different corresponding digital labels; the corresponding analysis set is obtained according to the predicted set label, and the analysis set is used as aquatic biological data; the analysis set is set in such a way that different biological species and biological abundances are randomly combined, and one combination corresponds to one analysis set; illustratively, one analysis set includes water plantain, water flea and snail, and the corresponding quantities are 10, 15, and 20, respectively, and the corresponding digital label is 1; another analysis set includes water plantain, diatom, snail and freshwater shrimp, and the corresponding quantities are 20, 10, 18, and 30, respectively, and the corresponding digital label is 2.

[0069] The training process of the biological prediction model includes:

[0070] A group of acoustic feature data is collected in advance, and corresponding set labels are set for the group of acoustic feature data, where a is an integer greater than 1, and the acoustic feature data and the corresponding set labels are converted into a corresponding set of feature vectors; the set labels corresponding to the acoustic feature data are collected by technicians in this field in the historical aquatic biological monitoring process, and each group of acoustic feature data is analyzed in turn based on actual experience to determine the set labels corresponding to each group of acoustic feature data, and the corresponding set labels are set for the group of acoustic feature data in turn;

[0071] Each set of feature vectors is used as the input of the biological prediction model. The biological prediction model uses a set of predicted set labels corresponding to each set of acoustic feature data as output, and uses the actual set labels corresponding to each set of acoustic feature data as the prediction target. The actual set labels are pre-set set labels corresponding to the acoustic feature data. Minimizing the sum of prediction errors of all acoustic feature data is used as the training target. The calculation formula of the prediction error is: ,in is the prediction error, K is the group number of the feature vector corresponding to the acoustic feature data, is the predicted set label corresponding to the Kth group of acoustic feature data, is the actual set label corresponding to the Kth group of acoustic feature data; the biological prediction model is trained until the sum of the prediction errors reaches convergence and the training is stopped.

[0072] The above-mentioned biological prediction model is specifically a deep neural network model; it includes an input layer, a hidden layer and an output layer; each hidden layer includes multiple neurons, each neuron is connected to the neurons in the next layer, and the connection contains weights, which determine the importance and influence of data transmission in the neural network; an activation function is applied to each neuron between the hidden layer and the output layer, and the activation function introduces nonlinearity, allowing the network to learn more complex patterns and features.

[0073] It should be noted that changes in sound wave intensity and sound wave scattering intensity reflect the number of aquatic organisms. The more aquatic organisms there are, the greater the sound wave intensity and sound wave scattering intensity, and vice versa. Sound wave frequency is used to distinguish different types of aquatic organisms. Different types of aquatic organisms have different reflection characteristics of sound waves, and the corresponding sound wave frequencies are also different. Echo duration provides information on the size of aquatic organisms. Larger aquatic organisms will produce longer echo duration, and vice versa. Echo duration can be used as auxiliary data for inferring the abundance and species of aquatic organisms.

[0074] Methods for obtaining acoustic feature data include:

[0075] According to the sound wave data, the sound pressure value within time T is obtained, and the root mean square sound pressure is calculated. The expression of the root mean square sound pressure is: ; In the formula, is the root mean square sound pressure, is the sound pressure value at time t, To integrate the square of the sound pressure value from time 0 to time t, ; Calculate the sound wave intensity based on the root mean square sound pressure; The expression of sound wave intensity is: ; In the formula is the sound wave intensity, is the water density, is the propagation speed of sound waves in water; , ;

[0076] The sound wave data is converted from a time domain signal into a frequency signal by using a fast Fourier transform to obtain the sound wave frequency; the fast Fourier transform is an existing technology and will not be described in detail here;

[0077] Obtain the sound wave emission time and the sound wave reception time, subtract the sound wave emission time from the sound wave reception time, and obtain the echo duration signal; the sound wave emission time and the sound wave reception time are both obtained by the timer built into the sonar sensor;

[0078] Get the emission intensity, which is the intensity of the sound wave when the sonar sensor emits the sound wave; calculate the sound wave scattering intensity based on the emission intensity and the sound wave intensity; the expression of the sound wave scattering intensity is: ; In the formula, is the sound wave scattering intensity, is the emission intensity.

[0079] S3: Develop a restoration strategy based on water quality data, bottom quality data, and aquatic biological data.

[0080] like Figure 2 As shown, the steps to develop a proliferation restoration strategy include:

[0081] Step S301: construct m restoration sets, where m is an integer greater than 1, and the restoration sets include the required introduction quantity of each aquatic plant and benthic organism, and the m restoration sets are all different; the range of the required introduction quantity of each aquatic plant and benthic organism is preset by those skilled in the art according to actual conditions;

[0082] Step S302: Randomly select from m recovery sets Different recovery sets are marked as selection sets. Each selection set is taken as an individual to generate a population. ;

[0083] Step S303: mutate each individual in the population to generate a mutant individual; the expression of the mutant individual is: ; In the formula is the jth mutant individual, is the jth individual, , are two individuals randomly selected from the population, is the variation factor, , ;

[0084] Step S304: Use crossover operation to combine each individual in the population with the corresponding variant individual to generate a new individual; the method of generating a new individual is: preset crossover probability , , crossover probability It is pre-set by technicians in this field according to the actual situation; a The random number between the individual variables is the number of introductions required for each aquatic plant or benthic organism, and the variables correspond to the number of introductions one by one; the random number corresponding to each variable in the individual is compared with the crossover probability. If the random number is less than the crossover probability, the corresponding variable is replaced by the corresponding variable in the corresponding variant individual; if the random number is greater than or equal to the crossover probability, the corresponding variable is not replaced; a variable is randomly selected from all the variables in the individual, and the selected variable is replaced by the corresponding variable in the corresponding variant individual to ensure that at least one variable in the new individual comes from the variant individual;

[0085] Step S305: Calculate the recovery degree of each individual in the population and the new individual respectively, and update the individuals in the population; the updating method is: if the recovery degree of the individual is less than the recovery degree of the corresponding new individual, then update the individual to the corresponding new individual; if the recovery degree of the individual is greater than the recovery degree of the corresponding new individual, then do not update the corresponding individual; the calculation method of the recovery degree is: take the water quality data, bottom quality data and aquatic biological data as research data, take the research data and the recovery set corresponding to the individual as calculation data, input the calculation data into the trained recovery assessment model, evaluate the corresponding recovery degree, and the recovery degree is the recovery effect of the ecological health state of the water area; the training process of the recovery assessment model is consistent with the training process of the biological prediction model, and both are neural network models;

[0086] Step S306: Obtain the maximum recovery degree corresponding to the updated population, and subtract the maximum recovery degree corresponding to the updated population to obtain the recovery degree difference; preset a difference threshold, if the recovery degree difference is less than the difference threshold, generate a slow instruction, if the recovery degree difference is greater than or equal to the difference threshold, do not generate a slow instruction; preset a quantity threshold, count the number of slow instructions generated continuously, and mark it as the instruction quantity, if the instruction quantity is greater than or equal to the quantity threshold, enter step S307, if the instruction quantity is less than the quantity threshold, return to step S303;

[0087] Step S307: Obtain the recovery set corresponding to the maximum recovery degree.

[0088] S4: Conduct fish resource assessment based on water area images and obtain fish resource data.

[0089] Fish resource data include fish abundance and fish species; fish abundance refers to the number of each type of fish in the fish species; fish species include silver carp, bighead carp, carp, perch, etc.

[0090] Methods for obtaining fish stock data include:

[0091] The target detection methods (such as YOLO, SSD, HOG + SVM, etc.) are used to mark the fish in the water images with rectangular frames, and the rectangular frames correspond to the fish one-to-one. According to the marked rectangular frames, the water images are segmented (such as U-Net, FCN, DeepLab, etc.) to segment the water sub-images, and the water sub-images correspond to the rectangular frames one-to-one. The trained fish recognition model is used to identify the fish in the water sub-images and output the recognition results, which are the fish species. According to the identified fish species, the number of each fish species is counted to obtain the fish abundance.

[0092] The specific training process of the fish recognition model includes:

[0093] Collect multiple water sub-images in advance, annotate the fish in each water sub-image, and annotate them as fish species; convert different fish species into different digital annotations, for example, convert silver carp into 0.1 and bighead carp into 0.2; divide the annotated water sub-images into a training set and a test set, use 70% of the water sub-images as the training set and 30% of the water sub-images as the test set; use the training set to train the fish recognition model, and use the test set to test the fish recognition model; preset an error threshold, and when the mean of the prediction errors of all water sub-images in the test set is less than the error threshold, output the fish recognition model; wherein the calculation formula for the mean prediction error is: ,in is the prediction error, b is the number of the water area sub-image, is the predicted annotation corresponding to the b-th group of water sub-images, is the actual annotation corresponding to the water sub-image of group b, and U is the number of water sub-images in the test set; the error threshold is preset according to the accuracy required by the fish recognition model;

[0094] The above-mentioned fish recognition model is specifically a convolutional neural network model.

[0095] S5: Based on the water quality data, bottom quality data, aquatic organism data and fish resource data, select the objects for reproduction.

[0096] The objects of enhancement are the fish species that are released into the waters.

[0097] Preferred methods of proliferating a subject include:

[0098] Step S501: Establish fuzzy sets, divide each data in the preferred data into multiple fuzzy sets, the preferred data includes research data and fish resource data; for example, the fuzzy sets corresponding to organic matter content are low organic matter content, medium organic matter content, high organic matter content, etc., and the fuzzy sets corresponding to silver carp are small silver carp number, medium silver carp number, large silver carp number, etc.;

[0099] Step S502: convert each preferred data into the membership of each corresponding fuzzy set through fuzzification technology; fuzzification is the process of converting an exact value into the membership corresponding to a fuzzy set, and fuzzification technologies include triangular membership function, trapezoidal membership function, etc.; for example, if the value of organic matter content is low, it is inferred that the membership of low organic matter content is 0.8, the membership of medium organic matter content is 0.1, and the membership of high organic matter content is 0.1;

[0100] Step S503: define fuzzy rules, which are defined based on expert knowledge or literature; for example, if the dissolved oxygen concentration is high, the ammonia nitrogen concentration is low, the number of silver carp is large, and the number of grass carp is small, then it is inferred that the object of proliferation is grass carp; if the number of silver carp is medium, the organic matter content is low, and the number of bighead carp is small, then it is inferred that the object of proliferation is bighead carp;

[0101] Step S504: matching the fuzzified preferred data with the fuzzy rules, performing fuzzy reasoning, and obtaining a fuzzy reasoning result, which is the membership of the proliferation object; the fuzzy reasoning method is, for example, the Mamdani or Sugeno fuzzy reasoning method; the fuzzy reasoning result is, for example, the membership of the proliferation object being grass carp is 0.2, and the membership of the proliferation object being bighead carp is 0.8;

[0102] Step S505: Defuzzify the fuzzy reasoning result to obtain the defuzzified result, which is the proliferation object; defuzzification is to convert the fuzzy reasoning result into a specific recognition result, and the defuzzification operation is: according to the fuzzy reasoning result, respectively obtain the membership corresponding to each fish; preset the membership threshold, which is pre-set by technical personnel in this field according to actual conditions; compare the membership corresponding to each fish with the membership threshold, and take the fish with a membership greater than the membership threshold as the proliferation object.

[0103] S6: Determine the proliferation and release capacity based on the proliferation objects.

[0104] Methods for determining the stocking capacity include:

[0105] The objects of enhancement are divided according to their feeding habits to obtain their feeding types; the feeding types include filter-feeding fish, herbivorous fish, benthic animal-feeding fish and detritus-feeding fish; filter-feeding fish feed on plankton in the water, such as silver carp, bighead carp, etc.; herbivorous fish feed on aquatic plants, such as grass carp, black carp, etc.; benthic animal-feeding fish feed on benthic animals, such as catfish and yellow eel, etc.; detritus-feeding fish feed on organic debris (such as dead plants, animal remains, etc.) deposited in the water, such as carp and crucian carp, etc.; the enhancement and release capacity includes the filter-feeding fish enhancement capacity, herbivorous fish enhancement capacity, benthic animal-feeding fish enhancement capacity and detritus-feeding fish enhancement capacity.

[0106] Obtain the number of aquatic plants, plankton, and benthic organisms based on biological abundance;

[0107] The expression for the proliferation capacity of filter-feeding fish is: ; In the formula, is the proliferation capacity of filter-feeding fish, is the number of plankton, is the first biological coefficient, Water area volume: The water area volume is obtained by technical personnel in this field through measurement.

[0108] The expression of the proliferation capacity of herbivorous fish is: ; is the proliferation capacity of herbivorous fish, The number of aquatic plants.

[0109] The expression for the proliferation capacity of benthic feeding fish is: ; The capacity for the proliferation of benthic feeding fish is is the number of benthic organisms, is the second biological coefficient, is the water area; the water area is obtained by technical personnel in this field through measurement; the region where the water area is located is obtained, such as the East China region, the Northeast China region, the Yunnan-Guizhou region, etc., and the region where the water area is located is manually input and obtained by technical personnel in this field before fishery enhancement; the first biological coefficient and the second biological coefficient are obtained through a preset biological coefficient set, and the biological coefficient set includes the values ​​of the first biological coefficient and the second biological coefficient in different regions; the biological coefficient set is pre-set by technical personnel in this field by referring to relevant literature, for example, the first biological coefficient corresponding to the East China region is 80, and the corresponding second biological coefficient is 5.

[0110] The expression for the proliferation capacity of detritus-feeding fish is:

[0111] ;

[0112] In the formula, is the proliferation capacity of detritus-feeding fish, is the organic carbon content, is the proportion of silver carp, is the proportion of bighead carp; the organic carbon content is obtained through the above-mentioned water quality monitoring sensor; the method for obtaining the proportion of silver carp and bighead carp is: according to the fish resource data, the number of silver carp and bighead carp is obtained; the number of silver carp is added to the number of bighead carp to obtain the total number of silver carp and bighead carp; the number of silver carp is divided by the total number of silver carp and bighead carp to obtain the proportion of silver carp; the number of bighead carp is divided by the total number of silver carp and bighead carp to obtain the proportion of bighead carp.

[0113] This embodiment is based on a large water surface environment. By collecting and analyzing water quality, bottom quality, sound waves and water image data, the ecological health status of the water body and the fishery resources are comprehensively evaluated. Deep learning technology is used to accurately obtain the types and abundance of aquatic organisms, and scientific reproduction and restoration strategies are formulated in combination with water quality data and bottom quality data. Fuzzy reasoning and defuzzification methods are used to select suitable fish species for reproduction, and reasonable reproduction and release capacity is calculated based on their feeding habits and ecological needs to ensure the reasonable reproduction of fishery resources. This not only improves the accuracy and intelligence level of fishery resource management, but also can dynamically adjust strategies to cope with environmental changes, balance population relationships in the ecosystem, and effectively improve the sustainable development capacity of fishery resources, thereby achieving a dual improvement in ecological and economic benefits. Example 2

[0114] The present application also provides an electronic device. The electronic device may include one or more processors and one or more memories. The memories store computer-readable codes, and when the computer-readable codes are run by the one or more processors, the method for coordinated proliferation of fishery resources based on a large water surface environment as described above may be executed.

[0115] The method or system according to the implementation mode of the present application can also be implemented with the aid of the architecture of the electronic device shown in the present application. The electronic device may include a bus, one or more CPUs, ROM, RAM, a communication port connected to a network, input / output, a hard disk, etc. A storage device in the electronic device, such as a ROM or a hard disk, can store the method for collaborative proliferation of fishery resources based on a large water surface environment provided by the present application. Furthermore, the electronic device may also include a user interface. Of course, the architecture shown in the present application is only exemplary. When implementing different devices, one or more components in the electronic device shown in the present application may be omitted according to actual needs. Example 3

[0116] As shown, one embodiment of the present application discloses a computer-readable storage medium. Computer-readable instructions are stored on the computer-readable storage medium. When the computer-readable instructions are executed by the processor, the method for coordinated proliferation of fishery resources based on a large water surface environment according to the embodiment of the present application described with reference to the above figures can be executed. The storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0117] In addition, according to the embodiments of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, which stores machine-readable instructions, and the machine-readable instructions can be executed by a processor to execute instructions corresponding to the method steps provided in the present application, for example: a method for the coordinated proliferation of fishery resources based on a large water surface environment. When the computer program is executed by a central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0118] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art 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.

[0119] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for the coordinated proliferation of fishery resources based on a large water surface environment, characterized in that: include: S1: Collect water area characteristic data; water area characteristic data includes water quality data, bottom quality data, sound wave data and water area image; S2: Monitor aquatic organisms based on acoustic wave data and obtain aquatic organism data; S3: Combine water quality data, bottom quality data and aquatic biological data to develop a restoration strategy; Steps in developing a restoration strategy include: Step S301: construct m restoration sets, where m is an integer greater than 1, and the restoration sets include the required introduction quantity of each aquatic plant and benthic organism, and the m restoration sets are all different; Step S302: Randomly select from m recovery sets Different recovery sets are marked as selection sets. Each selection set is taken as an individual to generate a population. ; Step S303: mutate each individual in the population to generate a mutant individual; Step S304: using a crossover operation to combine each individual in the population with the corresponding variant individual to generate a new individual; Step S305: Calculate the recovery degree corresponding to each individual in the population and the new individual respectively, and update the individuals in the population; the recovery degree is calculated by taking water quality data, bottom quality data and aquatic biological data as research data, taking the research data and the recovery set corresponding to the individual as calculation data, inputting the calculation data into the trained recovery evaluation model, and evaluating the corresponding recovery degree, which is the recovery effect of the ecological health state of the water area; Step S306: Obtain the maximum recovery degree corresponding to the updated population, and subtract the maximum recovery degree corresponding to the updated population to obtain the recovery degree difference; preset a difference threshold, if the recovery degree difference is less than the difference threshold, generate a slow instruction, if the recovery degree difference is greater than or equal to the difference threshold, do not generate a slow instruction; preset a quantity threshold, count the number of slow instructions generated continuously, and mark it as the instruction quantity, if the instruction quantity is greater than or equal to the quantity threshold, enter step S307, if the instruction quantity is less than the quantity threshold, return to step S303; Step S307: Obtain the recovery set corresponding to the maximum recovery degree; S4: Evaluate fish resources based on water area images and obtain fish resource data; S5: combining water quality data, bottom quality data, aquatic organism data and fish resource data, selecting the proliferation object; the proliferation object is the type of fish to be released in the water area; The method of preferably proliferating an object comprises: Step S501: establishing a fuzzy set, dividing each data in the preferred data into multiple fuzzy sets, the preferred data including research data and fish resource data; Step S502: convert each preferred data into the membership degree of each corresponding fuzzy set through fuzzification technology; fuzzification is the process of converting an exact numerical value into the membership degree corresponding to a fuzzy set; Step S503: define fuzzy rules; Step S504: matching the fuzzified preferred data with the fuzzy rules, performing fuzzy reasoning, and obtaining the fuzzy reasoning result, which is the membership degree of the proliferation object; Step S505: performing a defuzzification operation on the fuzzy reasoning result to obtain a defuzzification result, which is the proliferation object; the defuzzification operation is: according to the fuzzy reasoning result, respectively obtaining the membership corresponding to each fish; presetting a membership threshold, respectively comparing the membership corresponding to each fish with the membership threshold, and taking the fish with a membership greater than the membership threshold as the proliferation object; S6: Determine the proliferation and release capacity based on the proliferation objects.

2. The method for coordinated proliferation of fishery resources based on a large water surface environment according to claim 1, characterized in that: The aquatic biological data include biological abundance and biological species; the biological species include aquatic plant species, plankton species and benthic species; the biological abundance is the number of each aquatic organism in the biological species; The method for obtaining aquatic organism data comprises: Process the acoustic wave data to obtain corresponding acoustic feature data; the acoustic feature data includes acoustic wave intensity, acoustic wave frequency, echo duration and acoustic wave scattering intensity; input the acoustic feature data into the trained biological prediction model to predict the corresponding set label, which is a digital label corresponding to the analysis set, and different analysis sets have different corresponding digital labels; obtain the corresponding analysis set according to the predicted set label, and use the analysis set as aquatic biological data; the analysis set is set by randomly combining different biological species and biological abundances, and one combination corresponds to one analysis set; The training process of the biological prediction model includes: A group of acoustic feature data is collected in advance, and corresponding set labels are set for the a group of acoustic feature data, where a is an integer greater than 1, and the acoustic feature data and the corresponding set labels are converted into a corresponding set of feature vectors; each set of feature vectors is used as the input of a biological prediction model, and the biological prediction model uses a set of predicted set labels corresponding to each group of acoustic feature data as output, and uses the actual set labels corresponding to each group of acoustic feature data as a prediction target, where the actual set labels are pre-set set labels corresponding to the acoustic feature data; minimizing the sum of prediction errors of all acoustic feature data is used as a training target; the biological prediction model is trained until the sum of prediction errors converges and the training is stopped; the biological prediction model is a deep neural network model.

3. The method for coordinated proliferation of fishery resources based on a large water surface environment according to claim 2, characterized in that: Methods for obtaining acoustic feature data include: According to the sound wave data, the sound pressure value within time T is obtained, and the root mean square sound pressure is calculated. The expression of the root mean square sound pressure is: ; In the formula, is the root mean square sound pressure, is the sound pressure value at time t, To integrate the square of the sound pressure value from time 0 to time t, ; Calculate the sound wave intensity based on the root mean square sound pressure; The expression of sound wave intensity is: ; In the formula is the sound wave intensity, is the water density, is the speed of sound waves in water; Fast Fourier transform is used to convert the acoustic wave data from a time domain signal into a frequency signal to obtain the acoustic wave frequency; Obtain the sound wave emission time and the sound wave reception time, subtract the sound wave emission time from the sound wave reception time, and obtain the echo duration signal; Get the emission intensity, which is the intensity of the sound wave when the sonar sensor emits the sound wave; calculate the sound wave scattering intensity based on the emission intensity and the sound wave intensity; the expression of the sound wave scattering intensity is: ; In the formula, is the sound wave scattering intensity, is the emission intensity.

4. The method for coordinated proliferation of fishery resources based on a large water surface environment according to claim 3, characterized in that: The expression of the variant individual is: ; In the formula is the jth mutant individual, is the jth individual, , are two individuals randomly selected from the population, is the variation factor, , .

5. The method for coordinated proliferation of fishery resources based on a large water surface environment according to claim 4, characterized in that: In step S304, the method for generating new individuals is: preset crossover probability , ; For each variable in an individual, a The random number between the individual and the variable in the individual is the number of introductions required for each aquatic plant or benthic organism, and the variable corresponds to the introduction number one by one; the random number corresponding to each variable in the individual is compared with the crossover probability. If the random number is less than the crossover probability, the corresponding variable is replaced by the corresponding variable in the corresponding variant individual; if the random number is greater than or equal to the crossover probability, the corresponding variable is not replaced; a variable is randomly selected from all the variables in the individual, and the selected variable is replaced by the corresponding variable in the corresponding variant individual.

6. The method for coordinated proliferation of fishery resources based on a large water surface environment according to claim 5, characterized in that: In step S305, the updating method is: if the recovery degree of the individual is less than the recovery degree of the corresponding new individual, the individual is updated to the corresponding new individual; if the recovery degree of the individual is greater than the recovery degree of the corresponding new individual, the corresponding individual is not updated; the training process of the recovery assessment model is consistent with the training process of the biological prediction model, and both are neural network models.

7. The method for coordinated proliferation of fishery resources based on a large water surface environment according to claim 6, characterized in that: The fish resource data includes fish abundance and fish species; the fish abundance is the number of each type of fish in the fish species; The method for obtaining fish resource data comprises: The target detection method is used to mark the fish in the water image with rectangular frames, and the rectangular frames correspond to the fish one by one; the water image is segmented according to the marked rectangular frames to segment the water sub-images, and the water sub-images correspond to the rectangular frames one by one; the trained fish recognition model is used to identify the fish in the water sub-images, and the recognition results are output, and the recognition results are the types of fish; according to the identified types of fish, the number of each type of fish is counted to obtain the abundance of fish; The fish identification model training process includes: Collect multiple water sub-images in advance, and annotate the fish in each water sub-image as fish species; convert different fish species into different digital annotations; divide the annotated water sub-images into a training set and a test set; use the training set to train a fish recognition model, and use the test set to test the fish recognition model; preset an error threshold, and when the mean of the prediction errors of all water sub-images in the test set is less than the error threshold, output a fish recognition model; the fish recognition model is a convolutional neural network model.

8. The method for coordinated proliferation of fishery resources based on a large water surface environment according to claim 7, characterized in that: Methods for determining the stocking capacity include: The objects of enhancement are divided according to their feeding habits to obtain feeding types; feeding types include filter-feeding fish, herbivorous fish, benthic fish and detritus-feeding fish; enhancement and release capacity includes filter-feeding fish enhancement capacity, herbivorous fish enhancement capacity, benthic fish enhancement capacity and detritus-feeding fish enhancement capacity; Obtain the number of aquatic plants, plankton, and benthic organisms based on biological abundance; The expression for the proliferation capacity of filter-feeding fish is: ; In the formula, is the proliferation capacity of filter-feeding fish, is the number of plankton, is the first biological coefficient, Water volume; The expression of the proliferation capacity of herbivorous fish is: ; is the proliferation capacity of herbivorous fish, The number of aquatic plants.

9. The method for coordinated proliferation of fishery resources based on a large water surface environment according to claim 8, characterized in that: The expression for the proliferation capacity of benthic feeding fish is: ; The capacity for the proliferation of benthic feeding fish is is the number of benthic organisms, is the second biological coefficient, is the water area; obtain the region where the water area is located, and obtain the first biological coefficient and the second biological coefficient through a preset biological coefficient set, wherein the biological coefficient set includes the values ​​of the first biological coefficient and the second biological coefficient in different regions; The expression for the proliferation capacity of detritus-feeding fish is: ； In the formula, is the proliferation capacity of detritus-feeding fish, is the organic carbon content, The proportion of silver carp is is the proportion of bighead carp; the method for obtaining the proportion of silver carp and bighead carp is: according to the fish resource data, obtain the number of silver carp and bighead carp; add the number of silver carp to the number of bighead carp to obtain the total number of silver carp and bighead carp; divide the number of silver carp by the total number of silver carp and bighead carp to obtain the proportion of silver carp; divide the number of bighead carp by the total number of silver carp and bighead carp to obtain the proportion of bighead carp.

Images