阅读说明：本技术 一种对虾草共生养殖方法 (Prawn grass symbiotic cultivation method ) 是由 熊建华 张彬 黎铭 韦信贤 童桂香 杨艳 林勇 于 2021-01-15 设计创作，主要内容包括：本发明公开了一种对虾草共生养殖方法,包括以下步骤：S1:搭建6X3X3的天然池塘,用生石灰化水后均匀泼洒消毒,10天后用氯硝柳胺340全塘消毒,再消毒5天后,设置进水阀门,选择茎长10cm以上的水草种植于浮排之上,持续培养3-4天后,放置对虾幼苗；S2:监控水的溶解氧含量、氨、氮含量,利用水下摄像头观察对虾、水草的数量以及对虾的脱壳情况、运动量和摄食量,采用强化学习算法根据水的溶解氧含量、氨、氮含量以及对虾、水草的数量添加抑制原生生物的药物或肥水培草产品和增氧剂的投放量以及光照程度。本发明利用深度学习模型和强化学习模型进行自动化的动态调整,具有宏观调控和实时调整,不需要人工进行控制,可以极大的提高对虾和水草的养殖成果。(The invention discloses a symbiotic cultivation method for prawns and weeds, which comprises the following steps: s1, building a natural pond of 6X3X3, dissolving water with quick lime, uniformly splashing and disinfecting, disinfecting with niclosamide 340 after 10 days, setting a water inlet valve after disinfecting for 5 days, selecting aquatic plants with stems longer than 10cm to plant on the floating raft, continuously culturing for 3-4 days, and placing prawn seedlings; s2, monitoring the dissolved oxygen content, ammonia and nitrogen content of water, observing the quantity of prawns and aquatic weeds and the shelling condition, exercise amount and food intake of the prawns by using an underwater camera, and adding a medicament for inhibiting protozoon or the adding amount of a water fertilizing and fertilizing culture product and an oxygenation agent and the illumination degree according to the dissolved oxygen content, the ammonia and nitrogen content of the water and the quantity of the prawns and the aquatic weeds by adopting a reinforcement learning algorithm. The invention utilizes the deep learning model and the reinforcement learning model to carry out automatic dynamic adjustment, has macroscopic regulation and real-time adjustment, does not need manual control, and can greatly improve the cultivation results of prawns and aquatic weeds.)

1. A symbiotic cultivation method for prawns and weeds is characterized by comprising the following steps:

s1, building a natural pond of 6X3X3, dissolving water with quick lime, uniformly splashing and disinfecting, disinfecting with niclosamide 340 after 10 days, setting a water inlet valve after disinfecting for 5 days, selecting aquatic plants with stems longer than 10cm to plant on the floating raft, continuously culturing for 3-4 days, and placing prawn seedlings;

s2, monitoring the dissolved oxygen content, ammonia content and nitrogen content of water, observing the quantity of prawns and aquatic weeds and the shelling condition, exercise amount and food intake of the prawns by using an underwater camera, adding a medicament for inhibiting protozoon or a water fertilizing and weed cultivating product and the putting amount and the illumination degree of an oxygenation agent according to the dissolved oxygen content, the ammonia content and the nitrogen content of the water and the quantity of the prawns and the aquatic weeds by using a reinforcement learning algorithm, and dynamically adjusting the feeding amount of the prawns according to the shelling condition, the exercise amount, the food intake and the quantity of the prawns by using an expert database and a deep learning model.

2. The symbiotic cultivation method for prawns and weeds as claimed in claim 1, wherein in step S2, the reinforcement learning algorithm takes the dissolved oxygen content and ammonia and nitrogen content of water, the quantity of prawns and the quantity of aquatic weeds as state data, takes the drug for inhibiting protists or the water-fertilizing and weed-cultivating product added and the illumination degree as strategies, takes the growth degree of aquatic weeds as a strategy evaluation value output strategy, and calculates the adding amount of the oxygen-increasing agent according to the strategies and the state data.

3. The symbiotic prawn and grass cultivation method as claimed in claim 1, wherein in step S2, shelling condition, exercise amount and food intake data of the prawns are extracted by using the features, clustering is performed by using a clustering algorithm in an expert database according to the data, the growth categories of the prawns are judged, and feeding amount of the prawns is adjusted according to the growth categories.

4. The symbiotic prawn and grass cultivation method according to claim 3, wherein the clustering algorithm adopts a kmeans clustering algorithm, and the shelling condition, the exercise amount and the food consumption data of the prawns are x^(m)Putting the shelling condition, the exercise amount and the food intake data of the shrimps into an expert database to obtain a training sample set { x⁽¹⁾、x⁽²⁾……x^(m)And setting k classes and the mass center of each class as mu₁，μ₂，...，μ_kRepeating the following process until convergence

For each sample i, calculate the class to which it should belong

For each class j, the centroid of the class is recalculated

}

x⁽ⁱ⁾Belong to a training sample set { x⁽¹⁾、x⁽²⁾……x^(m)}，c⁽ⁱ⁾Representing the class of sample i that is closest to the k classes, c⁽ⁱ⁾Is one of 1 to k. Centroid mu_jRepresenting our expectation for the center point of the samples belonging to the same class.

From the above calculation, C can be obtained^(m)I.e. the class with sample m closest to the k classes.

5. The symbiotic prawn and grass cultivation method as claimed in claim 1, wherein in step S2, the bait is live worm larvae and organic pellet feed.

Technical Field

The invention relates to the technical field of prawn symbiotic culture, in particular to a prawn grass symbiotic culture method.

Background

The carbon dioxide that protist and shrimp produced can increase the survival rate of pasture and water, simultaneously, the pasture and water can provide the nourishment for protist and shrimp, form ecological symbiosis, need restrain the protist with the medicine or utilize fertile aquatic products to cultivate grass usually in order to ensure the oxygen suppliment of shrimp and growth, if the quantity control is not enough, can lead to pasture and water quantity too much, cause the large tracts of land death of pasture and water, aggravate the growth of protist, consequently, the medicine that need restrain the protist and the input volume of fertile aquatic products are regulated and control, form ecological balance, be favorable to promoting the growth of shrimp, but present regulation and control is regulated and control for the breeder according to the experience usually, can't accomplish real-time regulation and control, very big influence the breed efficiency of shrimp and pasture and water.

Disclosure of Invention

In order to at least solve or partially solve the problems, the method for symbiotic cultivation of prawns and aquatic weeds is provided, a deep learning model and a reinforcement learning model are used for carrying out automatic dynamic adjustment, macroscopic regulation and real-time adjustment are achieved, manual control is not needed, and cultivation results of prawns and aquatic weeds can be greatly improved.

In order to achieve the purpose, the invention provides the following technical scheme:

the invention relates to a symbiotic cultivation method of prawns and weeds, which comprises the following steps:

s1, building a natural pond of 6X3X3, dissolving water with quick lime, uniformly splashing and disinfecting, disinfecting with niclosamide 340 after 10 days, setting a water inlet valve after disinfecting for 5 days, selecting aquatic plants with stems longer than 10cm to plant on the floating raft, continuously culturing for 3-4 days, and placing prawn seedlings;

s2, monitoring the dissolved oxygen content, ammonia content and nitrogen content of water, observing the quantity of prawns and aquatic weeds and the shelling condition, exercise amount and food intake of the prawns by using an underwater camera, adding a medicament for inhibiting protozoon or a water fertilizing and weed cultivating product and the putting amount and the illumination degree of an oxygenation agent according to the dissolved oxygen content, the ammonia content and the nitrogen content of the water and the quantity of the prawns and the aquatic weeds by using a reinforcement learning algorithm, and dynamically adjusting the feeding amount of the prawns according to the shelling condition, the exercise amount, the food intake and the quantity of the prawns by using an expert database and a deep learning model.

In a preferred embodiment of the present invention, in step S2, the reinforcement learning algorithm uses the dissolved oxygen content of water, the ammonia and nitrogen content, the prawn quantity, and the quantity of aquatic weeds as status data, uses the addition of a drug inhibiting protists, a hydroponic product, and the light level as a policy, outputs the policy using the growth level of aquatic weeds as a policy evaluation value, and calculates the amount of the oxygen-increasing agent to be added based on the policy and the status data.

As a preferred technical solution of the present invention, in step S2, the shelling condition, exercise amount and food intake data of the prawns are extracted by using feature extraction, clustering is performed by using a clustering algorithm in an expert database according to the data, and what growth category the prawns are in is determined, and the feeding amount of the prawns is adjusted according to the growth category.

As a preferred technical scheme of the invention, the clustering algorithm adopts a kmeans clustering algorithm, and the shelling condition, the exercise amount and the food intake data of the shrimps are x^(m)Putting the shelling condition, the exercise amount and the food intake data of the shrimps into an expert database to obtain a training sample set { x⁽¹⁾、x⁽²⁾……x^(m)And setting k classes and the mass center of each class as mu₁，μ₂，...，μ_kRepeating the following process until convergence

For each sample i, calculate the class to which it should belong

For each class j, the centroid of the class is recalculated

}

x⁽ⁱ⁾Belong to a training sample set { x⁽¹⁾、x⁽²⁾……x^(m)}，c⁽ⁱ⁾Representing the class of sample i that is closest to the k classes, c⁽ⁱ⁾Is one of 1 to k. Centroid mu_jRepresenting our expectation for the center point of the samples belonging to the same class.

From the above calculation, C can be obtained^(m)I.e. the class with sample m closest to the k classes.

In a preferred embodiment of the present invention, in step S2, the bait is live worm larvae and organic pellet feed.

Compared with the prior art, the invention has the following beneficial effects:

the invention utilizes the deep learning model and the reinforcement learning model to carry out automatic dynamic adjustment, realizes the macroscopic regulation and control of the whole ecological environment by adjusting the growth of the aquatic weeds, can carry out real-time adjustment according to the current state, has higher efficiency compared with the manual ecological balance, and can greatly improve the cultivation results of prawns and aquatic weeds.

Detailed Description

The following description of the preferred embodiments of the present invention is provided for the purpose of illustration and description, and is in no way intended to limit the invention.

In addition, if a detailed description of the known art is not necessary to show the features of the present invention, it is omitted.

Example 1

The invention provides a symbiotic cultivation method for prawns and weeds, which comprises the following steps:

s1, building a natural pond of 6X3X3, dissolving water with quick lime, uniformly splashing and disinfecting, disinfecting with niclosamide 340 after 10 days, setting a water inlet valve after disinfecting for 5 days, selecting aquatic plants with stems longer than 10cm to plant on the floating raft, continuously culturing for 3-4 days, and placing prawn seedlings;

s2, monitoring the dissolved oxygen content, ammonia content and nitrogen content of water, observing the quantity of prawns and aquatic weeds and the shelling condition, exercise amount and food intake of the prawns by using an underwater camera, adding a medicament for inhibiting protozoon or a water fertilizing and weed cultivating product and the putting amount and the illumination degree of an oxygenation agent according to the dissolved oxygen content, the ammonia content and the nitrogen content of the water and the quantity of the prawns and the aquatic weeds by using a reinforcement learning algorithm, and dynamically adjusting the feeding amount of the prawns according to the shelling condition, the exercise amount, the food intake and the quantity of the prawns by using an expert database and a deep learning model.

In step S2, the reinforcement learning algorithm uses the dissolved oxygen content and the ammonia and nitrogen content of water, the prawn quantity and the aquatic weed quantity as state data, uses the drug for inhibiting protist, the water culture product for fattening and the illumination degree as strategies, uses the growth degree of aquatic weed as a strategy evaluation value output strategy, and calculates the amount of the oxygenation agent according to the strategies and the state data.

In step S2, the shelling condition, the amount of exercise, and the amount of food intake of the prawns are extracted by using the feature extraction, clustering is performed by using a clustering algorithm in an expert database according to the data, the growth category of the prawns is judged, and the feeding amount of the prawns is adjusted according to the growth category.

The clustering algorithm adopts a kmeans clustering algorithm, and the shelling condition, the exercise amount and the food intake data of the shrimps are x^(m)Putting the shelling condition, the exercise amount and the food intake data of the shrimps into an expert database to obtain a training sample set { x⁽¹⁾、x⁽²⁾……x^(m)And setting k classes and the mass center of each class as mu₁，μ₂，...，μ_kRepeating the following process until convergence

For each sample i, calculate the class to which it should belong

For each class j, the centroid of the class is recalculated

}

x⁽ⁱ⁾Belong to a training sample set { x⁽¹⁾、x⁽²⁾……x^(m)}，c⁽ⁱ⁾Representing the class of sample i that is closest to the k classes, c⁽ⁱ⁾Is one of 1 to k. Centroid mu_jRepresenting our expectation for the center point of the samples belonging to the same class.

From the above calculation, C can be obtained^(m)I.e. the class with sample m closest to the k classes.

In the step S2, the bait is live worm larvae and organic pellet feed.

Specifically, the shelling condition, the exercise amount and the food intake of the prawns are subjected to data extraction by using an underwater camera in a characteristic extraction mode, then the clustering algorithm and the expert database are used for carrying out accurate control, the pollution of feed to a water body can be reduced while the feeding of the prawns can be guaranteed, meanwhile, the current state data is judged by using a reinforcement learning algorithm, medicines for inhibiting protozoon and a strategy for fertilizing the water culture product putting amount are output and added, the ecological dynamic balance is guaranteed, and the effect of stable growth of the yield of the prawns is achieved.

The invention utilizes the deep learning model and the reinforcement learning model to carry out automatic dynamic adjustment, realizes the macroscopic regulation and control of the whole ecological environment by adjusting the growth of the aquatic weeds, can carry out real-time adjustment according to the current state, has higher efficiency compared with the manual ecological balance, and can greatly improve the cultivation results of prawns and aquatic weeds.

Finally, it should be noted that: although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that changes may be made in the embodiments and/or equivalents thereof without departing from the spirit and scope of the invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.