Sound spectrum intelligent feeding method for penaeus vannamei boone

Abstract

The invention discloses a sound spectrum intelligent feeding method for penaeus vannamei boone, which generally comprises the following steps: a signal acquisition step: acquiring signal items including penaeus vannamei boone ingestion sound signals and environmental parameters; a signal analysis step: performing feature extraction and analysis on the prawn ingestion sound spectrum to obtain real-time ingestion state information of a prawn group, and grading healthy hunger states and non-healthy hunger states of the prawns in each time period; and a bait casting control step: controlling the bait casting amount of a bait casting machine according to the healthy and hungry states of the prawns. The state of the prawns is judged in real time by collecting and analyzing the sound signals generated when the prawns eat food, so that the automatic feeding strategy is adjusted according to the monitoring condition, the cultured prawns can eat food according to needs, and the feeding strategy can be adaptively adjusted according to the monitoring result.

Description

Technical Field

[0001] The invention relates to the technical field of aquaculture, in particular to a method for intelligently feeding white shrimp using acoustic spectrum. Background Art

[0002] Being able to monitor the living conditions of shrimps in underwater culture in real time is the key to aquaculture informatization. Due to the poor optical visibility underwater in aquaculture environments, and the limitation of underwater light sources at night, effective observation and monitoring cannot be achieved only by video. Currently, there are methods that combine the use of hydrophones for acoustic monitoring, but the current methods are easily affected by uncontrollable factors such as space and weather.

[0003] The Chinese patent with authorization announcement number CN115136912B provides a method for predicting the disease of farmed shrimps by combining water quality parameters and behavioral vocalization. Nine independent signal acquisition units are set up, and each independent unit is equipped with a microprocessor, a water quality detector and a hydrophone to establish a water quality and acoustic parameter monitoring channel and analyze the information; the signal acquisition unit records water quality parameters through the water quality detector, and collects behavioral vocalization signals of shrimp aquatic organisms when the water quality does not change through the hydrophone, and the signal analysis unit judges the water quality changes at nine locations; the signal analysis unit includes a dual-channel water quality acoustic comprehensive algorithm model and a multi-channel water quality acoustic comprehensive algorithm model for analysis and judgment.

[0004] Although the existing technology can well utilize the hydroacoustic principle to monitor the living status (including feeding status) of shrimps, it does not consider how to improve the feeding strategy based on the monitoring results. Summary of the invention

[0005] The purpose of the present invention is to provide a method for intelligent feeding of white shrimp using acoustic spectrum, which can adaptively adjust the feeding strategy according to monitoring results.

[0006] The object of the present invention is achieved in this way: a method for intelligent feeding of white shrimp by sound spectrum, comprising the following steps:

[0007] A signal collection step, wherein the collected signal items include shrimp feeding sound signals and environmental parameters;

[0008] The signal analysis step extracts and analyzes the characteristics of the shrimp feeding sound spectrum to obtain the real-time feeding status information of the shrimp group and classify the hunger status of the shrimp in healthy and unhealthy states at different time periods;

[0009] The feeding step is controlled, and the feeding amount of the feeding machine is controlled according to the hunger state information of the shrimp in the healthy and unhealthy states at different time periods obtained in the signal analysis step.

[0010] Furthermore, in the signal analysis step, several indoor experiments are first carried out to simulate the conditions of shrimp farming, collect and process sound signals, establish a shrimp feeding sound signal database and indoor behavior samples, and then train and extract features to establish a corresponding deep learning model; when it is officially used, the real-time collected feeding sound signals are input into the established deep learning model to classify the shrimp feeding sound spectrum.

[0011] Furthermore, in the signal analysis step, the hunger state of the shrimp is classified into a healthy hunger state, a healthy semi-hunger state, a healthy satiety state, an unhealthy hunger state, an unhealthy semi-hunger state, and an unhealthy satiety state.

[0012] Further, in the step of controlling the feeding:

[0013] For the situation that the shrimp is in a healthy state: when the shrimp is in a healthy state of satiation, the switch of the feeder is in the off state; when the shrimp is in a healthy state of hunger, the switch of the feeder is in the on state, and the amount of feed fed by the feeder is relatively large; when the shrimp is in a healthy state of semi-starvation, the switch of the feeder is in the on state, and the amount of feed fed by the feeder is relatively small compared to the healthy state of hunger;

[0014] For the situation where the shrimp is in an unhealthy state: when the shrimp is in an unhealthy state of satiation, the switch of the feeding machine is in the off state; when the shrimp is in an unhealthy state of hunger, the switch of the feeding machine is in the on state, and the amount of feed fed by the feeding machine is smaller than that in the healthy state of hunger; when the shrimp is in an unhealthy state of semi-hunger, the switch of the feeding machine is in the on state, and the amount of feed fed by the feeding machine is less than that in the unhealthy state of hunger.

[0015] Furthermore, in the signal collection step, the optimal sampling time and sampling interval are selected to perform timed and fixed-length sampling of the feeding sound signal of the white shrimp.

[0016] Furthermore, in the signal collection step, the environmental parameters include mechanical noise, water flow noise, wind and rain noise and other environmental noises.

[0017] Furthermore, relevant data on the feeding amount of the feeding machine, feeding sound signals, and information on the healthy and unhealthy hunger status of the shrimp are all displayed on the computer screen.

[0018] The beneficial effects of the present invention are:

[0019] 1. By collecting and analyzing the sound signals generated by shrimps when feeding, the health status of shrimps can be judged in real time, so as to adjust the automatic feeding strategy according to the monitoring situation, so that the farmed shrimps can be fed on demand, and the feeding strategy can be adaptively adjusted according to the monitoring results;

[0020] 2. The healthy and unhealthy hunger states of the shrimps can be classified, thereby driving the feeding machine to determine the amount of feeding according to the healthy and unhealthy hunger states of the shrimps. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a schematic diagram of the overall steps of the present invention.

[0022] Figure 2 It is a schematic diagram for judging the feeding sound signals of shrimp in different states.

[0023] Figure 3 This is a schematic diagram of feeding according to the different states of shrimp. DETAILED DESCRIPTION

[0024] The following will be combined with the attached embodiment of the present invention Figure 1-3 , the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. 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.

[0025] refer to Figure 1 As shown in the figure, a sound spectrum intelligent feeding method for white shrimp is proposed. The hydrophone in the breeding pond is connected to the collector and the preamplifier through a data cable, and then the collector is connected to the computer, and the software of the collector in the computer is opened; the hydrophone and the camera are placed in the glass tank, and then the hydrophone is turned on to collect sound signals, and the camera is used to collect image signals for feeding. The sound signal of the shrimp eating feed is obtained by combining sound and image, and then the shrimp's demand for feed is calculated through a specific algorithm, so that automatic feeding can be performed as needed later.

[0026] according to Figure 1 It can be seen that the method generally includes three links. The first link is the extraction of breeding environment characteristics, that is, the collection of sound signals. The second link is the recognition and analysis of sound signal behavior characteristics. The third link is the feeding decision control, thereby controlling the amount of feed fed by the feeding machine.

[0027] Further expansion is given based on the above content.

[0028] In the first stage, taking into account the noise interference factors caused by the living environment of shrimps during the sound signal collection process, this embodiment not only collects the sound signals of the shrimp group (including healthy and unhealthy states), but also collects the natural noise and other interference noises in the pond, and uses natural noise and other interference noises as environmental parameters; among which, natural noise includes mechanical noise, water flow noise, and wind and rain noise (not limited to the types listed in this embodiment).

[0029] Targeting underwater whiteleg shrimp groups, hydrophones and other environmental sensors are used to accurately collect shrimp feeding sound signals and related environmental information under different time periods and environmental factors, and record the status of the shrimp; shrimp feeding sound signals under multiple environmental parameters are collected to build a sample library.

[0030] refer to Figure 2 As shown in the figure, in view of the lack of samples, several indoor experiments were conducted to simulate the situation of shrimp farming, collect and process sound signals, establish a shrimp feeding sound signal database and indoor behavior samples, and then train and extract features to establish a corresponding deep learning model; when it is officially used, the real-time collected feeding sound signals are input into the established deep learning model to classify the shrimp feeding sound spectrum quickly and accurately, improve the efficiency of shrimp feeding sound signal collection, and more accurately judge whether to put shrimp bait. By measuring the sound of shrimp feeding as a feedback signal, the feeding time and amount are controlled.

[0031] In the second stage, the algorithm for extracting the frequency spectrum features of shrimp feeding sounds and intelligently identifying the behavior patterns is studied. The characteristics of the shrimp feeding sound spectrum are extracted, and the behavior patterns are intelligently identified and analyzed to obtain the characteristic changes of the shrimp feeding sound signal under different environmental noises, and the healthy and unhealthy hunger status of the shrimp at different times is effectively monitored. In this step, it is necessary to clearly distinguish between the shrimp feeding sound signal and the environmental parameters, and denoise the signal to accurately extract the shrimp feeding sound signal.

[0032] comprehensive Figure 2-3 As shown in the figure, a deep neural network is used as a classifier for the shrimp feeding sound spectrum. The extracted shrimp feeding sound spectrum and its various features are first trained to obtain the features decomposed in the deep neural network. According to the various knowledge and methods used therein, the actual shrimp feeding sound signal is input into the deep neural network, so as to classify the shrimp feeding sound spectrum based on the healthy state and the unhealthy state. The healthy state mainly includes the hungry state ( Figure 3 The middle mark is S1); semi-starved state ( Figure 3 The middle mark is S2); satiety state ( Figure 3 The middle mark is S3); In addition, the unhealthy state mainly includes the hungry state ( Figure 3 The middle mark is F1); semi-starved state ( Figure 3 The middle mark is F2); satiety state ( Figure 3 The winning number is F3).

[0033] In these six states, the feeding sound signals generated by shrimps are obviously different. These state information is sent to the decision-making system for feeding decisions, that is, the hunger state of shrimps can be classified (classification), so as to drive the feeding machine to decide the amount of feeding according to the hunger state of shrimps.

[0034] In the third stage, the trained deep discrimination model of feeding sound signals and behavior association of white shrimp is embedded in the control system to achieve the timing and quantitative control of shrimp feed delivery. The experiment selects the optimal sampling time and sampling interval, and samples the feeding sound signals of white shrimp at a fixed time and length to achieve accurate classification of the characteristics of shrimp feeding sound signals. The changes in the characteristics of shrimp feeding sound signals during the feeding period are recorded, and the hunger state under healthy and unhealthy conditions is refined and classified to achieve real-time feedback of feed delivery results and accurate quantitative feed delivery.

[0035] like Figure 3 As shown, through analysis, it was found that when in a healthy state, the shrimp in a full state ( Figure 3 When the bait throwing machine is in the off state, the bait throwing machine is in the hungry state ( Figure 3 When the bait throwing machine is in the on state, the bait throwing machine throws a relatively large amount of bait; when the bait is in the semi-starved state ( Figure 3 When the winning bid is S2), the bait feeding amount of the bait feeding machine is relatively small.

[0036] When in an unhealthy condition, the shrimp in a well-fed state ( Figure 3 When the bait throwing machine is in the off state, the bait throwing machine is in the hungry state ( Figure 3 When the bait throwing machine is in the on state, the bait throwing machine's bait throwing amount is relatively small, that is, the bait throwing machine's bait throwing amount is smaller than that of the healthy hunger state; when the bait throwing machine is in the semi-starved state ( Figure 3 When the winning bid is F2), the bait throwing machine throws even less bait.

[0037] As the shrimps continue to eat, the system makes real-time judgments. If the shrimps are found to be full, the switch of the feeder is turned off. Later, as long as the shrimps are found to be hungry, the automatic feeding operation will continue.

[0038] It is worth mentioning that the above-mentioned signal collection, analysis, and baiting decision-making processes are carried out in real time.

[0039] Generally speaking, when casting bait, the process is roughly divided into the following:

[0040] The remote PC sends a signal;

[0041] The PLC device receives the signal;

[0042] Check whether there is enough feed in the feed box: if there is not enough feed, the indicator light will light up to remind the operator to add feed in time;

[0043] If there is enough feed, the process of determining the feeding time and feeding amount begins;

[0044] According to the results of the previous classification judgment (health status S mainly includes hunger status ( Figure 3 The middle mark is S1); semi-starved state ( Figure 3 The middle mark is S2); satiety state ( Figure 3 The middle mark is S3); the unhealthy state F mainly includes the hungry state ( Figure 3 The middle mark is F1); semi-starved state ( Figure 3 The middle mark is F2); satiety state ( Figure 3 The middle mark is F3), and the corresponding feeding strategy is selectively output. In this step, the feeding strategy is matched with the specific state of the shrimp;

[0045] At the same time, the bait feeding process is monitored in real time and the monitoring signal is fed back to the remote PC.

[0046] like Figure 3 As shown, the relevant data of the feeding amount of the feeding machine, the feeding sound signal and the rating information of the health and hunger status of the shrimp will be displayed in real time on the professional software of the computer, so that the operator can know the automatic feeding status and the real-time living status of the shrimp in real time.

[0047] The advantages of this embodiment are: low cost, little environmental impact, wide application range and long working cycle. By collecting and analyzing the sound signals generated by the shrimp when feeding, the shrimp's food intake is measured. By using this technology to analyze and evaluate the basic changes in shrimp feeding, such as shrimp size, food characteristics, feed preference and food consumption rate, the technical solution is combined with a feeding machine to adjust the automatic feeding strategy according to the monitoring situation, so that the farmed shrimp can be fed on demand.

[0048] In the description of the present invention, it should be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "front", "back" and the like indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as a limitation on the present invention; in the present invention, it should also be noted that the terms "installation" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, an integrally formed connection, a mechanical connection, or an indirect connection through an intermediate connecting component, and the specific meaning of the terms in this utility model can be understood according to the specific circumstances.

[0049] It is obvious to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential features of the present invention. Therefore, the embodiments should be regarded as exemplary and non-limiting from any point of view, and the scope of the present invention is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention, and any reference numerals in the claims should not be regarded as limiting the claims involved.

[0050] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. A method for intelligent feeding of white shrimp using acoustic spectrum, characterized in that: The steps include: A signal collection step, wherein the collected signal items include shrimp feeding sound signals and environmental parameters; The signal analysis step extracts and analyzes the characteristics of the shrimp feeding sound spectrum to obtain the real-time feeding status information of the shrimp group and classify the hunger status of the shrimp in healthy and unhealthy states at different time periods; The feeding step is controlled, and the feeding amount of the feeding machine is controlled according to the hunger state information of the shrimp in the healthy and unhealthy states at different time periods obtained in the signal analysis step.

2. The method for feeding white shrimp with sound spectrum according to claim 1, characterized in that: In the signal analysis step, several indoor experiments are first carried out to simulate the conditions of shrimp farming, collect and process sound signals, establish a shrimp feeding sound signal database and indoor behavior samples, and then train and extract features to establish a corresponding deep learning model. When it is officially used, the real-time collected feeding sound signals are input into the established deep learning model to classify the shrimp feeding sound spectrum.

3. The method for intelligent feeding of Penaeus vannamei using acoustic spectrum according to claim 2, characterized in that: In the signal analysis step, the hunger state of the shrimps is classified into a healthy hunger state, a healthy semi-hunger state, a healthy satiety state, an unhealthy hunger state, an unhealthy semi-hunger state, and an unhealthy satiety state.

4. The method for intelligent feeding of Penaeus vannamei using acoustic spectrum according to claim 3, characterized in that: In the controlled feeding step: For the situation that the shrimp is in a healthy state: when the shrimp is in a healthy state of satiation, the switch of the feeder is in the off state; when the shrimp is in a healthy state of hunger, the switch of the feeder is in the on state, and the amount of feed fed by the feeder is relatively large; when the shrimp is in a healthy state of semi-starvation, the switch of the feeder is in the on state, and the amount of feed fed by the feeder is relatively small compared to the healthy state of hunger; For the situation where the shrimp is in an unhealthy state: when the shrimp is in an unhealthy state of satiation, the switch of the feeding machine is in the off state; when the shrimp is in an unhealthy state of hunger, the switch of the feeding machine is in the on state, and the amount of feed fed by the feeding machine is smaller than that in the healthy state of hunger; when the shrimp is in an unhealthy state of semi-hunger, the switch of the feeding machine is in the on state, and the amount of feed fed by the feeding machine is less than that in the unhealthy state of hunger.

5. The method for intelligent feeding of Penaeus vannamei using acoustic spectrum according to claim 2, characterized in that: In the signal collection step, the optimal sampling time and sampling interval are selected to perform timed and fixed-length sampling of the feeding sound signals of the white shrimp.

6. A method for feeding white shrimp with sound spectrum intelligence according to any one of claims 1 to 5, characterized in that: In the signal collection step, the environmental parameters include mechanical noise, water flow noise, wind and rain noise and other environmental noises.

7. The method for intelligent feeding of Penaeus vannamei using acoustic spectrum according to claim 4, characterized in that: The relevant data of the feeding amount of the feeding machine, the feeding sound signal and the information on the healthy and unhealthy hunger status of the shrimp are all displayed on the computer screen.

Images