CN115336549B - Intelligent fish culture feeding system and method - Google Patents
Intelligent fish culture feeding system and method Download PDFInfo
- Publication number
- CN115336549B CN115336549B CN202211057986.7A CN202211057986A CN115336549B CN 115336549 B CN115336549 B CN 115336549B CN 202211057986 A CN202211057986 A CN 202211057986A CN 115336549 B CN115336549 B CN 115336549B
- Authority
- CN
- China
- Prior art keywords
- fish
- feeding
- information
- sample set
- feed
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01K—ANIMAL HUSBANDRY; CARE OF BIRDS, FISHES, INSECTS; FISHING; REARING OR BREEDING ANIMALS, NOT OTHERWISE PROVIDED FOR; NEW BREEDS OF ANIMALS
- A01K61/00—Culture of aquatic animals
- A01K61/80—Feeding devices
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A40/00—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production
- Y02A40/80—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production in fisheries management
- Y02A40/81—Aquaculture, e.g. of fish
Abstract
The invention discloses an intelligent fish culture feeding system and method, wherein the feeding system comprises a discharging bin, the lower end of the discharging bin is provided with an opening, the upper end of the discharging bin is connected with a collecting pipe, and a first valve is arranged at the joint of the collecting pipe and the discharging bin; the upper end of the collecting pipe is connected with three conveying pipes for conveying feed in juvenile fish stage, feed in sub-adult fish stage and feed in adult fish stage respectively; the upper ends of the three conveying pipes are connected with conveying pipes, a second valve is arranged between the conveying pipes, and a feeding mechanism is arranged in the conveying pipes. The method comprises steps S1-S15. According to the invention, parameters such as environmental temperature, water quality PH value data information and the like are acquired through the system, the feeding history data is acquired through the monitoring platform, a multi-element secondary discrimination model is selected for the data in the controller, and the built intelligent feeding system for aquaculture can reduce bait waste, reduce cost, protect water ecological environment and reduce fish disease risk, so that the intelligent feeding system has great significance in promoting aquaculture industry.
Description
Technical Field
The invention relates to the technical field of aquaculture, in particular to an intelligent fish culture feeding system and method.
Background
The bait cost in grass carp culture accounts for a large proportion of the total culture cost, and an effective feeding strategy is an important way for reducing the production cost of aquaculture, so that the nutrition requirement of fish shoal growth is met, excessive feeding is avoided, and the pollution to the environment is reduced. The intelligent and accurate control of the feeding mode is a key for improving the feeding efficiency of aquaculture. The aquaculture in China starts later, and the automation degree in the current stage is relatively low, especially in the aspect of bait feeding. However, in recent decades, related researches on related theories, technologies, equipment, facilities and the like of bait feeding by scientific researchers in China never stop, and at present, although feeding strategies mainly quantitatively feed according to manual experience, the accumulation of deviation of each feeding affects the economic benefit of fishing in the whole period, certain research results are obtained, such as the industrial automatic feeding system is developed by utilizing a track transmission technology and a sensor technology, and meanwhile, a precise bait feeding system is developed by utilizing a machine vision technology and a fuzzy control technology. Therefore, the scientific feeding strategy can effectively reduce the cultivation cost and improve the cultivation benefit.
At present, the field of aquaculture feeding is mainly limited to traditional artificial experience quantitative feeding, the intelligent level of aquaculture feeding equipment is lower, the quantitative feeding of artificial experience shows that the bait cost in aquaculture is about 40% of the total cost of aquaculture, and therefore the fishery production value and profit are often influenced due to the accumulation of deviation of feeding amount each time. Therefore, how to make the cultivation process automatic, intelligent and accurate is the key to improve the feeding efficiency of aquaculture.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides an intelligent fish culture feeding system and method capable of automatically collecting data and calculating feeding amount.
In order to achieve the aim of the invention, the invention adopts the following technical scheme:
the intelligent fish culture feeding system comprises a discharging bin, wherein the lower end of the discharging bin is provided with an opening, the upper end of the discharging bin is connected with a collecting pipe, and a first valve is arranged at the joint of the collecting pipe and the discharging bin; the upper end of the collecting pipe is connected with three conveying pipes for conveying feed in juvenile fish stage, feed in sub adult fish stage and feed in adult fish stage respectively, and the collecting pipe is connected with the three conveying pipes in a four-fork structure; the upper ends of three conveyer pipes all are connected with the conveying pipe, and three conveying pipes are connected with the feed box that stores juvenile fish stage fodder, sub-adult fish stage fodder and adult fish stage fodder respectively, are provided with the second valve between conveying pipe and the conveyer pipe, are provided with feeding mechanism in the conveying pipe, are provided with PH sensor and temperature detection sensor in the blowing storehouse, and PH sensor and temperature detection sensor all are connected with the controller electricity.
Further, the first valve comprises a first circular baffle plate, the first circular baffle plate is fixed on a first rotating shaft, a gravity sensor is arranged between the first rotating shaft and the first circular baffle plate, and the first circular baffle plate is rotatably arranged in the collecting pipe; the first rotating shaft penetrates through the side wall of the collecting pipe and is in transmission connection with the first motor, the first motor is fixed on the collecting pipe through the support, and the gravity sensor and the first motor are electrically connected with the controller.
Further, the second valve comprises a second circular baffle plate, the second circular baffle plate is fixed on the second rotating shaft, and the second circular baffle plate is rotatably arranged in the conveying pipe; the second rotating shaft penetrates through the side wall of the conveying pipe and is in transmission connection with the second motor, the second motor is fixed on the conveying pipe through the support, and the second motor is electrically connected with the controller.
Further, the feeding mechanism comprises a screw rod, the screw rod is arranged in the feeding pipe, and the feeding pipe is coaxial with the screw rod; the upper end of the screw rod is in transmission connection with a third motor, the third motor is fixed at the upper end of the feeding pipe through a bracket, and the third motor is electrically connected with the controller.
Further, the upper end of the feed pipe is provided with a connecting ring which is connected with the feed box.
Further, the feeding pipe is connected with the conveying pipe through threads.
The feeding method of the intelligent fish culture feeding system comprises the following steps:
s1: data information of the fish ponds is input, N fish ponds are randomly selected as sample sets, and a fish pond total data information matrix X is input N ；
Wherein x is 11 、x 21 、···、x N1 Is the temperature information of water in N fish ponds, x 12 、x 22 、···、x N2 For the dissolved oxygen content, x in N fish pond sample sets 13 、x 23 、···、x N3 For fish weight information in N fish pond sample sets, x 14 、x 24 、···、x N4 The fish tail number information in the N fish pond sample sets;
s2: randomly selecting M fish ponds from N fish ponds as a training sample set X M The remaining N-M fish ponds are used as a test sample set X N-M ，N＞M；
Wherein x is 11 、x 21 、...、x M1 For training sample set X M Temperature information of water in M fish ponds, x 12 、x 22 、...、x M2 For training sample set X M Dissolved oxygen content in M fish ponds, x 13 、x 23 、...、x M3 For training sample set X M Fish weight information, x, in M fish ponds 14 、x 24 、...、x M4 For training sample set X M The number information of the fish tails in the M fish ponds;
x (M+1)1 、x (M+2)1 、...、x (N-M)1 for testing sample set X N-M Temperature information of water in N-M fish ponds; x is x (M+1)2 、x (M+2)2 、...、x (N-M)2 For testing sample set X N-M The information of the content of dissolved oxygen in N-M fish ponds; x is x (M+1)3 、x (M+2)3 、...、x (N-M)3 For testing sample set X N-M Fish weight information in N-M fish ponds; x is x (M+1)4 、x (M+2)4 、...、x (N-M)4 For testing sample set X N-M The number information of the fish tails in the N-M fish ponds;
s3: construction of feeding vector information y N Selecting feeding amount information of M fish ponds from N fish ponds as training vector information y M Feeding amount information of N-M fish ponds is used as vector information y of a test sample set N-M
y N ＝[y 1 ，y 2 ，...，y N ]
y M ＝[y 1 ，y 2 ，...，y M ]
y N-M ＝[y M+1 ，y M+2 ，...，y N-M ]
Wherein y is 1 ，y 2 ，...，y N Feeding amount information, y, in N fish ponds 1 ，y 2 ，...，y M Feeding amount information in M fish ponds; y is M+1 ，y M+2 ，...，y N-M Feeding amount information in N-M fish ponds;
s4: constructing a prediction function, and establishing a predicted feeding quantity f i Is a multiple quadratic model of (a):
defining the prediction function coefficient vector a= [ a ] 0 ，a 1 ，a 2 ，a 3 ，a 4 ，a 5 ，a 6 ，a 7 ，a 8 ]
Extracting unknowns about X 1 、X 2 ，X 3 And X 4 Is a vector of information:
X 1 ＝[x 11 x 21 … x M1 ] T
X 2 ＝[x 12 x 22 … x M2 ] T
X 3 ＝[x 13 x 23 … x M3 ] T
X 4 ＝[x 14 x 24 … x M4 ] T
s5: parameter X 1 、X 2 ，X 3 ，X 4 Respectively substituting information vectors related to the predicted feeding quantity f i In the multi-element quadratic model of (2), the least square method is used for solving optimization, the final fitting parameters are calculated, and the loss function is minimized:
since the loss function is minimum, the loss function is derived, X is X 1 、X 2 ，X 3 And X 4 An information matrix formed by information vectors, i is an ith fish pond in the M fish ponds; y is i In M fish pondsFeeding amount of the ith fish pond; f (f) i The method comprises the steps of predicting feeding quantity of an ith fish pond in M fish ponds; a is a prediction function coefficient vector;
solving to obtain coefficient matrix A' = [ a ] 0 ′，a 1 ′，a 2 ′，a 3 ′，a 4 ′，a 5 ′，a 6 ′，a 7 ′，a 8 ′]，X T Is the transposition of the information matrix X;
s6: outputting fitting parameters to obtain feeding quantity f i Is a relational model of (a);
f i ＝a′ 0 +a′ 1 X 1 2 +a′ 2 X 1 +a′ 3 X 2 2 +a′ 4 X 2 +a′ 5 X 3 2 +a′ 6 X 3 +a′ 7 X 4 2 +a′ 8 X 4
s7: the obtained feeding quantity f i Application of relational model to test sample set X N-M The test sample set X is output N-M Corresponding predicted feeding amount;
f N-M ＝X N-M ·A′
wherein X is N-M For testing the matrix information of the sample set, A' is the vector coefficient of the coefficient matrix, f N-M For testing sample set X N-M The predicted feeding amount information in (a);
s8: calculating the relation f N-M Root mean square error RMSE, mean absolute percentage error MAPE and a decision coefficient R 2 ：
Wherein N-M is the test sample set X N-M Number of fish ponds, y N-M For testing sample set X N-M The actual feeding amount of the feed;
for testing sample set X N-M An average value of actual feeding amounts;
s9: the root mean square error RMSE, the mean absolute percentage error MAPE and the determination coefficient R 2 Compare to corresponding thresholds:
if all meet
RMSE≤RMSE Threshold value 、MAPE≤MAPE Threshold value Then the fitted feeding quantity f is judged i The relation model of (3) is accurate, and the step S13 is executed;
otherwise, executing step S10;
s10: comparing the predicted feed to the minimum and maximum of the actual feeds in the N fish ponds:
if the predicted feeding amount is in the range between the minimum value and the maximum value, judging that the fitted relation model of the feeding amount fi is accurate, and executing the step S13;
otherwise, step S11 is performed
S11: returning to the step S6, and re-fitting the feeding quantity f i Outputting the re-fitted feeding quantity f by the coefficients in the relation model of (2) i And optimizing fitting parameters by using a standard deviation model, wherein the standard deviation model is as follows:
wherein y is i Pre-fitting for re-fittingThe output value of the feeding amount information is measured,
for training sample set X M An average value of actual feeding amounts;
s12: the output Avg value is compared with the Avg Threshold value Comparing, if Avg is less than or equal to Avg Threshold value Outputting the re-fitted feeding quantity f i Is entered into step S13;
otherwise, returning to the step S11, continuing to re-fit the feeding amount f i Coefficients in a relational model of (a);
s13: according to the average fish weight in the pre-feeding fish pond, the PH value in the fish pond is collected by the PH sensor, the water temperature value in the fish pond, the average fish weight and the fish tail number in the fish pond are collected by the temperature detection sensor, and the fitted feeding amount f is input i Obtaining the feeding amount of the time in the relation model of (2);
s14: opening a second valve for correspondingly controlling feed in the juvenile stage, the sub-adult stage or the adult stage, detecting the weight of the feed falling into the collecting pipe by a gravity sensor, and closing the second valve when the feeding amount is reached;
s15: and then opening the first valve, discharging all the feed in the collecting pipe into the discharging bin, and discharging the feed into the fish pond through the discharging bin.
The beneficial effects of the invention are as follows: according to the invention, parameters such as environmental temperature, water quality PH value data information and the like are acquired through the system, feeding history data is acquired through the monitoring platform, a multi-element secondary discrimination model is selected for the data in the controller, experimental verification is carried out on the data, the feeding quantity of the fish shoals is predicted, and an artificial intelligent support is provided for the intelligent feeding device by the model output result. The intelligent feeding system can be interconnected through a 5G transmission technology, and the feeding system can monitor water quality, environment, feeding quantity and the like in real time. The intelligent feeding system for aquaculture can reduce bait waste, reduce cost, protect water ecological environment and reduce fish disease risk, and has important significance for pushing aquaculture industry.
Predicting the feeding of the fish shoal by using a multi-element secondary discrimination model, matching the data with the feeding quantity according to the obtained environmental temperature, the water quality PH value, the fish tail number and the average fish weight as data input quantity, and establishing a feeding prediction model to provide theoretical and technical support for intelligent feeding, so that the feeding quantity can be accurately controlled, the feeding waste or shortage is reduced, and the problems of environmental pollution or insufficient nutrition of the fish shoals are caused; different fishing periods are distinguished, a single independent feeding pipeline is used for connecting three bait storage boxes, and different baits are provided for fishes in different fishing periods.
Drawings
FIG. 1 is a block diagram of an intelligent fish farming feeding system.
Fig. 2 is a structural view of the first valve.
Fig. 3 is a structural view of the feeding mechanism.
The automatic feeding device comprises a feeding pipe, a connecting ring, a feeding pipe, a feeding hopper, a first rotary shaft, a gravity sensor, a first round baffle, a third motor, a screw rod and a screw rod, wherein the feeding pipe is arranged at the bottom of the feeding pipe, the feeding pipe is connected with the feeding pipe, the feeding valve is arranged at the bottom of the feeding pipe, the feeding pipe is connected with the feeding pipe, the feeding hopper is connected with the feeding pipe, the feeding pipe is connected with the feeding pipe through the feeding pipe, the feeding pipe is arranged at the feeding hopper, the feeding pipe is connected with the first valve, the first.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and all the inventions which make use of the inventive concept are protected by the spirit and scope of the present invention as defined and defined in the appended claims to those skilled in the art.
As shown in fig. 1 to 3, the intelligent fish culture feeding system of the scheme comprises a discharging bin 5, wherein the lower end of the discharging bin 5 is opened, the upper end of the discharging bin 5 is connected with a collecting pipe, and a first valve 6 is arranged at the joint of the collecting pipe and the discharging bin 5; the upper end of the collecting pipe is connected with three conveying pipes 4 for conveying feed in juvenile fish stage, feed in sub adult fish stage and feed in adult fish stage respectively, and the collecting pipe is connected with the three conveying pipes 4 in a four-fork structure; the upper end of three conveyer pipes 4 all is connected with conveying pipe 2, and three conveying pipe 2 are connected with the feed box that stores juvenile stage fodder, sub-adult stage fodder and adult stage fodder respectively, are provided with second valve 3 between conveying pipe 2 and the conveyer pipe 4, are provided with feeding mechanism in the conveying pipe 2, are provided with PH sensor and temperature-detecting sensor in the blowing storehouse 5, and PH sensor and temperature-detecting sensor all are connected with the controller electricity.
In this embodiment, the first valve 6 includes a first circular baffle 9, the first circular baffle 9 is fixed on the first rotating shaft 7, a gravity sensor 8 is disposed between the first rotating shaft 7 and the first circular baffle 9, and the first circular baffle 9 is rotatably disposed in the collecting pipe; the first rotating shaft 7 penetrates through the side wall of the collecting pipe and is in transmission connection with the first motor, the first motor is fixed on the collecting pipe through a support, and the gravity sensor 8 and the first motor are electrically connected with the controller.
In this embodiment, the second valve 3 includes a second circular baffle plate, the second circular baffle plate is fixed on the second rotating shaft, and the second circular baffle plate is rotatably disposed in the conveying pipe 4; the second rotating shaft penetrates through the side wall of the conveying pipe 4 and is in transmission connection with a second motor, the second motor is fixed on the conveying pipe 4 through a support, and the second motor is electrically connected with the controller.
When the first valve 6 and the second valve 3 are opened and closed, the feeding amount can be accurately controlled in each section only by rotating the first motor or the second motor.
In this embodiment, the feeding mechanism includes a screw rod 11, the screw rod 11 is disposed in the feeding pipe 2, and the feeding pipe 2 is coaxial with the screw rod 11; the upper end of the screw rod 11 is in transmission connection with a third motor 10, the third motor 10 is fixed at the upper end of the feeding pipe 2 through a bracket, and the third motor 10 is electrically connected with a controller. The screw rod 11 can promote the feed delivery, avoids the condition of pipe clamping.
In the embodiment, the upper end of a feed pipe 2 is provided with a connecting ring 1 which is connected with a feed box; the feeding pipe 2 is in threaded connection with the conveying pipe 4, so that the assembly and the connection are convenient.
The feeding method of the intelligent fish culture feeding system comprises the following steps of:
s1: data information of the fish ponds is input, N fish ponds are randomly selected as sample sets, and a fish pond total data information matrix X is input N ；
Wherein x is 11 、x 21 、···、x N1 Is the temperature information of water in N fish ponds, x 12 、x 22 、···、x N2 For the dissolved oxygen content, x in N fish pond sample sets 13 、x 23 、···、x N3 For fish weight information in N fish pond sample sets, x 14 、x 24 、···、x N4 The fish tail number information in the N fish pond sample sets;
s2: randomly selecting M fish ponds from N fish ponds as a training sample set X M The remaining N-M fish ponds are used as a test sample set X N-M ，N＞M；
Wherein x is 11 、x 21 、...、x M1 For training sample set X M Temperature information of water in M fish ponds, x 12 、x 22 、...、x M2 For training sample set X M Dissolved oxygen content in M fish ponds, x 13 、x 23 、...、x M3 For training sample set X M Fish weight information, x, in M fish ponds 14 、x 24 、...、x M4 For training sample set X M The number information of the fish tails in the M fish ponds;
x (M+1)1 、x (M+2)1 、...、x (N-M)1 for testing sample set X N-M Temperature information of water in N-M fish ponds; x is x (M+1)2 、x (M+2)2 、...、x (N-M)2 For testing sample set X N-M The information of the content of dissolved oxygen in N-M fish ponds; x is x (M+1)3 、x (M+2)3 、...、x (N-M)3 For testing sample set X N-M Fish weight information in N-M fish ponds; x is x (M+1)4 、x (M+2)4 、...、x (N-M)4 For testing sample set X N-M The number information of the fish tails in the N-M fish ponds;
s3: construction of feeding vector information y N Selecting feeding amount information of M fish ponds from N fish ponds as training vector information y M Feeding amount information of N-M fish ponds is used as vector information y of a test sample set N-M
y N ＝[y 1 ，y 2 ，...，y N ]
y M ＝[y 1 ，y 2 ，...，y M ]
y N-M ＝[y M+1 ，y M+2 ，...，y N-M ]
Wherein y is 1 ，y 2 ，...，y N Feeding amount information, y, in N fish ponds 1 ，y 2 ，...，y M Feeding amount information in M fish ponds; y is M+1 ，y M+2 ，...，y N-M Feeding amount information in N-M fish ponds;
s4: constructing a prediction function, and establishing a predicted feeding quantity f i Is a multiple quadratic model of (a):
defining the prediction function coefficient vector a= [ a ] 0 ，a 1 ，a 2 ，a 3 ，a 4 ，a 5 ，a 6 ，a 7 ，a 8 ]
Extracting unknowns about X 1 、X 2 ，X 3 And X 4 Is a vector of information:
X 1 ＝[x 11 x 21 … x M1 ] T
X 2 ＝[x 12 x 22 … x M2 ] T
X 3 ＝[x 13 x 23 … x M3 ] T
X 4 ＝[x 14 x 24 … x M4 ] T
s5: parameter X 1 、X 2 ，X 3 ，X 4 Respectively substituting information vectors related to the predicted feeding quantity f i In the multi-element quadratic model of (2), the least square method is used for solving optimization, the final fitting parameters are calculated, and the loss function is minimized:
since the loss function is minimum, the loss function is derived, X is X 1 、X 2 ，X 3 And X 4 An information matrix formed by information vectors, i is an ith fish pond in the M fish ponds; y is i Feeding the ith fish pond in the M fish ponds; f (f) i The method comprises the steps of predicting feeding quantity of an ith fish pond in M fish ponds; a is a prediction function coefficient vector;
solving to obtain coefficient matrix A' = [ a ] 0 ′，a 1 ′，a 2 ′，a 3 ′，a 4 ′，a 5 ′，a 6 ′，a 7 ′，a 8 ′]，X T Is the transposition of the information matrix X;
s6: outputting fitting parameters to obtain feeding quantity f i Is a relational model of (a);
f i ＝a′ 0 +a′ 1 X 1 2 +a′ 2 X 1 +a′ 3 X 2 2 +a′ 4 X 2 +a′ 5 X 3 2 +a′ 6 X 3 +a′ 7 X 4 2 +a′ 8 X 4
s7: the obtained feeding quantity f i Application of relational model to test sample set X N-M The test sample set X is output N-M Corresponding predicted feeding amount;
f N-M ＝X N-M ·A′
wherein X is N-M For testing the matrix information of the sample set, A' is the vector coefficient of the coefficient matrix, f N-M For testing sample set X N-M The predicted feeding amount information in (a);
s8: calculating the relation f N-M Root mean square error RMSE, mean absolute percentage error MAPE and a decision coefficient R 2 ：
Wherein N-M is the test sample set X N-M Number of fish ponds, y N-M For testing sample set X N-M The actual feeding amount of the feed;
for testing sample set X N-M An average value of actual feeding amounts;
s9: the root mean square error RMSE, the mean absolute percentage error MAPE and the determination coefficient R 2 Compare to corresponding thresholds:
if all meet
RMSE≤RMSE Threshold value 、MAPE≤MAPE Threshold value Then the fitted feeding quantity f is judged i The relation model of (3) is accurate, and the step S13 is executed;
otherwise, executing step S10;
s10: comparing the predicted feed to the minimum and maximum of the actual feeds in the N fish ponds:
if the predicted feeding amount is within the range between the minimum value and the maximum value, the fitted feeding amount f is determined i The relation model of (3) is accurate, and the step S13 is executed;
otherwise, step S11 is performed
S11: returning to the step S6, and re-fitting the feeding quantity f i Outputting the re-fitted feeding quantity f by the coefficients in the relation model of (2) i And optimizing fitting parameters by using a standard deviation model, wherein the standard deviation model is as follows:
wherein y is i Outputting a value for the re-fitted predicted feeding amount information,
for training sample set X M An average value of actual feeding amounts;
s12: the output Avg value is compared with the Avg Threshold value Comparing, if Avg is less than or equal to Avg Threshold value Outputting the re-fitted feeding quantity f i Is entered into step S13;
otherwise, returning to the step S11, continuing to re-fit the feeding amount f i Coefficients in a relational model of (a);
s13: according to the average fish weight in the pre-feeding fish pond, the PH value in the fish pond is collected by the PH sensor, the water temperature value in the fish pond, the average fish weight and the fish tail number in the fish pond are collected by the temperature detection sensor, and the fitted feeding amount f is input i Obtaining the feeding amount of the time in the relation model of (2);
s14: opening a second valve 3 correspondingly controlling feed in the juvenile stage, the sub-adult stage or the adult stage, detecting the weight of the feed falling into the collecting pipe by a gravity sensor 8, and closing the second valve 2 when the feeding amount is reached;
s15: then the first valve 6 is opened, all the feed in the collecting pipe is discharged into the discharging bin 5, and the feed is discharged into the fish pond through the discharging bin 5.
According to the invention, parameters such as environmental temperature, water quality PH value data information and the like are acquired through the system, feeding history data is acquired through the monitoring platform, a multi-element secondary discrimination model is selected for the data in the controller, experimental verification is carried out on the data, the feeding quantity of the fish shoals is predicted, and an artificial intelligent support is provided for the intelligent feeding device by the model output result. The intelligent feeding system can be interconnected through a 5G transmission technology, and the feeding system can monitor water quality, environment, feeding quantity and the like in real time. The intelligent feeding system for aquaculture can reduce bait waste, reduce cost, protect water ecological environment and reduce fish disease risk, and has important significance for pushing aquaculture industry.
Predicting the feeding of the fish shoal by using a multi-element secondary discrimination model, matching the data with the feeding quantity according to the obtained environmental temperature, the water quality PH value, the fish tail number and the average fish weight as data input quantity, and establishing a feeding prediction model to provide theoretical and technical support for intelligent feeding, so that the feeding quantity can be accurately controlled, the feeding waste or shortage is reduced, and the problems of environmental pollution or insufficient nutrition of the fish shoals are caused; different fishing periods are distinguished, a single independent feeding pipeline is used for connecting three bait storage boxes, and different baits are provided for fishes in different fishing periods.
Claims (1)
1. The intelligent fish culture feeding system comprises a discharging bin, wherein the lower end of the discharging bin is provided with an opening, the upper end of the discharging bin is connected with a collecting pipe, and a first valve is arranged at the joint of the collecting pipe and the discharging bin; the upper end of the collecting pipe is connected with three conveying pipes for conveying feed in the juvenile fish stage, feed in the sub-adult fish stage and feed in the adult fish stage respectively, and the collecting pipe is connected with the three conveying pipes in a four-fork structure; the upper ends of the three conveying pipes are connected with conveying pipes, the three conveying pipes are respectively connected with feed boxes for storing feed in juvenile fish period, feed in sub-adult fish period and feed in adult fish period, a second valve is arranged between the conveying pipes, a feeding mechanism is arranged in the conveying pipes, a PH sensor and a temperature detection sensor are arranged in the discharging bin, and the PH sensor and the temperature detection sensor are electrically connected with a controller;
the first valve comprises a first circular baffle plate, the first circular baffle plate is fixed on a first rotating shaft, a gravity sensor is arranged between the first rotating shaft and the first circular baffle plate, and the first circular baffle plate is rotatably arranged in the collecting pipe; the first rotating shaft penetrates through the side wall of the collecting pipe and is in transmission connection with the first motor, the first motor is fixed on the collecting pipe through a bracket, and the gravity sensor and the first motor are electrically connected with the controller;
the second valve comprises a second circular baffle plate, the second circular baffle plate is fixed on a second rotating shaft, and the second circular baffle plate is rotatably arranged in the conveying pipe; the second rotating shaft penetrates through the side wall of the conveying pipe and is in transmission connection with a second motor, the second motor is fixed on the conveying pipe through a bracket, and the second motor is electrically connected with the controller;
the feeding mechanism comprises a screw rod, the screw rod is arranged in the feeding pipe, and the feeding pipe is coaxial with the screw rod; the upper end of the screw rod is in transmission connection with a third motor, the third motor is fixed at the upper end of the feeding pipe through a bracket, and the third motor is electrically connected with the controller;
the upper end of the feed pipe is provided with a connecting ring which is connected with the feed box;
the feeding pipe is in threaded connection with the conveying pipe;
the method is characterized by comprising the following steps of:
s1: data information of the fish ponds is input, N fish ponds are randomly selected as sample sets, and a fish pond total data information matrix X is input N ；
Wherein x is 11 、x 21 、···、x N1 Is the temperature information of water in N fish ponds, x 12 、x 22 、···、x N2 For the dissolved oxygen content, x in N fish pond sample sets 13 、x 23 、···、x N3 For fish weight information in N fish pond sample sets, x 14 、x 24 、···、x N4 The fish tail number information in the N fish pond sample sets;
s2: randomly selecting M fish ponds from N fish ponds as a training sample set X M The remaining N-M fish ponds are used as a test sample set X N-M ，N＞M；
Wherein x is 11 、x 21 、…、x M1 For training sample set X M Temperature information of water in M fish ponds, x 12 、x 22 、…、x M2 For training sample set X M Dissolved oxygen content in M fish ponds, x 13 、x 23 、…、x M3 For training sample set X M Fish weight information, x, in M fish ponds 14 、x 24 、…、x M4 For training sample set X M The number information of the fish tails in the M fish ponds;
x (M+1)1 、x (M+2)1 、…、x (N-M)1 for testing sample set X N-M Temperature information of water in N-M fish ponds; x is x (M+1)2 、x (M+2)2 、…、x (N-M)2 For testing sample set X N-M The information of the content of dissolved oxygen in N-M fish ponds; x is x (M+1)3 、x (M+2)3 、…、x (N-M)3 For testing sample set X N-M Fish weight information in N-M fish ponds; x is x (M+1)4 、x (M+2)4 、…、x (N-M)4 For testing sample set X N-M The number information of the fish tails in the N-M fish ponds;
s3: construction of feeding vector information y N Selecting feeding amount information of M fish ponds from N fish ponds as training vector information y M Feeding amount information of N-M fish ponds is used as vector information y of a test sample set N-M
y N ＝[y 1 ,y 2 ,…,y N ]
y M ＝[y 1 ,y 2 ,…,y M ]
y N-M ＝[y M+1 ,y M+2 ,…,y N-M ]
Wherein y is 1 ,y 2 ,…,y N Feeding amount information, y, in N fish ponds 1 ,y 2 ,…,y M Feeding amount information in M fish ponds; y is M+1 ,y M+2 ,…,y N-M Feeding amount information in N-M fish ponds;
s4: constructing a prediction function, and establishing a predicted feeding quantity f i Is a multiple quadratic model of (a):
defining the prediction function coefficient vector a= [ a ] 0 ,a 1 ,a 2 ,a 3 ,a 4 ,a 5 ,a 6 ,a 7 ,a 8 ]
Extracting unknowns about X 1 、X 2 ，X 3 And X 4 Is a vector of information:
X 1 ＝[x 11 x 21 … x M1 ] T
X 2 ＝[x 12 x 22 … x M2 ] T
X 3 ＝[x 13 x 23 … x M3 ] T
X 4 ＝[x 14 x 24 … x M4 ] T
s5: parameter X 1 、X 2 ，X 3 ，X 4 Respectively substituting information vectors related to the predicted feeding quantity f i In the multi-element quadratic model of (2), the least square method is used for solving optimization, the final fitting parameters are calculated, and the loss function is minimized:
since the loss function is minimum, the loss function is derived, X is X 1 、X 2 ，X 3 And X 4 An information matrix formed by information vectors, i is an ith fish pond in the M fish ponds; y is i Feeding the ith fish pond in the M fish ponds; f (f) i The method comprises the steps of predicting feeding quantity of an ith fish pond in M fish ponds; a is a prediction function coefficient vector;
solving to obtain coefficient matrix A' = [ a ] 0 ′, 1 ′, 2 ′, 3 ′, 4 ′, 5 ′, 6 ′, 7 ′, 8 ′]，X T Is the transposition of the information matrix X;
s6: outputting fitting parameters to obtain feeding quantity f i Is a relational model of (a);
f i ＝a′ 0 +a 1 ′X 1 2 +a′ 2 X 1 +a 3 ′X 2 2 +a′ 4 X 2 +a 5 ′X 3 2 +a′ 6 X 3 +a′ 7 X 4 2 +a 8 ′X 4
s7: the obtained feeding quantity f i Application of relational model to test sample set X N-M The test sample set X is output N-M Corresponding predicted feeding amount;
f N-M ＝ N-M ·A'
wherein X is N-M For testing the matrix information of the sample set, A' is the vector coefficient of the coefficient matrix, f N-M For testing sample set X N-M The predicted feeding amount information in (a);
s8: calculating the relation f N-M Root mean square error RMSE, mean absolute percentage error MAPE and a decision coefficient R 2 ：
Wherein N-M is the test sample set X N-M Number of fish ponds, y N-M For testing sample set X N-M The actual feeding amount of the feed;
for testing sample set X N-M An average value of actual feeding amounts;
s9: the root mean square error RMSE, the mean absolute percentage error MAPE and the determination coefficient R 2 Compare to corresponding thresholds:
if all meet
RMSE≤RMSE Threshold value 、MAPE≤MAPE Threshold value Then the fitted feeding quantity f is judged i The relation model of (3) is accurate, and the step S13 is executed;
otherwise, executing step S10;
s10: comparing the predicted feed to the minimum and maximum of the actual feeds in the N fish ponds:
if the predicted feeding amount is within the range between the minimum value and the maximum value, the fitted feeding amount f is determined i The relation model of (3) is accurate, and the step S13 is executed;
otherwise, step S11 is performed
S11: returning to the step S6, and re-fitting the feeding quantity f i Outputting the re-fitted feeding quantity f by the coefficients in the relation model of (2) i And optimizing fitting parameters by using a standard deviation model, wherein the standard deviation model is as follows:
wherein y is i Outputting a value for the re-fitted predicted feeding amount information,
for training sample set X M An average value of actual feeding amounts;
s12: the output Avg value is compared with the Avg Threshold value Comparing, if Avg is less than or equal to Avg Threshold value Outputting the re-fitted feeding quantity f i Is entered into step S13;
otherwise, returning to the step S11, continuing to re-fit the feeding amount f i Coefficients in a relational model of (a);
s13: according to the average fish weight in the pre-feeding fish pond, the PH value in the fish pond is collected by the PH sensor, the water temperature value in the fish pond, the average fish weight and the fish tail number in the fish pond are collected by the temperature detection sensor, and the fitted feeding amount f is input i Obtaining the feeding amount of the time in the relation model of (2);
s14: opening a second valve for correspondingly controlling feed in the juvenile stage, the sub-adult stage or the adult stage, detecting the weight of the feed falling into the collecting pipe by a gravity sensor, and closing the second valve when the feeding amount is reached;
s15: and then opening the first valve, discharging all the feed in the collecting pipe into the discharging bin, and discharging the feed into the fish pond through the discharging bin.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211057986.7A CN115336549B (en) | 2022-08-30 | 2022-08-30 | Intelligent fish culture feeding system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211057986.7A CN115336549B (en) | 2022-08-30 | 2022-08-30 | Intelligent fish culture feeding system and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115336549A CN115336549A (en) | 2022-11-15 |
| CN115336549B true CN115336549B (en) | 2023-06-20 |
Family
ID=83955962
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211057986.7A Active CN115336549B (en) | 2022-08-30 | 2022-08-30 | Intelligent fish culture feeding system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115336549B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103070126A (en) * | 2013-01-17 | 2013-05-01 | 中国水产科学研究院渔业机械仪器研究所 | Pond culture sound control feeding method and device based on same |
| CN103548721A (en) * | 2013-10-17 | 2014-02-05 | 宁波大学 | Method for cultivating large-size offspring seeds of pharaoh cuttlefish |
| CN104782552A (en) * | 2015-04-20 | 2015-07-22 | 中国水产科学研究院渔业机械仪器研究所 | Centralized automatic feeding and oxygenation system for aquaculture |
| CN109625218A (en) * | 2018-12-29 | 2019-04-16 | 浙江大学 | A kind of no worker monitor underwater culture robot system based on solar recharging |
| CN111557265A (en) * | 2020-05-18 | 2020-08-21 | 中国水产科学研究院渔业机械仪器研究所 | Automatic feeding device of large-scale waters cultivation management platform |
| CN113724255A (en) * | 2021-10-25 | 2021-11-30 | 鲁东大学 | Counting method for abalones in seedling raising period |
| CN114503946A (en) * | 2022-01-24 | 2022-05-17 | 海南大学 | Fishing ground feeding system and method based on high frame rate dynamic frame difference accurate identification |
Family Cites Families (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB436192A (en) * | 1935-04-13 | 1935-10-07 | Frederic Ernest Webb | Improvements in or relating to apparatus for feeding paints, enamels and the like tospray guns |
| JP4081092B2 (en) * | 2005-01-21 | 2008-04-23 | 学校法人近畿大学 | Method for preventing abnormal behavior of tuna by controlling illuminance |
| CN103210868A (en) * | 2013-02-05 | 2013-07-24 | 大连海洋大学 | Digital bait feeding machine based on 3G (3rd Generation) network |
| CN103609839B (en) * | 2013-11-26 | 2016-05-11 | 中国水产科学研究院渔业机械仪器研究所 | A kind of Feed Manufacturing for marine seining cultivation and the system and method for throwing something and feeding |
| CN104335948B (en) * | 2014-10-17 | 2017-01-11 | 浙江海洋学院 | Intensive cultivation device with intelligent feed supply |
| CN206586189U (en) * | 2017-02-07 | 2017-10-27 | 泉州惠安长圣生物科技有限公司 | One kind is for fishery cultivating full-automatic charging machine |
| CN107656900B (en) * | 2017-09-01 | 2021-09-10 | 武汉钢铁有限公司 | Method for determining compensation angles of different stocklines in material distribution process of blast furnace |
| CN108812479A (en) * | 2018-05-21 | 2018-11-16 | 江苏科技大学 | Multifunctional marine pasture assists unmanned boat |
| DK201970599A1 (en) * | 2019-09-09 | 2021-05-17 | Apple Inc | Techniques for managing display usage |
| CN113841649A (en) * | 2021-09-24 | 2021-12-28 | 淮阴工学院 | Pond feeding system |
-
2022
- 2022-08-30 CN CN202211057986.7A patent/CN115336549B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103070126A (en) * | 2013-01-17 | 2013-05-01 | 中国水产科学研究院渔业机械仪器研究所 | Pond culture sound control feeding method and device based on same |
| CN103548721A (en) * | 2013-10-17 | 2014-02-05 | 宁波大学 | Method for cultivating large-size offspring seeds of pharaoh cuttlefish |
| CN104782552A (en) * | 2015-04-20 | 2015-07-22 | 中国水产科学研究院渔业机械仪器研究所 | Centralized automatic feeding and oxygenation system for aquaculture |
| CN109625218A (en) * | 2018-12-29 | 2019-04-16 | 浙江大学 | A kind of no worker monitor underwater culture robot system based on solar recharging |
| CN111557265A (en) * | 2020-05-18 | 2020-08-21 | 中国水产科学研究院渔业机械仪器研究所 | Automatic feeding device of large-scale waters cultivation management platform |
| CN113724255A (en) * | 2021-10-25 | 2021-11-30 | 鲁东大学 | Counting method for abalones in seedling raising period |
| CN114503946A (en) * | 2022-01-24 | 2022-05-17 | 海南大学 | Fishing ground feeding system and method based on high frame rate dynamic frame difference accurate identification |
Non-Patent Citations (2)
| Title |
|---|
| 刘超.《 不同方式饲养的七星瓢虫对茶蚜捕食功能研究》.茶叶学报.2021,第62卷(第3期),150-155. * |
| 鸡数字化精细养殖中营养需要量预测模型;王文杰;穆淑琴;张敬;葛翔;;天津农业科学(第02期);1-4 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115336549A (en) | 2022-11-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106773721B (en) | A kind of aquaculture intelligence feeding system and its control method based on fuzzy control | |
| CN114467824B (en) | Intelligent bait casting boat | |
| CN103875578A (en) | Automatic bait casting machine | |
| CN204599016U (en) | A kind of feeding device of automatic feeding system | |
| CN108375548B (en) | Fish pond culture feeding evaluation system based on big data | |
| CN211379204U (en) | Intelligent accurate feeder for fattening pigs | |
| CN107765665A (en) | Based on internet breeding soft-shell turtle management system | |
| CN115336549B (en) | Intelligent fish culture feeding system and method | |
| TW202209965A (en) | A smart shrimp and/or crab feeding management system and the method thereof | |
| CN111387064A (en) | Special device for captive livestock automatic breeding and breeding method | |
| CN114895731A (en) | Breeding monitoring system and method based on digital twins | |
| CN208317891U (en) | A kind of laying hen feed for nursing circulation supply equipment | |
| CN207978656U (en) | A kind of pig raising device for feeding | |
| CN206933023U (en) | Intelligent fish pond feeding mixes supply system | |
| CN202773747U (en) | Programmable automatic screw feeder utilizing pipe | |
| KR20220079115A (en) | Intelligent automatic feeding system for fish farms using artificial intelligence | |
| CN111758644A (en) | Intelligent automatic emptying equipment for aquaculture | |
| CN111713442A (en) | Feed delivery ship | |
| CN112586398A (en) | Automatic feeder for livestock and poultry granulated feed | |
| CN2836460Y (en) | Automatic feeding machine | |
| CN212697251U (en) | A aquatic products bait nutrition reinforcement case for schizochytrium limacinum powder is fed | |
| CN206178513U (en) | Intelligence aquaculture system based on LABVIEW remote control and measurement technique | |
| CN211510270U (en) | Intelligent aquaculture pond based on big data analysis | |
| CN220654457U (en) | Accurate control device for baiting amount of bait casting machine | |
| CN216146980U (en) | Large-scale meat rabbit breeding equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |