CN115497013A - Method, system, equipment and medium for monitoring stress state of industrially cultured fishes - Google Patents
Method, system, equipment and medium for monitoring stress state of industrially cultured fishes Download PDFInfo
- Publication number
- CN115497013A CN115497013A CN202110615931.2A CN202110615931A CN115497013A CN 115497013 A CN115497013 A CN 115497013A CN 202110615931 A CN202110615931 A CN 202110615931A CN 115497013 A CN115497013 A CN 115497013A
- Authority
- CN
- China
- Prior art keywords
- fish
- fishes
- video data
- stress
- monitoring
- 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.)
- Pending
Links
- 241000251468 Actinopterygii Species 0.000 title claims abstract description 181
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000012544 monitoring process Methods 0.000 title claims abstract description 34
- 238000009360 aquaculture Methods 0.000 claims abstract description 21
- 244000144974 aquaculture Species 0.000 claims abstract description 21
- 238000001514 detection method Methods 0.000 claims abstract description 21
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims abstract description 21
- 238000013528 artificial neural network Methods 0.000 claims abstract description 13
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 12
- 238000004590 computer program Methods 0.000 claims description 13
- 230000004927 fusion Effects 0.000 claims description 7
- 238000009313 farming Methods 0.000 claims description 5
- 238000005070 sampling Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000003064 k means clustering Methods 0.000 claims description 3
- 238000012935 Averaging Methods 0.000 claims description 2
- 238000000605 extraction Methods 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 7
- 238000004891 communication Methods 0.000 description 5
- 230000006399 behavior Effects 0.000 description 4
- 230000008859 change Effects 0.000 description 4
- 208000023414 familial retinal arterial macroaneurysm Diseases 0.000 description 4
- 239000000126 substance Substances 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- JYGXADMDTFJGBT-VWUMJDOOSA-N hydrocortisone Chemical compound O=C1CC[C@]2(C)[C@H]3[C@@H](O)C[C@](C)([C@@](CC4)(O)C(=O)CO)[C@@H]4[C@@H]3CCC2=C1 JYGXADMDTFJGBT-VWUMJDOOSA-N 0.000 description 2
- 230000000366 juvenile effect Effects 0.000 description 2
- 238000011176 pooling Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- SNICXCGAKADSCV-JTQLQIEISA-N (-)-Nicotine Chemical compound CN1CCC[C@H]1C1=CC=CN=C1 SNICXCGAKADSCV-JTQLQIEISA-N 0.000 description 1
- 241001465754 Metazoa Species 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000001066 destructive effect Effects 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000013632 homeostatic process Effects 0.000 description 1
- 229960000890 hydrocortisone Drugs 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 229960002715 nicotine Drugs 0.000 description 1
- SNICXCGAKADSCV-UHFFFAOYSA-N nicotine Natural products CN1CCCC1C1=CC=CN=C1 SNICXCGAKADSCV-UHFFFAOYSA-N 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01K—ANIMAL HUSBANDRY; AVICULTURE; APICULTURE; PISCICULTURE; FISHING; REARING OR BREEDING ANIMALS, NOT OTHERWISE PROVIDED FOR; NEW BREEDS OF ANIMALS
- A01K61/00—Culture of aquatic animals
- A01K61/10—Culture of aquatic animals of fish
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
Landscapes
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Environmental Sciences (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Animal Husbandry (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Zoology (AREA)
- Data Mining & Analysis (AREA)
- Biodiversity & Conservation Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Marine Sciences & Fisheries (AREA)
- Farming Of Fish And Shellfish (AREA)
Abstract
The invention relates to a method, a system, equipment and a medium for monitoring fish stress state for industrial aquaculture. The method comprises the following steps: acquiring video data of industrially cultured underwater fishes through a camera device; detecting and identifying fish in the video data by using a full convolution neural network algorithm; classifying the fishes according to the relative pixel positions of the fishes in the video data; and quantifying the water layer distribution of the fishes, and calculating according to the water layer distribution of the fishes to obtain the stress line. According to the invention, the first thirteen layers of VGG-16 are used as a feature extraction backbone network, a multi-scale convolution kernel-based detection module is established for target positioning and identification, deep information and shallow information are fused, and the multi-scale and shielding problems in the culture environment are solved.
Description
Technical Field
The invention relates to the technical field of aquatic animal state monitoring methods, in particular to a method, a system, equipment and a medium for monitoring a fish stress state for industrial aquaculture.
Background
With the continuous development of aquaculture industry, each link of industrial aquaculture is automatically monitored, which is particularly important for ensuring safe aquaculture environment, healthy living state of fishes and quality of fishes, especially for monitoring and evaluating fish behaviors.
In aquaculture, the stress to which fish are subjected mainly includes: factors such as improper operation of cultivation personnel in the pond separation process; factors such as deterioration of water quality and environment and overhigh culture density in the transportation process can bring huge stress to the fishes, so that the fishes have the phenomenon of bottom layer life. According to the behavior change of the fish, the state of the fish can be judged by combining the water layer distribution change of the fish.
Conventional ways of observing the stress state of fish, such as calculating the time required for the fish to enter the top from the bottom, measuring the concentration of cortisol, etc., consume manpower and material resources, or cause damage to the fish. Therefore, a low-cost, non-destructive method for monitoring stress status of fish is needed.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides a method, a system, equipment and a medium for monitoring the stress state of fishes for industrial aquaculture.
In a first aspect, the invention provides a fish stress state monitoring method for industrial aquaculture, which comprises the following steps:
acquiring video data of industrially cultured underwater fishes through a camera device;
detecting and identifying fish in the video data by using a full convolution neural network algorithm;
classifying the fish according to their relative pixel positions in the video data; and
and quantifying the water layer distribution of the fishes, and calculating according to the water layer distribution of the fishes to obtain a stress line.
Further, the detecting and identifying the fish in the video data by using the full convolution neural network algorithm includes:
extracting image features by taking the front 13 layers of VGG-16 as a backbone network; and
and establishing a detection module based on the multi-scale convolution kernel to carry out target positioning and identification.
Further, the establishing of the detection module based on the multi-scale convolution kernel for target positioning and identification includes:
respectively extracting feature maps from the downsampling positions to carry out splicing and fusion to obtain feature maps of 3 scales; and
inputting the feature maps of 3 scales into corresponding three rows of multi-scale convolution kernels for fusion, and finally obtaining detection results of the targets of three scales, namely large scale, medium scale and small scale.
Further, the classifying the fish according to their relative pixel positions in the video data comprises:
classifying the fishes into upper fishes aiming at the pixel positions of 20% -50% of the video data of the detection result; and
and classifying the fishes into lower fishes aiming at the detection results located at 80% -50% of pixel positions of the video data.
Further, the quantifying the water layer distribution of the fishes and calculating the stress line according to the water layer distribution of the fishes comprises:
acquiring the number of each type of fish in the video data within unit time;
obtaining the optimal lower-layer fish proportion under the stress state and the optimal lower-layer fish proportion under the non-stress state by a K-means clustering method; and
and calculating to obtain the stress line.
Further, the obtaining the number of each type of fish in the video data per unit time includes:
determining the total frame number in unit time;
summing the number of each type of fish in each frame of image in unit time; and
and averaging the summation result to obtain the number of each type of fish in the current unit time.
Further, the calculating the duress line includes:
the stress line is obtained by the following formula:
wherein x is the stress line, x 1 Is the optimal lower layer fish proportion in the non-stressed state, x 2 The optimal lower layer fish ratio under the stress state.
In a second aspect, the present invention provides a fish stress condition monitoring system for industrial aquaculture, comprising:
the video data acquisition module is used for acquiring video data of the industrially cultured underwater fishes through the camera device;
the detection and identification module is used for detecting and identifying the fish in the video data by utilizing a full convolution neural network algorithm;
the classification module is used for classifying the fishes according to the relative pixel positions of the fishes in the video data; and
and the stress line calculation module is used for quantifying the water layer distribution of the fishes and calculating the stress line according to the water layer distribution of the fishes.
In a third aspect, the present invention further provides an electronic device, which includes a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method for monitoring the stress status of fish for industrial aquaculture according to the first aspect.
In a fourth aspect, the present invention further provides a non-transitory computer readable storage medium, on which a computer program is stored, wherein the computer program, when executed by a processor, implements the steps of the method for monitoring the stress status of fish for industrial farming according to the first aspect.
The invention provides a method, a system, equipment and a medium for monitoring the stress state of fishes for industrial culture, wherein the first thirteen layers of VGG-16 are used as a characteristic extraction backbone network, a multi-scale convolution kernel-based detection module is established for target positioning and identification, deep information and shallow information are fused, the multi-scale and shielding problems in a culture environment are solved, meanwhile, the method is visual in the number of each kind of fishes of input videos, so that culture personnel can analyze the dynamic change of the behaviors of the fishes conveniently, and a non-invasive, efficient and intelligent tool is provided for monitoring the fishes in real time.
Drawings
A more complete understanding of exemplary embodiments of the present invention may be had by reference to the following drawings in which:
FIG. 1 is a flow chart of a method for monitoring stress status of fish in industrial aquaculture according to an embodiment of the invention;
FIG. 2 is a schematic structural diagram of a full convolution neural network unit according to an embodiment of the present invention;
FIGS. 3 and 4 are graphs of the number of each type of fish and the proportion of the fish in the lower layer as a function of time provided by an embodiment of the present invention;
FIG. 5 is a block diagram of a fish stress condition monitoring system for industrial aquaculture according to an embodiment of the present invention; and
fig. 6 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention are clearly and completely described below with reference to the drawings in the embodiments of the present invention. It is to be understood that the embodiments described are only a few embodiments of the present invention, and not all embodiments. All other embodiments, which can be obtained by a person skilled in the art without inventive step based on the embodiments of the present invention, are within the scope of protection of the present invention.
Fig. 1 is a flowchart of a method for monitoring stress status of fish in industrial aquaculture according to an embodiment of the present invention.
Referring to fig. 1, the method for monitoring the stress state of fish for industrial farming includes the steps of:
s101: acquiring video data of industrially cultured underwater fishes through a camera device;
s103: detecting and identifying fish in the video data by using a full convolution neural network algorithm;
s105: classifying the fishes according to the relative pixel positions of the fishes in the video data; and
s107: and quantifying the water layer distribution of the fishes, and calculating according to the water layer distribution of the fishes to obtain a stress line.
In the embodiment, the stress line is obtained by clustering the lower layer fish proportion of the fish under the stress state according to the water layer distribution of the fish under the non-stress state and the stress state.
In an embodiment, specifically, step S101 includes acquiring video data of industrially-cultivated underwater fish, including video data of different fish pond environments, different light conditions, and the like, by using a camera device.
In an embodiment, specifically, the step S103 includes extracting image features with the first 13 layers of VGG-16 as a backbone network; and establishing a detection module based on a multi-scale convolution kernel for target positioning and identification: respectively extracting feature maps from the down-sampling positions to carry out splicing and fusion to obtain feature maps of 3 scales; inputting the feature maps of 3 scales into corresponding three rows of multi-scale convolution kernels for fusion, and finally obtaining detection results of the targets of three scales, namely large scale, medium scale and small scale.
Referring to fig. 2, wherein k =3, c =64, d =1, and k =2 of max power of the convolutional layer of layer 1-2 of the backbone network; k =3, c =128, d =1, and k =2 for max pooling for the convolutional layers of layers 3-4; k =3, c =256, d =1, and k =2 for max pooling for the convolutional layers of layers 5-7; k =3, c =512, d =1 for the 8-10 th convolutional layer; k =3, c =512, d =1, and k =2 for max power of the convolutional layers of layers 11-13; where k is the convolution kernel size, c is the number of channels, and d is the expansion ratio.
And a detection module based on the multi-scale convolution kernels extracts the characteristic diagrams from the 8-time, 16-time and 32-time downsampling positions of the backbone network respectively and inputs the three rows of multi-scale convolution kernels corresponding to the characteristic diagrams for fusion. And for the feature maps subjected to 8-time and 16-time down-sampling, splicing and fusing the feature maps with the feature map of the up-sampling result to obtain a new feature map.
Wherein, the three-row multi-scale convolution kernels comprise three parallel convolution networks, which are respectively L rows, and the large-scale convolution kernels are used: 7 × 7, 5 × 5; m columns, using a mesoscale convolution kernel: 5 × 5, 3 × 3; column S, using a small scale convolution kernel: 3 × 3, 1 × 1; inputting the spliced characteristic diagram into three rows of convolutional neural networks to finally obtain detection results of the large, medium and small scale targets.
In an embodiment, the full convolution neural network training approach is as follows:
(1) Scaling input video data to a video image with the size of 416 multiplied by 416, and connecting a ReLU layer after each convolution layer of the constructed backbone network;
(2) The texture features of the full convolution neural network learning image are finally output to feature maps of three scales through three rows of convolution neural networks, wherein the feature maps are 13 multiplied by 13, 26 multiplied by 26 and 52 multiplied by 52 respectively;
(3) And training and testing to obtain the full convolution neural network algorithm.
In an embodiment, the sample video data is video data of one hour each obtained after moving fish from a culture pond into two new culture ponds, one of which is filled with nicotine drug and the other is not processed.
In an embodiment, specifically, step S105 includes determining the relative pixel positions of the detection results in the input video and classifying the fishes. Determining that the detected result is an upper-layer fish and marking the detected result as an A-type fish according to the pixel positions of 20% -50% of the input video; and determining that the detected result is 80% -50% of the pixel positions of the input video, and marking the detected result as a fish of the lower layer and a fish of the B type.
In an embodiment, specifically, step S107 includes calculating the number of each type of fish and the ratio of the lower layer fish in the video data per unit time, and calculating the optimal ratio of the lower layer fish in the stressed state to the non-stressed state by using a K-means algorithm, so as to obtain the stress line size and the time from the lower layer to the upper layer of the fish. The method comprises the following specific steps:
first, based on the fish classification result obtained in step S105, the number of each type of fish per unit time in the video data is calculated as follows:
(1) The total number of frames per unit time is determined as follows:
frams=fps*60
wherein frams is the total number of frames per minute (unit time); fps is the frame rate of the video data.
(2) The number of fish of each type per image per minute is summed as follows:
wherein the content of the first and second substances,the sum of the number of A-type fishes in each frame of image in the jth minute;the number of the A-type fish in the j minute and i frame image.
Wherein the content of the first and second substances,the sum of the number of B-type fishes in each frame of image in the jth minute;the number of the B-type fishes in the jth minute and ith frame image.
(3) The average is taken as the number of fish of each type in the current minute as follows:
wherein the content of the first and second substances,is the average number of class a fish in the jth minute;the sum of the number of A-type fishes in each frame of image in the jth minute; frams is the total number of frames per minute.
Wherein the content of the first and second substances,the average number of the B-type fishes in the h minute;the sum of the number of B-type fishes in each frame of image in the jth minute; frams is the total number of frames per minute.
Then, based on the number of each type of fish in the video data per unit time (i.e., the number of the type a fish and the number of the type B fish) the proportion of the lower layer fish (i.e., the type B fish) in the unit time is obtained, which is as follows:
wherein P is the proportion of the class B fish to all the fishes at the jth minute;is the average of the number of class a fish at minute j;the average of the number of class B fish at minute j.
And finally, calculating the optimal lower-layer fish proportion under the stress state and the optimal lower-layer fish proportion under the non-stress state by a K-means algorithm based on the proportion of the B-type fish in unit time, so as to obtain the stress line size and the time from the lower layer to the upper layer of the fish, wherein the method specifically comprises the following steps:
(1) Forming a key value pair < k, v > by the proportion of the current unit time and the lower layer fish, wherein k represents the kth minute, and v represents the proportion of the lower layer fish corresponding to the kth minute;
(2) Obtaining the optimal lower layer fish proportion under the non-stressed state through a K-means clustering method, and recording the optimal lower layer fish proportion as a seed point x 1 And taking the seed point x or more 1 The maximum k value of (A) is marked as the cost of fish from lower to upper layers in a stress-free stateTime t of 1 (ii) a Similarly, the optimum ratio of fish in the lower layer under stress can be obtained and recorded as seed point x 2 And is taken to be less than or equal to x 2 The minimum k value at the seed point is recorded as the time t taken by the fish from the lower layer to the upper layer under the stress condition 2 ;
(3) Stress line size was calculated as follows:
by quantifying the underwater distribution of fish, the ratio of the fish in the lower layer exceeds the stress line, and the time exceeds t 1 At t 1 To t 2 During the time period, the farmer can perform related operations to relieve the pressure on the juvenile fish, so that the juvenile fish can keep the homeostasis and the fish body healthy.
In order to enable the farmer to observe and analyze the stress state of the fish more clearly, the stress state can be displayed in a visual mode. In the examples, fig. 3 and 4 show graphs of the number of each type of fish and the proportion of fish in the lower layer as a function of time, provided by the examples of the present invention.
In fig. 3 and 4, the abscissa represents time, i.e., the total duration of video data, which can be obtained by the following formula:
the unit of the total duration of the video data is minutes, nframs is the total number of frames of the video data, fps is the frame rate of the video data, and the video data of a frame image less than one minute is discarded.
In addition, the ordinate of fig. 3 and 4 represents the number of a-type fishes and B-type fishes per unit time in the video data, and the proportion of the B-type fishes per unit time in the video data, respectively. The numbers of the type A fish and the type B fish per unit time in the video data correspond to those obtained in step S107 described above with reference to the embodimentAndand the proportion of the class B fish per unit time in the video data corresponds to P obtained in step S107 described above with reference to the embodiment.
Fig. 5 is a block diagram illustrating a fish stress condition monitoring system for industrial farming according to an embodiment of the present invention. Referring to fig. 5, the system 500 includes:
a video data acquisition module 501, configured to acquire video data of industrially-cultured underwater fishes through a camera device;
a detection and identification module 503, configured to detect and identify fish in the video data by using a full convolutional neural network algorithm;
a classification module 505, configured to classify the fish according to their relative pixel positions in the video data; and
and a stress line calculation module 507, configured to quantify the water layer distribution of the fish, and calculate a stress line according to the water layer distribution of the fish.
As can be seen from the above, the respective modules 501 to 507 of the system 500 can respectively perform the steps of the monitoring method described with reference to the above embodiments, and the details thereof will not be described here.
Therefore, the invention provides a method and a system for monitoring the fish coerce state for industrial culture, wherein the first thirteen layers of VGG-16 are used as a characteristic extraction backbone network, a multi-scale convolution kernel-based detection module is established for target positioning and identification, deep layer information and shallow layer information are fused, the multi-scale and shielding problems occurring in a culture environment are solved, meanwhile, the invention visualizes the number of each type of fish in input video, is convenient for culture personnel to analyze the dynamic change of fish behaviors, and provides a non-invasive, efficient and intelligent tool for monitoring the fish in real time.
In another aspect, the present invention provides an electronic device. As shown in fig. 6, electronic device 600 includes a processor 601, a memory 602, a communication interface 603, and a communication bus 604;
the processor 601, the memory 602, and the communication interface 603 complete communication with each other through the communication bus 304;
the processor 601 is used for calling the computer program in the memory 602, and the steps of the fish stress condition monitoring method for factory culture provided by the embodiment of the invention as described above are realized when the processor 601 executes the computer program.
Further, the computer program in the memory may be implemented in the form of a software functional unit and may be stored in a computer-readable storage medium when sold or used as a stand-alone product. Based on such understanding, the technical solution of the present invention or a part thereof, which essentially contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium and includes several computer programs to make a computer device (which may be a personal computer, a server, or a network device) execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In another aspect, the present invention provides a non-transitory computer readable storage medium having a computer program stored thereon, which when executed by a processor, implements the steps of the method for monitoring the stress status of fish for industrial farming provided by the embodiment of the present invention as described above.
The above-described system embodiments are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A method for monitoring the stress state of fishes for industrial culture is characterized by comprising the following steps:
acquiring video data of industrially cultured underwater fishes through a camera device;
detecting and identifying fish in the video data by using a full convolution neural network algorithm;
classifying the fish according to their relative pixel positions in the video data; and
and quantifying the water layer distribution of the fishes, and calculating to obtain a stress line according to the water layer distribution of the fishes.
2. The method for monitoring the stress state of fish in industrial aquaculture according to claim 1, wherein the detecting and identifying the fish in the video data by using a full convolution neural network algorithm comprises:
extracting image features by taking the front 13 layers of VGG-16 as a backbone network; and
and establishing a detection module based on the multi-scale convolution kernel to carry out target positioning and identification.
3. The method for monitoring the stress state of fish for industrial aquaculture according to claim 2, wherein the establishing of the multi-scale convolution kernel-based detection module for target location and identification comprises:
respectively extracting feature maps from the down-sampling positions to carry out splicing and fusion to obtain feature maps of 3 scales; and
inputting the feature maps of 3 scales into corresponding three rows of multi-scale convolution kernels for fusion, and finally obtaining detection results of the targets of three scales, namely large scale, medium scale and small scale.
4. The fish stress situation monitoring method for industrial aquaculture of claim 1, wherein the classifying the fish according to their relative pixel positions in the video data comprises:
classifying the fishes into upper fishes aiming at the 20% -50% pixel positions of the video data of the detection result; and
and classifying the fishes into lower fishes aiming at the detection results located at 80% -50% of pixel positions of the video data.
5. The method for monitoring the stress state of fish for industrial aquaculture according to claim 1, wherein the quantifying the water layer distribution of the fish and calculating the stress line according to the water layer distribution of the fish comprises:
acquiring the number of each type of fish in the video data within unit time;
obtaining the optimal lower-layer fish proportion under the stress state and the optimal lower-layer fish proportion under the non-stress state by a K-means clustering method; and
and calculating to obtain the stress line.
6. The fish stress condition monitoring method for industrial aquaculture of claim 5, wherein the acquiring of the number of each type of fish per unit time in the video data comprises:
determining the total frame number in the unit time;
summing the number of each type of fish in each frame of image in the unit time; and
and averaging the summation result to obtain the number of each type of fish in the current unit time.
7. The method for monitoring the stress state of fishes in industrial farming according to claim 5, wherein the calculating the stress line comprises:
the stress line is obtained by the following formula:
wherein x is the stress line, x 1 Is the optimum ratio of fish in the lower layer under the stress-free state, x 2 Is the optimal ratio of the fish in the lower layer under the stress state.
8. A fish stress condition monitoring system for industrial aquaculture, comprising:
the video data acquisition module is used for acquiring video data of the industrially cultured underwater fishes through the camera device;
the detection and identification module is used for detecting and identifying the fish in the video data by utilizing a full convolution neural network algorithm;
the classification module is used for classifying the fishes according to the relative pixel positions of the fishes in the video data; and
and the stress line calculation module is used for quantifying the water layer distribution of the fishes and calculating the stress line according to the water layer distribution of the fishes.
9. An electronic device, comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the fish stress condition monitoring method for industrial aquaculture according to any one of claims 1 to 7.
10. A non-transitory computer-readable storage medium, on which a computer program is stored, wherein the computer program, when being executed by a processor, implements the steps of the method for monitoring stress status of fish for industrial aquaculture according to any one of claims 1 to 7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110615931.2A CN115497013A (en) | 2021-06-02 | 2021-06-02 | Method, system, equipment and medium for monitoring stress state of industrially cultured fishes |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110615931.2A CN115497013A (en) | 2021-06-02 | 2021-06-02 | Method, system, equipment and medium for monitoring stress state of industrially cultured fishes |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115497013A true CN115497013A (en) | 2022-12-20 |
Family
ID=84465235
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110615931.2A Pending CN115497013A (en) | 2021-06-02 | 2021-06-02 | Method, system, equipment and medium for monitoring stress state of industrially cultured fishes |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115497013A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117172598A (en) * | 2023-09-05 | 2023-12-05 | 中国长江电力股份有限公司 | Basin water ecology fish monitoring management system based on cloud computing |
-
2021
- 2021-06-02 CN CN202110615931.2A patent/CN115497013A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117172598A (en) * | 2023-09-05 | 2023-12-05 | 中国长江电力股份有限公司 | Basin water ecology fish monitoring management system based on cloud computing |
CN117172598B (en) * | 2023-09-05 | 2024-05-28 | 中国长江电力股份有限公司 | Basin water ecology fish monitoring management system based on cloud computing |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10747999B2 (en) | Methods and systems for pattern characteristic detection | |
Blok et al. | The effect of data augmentation and network simplification on the image‐based detection of broccoli heads with Mask R‐CNN | |
Mishra et al. | A Deep Learning-Based Novel Approach for Weed Growth Estimation. | |
CN114494837B (en) | Intelligent density identification method and system for fishery resources | |
CN110349153B (en) | Image method tobacco peduncle content detection method based on multi-feature fusion | |
SE1930281A1 (en) | Method for calculating deviation relations of a population | |
Herzog | Initial steps for high-throughput phenotyping in vineyards | |
Costa et al. | Measuring pecan nut growth utilizing machine vision and deep learning for the better understanding of the fruit growth curve | |
US11798262B2 (en) | Methods and systems for identifying internal conditions in juvenile fish through non-invasive means | |
Strange et al. | Automatic estimation of wheat grain morphometry from computed tomography data | |
CN111507227B (en) | Multi-student individual segmentation and state autonomous identification method based on deep learning | |
Latif et al. | Deep learning based intelligence cognitive vision drone for automatic plant diseases identification and spraying | |
CN112016400A (en) | Single-class target detection method and device based on deep learning and storage medium | |
US20240104900A1 (en) | Fish school detection method and system thereof, electronic device and storage medium | |
Atanbori et al. | Convolutional neural net-based cassava storage root counting using real and synthetic images | |
Yudhana et al. | Fish Freshness Identification Using Machine Learning: Performance Comparison of k-NN and Naïve Bayes Classifier. | |
CN115497013A (en) | Method, system, equipment and medium for monitoring stress state of industrially cultured fishes | |
CN115240060A (en) | Starfish disaster early warning method and system | |
Stasenko et al. | Deep Learning for improving the storage process: Accurate and automatic segmentation of spoiled areas on apples | |
CN115526852A (en) | Molten pool and splash monitoring method in selective laser melting process based on target detection and application | |
CN111898677A (en) | Plankton automatic detection method based on deep learning | |
CN115512215A (en) | Underwater biological monitoring method and device and storage medium | |
CN113326743B (en) | Method for extracting and analyzing fish school movement behavior parameters under cultivation background condition | |
CN116363494B (en) | Fish quantity monitoring and migration tracking method and system | |
Chen et al. | High-throughput phenotyping of nematode cysts |
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 |