CN114586728A - Fry sorting system and sorting method - Google Patents
Fry sorting system and sorting method Download PDFInfo
- Publication number
- CN114586728A CN114586728A CN202210226859.9A CN202210226859A CN114586728A CN 114586728 A CN114586728 A CN 114586728A CN 202210226859 A CN202210226859 A CN 202210226859A CN 114586728 A CN114586728 A CN 114586728A
- Authority
- CN
- China
- Prior art keywords
- fry
- identification
- pool
- sorting
- pipeline
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 48
- 230000000694 effects Effects 0.000 claims abstract description 185
- 238000011156 evaluation Methods 0.000 claims description 64
- 238000012549 training Methods 0.000 claims description 45
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 44
- 238000007664 blowing Methods 0.000 claims description 21
- 238000000926 separation method Methods 0.000 claims description 19
- 238000003062 neural network model Methods 0.000 claims description 18
- 230000008569 process Effects 0.000 claims description 12
- 230000000007 visual effect Effects 0.000 claims description 11
- 238000010276 construction Methods 0.000 claims description 5
- 241000251468 Actinopterygii Species 0.000 description 30
- 238000001514 detection method Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 7
- 238000004891 communication Methods 0.000 description 5
- 238000003556 assay Methods 0.000 description 4
- 238000009395 breeding Methods 0.000 description 3
- 230000001488 breeding effect Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000001174 ascending effect Effects 0.000 description 2
- 239000012141 concentrate Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000000630 rising effect Effects 0.000 description 2
- 238000003756 stirring Methods 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013210 evaluation model Methods 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000011435 rock Substances 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000000638 stimulation Effects 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
- A01K61/90—Sorting, grading, counting or marking live aquatic animals, e.g. sex determination
- A01K61/95—Sorting, grading, counting or marking live aquatic animals, e.g. sex determination specially adapted for fish
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- 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
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Zoology (AREA)
- Environmental Sciences (AREA)
- Marine Sciences & Fisheries (AREA)
- Biodiversity & Conservation Biology (AREA)
- Animal Husbandry (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Farming Of Fish And Shellfish (AREA)
Abstract
The invention discloses a fry sorting system and a fry sorting method, which comprise an original pool, a high-quality pool, an inferior pool and a re-identification device, wherein the original pool is respectively connected to the high-quality pool and the re-identification device through sorting pipelines; the sorting pipeline is provided with a first activity identification part for carrying out primary activity identification, and the sorting pipeline is controlled to be communicated with the high-quality pool or the re-identification pipeline device according to an identification result; the re-identification device is connected to the original pool and the inferior pool through pipelines respectively, is used for carrying out secondary activity identification, and is controlled to be communicated with the original pool or the inferior pool according to an identification result. The fry sorting machine can quickly and accurately automatically sort fry, and greatly improves the fry sorting efficiency.
Description
Technical Field
The invention relates to the technical field of fishery breeding equipment, in particular to a fry sorting system and a fry sorting method.
Background
The breeding and selling of the fry are common economic behaviors in the market, and both require breeding personnel to know the number and the activity of the fry more accurately. At present adopt manual observation's mode to concentrate on the fry in the bucket more, observe through stirring the water and detect the fry activity, the high-quality fry of activity can move about along the vortex edge, and the fry that activity is low or dead just can be drawn into the vortex center, and this kind of detection mode intensity of labour is big, and can't carry out the continuous detection operation of large batch.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a fry sorting system, which can quickly and accurately automatically sort fry, greatly improve the fry sorting efficiency and solve the problems of high labor intensity and incapability of performing large-batch continuous detection operation caused by manual sorting in the prior art.
The second purpose of the invention is to provide a fry sorting method, which combines an artificial intelligence algorithm to sort fries, thereby effectively improving the fry sorting efficiency.
A third object of the present invention is to provide a storage medium.
A fourth object of the invention is to provide a computing device.
The first purpose of the invention is realized by the following technical scheme: a fry sorting system comprises an original pool, a high-quality pool, an inferior pool and a re-identification device, wherein the original pool is respectively connected to the high-quality pool and the re-identification device through sorting pipelines;
the sorting pipeline is provided with a first activity identification part for carrying out primary activity identification, and the sorting pipeline is controlled to be communicated with the high-quality pool or the re-identification pipeline device according to an identification result;
the re-identification device is connected to the original pool and the inferior pool through pipelines respectively and is used for carrying out secondary activity identification, and the re-identification device is controlled to be communicated with the original pool or the inferior pool according to an identification result.
Preferably, the sorting pipeline is further provided with a counting part for counting the fry, and the counting part is positioned between the original pool and the first activity identification part;
the counting part comprises a first cavity and a first camera device on the sorting pipeline;
a first movable piece is arranged on one side of an inlet or an outlet of the first cavity and controls the working state through a first motor;
the first camera device is arranged on the first cavity, identifies the passing fries in the range of the shot visual field and counts;
the first motor is connected with the main control equipment through a motor driver;
the first camera device is connected with the main control equipment.
Preferably, the first activity assay site comprises a second cavity on the sorting pipeline, a second camera device and an air blower arranged in the second cavity;
a mesh movable plate controlled to ascend and descend by a sixth motor is arranged on the outlet side of the second cavity; a second movable part is arranged on a pipeline of the sorting pipeline connected to the re-identification device, and a third movable part is arranged on a pipeline of the sorting pipeline connected to the high-quality pool; the second movable part and the third movable part respectively control the working state through a second motor and a third motor so as to control the separation pipeline to be communicated to the re-appraisal device or the high-quality pool;
the second camera device is arranged on the second cavity and used for recording a fry movement video in the second cavity and performing primary activity identification on the fry according to the fry image;
the second motor, the third motor and the sixth motor are respectively connected with the main control equipment through motor drivers; the air blower is connected with the main control equipment;
the second camera device is connected with the main control equipment.
Preferably, the re-identification device comprises an inclined slope pipeline and a third camera device which are obliquely arranged, the sorting pipeline is connected to an inlet at the high position of the inclined slope pipeline through a pipeline, and an outlet at the low position of the inclined slope pipeline is respectively connected to the original pool and the poor pool through pipelines;
the third camera device is arranged above the slope pipeline and used for shooting a fry moving image in the slope pipeline and carrying out secondary activity identification on the fry based on the fry image;
a fourth movable part is arranged on a pipeline, the outlet at the lower part of the slope pipeline is connected to the inferior pool, a fifth movable part is arranged on a pipeline, the outlet at the lower part of the slope pipeline is connected to the original pool, and the fourth movable part and the fifth movable part respectively and correspondingly control the working state through a fourth motor and a fifth motor so as to control the slope pipeline to be communicated to the inferior pool or the original pool;
the fourth motor and the fifth motor are respectively connected with the master control equipment through motor drivers;
the third camera device is connected with the main control equipment.
The second purpose of the invention is realized by the following technical scheme: a fry sorting method realized by the fry sorting system based on the first object of the invention comprises the following steps:
controlling the water in the original pool to generate vortex;
controlling a first activity identification part on the sorting pipeline to perform activity identification work;
controlling the separation pipeline to be communicated with a high-quality pool or a re-identification device according to the activity identification result of the first activity identification part; wherein:
when the activity identification result of the first activity identification part is that the fry is high-quality, controlling the sorting pipeline to be communicated to a high-quality pool, or controlling the sorting pipeline to be communicated to a re-identification device;
when the sorting pipeline is communicated with the re-identification device, controlling the re-identification device to perform secondary activity identification on the fries entering the re-identification device; controlling the re-identification device to be communicated with the original pool or the inferior pool according to the secondary activity identification result or combining the primary activity identification result and the secondary activity identification result; wherein:
and when the fry is inferior in identification result, controlling the re-identification device to be communicated to the inferior pool, otherwise, controlling the re-identification device to be communicated to the original pool.
Preferably, the method further comprises the following steps:
after the water in the original pool is controlled to generate vortex, the first movable piece of the counting part is controlled by the first motor to act, so that the water pool in the original pool enters the counting part of the sorting pipeline and further reaches the first activity identification part;
controlling a first camera device at the counting part of the sorting pipeline to start shooting work;
recognizing the passing fries in the visual field range shot by the first camera device through a fry recognition model and counting the fries;
when the number of the detected fries reaches a first preset value, the state of a first moving part of a counting part is controlled through a first motor, and the fries in the original pond are prevented from entering a first activity identification part;
controlling a blower of the first activity identification part to work, enabling ripples to appear on the water surface of the second cavity, and simultaneously controlling a second camera device of the first activity identification part of the sorting pipeline to record videos of the conditions of the second cavity type fries;
acquiring a fry condition video shot by a second camera device, judging and evaluating fry activity by an activity evaluation blowing observation model according to the fry condition video, and obtaining a first evaluation result;
when the first evaluation result exceeds a second preset value, controlling the sorting channel to be communicated with the high-quality pool, so that the fry in the second cavity enters the high-quality pool; otherwise, controlling the sorting channel to be communicated with the re-identification device, and enabling the fry in the second cavity to enter the re-identification device.
Further, the process of controlling the re-identification device to perform the secondary activity identification is as follows:
when the separation channel is controlled to be communicated with the re-identification device, a third camera device in the re-identification device is controlled to shoot a picture of the fry state on the slope pipeline to obtain a fry pouring state image, an activity evaluation pouring observation model is used for evaluating the activity of the fry according to the fry state image, and a second evaluation result is obtained;
taking the second evaluation result or the average evaluation result as a final evaluation result; wherein the average assessment result is an average of the first assessment result and the second assessment result;
and when the final evaluation result exceeds a third preset value, controlling the re-identification device to be communicated with the original pool to enable the fries to enter the original pool, otherwise, controlling the re-identification device to be communicated with the inferior pool to enable the fries to enter the inferior pool.
Further, the fry identification model is constructed by the following process:
shooting a fry photo based on existing hardware and collecting the fry photo as a training sample;
marking the collected training samples, and setting the labels as fry;
constructing a neural network model, and training the neural network model through the labeled training samples to obtain a fry identification model;
the fry identification method comprises the steps that fry identification is carried out in real time on the basis of a fry identification model, and when the fry touches a side boundary of a downstream flow, which is vertical to the flow direction of water flow and is positioned at a counting part of a sorting pipeline, in a visual field of a first camera device, the fry is counted;
the construction process of the activity evaluation blowing observation model is as follows:
recording a video of the fry in a blowing environment based on the existing hardware, and collecting the video as a training sample;
marking the collected training samples, wherein the labels are high-quality fry and poor-quality fry;
tracking the fry in the training sample to determine the motion state of the fry;
constructing a neural network model, and training the neural network model according to the motion state of the fry of the training sample and the label to obtain an activity evaluation blowing observation model;
evaluating the activity of the fry in the fry condition video shot by the second camera device based on an activity evaluation blowing observation model, and taking the ratio of the number of high-quality fries to the total number of fries as an evaluation result;
the construction process of the activity evaluation pouring water observation model is as follows:
shooting a picture of the fry in a water pouring environment based on the existing hardware, and collecting the picture as a training sample;
marking the collected training samples, and setting the labels as high-quality fry and poor-quality fry;
acquiring the state of the fry in a water pouring environment in a training sample;
and (3) constructing a neural network model, and training the neural network model according to the red fry state and the label of the training sample to obtain an activity evaluation pouring observation model.
The third purpose of the invention is realized by the following technical scheme: a storage medium storing a program, wherein the program when executed by a processor realizes the fry sorting method according to the first object of the invention.
The fourth purpose of the invention is realized by the following technical scheme: a computing device comprises a processor and a memory for storing processor executable programs, and when the processor executes the programs stored in the memory, the fry sorting method achieves the fry sorting method.
Compared with the prior art, the invention has the following advantages and effects:
(1) the fry sorting system comprises an original pool, a high-quality pool, an inferior pool and a re-identification device, wherein the original pool is respectively connected to the high-quality pool and the re-identification device through sorting pipelines; the sorting pipeline is provided with a first activity identification part for carrying out primary activity identification, and the identification result controls the sorting pipeline to be communicated with the high-quality pool or the re-identification pipeline device; the re-identification device is connected to the original pool and the poor pool through pipelines respectively, the re-identification device is used for carrying out secondary activity identification, and the re-identification device is controlled to be communicated with the original pool or the poor pool according to an identification result. Based on the first activity identification part in the fry sorting system, part of high-quality fries which are determined by comparison can be sorted, the fries which are not judged to be high-quality by the first activity identification part are re-identified by the re-identification device and then determined to be poor, and if the fries are determined to be poor, the fries are sorted into a poor pond, and the fries which are not determined to be poor are returned to the original pond again to wait for next sorting. Therefore, the fry can be quickly and accurately automatically sorted by the system, the fry sorting efficiency is greatly improved, and the problems that labor intensity is high and large-batch continuous detection operation cannot be performed due to manual sorting in the prior art are solved.
(2) In the fry sorting system, the sorting pipeline is also provided with a counting part for counting the fries, the counting part is positioned between the original pool and the first activity identification part, and the fry sorting system comprises a first cavity and a first camera device on the sorting pipeline. In the invention, the first movable part is arranged on the first cavity, the first movable part can be controlled to ascend and descend by the first motor so as to play a role of controlling water flow, so that the separation work is controlled to start, and in addition, the fry can enter a first activity identification part behind the first movable part by the ascending of the baffle plate so as to lay an identification space for the first activity identification part. The counting part is provided with the first camera device, and the fry passing through the visual field range is identified and counted based on the first camera device.
In addition, in the sorting system, based on the combination of the first moving part and the counting function, the first activity identification part and the re-identification device can be controlled to perform batch screening on the fish fries with the preset number, namely after the fish fries with the preset number are obtained, the first activity identification part is controlled by the first moving part to perform activity identification on the fish school each time, when the activity identification result is high-quality fish fries, the fish schools formed by the fish fries with the preset number are all placed into a high-quality pond, otherwise, the fish schools are all placed into the re-identification device, the re-identification device performs secondary activity identification on the fish schools, and the fish schools are further placed into an original pond or an inferior pond according to the identification result.
(3) In the fry sorting system, the second movable part and the third movable part which are connected with the motor are respectively and correspondingly arranged on the pipelines which are connected with the sorting pipeline, the high-quality pool and the re-identification device, so that the fry can be automatically sorted according to the identification result of the first activity identification part, and the fry can automatically enter the high-quality pool or the re-identification device. In addition, a fourth moving part and a fifth moving part are also arranged on pipelines connected with the inferior pool and the original pool respectively, and the states of the fourth moving part and the fifth moving part can be automatically controlled through a motor according to an identification result, so that the fry can automatically enter the inferior pool or the original pool. Therefore, the sorting system can realize the full-automatic sorting of the fry without any manual interference.
(4) The fry sorting method of the invention comprises the following steps of firstly, controlling water in an original pool to generate vortex; controlling a first activity identification part on the sorting pipeline to perform activity identification work; controlling the separation pipeline to be communicated with a high-quality pool or a re-identification device according to the activity identification result of the first activity identification part; when the sorting pipeline is communicated with the re-identification device, controlling the re-identification device to perform secondary activity identification on the fries entering the re-identification device; controlling the re-identification device to be communicated with the original pool or the inferior pool according to the secondary activity identification result or combining the primary activity identification result and the secondary activity identification result; according to the method, high-quality fry directly enters a high-quality pond based on a primary activity evaluation result, and the fry of medium-grade and high-grade fishes enter an original pond and the fry of poor-quality fishes enter an inferior pond based on two activity evaluation results; through the mode, the activity of the original pond fry can be rapidly and accurately detected, and the working efficiency is improved.
(5) In the fry sorting method, the fry passing through the range of the visual field shot by the first camera at the counting part is identified and counted by the fry identification model, so that the number of the fries is obtained. When the fry activity is identified once, the activity evaluation blowing observation model evaluates the fry activity according to the motion state fed back by the lock in the fry condition video. And when the secondary activity evaluation is carried out on the fry, the activity evaluation water pouring observation model carries out secondary evaluation on the fry activity according to the fry state fed back in the image. The three models are obtained based on machine learning, artificial intelligence is combined, and therefore the fry identification, counting and activity identification speed is higher, and accuracy is higher.
Drawings
FIG. 1 is a schematic diagram of the sorting system of the present invention.
FIG. 2 is a schematic diagram of the cell configuration in the sorting system of the present invention.
FIG. 3 is a schematic view of the structure of the counting part on the sorting pipeline of the sorting system of the invention.
FIG. 4 is a schematic diagram of a first active identification site on a sorting channel of the sorting system of the present invention.
FIG. 5 is a schematic view of the structure of the re-authentication device in the sorting system of the present invention.
In the figure: 1. the device comprises an original pool, a high-quality pool, a low-quality pool, a rotor, a separation pipeline, a first cavity, a first camera device, a second camera device, a third camera device, a first cavity, a second cavity, a blower, a second cavity, a blower, a second cavity, a re-identification box, a slope pipeline, a first movable baffle, a second movable baffle, a third movable baffle, a slope pipeline, a first movable baffle, a slope pipeline, a fourth movable baffle, a slope pipeline, a fourth movable baffle, a slope pipeline, a slope baffle, a slope, a second movable baffle, a slope, a fourth movable baffle, a rotor, a fourth movable baffle, a slope, a fourth movable baffle, a rotor, a second movable baffle, a rotor, a fourth movable baffle, a rotor, a second movable baffle, a rotor, a fourth movable baffle, a rotor, a second cavity, a fourth movable baffle, a rotor and a rotor, a third movable baffle, a rotor, a fourth movable baffle, a third.
Detailed Description
The present invention will be described in further detail with reference to examples and drawings, but the present invention is not limited thereto.
Example 1
Present fry is selected separately and is adopted manual observation's mode to select separately more, concentrates on the fry in the bucket, observes through stirring the water and detects fry activity, and active high quality good fry can move about along the vortex edge, and active low or dead fry just can be drawn into the vortex center, and it is big to have intensity of labour, and can't carry out the problem of large batch continuous detection operation. Based on this, this embodiment discloses a fry sorting system, and this system can be fast and accurate with the good and bad quality fry sorting out, can realize the count of fry simultaneously.
For the convenience of understanding the present embodiment, a fry sorting system disclosed in the embodiments of the present application will be described in detail first.
Referring to fig. 1, a fry sorting system comprises a primary pond, a quality pond, a poor pond and a re-identification device, wherein the primary pond is respectively connected to the quality pond and the re-identification device through sorting pipelines.
Wherein:
the sorting pipeline is provided with a counting part and a first activity identification part. Wherein the counting part is used for counting the fry, and the first activity identification part is used for continuously carrying out activity identification on the fry once. The counting site is located between the cuvette and the first active identification site. The sort conduit after the first active assay site is connected to a quality cell and a re-identification device. In this embodiment, the sorting channel may be controlled to communicate with the quality pool or the re-identification channel device according to the identification result of the first active identification site. The re-identification device is connected to the original pool and the inferior pool through pipelines respectively and used for secondary activity identification, the re-identification device is controlled to be communicated with the original pool or the inferior pool according to the identification result, and the re-identification device is controlled to be communicated with the original pool or the inferior pool specifically according to the secondary activity identification result or by combining the primary activity identification result and the secondary activity identification result.
In this embodiment, the original pond is a pond filled with fries to be counted and sorted; as shown in figure 2, install the rotor through motor control operation at former pond bottom, former pond upper end is connected and is selected separately the pipeline, and the motor passes through motor drive and connects master control equipment, when through the motor control rotor operation, can produce the swirl in the former pond, high-quality fry along the marginal high in the clouds of swirl in the former pond, and the seedling of being poor quality is drawn into in the swirl.
In this embodiment, as shown in fig. 3, the counting portion includes a first cavity and a first camera device on the sorting pipeline, and the first cavity may be specifically a box; a first moving part is arranged on one side of an inlet or an outlet of the first cavity and controls the working state through a first motor so as to control the water flow of the sorting pipeline. In this embodiment, the first movable machine may be a movable baffle, that is, a first movable baffle, the first movable baffle is disposed at an inlet or an outlet side of the first chamber, as shown in fig. 3, the first movable baffle rotating shaft is installed at the outlet side of the first chamber, that is, between the counting portion and the primary activity identification portion, and the first movable baffle can be raised or lowered under the control of the first motor. Under the condition that former pond rotor operates, former pond inside forms the swirl environment, treats about 5 seconds, and the swirl environment is stable, and first adjustable fender descends, then plays the guide rivers, forms the effect of system hydrologic cycle for the fry enters into activity appraisal part once along with the rivers. And when the baffle rises, the function of blocking water flow and forming a counting table of the fry identification part is achieved. In this embodiment, the first movable member may also be an electromagnetic valve or other devices, and in the case of an electromagnetic valve, the first movable member may be disposed on a conduit of the sorting conduit between the first chamber and the primary activity assay site.
In the embodiment, as shown in fig. 3, a first camera device of the counting part is installed on the first cavity, and is used for identifying and counting fish fries passing through the range of the shooting field; the first camera device may be a camera. Specifically, when first adjustable fender descends, the fry gets into with the rivers and selects separately pipeline count part, and first camera device begins work, observes the fry circumstances of flowing into based on the camera to can discern the fry through the camera field of vision through software, realize the count of fry. Based on the real-time counting result of the software, when the number of the fries is less than the preset number, for example, 100, the fries enter a counting part of a sorting pipeline along with water flow and enter a primary activity identification part of the sorting pipeline; based on the real-time counting result of the software, the number of the fries reaches 100, the first movable baffle is driven to ascend through the first motor, the software stops working, and the fry activity evaluation function is carried out.
In this embodiment, the first motor is connected to the main control device through the motor driver, and the main control device controls the operating state of the first motor. First motor and first adjustable fender's be connected the mounting means, can select according to actual environment, for example first adjustable fender one side sets up the rotation axis, and the rotation axis is installed on the output shaft of first motor, and it is rotatory to drive first adjustable fender through the rotation of first motor, realizes the rising and the decline of first adjustable fender. Of course, the specific mounting structure may be other, and is not limited thereto. Hereinafter, the same is true of the installation relationship of each motor with the second to fifth flappers and the mesh flappers.
The first camera device is connected with the main control equipment, and the main control equipment controls the first camera device to start to carry out camera shooting when the first movable baffle plate is controlled to descend by the motor.
As shown in fig. 4, the first activity assay site in this embodiment includes a second chamber on the sorting channel, a second camera disposed on the second chamber, and an air blower disposed in the second chamber; the second chamber may be a box.
A net movable plate controlled to ascend and descend by a sixth motor is arranged at the outlet side of the second cavity, and in the sorting starting stage, the website movable plate is in an ascending state, namely, the fry entering the second cavity cannot pass through the net movable plate, and only water can pass through the net movable plate; a second movable part is arranged on a pipeline of the sorting pipeline connected to the re-identification device, and a third movable part is arranged on a pipeline of the sorting pipeline connected to the high-quality pool; the second movable part and the third movable part respectively control the working state through a second motor and a third motor so as to control the separation pipeline to be communicated to the re-appraisal device or the high-quality pool. In this embodiment, like the first movable member, the second movable member and the third movable member may be movable baffles, corresponding to the second movable baffle and the third movable baffle, or may be electromagnetic valves, and the like, and mainly play a role in communication control.
In this embodiment, when the first movable baffle rises, a preset number of 100 fish fries enter the second cavity of the first activity identification part, the blower is controlled to start to operate, based on the operation of the blower, ripples appear on the water surface of the second cavity, a high-quality fish fry can push the ripples to move, and a poor-quality fish fry can gradually flow along with the ripples; recording a video of the condition of the fry of the second cavity based on a second camera device such as a camera, and performing activity evaluation analysis based on software to obtain a fry activity identification result; when the fish fry is high-quality fish fry, controlling the third movable part to descend like a third movable baffle, controlling the second movable part to ascend like a second movable baffle, and controlling the net baffle to descend so that the fish fry enters a high-quality pool; otherwise, the second movable baffle is controlled to descend, the third movable baffle ascends, and the mesh baffle descends, so that the fry enters the re-identification device.
In this embodiment, the second camera device is arranged on the second cavity, the second camera device is connected with the main control device, the main control device controls the second camera device to start shooting, and the video of the fry status in the shot second cavity is sent to the main control device, and the fry activity is analyzed through software on the main control device.
In this embodiment, the second motor, the third motor and the sixth motor are respectively connected to the main control device through the motor drivers; the air blower is connected with the main control equipment. The main control equipment controls the second movable part, the third movable part and the network movable plate through the motors.
In this embodiment, as shown in fig. 5, the re-identification apparatus includes a re-identification box, an inclined slope pipeline and a third camera device, the sorting pipeline is connected to the high inlet of the inclined slope pipeline, and the low outlet of the inclined slope pipeline is connected to the original pool and the poor pool through pipelines. The slope pipeline passes through the interior of the re-identification box, and the third camera device is arranged at the top of the re-identification box.
The third camera device is arranged above the slope pipeline and used for shooting fry images in the slope pipeline and carrying out secondary activity identification on the fries based on the fry images; the re-identification device of the embodiment utilizes the slope pipeline to create a water pouring environment, the body of the high-quality fry can be severely bent, and the poor-quality fry only shakes head and tail; shooting a picture of the slope fry condition based on a camera, uploading the picture to software for activity secondary evaluation analysis, and obtaining fry activity secondary evaluation; and obtaining a final identification result based on the two fish fry activity evaluation results.
The lower outlet of the slope pipeline is connected to a pipeline of the inferior pool and is provided with a fourth moving part, the lower outlet of the slope pipeline is connected to a pipeline of the original pool and is provided with a fifth moving part, and the fourth moving part and the fifth moving part respectively and correspondingly control the working state through a fourth motor and a fifth motor so as to control the slope pipeline to be communicated to the inferior pool or the original pool. In this embodiment, like the first movable member, the fourth movable member and the fifth movable member may be movable baffles or electromagnetic valves. And when the identification result is the inferior fry, the fourth movable part, namely the fourth movable baffle plate, is controlled to descend, the fifth movable part, namely the fifth movable baffle plate, is controlled to ascend, so that the fry enters the inferior pond, otherwise, the fifth movable baffle plate is controlled to descend, and the fourth movable baffle plate ascends, so that the fry enters the original pond.
In this embodiment, the fourth motor and the fifth motor are respectively connected to the main control device through the motor drivers, and the main control device controls the fourth movable baffle and the fifth movable baffle through the motors. The third camera device is connected with the main control equipment; and controlling the third camera device to start shooting through the main control equipment, acquiring the fry moving images shot by the third camera device, and identifying the fry activity by software according to the fry moving images.
In this embodiment, in the initial state of the entire sorting system: the first movable baffle, the net movable baffle, the third movable baffle and the fourth movable baffle are all in a vertical state, namely, all in a rising state, and the second movable baffle and the fifth movable baffle are all in a horizontal state, namely, all in a falling state. At this time, the water flow in the original pool is in a static state, and no vortex is generated and no flow is generated in the pipeline.
When the separation starts, the first movable baffle is controlled to descend through the first motor, the water flow of the original pool sequentially passes through the separation pipeline, the counting part, the first movable baffle, the first activity identification part, the second movable baffle, the re-identification device and the fifth movable baffle and finally returns to the original pool, and the fry passes through the separation pipeline, the counting part and the first movable baffle and is blocked at the first activity identification part by the net movable plate. And then respectively controlling the states of the second movable baffle and the third movable baffle according to the primary activity identification result, when the primary activity identification result is high-quality fry, controlling the mesh movable plate and the third movable plate to descend, and controlling the second movable plate to ascend, so that the separation pipeline is communicated with the high-quality pool, or controlling the mesh movable plate to descend, so that the separation pipeline is communicated with the re-identification device. And controlling the states of the fourth movable baffle and the fifth movable baffle according to the secondary activity identification result or combining the primary activity identification result and the secondary activity identification result, controlling the re-identification device to be directly communicated to the poor-quality pool based on the fifth movable baffle in the horizontal state when the identification result has poor-quality fry, and otherwise controlling the fourth movable baffle to descend and the fifth movable baffle to ascend so that the re-identification device is communicated to the original pool.
After the sorting system finishes the primary sorting work, no matter the fry flows into an original pond, a high-quality pond or an inferior pond, each movable baffle and the net mass baffle are controlled to return to the initial state, the first movable baffle, the net mass movable plate, the third movable baffle and the fourth movable baffle are all in the vertical state, and the second movable baffle and the fifth movable baffle are all in the horizontal state. This way, the control of each sorting can be made simpler and more convenient. Avoiding the need to consider the status of the network movable plate and each movable baffle after each sorting.
In this embodiment, the master control device connected to each motor, blower and each camera device may be a device having a control function, such as an industrial personal computer, a controller, and the like, and the master control device may be connected to an upper computer, and the upper computer is connected to the master control device in a wired or wireless manner, and the upper computer may be a user mobile phone, a computer, and the like, and the user may send a corresponding instruction to the master control device based on the upper computer app, so as to implement remote control of the sorting system, such as sending a sorting start instruction to the master control device, and after receiving the start instruction sent by the upper computer, the master control device first controls the operation of the rotor at the bottom of the original pool through the motor, and starts the sorting operation.
Example 2
The embodiment discloses a fry sorting method implemented based on the fry sorting system of embodiment 1, which comprises the following steps:
s1, controlling the water in the original pool to generate vortex; in the embodiment, the rotor is arranged at the bottom of the original pool, the rotor operates by controlling the motor of the rotor to work, the operation of the rotor can enable the water in the original pool to generate vortex, and when the vortex is generated in the original pool, the high-quality fry moves upwards based on stimulation, and the poor-quality fry does not have obvious reaction when being stimulated; high-quality fry can move along the edge of the vortex, and poor-quality fry can be involved in the vortex.
S2, when the water in the raw pool is controlled to generate vortex, the first motor is used for controlling the first movable piece to move, for example, the first movable baffle plate descends, so that the water pool in the raw pool enters the sorting pipeline counting part; then, starting counting work, which is concretely as follows:
and S21, controlling the first camera device at the counting part of the sorting pipeline to start shooting control.
S22, recognizing the passing fries in the visual field range shot by the first camera device through the fry recognition model and counting; when the number of the detected fries reaches a first preset value, for example 100 tails, a first movable part of the counting part is controlled to ascend through a first motor; controlling that no more fish fry can enter the first activity identification part.
In this step, the fry identification model construction method is as follows:
s221, shooting a fry photo based on existing hardware, and collecting the fry photo as a training sample;
s222, marking the collected training samples by using labelme software, and setting labels as fries;
s223, constructing a neural network model, and training the neural network model through the labeled training samples to obtain a fry identification model; in this embodiment, the training sample labeled in step S223 is uploaded to a BML object detection function part of a Baidu cloud full-function AI development platform, and a fry identification model is obtained through training.
Wherein, the fry counting is realized by the following steps: the fry is recognized in real time based on the fry recognition model, but when the fry recognition model recognizes the fry in the visual field range of the first camera device, a square frame is arranged around the fry, and the square frame moves along with the movement of the fish. And counting when the fry touches the boundary of the downstream side, which is vertical to the flow direction of the water flow and is positioned at the counting part of the sorting pipeline, in the visual field of the first camera device. In this embodiment, a sensing line is arranged on the rightmost end boundary of the field of view of the first camera device, and when a square frame around a fish touches the sensing line, software records the sensing line once so as to count the number of fries.
And S3, controlling the first activity identification part on the sorting pipeline to perform activity identification work. The specific process is as follows:
and S31, controlling the blower at the first activity identification part to work, enabling the water surface of the second cavity to ripple, and controlling the second camera device at the first activity identification part of the sorting pipeline to record videos of the fry conditions of the second cavity.
And S32, acquiring the fry condition video shot by the second camera device, evaluating the fry activity by the activity evaluation blowing observation model according to the fry condition video, and obtaining a first evaluation result. In this embodiment, based on the operation of the blower, the water surface of the second cavity of the first activity identification part ripples, the high-quality fry can push the ripple motion, the poor-quality fry can flow gradually along with the ripples, the motion state of each fry is shot through the second camera device in this step, and based on this, the activity of the fry can be determined by observing the motion state of the fry in the air blowing environment. In the implementation, the activity evaluation blowing observation model evaluates the activity of the fry according to the fry condition video, so as to obtain an identification result. Tracking and observing the motion state of each fish in the video based on an activity evaluation blowing observation model, and determining whether the fish is a high-quality fry; further, regarding the fish school in the second cavity, the ratio of the number of the high-quality fries to the total number of the fries is used as a final score, for example, the second preset value is set to 85 minutes, but when the final score is more than 85 minutes, the identification result is as follows: high-quality fry.
In this example, the activity evaluation air blowing observation model was constructed as follows:
s321, recording a video of the fry in a blowing environment based on existing hardware, and collecting the video serving as a training sample;
s322, labeling the collected training samples by using labelme labeling software, wherein the labels are high-quality fry and poor-quality fry;
s323, tracking the fry in the training sample to determine the motion state of the fry, wherein the fry pushes ripple motion in a blowing environment or gradually flows along with ripples; and (3) constructing a neural network model, and training the neural network model according to the motion state of the fry of the training sample and the label to obtain an activity evaluation blowing observation model. In this embodiment, based on the video uploading Baidu cloud zero threshold AI development platform easy DL target tracking function part which is labeled in step S322, an activity assessment blowing observation model is obtained through training.
S4, controlling the sorting pipeline to be communicated with a high-quality pool or a re-identification device according to the activity identification result of the first activity identification part; wherein: and when the activity identification result of the first activity identification part is that the fry is high-quality, controlling the separation pipeline to be communicated with the high-quality pool, or controlling the separation pipeline to be communicated with the re-identification device. Specifically, the method comprises the following steps: when a first evaluation result obtained based on the activity evaluation blowing observation model exceeds a second preset value, controlling the sorting channel to be communicated with the high-quality pool, and specifically controlling the mesh baffle to descend, the third movable baffle to descend and the second movable baffle to ascend so that the fry in the second cavity enters the high-quality pool; otherwise, the sorting channel is controlled to be communicated with the re-identification device, the mesh baffle is controlled to descend, the second movable baffle descends and the third movable baffle ascends, so that the fry in the second cavity enters the re-identification device.
S5, controlling the re-identification device to perform secondary activity identification on the fries entering the re-identification device when the sorting pipeline is communicated with the re-identification device; controlling the re-identification device to be communicated with the original pool or the inferior pool according to the secondary activity identification; wherein: and when the secondary activity identification result is that the fry is poor, controlling the re-identification device to be communicated to the poor pond, or controlling the re-identification device to be communicated to the original pond. In this embodiment, the process of controlling the re-identification device to perform the secondary activity identification is as follows:
s51, when the sorting channel is controlled to be communicated with the re-identification device, a third camera in the re-identification device is controlled to shoot a picture of the fry state on the slope pipeline to obtain a fry water pouring state image, the activity of the fry is evaluated according to the fry state image by the activity evaluation water pouring observation model, and a second evaluation result is obtained; and judging whether the fish fries are high-quality fish fries or not by the activity evaluation pouring observation model based on the states of the fish fries in the fish fry state images. And aiming at the fish school entering the re-identification device, taking the ratio of the number of the high-quality fries to the total number of the fries as a final score, and taking the final score as a second evaluation result.
S52, taking the second evaluation result or the average evaluation result as a final evaluation result; wherein the average assessment result is an average of the first assessment result and the second assessment result.
And S53, when the final evaluation result exceeds a third preset value, controlling the re-identification device to be communicated with the original pool, specifically controlling the fourth movable baffle to ascend and the fifth movable baffle to descend to enable the fry to enter the original pool, otherwise controlling the re-identification device to be communicated with the inferior pool, specifically controlling the fourth movable baffle to descend and the fifth movable baffle to ascend to enable the fry to enter the inferior pool.
In the present embodiment, the third preset value is set to 40 minutes. And when the final evaluation result is the average value of the first evaluation result and the second evaluation result, when the final evaluation result is between 40 and 85 points, judging that the fry is a medium-grade fry, and controlling the fry to enter the original pool, and when the final evaluation result is below 40 points, judging that the fry is a poor-grade fry, and controlling the fry to enter the poor-grade pool.
In this embodiment, including the slope pipeline in the heavy identification equipment, utilize the slope to build the environment of pouring water, high-quality fry health can acutely be crooked, and inferior fry only rocks end to end. The activity evaluation pouring observation model identifies the activity of the fry based on the state of the fry in a pouring environment. In this embodiment, the process of constructing the activity evaluation pouring observation model is as follows:
s531, shooting a picture of the fry in a water pouring environment based on existing hardware, and collecting the picture as a training sample;
s532, marking the collected training samples by using labelme software, and setting labels as high-quality fry and poor-quality fry;
s533, obtaining the state of the fry in the water pouring environment in the training sample to construct a neural network model, and training the neural network model according to the state of the fry in the training sample and the label to obtain an activity assessment water pouring observation model. In this embodiment, a fry activity evaluation model is trained based on the BML object detection function part of the Baidu cloud-full-function AI development platform on which the annotated photos are uploaded.
The method can be executed on a main control device such as an industrial personal computer, and the industrial personal computer runs software to realize the steps so as to control the sorting system to realize the sorting of the fry. In addition, the industrial personal computer can be connected with an upper computer in a wired or wireless mode, for example, a user terminal, and the user can start a sorting instruction to the industrial personal computer based on an APP interface on the terminal, so that remote control over the sorting system is realized. Or the software running in the method of the embodiment may be deployed in a cloud, the execution of the software running is realized by accessing the cloud, and at this time, the main control device receives a corresponding control instruction from the cloud, so that the control of each motor and each camera in the sorting system is realized.
Those skilled in the art will appreciate that all or part of the steps in the method according to the present embodiment may be implemented by a program to instruct the relevant hardware, and the corresponding program may be stored in a computer-readable storage medium. It should be noted that although the method operations of embodiment 1 are depicted in the drawings in a particular order, this does not require or imply that these operations must be performed in this particular order, or that all of the illustrated operations must be performed, to achieve desirable results. Rather, the depicted steps may change the order of execution, some steps may be performed concurrently, some steps may additionally or alternatively be omitted, multiple steps may be combined into one step execution, and/or a step may be broken down into multiple step executions.
Example 3
The embodiment discloses a storage medium storing a program, wherein the program is executed by a processor to implement the fry sorting method of embodiment 2, and the method specifically comprises the following steps:
controlling the water in the original pool to generate vortex;
controlling a first activity identification part on the sorting pipeline to perform activity identification work;
controlling the separation pipeline to be communicated with a high-quality pool or a re-identification device according to the activity identification result of the first activity identification part; wherein:
when the activity identification result of the first activity identification part is that the fry is high-quality, controlling the sorting pipeline to be communicated to a high-quality pool, or controlling the sorting pipeline to be communicated to a re-identification device;
when the sorting pipeline is communicated with the re-identification device, controlling the re-identification device to perform secondary activity identification on the fries entering the re-identification device; controlling the re-identification device to be communicated with the original pool or the inferior pool according to the secondary activity identification result or combining the primary activity identification result and the secondary activity identification result; wherein:
and when the fry is inferior in identification result, controlling the re-identification device to be communicated to the inferior pool, otherwise, controlling the re-identification device to be communicated to the original pool.
In this embodiment, specific implementation of each process may refer to embodiment 2, which is not described herein again.
In this embodiment, the storage medium may be a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM), a Random Access Memory (RAM), a usb disk, a removable hard disk, or other media.
Example 4
The embodiment discloses a computing device, which comprises a processor and a memory for storing an executable program of the processor, wherein when the processor executes the program stored in the memory, the fry sorting method of the embodiment 2 is implemented, and the method comprises the following steps:
controlling the water in the original pool to generate vortex;
controlling a first activity identification part on the sorting pipeline to perform activity identification work;
controlling the separation pipeline to be communicated with a high-quality pool or a re-identification device according to the activity identification result of the first activity identification part; wherein:
when the activity identification result of the first activity identification part is that the fry is high-quality, controlling the sorting pipeline to be communicated to a high-quality pool, or controlling the sorting pipeline to be communicated to a re-identification device;
when the sorting pipeline is communicated with the re-identification device, controlling the re-identification device to perform secondary activity identification on the fries entering the re-identification device; controlling the re-identification device to be communicated with the original pool or the inferior pool according to the secondary activity identification result or combining the primary activity identification result and the secondary activity identification result; wherein:
and when the fry is inferior in identification result, controlling the re-identification device to be communicated to the inferior pool, otherwise, controlling the re-identification device to be communicated to the original pool.
In this embodiment, specific implementation of each process may refer to embodiment 2, which is not described herein again.
In this embodiment, the computing device may be a desktop computer, a notebook computer, a PDA handheld terminal, a tablet computer, or other terminal devices.
In this embodiment, the computing device includes: the system comprises a processor, a memory, a bus and a communication interface, wherein the processor, the communication interface and the memory are connected through the bus; the processor is configured to execute an executable module, such as a computer program, stored in the memory.
The Memory may include a high-speed Random Access Memory (RAM) and may also include a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. The communication connection between the network element of the system and at least one other network element is realized through at least one communication interface (which may be wired or wireless), and the internet, a wide area network, a local network, a metropolitan area network and the like can be used.
The bus may be an ISA bus, PCI bus, EISA bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, and the like.
The storage is configured to store a program, and the processor executes the program after receiving an execution instruction, and the method performed by the apparatus defined by the flow program disclosed in the foregoing embodiments of the present application may be applied to or implemented by the processor.
The processor may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware in a processor or instructions in the form of software. The Processor may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; but also Digital Signal Processors (DSP), Application Specific Integrated Circuits (ASIC), Field-Programmable Gate arrays (FPGA) or other Programmable logic devices, discrete Gate or transistor logic devices, and discrete hardware components. The various methods, steps, and logic blocks disclosed in the embodiments of the present application may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present application may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module may be located in a storage medium that is mature in the art, such as a random access memory, a flash memory and/or a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory and/or a register, and the storage medium is located in a memory, and a processor reads information in the memory and completes the steps of the method in combination with hardware thereof.
The above embodiments are preferred embodiments of the present invention, but the present invention is not limited to the above embodiments, and any other changes, modifications, substitutions, combinations, and simplifications which do not depart from the spirit and principle of the present invention should be construed as equivalents thereof, and all such changes, modifications, substitutions, combinations, and simplifications are intended to be included in the scope of the present invention.
Claims (10)
1. A fry sorting system is characterized by comprising an original pool, a high-quality pool, an inferior pool and a re-identification device, wherein the original pool is respectively connected to the high-quality pool and the re-identification device through sorting pipelines;
the sorting pipeline is provided with a first activity identification part for carrying out primary activity identification, and the sorting pipeline is controlled to be communicated with the high-quality pool or the re-identification pipeline device according to an identification result;
the re-identification device is connected to the original pool and the inferior pool through pipelines respectively and is used for carrying out secondary activity identification, and the re-identification device is controlled to be communicated with the original pool or the inferior pool according to an identification result.
2. The fry sorting system of claim 1, wherein the sorting conduit is further provided with a counting section for counting the fry, the counting section being located between the stock tank and the first active identification section;
the counting part comprises a first cavity and a first camera device on the sorting pipeline;
a first movable piece is arranged on one side of an inlet or an outlet of the first cavity and controls the working state through a first motor;
the first camera device is arranged on the first cavity, identifies the passing fries in the range of the shot visual field and counts;
the first motor is connected with the main control equipment through a motor driver;
the first camera device is connected with the main control equipment.
3. The fry sorting system of claim 1, wherein the first active identification site comprises a second cavity on the sorting conduit, a second camera, and an air blower disposed within the second cavity;
a mesh movable plate controlled to ascend and descend by a sixth motor is arranged on the outlet side of the second cavity; a second movable part is arranged on a pipeline of the sorting pipeline connected to the re-identification device, and a third movable part is arranged on a pipeline of the sorting pipeline connected to the high-quality pool; the second movable part and the third movable part respectively control the working state through a second motor and a third motor so as to control the separation pipeline to be communicated to the re-appraisal device or the high-quality pool;
the second camera device is arranged on the second cavity and used for recording a fry movement video in the second cavity and performing primary activity identification on the fry according to the fry image;
the second motor, the third motor and the sixth motor are respectively connected with the main control equipment through motor drivers; the air blower is connected with the main control equipment;
the second camera device is connected with the main control equipment.
4. The fry sorting system of claim 1, wherein the re-identification device comprises an inclined slope pipeline and a third camera device, the inclined slope pipeline is connected to an inlet at the high position of the inclined slope pipeline through a pipeline, and an outlet at the low position of the inclined slope pipeline is respectively connected to the original pool and the poor pool through pipelines;
the third camera device is arranged above the slope pipeline and used for shooting a fry moving image in the slope pipeline and carrying out secondary activity identification on the fry based on the fry moving image;
a fourth movable part is arranged on a pipeline, connected with the inferior pool, of the outlet at the lower part of the slope pipeline, a fifth movable part is arranged on a pipeline, connected with the original pool, of the outlet at the lower part of the slope pipeline, and the fourth movable part and the fifth movable part respectively control working states correspondingly through a fourth motor and a fifth motor so as to control the slope pipeline to be communicated to the inferior pool or the original pool;
the fourth motor and the fifth motor are respectively connected with the master control equipment through motor drivers;
the third camera device is connected with the main control equipment.
5. A fry sorting method realized based on the fry sorting system of any one of claims 1 to 4, characterized by comprising the following steps:
controlling the water in the original pool to generate vortex;
controlling a first activity identification part on the sorting pipeline to perform activity identification work;
controlling the separation pipeline to be communicated with a high-quality pool or a re-identification device according to the activity identification result of the first activity identification part; wherein:
when the activity identification result of the first activity identification part is that the fry is high-quality, controlling the sorting pipeline to be communicated to a high-quality pool, or controlling the sorting pipeline to be communicated to a re-identification device;
when the sorting pipeline is communicated with the re-identification device, controlling the re-identification device to perform secondary activity identification on the fries entering the re-identification device; controlling the re-identification device to be communicated with the original pool or the inferior pool according to the secondary activity identification result or combining the primary activity identification result and the secondary activity identification result; wherein:
and when the fry is inferior in identification result, controlling the re-identification device to be communicated to the inferior pool, otherwise, controlling the re-identification device to be communicated to the original pool.
6. The fry sorting method of claim 5, further comprising the steps of:
after the water in the original pool is controlled to generate vortex, the first movable piece of the counting part is controlled by the first motor to act, so that the water pool in the original pool enters the counting part of the sorting pipeline and further reaches the first activity identification part;
controlling a first camera device at the counting part of the sorting pipeline to start shooting work;
recognizing the passing fries in the visual field range shot by the first camera device through a fry recognition model and counting;
when the number of the detected fries reaches a first preset value, the state of a first moving part of a counting part is controlled through a first motor, and the fries in the original pond are prevented from entering a first activity identification part;
controlling a blower of the first activity identification part to work, enabling ripples to appear on the water surface of the second cavity, and simultaneously controlling a second camera device of the first activity identification part of the sorting pipeline to record videos of the conditions of the second cavity type fries;
acquiring a fry condition video shot by a second camera device, judging and evaluating fry activity by an activity evaluation blowing observation model according to the fry condition video, and obtaining a first evaluation result;
when the first evaluation result exceeds a second preset value, controlling the sorting channel to be communicated with the high-quality pool, so that the fry in the second cavity enters the high-quality pool; otherwise, controlling the sorting channel to be communicated with the re-identification device, and enabling the fry in the second cavity to enter the re-identification device.
7. The fry sorting method according to claim 6, wherein the process of controlling the re-identification device to perform the secondary activity identification is as follows:
when the separation channel is controlled to be communicated with the re-identification device, a third camera device in the re-identification device is controlled to shoot a picture of the fry state on the slope pipeline to obtain a fry pouring state image, an activity evaluation pouring observation model is used for evaluating the activity of the fry according to the fry state image, and a second evaluation result is obtained;
taking the second evaluation result or the average evaluation result as a final evaluation result; wherein the average assessment result is an average of the first assessment result and the second assessment result;
and when the final evaluation result exceeds a third preset value, controlling the re-identification device to be communicated with the original pool to enable the fry to enter the original pool, otherwise, controlling the re-identification device to be communicated with the inferior pool to enable the fry to enter the inferior pool.
8. The fry sorting method according to claim 7, wherein the fry identification model is constructed by the following steps:
shooting a fry photo based on existing hardware and collecting the fry photo as a training sample;
marking the collected training samples, and setting the labels as fry;
constructing a neural network model, and training the neural network model through the labeled training samples to obtain a fry identification model;
the fry identification method comprises the steps that fry identification is carried out in real time on the basis of a fry identification model, and when the fry touches a boundary of a downstream side, which is vertical to the flow direction of water flow and is positioned at a counting part of a sorting pipeline, in a visual field of a first camera device, the fry is counted;
the construction process of the activity evaluation blowing observation model is as follows:
recording a video of the fry in a blowing environment based on the existing hardware, and collecting the video as a training sample;
marking the collected training samples, wherein the labels are high-quality fry and poor-quality fry;
tracking the fry in the training sample to determine the motion state of the fry;
constructing a neural network model, and training the neural network model according to the motion state of the fry of the training sample and the label to obtain an activity evaluation blowing observation model;
evaluating the activity of the fry in the fry condition video shot by the second camera device based on an activity evaluation blowing observation model, and taking the ratio of the number of the high-quality fries to the total number of the fries as an evaluation result;
the construction process of the activity evaluation pouring water observation model is as follows:
shooting a picture of the fry in a water pouring environment based on the existing hardware, and collecting the picture as a training sample;
marking the collected training samples, and setting the labels as high-quality fry and poor-quality fry;
acquiring the state of the fry in a water pouring environment in a training sample;
and (3) constructing a neural network model, and training the neural network model according to the red fry state and the label of the training sample to obtain an activity evaluation pouring observation model.
9. A storage medium storing a program, wherein the program, when executed by a processor, implements the fry sorting method of any one of claims 5 to 8.
10. A computing device comprising a processor and a memory for storing processor-executable programs, wherein the processor, when executing a program stored in the memory, implements the fry sorting method of any one of claims 5 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210226859.9A CN114586728B (en) | 2022-03-08 | 2022-03-08 | Fry sorting system and sorting method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210226859.9A CN114586728B (en) | 2022-03-08 | 2022-03-08 | Fry sorting system and sorting method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114586728A true CN114586728A (en) | 2022-06-07 |
| CN114586728B CN114586728B (en) | 2022-12-20 |
Family
ID=81809695
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210226859.9A Active CN114586728B (en) | 2022-03-08 | 2022-03-08 | Fry sorting system and sorting method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114586728B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115624008A (en) * | 2022-07-08 | 2023-01-20 | 珠海科艺普检测科技有限公司 | Intelligent fry detection method for micropterus salmoides based on biological information technology |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201204865Y (en) * | 2008-04-17 | 2009-03-11 | 上海水产大学 | Endless track for fish offspring quality quick classification |
| JP2015104379A (en) * | 2013-12-03 | 2015-06-08 | 株式会社光コーポレーション | Breed discrimination device and breed discrimination method |
| CN105933652A (en) * | 2016-05-09 | 2016-09-07 | 江苏大学 | Apparatus and method for detecting sturgeon activity based on image identifying and positioning |
| WO2018011745A1 (en) * | 2016-07-13 | 2018-01-18 | Biosort As | Method and system for sorting live fish |
| CN110419484A (en) * | 2019-08-27 | 2019-11-08 | 弓可飞 | Fancy carp automates separation system |
| CN111275656A (en) * | 2018-11-16 | 2020-06-12 | 绍兴图聚光电科技有限公司 | Intelligent fish and shrimp fry counting device |
| CN211185413U (en) * | 2019-12-09 | 2020-08-07 | 广东香良水产有限公司 | Fry vitality detection equipment for industrial culture |
| CN213153509U (en) * | 2020-09-09 | 2021-05-11 | 宁波大学 | Separator of good and bad embryonated egg of silvery pomfret |
| US20210368747A1 (en) * | 2020-05-28 | 2021-12-02 | X Development Llc | Analysis and sorting in aquaculture |
| CN113793350A (en) * | 2021-09-24 | 2021-12-14 | 安徽工大信息技术有限公司 | Fry counting Internet of things device and fry condition statistical method |
-
2022
- 2022-03-08 CN CN202210226859.9A patent/CN114586728B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201204865Y (en) * | 2008-04-17 | 2009-03-11 | 上海水产大学 | Endless track for fish offspring quality quick classification |
| JP2015104379A (en) * | 2013-12-03 | 2015-06-08 | 株式会社光コーポレーション | Breed discrimination device and breed discrimination method |
| CN105933652A (en) * | 2016-05-09 | 2016-09-07 | 江苏大学 | Apparatus and method for detecting sturgeon activity based on image identifying and positioning |
| WO2018011745A1 (en) * | 2016-07-13 | 2018-01-18 | Biosort As | Method and system for sorting live fish |
| CN111275656A (en) * | 2018-11-16 | 2020-06-12 | 绍兴图聚光电科技有限公司 | Intelligent fish and shrimp fry counting device |
| CN110419484A (en) * | 2019-08-27 | 2019-11-08 | 弓可飞 | Fancy carp automates separation system |
| CN211185413U (en) * | 2019-12-09 | 2020-08-07 | 广东香良水产有限公司 | Fry vitality detection equipment for industrial culture |
| US20210368747A1 (en) * | 2020-05-28 | 2021-12-02 | X Development Llc | Analysis and sorting in aquaculture |
| CN213153509U (en) * | 2020-09-09 | 2021-05-11 | 宁波大学 | Separator of good and bad embryonated egg of silvery pomfret |
| CN113793350A (en) * | 2021-09-24 | 2021-12-14 | 安徽工大信息技术有限公司 | Fry counting Internet of things device and fry condition statistical method |
Non-Patent Citations (1)
| Title |
|---|
| 杜维军: "鉴别鱼苗种质量的方法", 《吉林农业》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115624008A (en) * | 2022-07-08 | 2023-01-20 | 珠海科艺普检测科技有限公司 | Intelligent fry detection method for micropterus salmoides based on biological information technology |
| CN115624008B (en) * | 2022-07-08 | 2023-09-05 | 珠海科艺普检测科技有限公司 | Intelligent detection method for fries of largemouth black weever based on biological information technology |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114586728B (en) | 2022-12-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106733701B (en) | Potato multi-stage combination dysnusia detecting system and detection method | |
| CN109657716B (en) | Vehicle appearance damage identification method based on deep learning | |
| CN109460753B (en) | Method for detecting floating object on water | |
| CN113971779B (en) | Water gauge automatic reading method based on deep learning | |
| CN103801520B (en) | The automatic well-chosen stage division of shrimp and device | |
| CN101456501B (en) | Method and apparatus for controlling elevator button | |
| CN114586728B (en) | Fry sorting system and sorting method | |
| CN109300111A (en) | A kind of chromosome recognition methods based on deep learning | |
| CN111339891A (en) | Target detection method of image data and related device | |
| CN103344583B (en) | A kind of praseodymium-neodymium (Pr/Nd) component concentration detection system based on machine vision and method | |
| CN109886928A (en) | A kind of target cell labeling method, device, storage medium and terminal device | |
| CN111161214B (en) | System and method for measuring pig weight and identifying drinking behavior based on binocular vision | |
| CN108229483A (en) | Based on the doorplate pressed characters identification device under caffe and soft triggering | |
| CN107330403A (en) | A kind of yak method of counting based on video data | |
| CN103177266A (en) | Intelligent stock pest identification system | |
| CN109961030A (en) | Pavement patching information detecting method, device, equipment and storage medium | |
| CN107292302A (en) | Detect the method and system of point of interest in picture | |
| CN112200011A (en) | Aeration tank state detection method and system, electronic equipment and storage medium | |
| CN111914657B (en) | Pet behavior detection method and device, electronic equipment and storage medium | |
| CN109876569A (en) | A kind of bag filter purification gas index on-line measuring device, system and method | |
| CN107273815A (en) | A kind of individual behavior recognition methods and system | |
| CN103646251B (en) | Apple postharvest field classification detection method and system based on embedded technology | |
| CN114359199A (en) | Fish counting method, device, equipment and medium based on deep learning | |
| CN117746077B (en) | Chip defect detection method, device, equipment and storage medium | |
| CN107292222A (en) | A kind of vehicle peccancy detection method and device |
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 |