CN117036921A - Method, device, equipment and storage medium for presuming number of cultured life bodies - Google Patents

Method, device, equipment and storage medium for presuming number of cultured life bodies Download PDF

Info

Publication number
CN117036921A
CN117036921A CN202311011575.9A CN202311011575A CN117036921A CN 117036921 A CN117036921 A CN 117036921A CN 202311011575 A CN202311011575 A CN 202311011575A CN 117036921 A CN117036921 A CN 117036921A
Authority
CN
China
Prior art keywords
cultured
different periods
water environment
remote sensing
life
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
Application number
CN202311011575.9A
Other languages
Chinese (zh)
Inventor
郑徵羽
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hanning Remote Sensing Technology Research Institute Nanjing Co ltd
Original Assignee
Hanning Remote Sensing Technology Research Institute Nanjing Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Hanning Remote Sensing Technology Research Institute Nanjing Co ltd filed Critical Hanning Remote Sensing Technology Research Institute Nanjing Co ltd
Priority to CN202311011575.9A priority Critical patent/CN117036921A/en
Publication of CN117036921A publication Critical patent/CN117036921A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/05—Underwater scenes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/20—Ensemble learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/01—Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/766—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using regression, e.g. by projecting features on hyperplanes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
    • G06V10/7747—Organisation of the process, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/776—Validation; Performance evaluation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G06V20/13—Satellite images
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G06V20/17—Terrestrial scenes taken from planes or by drones
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G06V20/194—Terrestrial scenes using hyperspectral data, i.e. more or other wavelengths than RGB

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Computing Systems (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Health & Medical Sciences (AREA)
  • Databases & Information Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Remote Sensing (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Astronomy & Astrophysics (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • Computational Linguistics (AREA)
  • Measuring Or Testing Involving Enzymes Or Micro-Organisms (AREA)

Abstract

The application relates to a method, a device, equipment and a storage medium for presuming the number of cultured life bodies, wherein the method comprises the following steps: remote sensing data acquisition is carried out on the sample culture pond mouth with a preset acquisition period, so that a plurality of remote sensing data in different periods are obtained; preprocessing the remote sensing data in different periods and inverting the remote sensing water environment to obtain water environment parameters in different periods; training and optimizing the established initial number of cultured life bodies presumption model according to the water environment parameters in different periods to obtain a target number of cultured life bodies presumption model; and inputting remote sensing data of different periods of the pool mouth to be presumed into the target aquaculture life body quantity presumption model to presume the quantity of the aquaculture life bodies under water. According to the application, the relationship between the water environmental parameters and the number of the cultured life bodies is analyzed through the remote sensing data, the number of the cultured life bodies is estimated, the influence of the severe underwater imaging environment on the estimation is eliminated, the cultured life bodies are not tracked, and the influence of the movement of the cultured life bodies is avoided.

Description

Method, device, equipment and storage medium for presuming number of cultured life bodies
Technical Field
The application relates to the technical field of aquaculture, in particular to a method, a device, equipment and a storage medium for estimating the number of cultured life bodies.
Background
In the aquatic product cultivation process, the feeding cost of the aquatic products accounts for the main part of the cultivation cost, timely grasping the growth and survival quantity of the aquatic products is a basis for realizing accurate feeding and scientific cultivation, and determining the quantity of the aquatic products is beneficial to controlling the feeding cost, so that the method is a key point for improving the cultivation benefit for determining the quantity of the underwater cultivation life bodies.
In the prior art, in the aquaculture process, the number of underwater aquaculture life bodies is mainly determined by adopting an underwater camera video monitoring mode, underwater environment videos are shot through an underwater camera, and then the shot underwater videos are analyzed, so that the number of target underwater aquaculture life bodies is estimated.
However, the method is limited by the poor visual imaging effect under water, and is very difficult to detect, identify and extract the objects of the underwater cultured life bodies, and in addition, the mouth area of a conventional cultured pond is relatively large, and the cultured life bodies can always move under water during the growth period, so that the method for extracting the underwater cultured life bodies one by one through a camera to obtain the number is difficult to be practically feasible at present.
Disclosure of Invention
In view of the foregoing, it is necessary to provide a method, apparatus, device and storage medium for estimating the number of cultured life bodies, which are used for solving the problems that the method for estimating the number of cultured life bodies by video monitoring with an underwater camera in the prior art is difficult to determine the number of cultured life bodies under water due to poor underwater visual imaging effect, and the cultured life bodies often move under water during growth, and the number of cultured life bodies under water cannot be determined by tracking one by one with the camera.
In order to achieve the technical purpose, the application adopts the following technical scheme:
in a first aspect, the present application provides a method for estimating the number of farmed life bodies, comprising:
remote sensing data acquisition is carried out on the sample culture pond mouth with a preset acquisition period, so that a plurality of remote sensing data in different periods are obtained;
preprocessing a plurality of remote sensing data in different periods and inverting the remote sensing water environment to obtain water environment parameters in different periods;
training and optimizing the established initial number of cultured life body presumption models according to water environment parameters in different periods to obtain target number of cultured life body presumption models;
and inputting water environment data of the mouth of the pond to be presumed into a target cultured life body quantity presumption model to presume the quantity of the cultured life bodies under water.
In some possible implementations, preprocessing and remote sensing water environment inversion are performed on multiple remote sensing data in different periods to obtain water environment parameters in different periods, including:
removing interference data in multiple remote sensing data in different periods to obtain multiple water body data in different periods;
determining a remote sensing water environment inversion formula, and carrying out water environment inversion on a plurality of pieces of water body data in different periods to obtain water environment parameters in different periods.
In some possible implementations, removing interference data in multiple remote sensing data of different periods to obtain multiple water body data of different periods includes:
correcting the remote sensing data in different periods to obtain undistorted remote sensing data in different periods;
carrying out radiation treatment on a plurality of remote sensing data without distortion in different periods to obtain a plurality of radiation calibrated data in different periods;
and screening the data calibrated by the plurality of radiation in different periods to obtain the data of the plurality of water bodies in different periods.
In some possible implementations, determining a remote sensing water environment inversion formula, performing water environment inversion on a plurality of pieces of water body data in different periods to obtain water environment parameters in different periods, including:
establishing a mathematical relationship between water body data and water environment parameters;
fitting the mathematical relationship to obtain a remote sensing water environment inversion formula;
and carrying out water environment inversion on the multiple water body data in different periods according to the remote sensing water environment inversion formula to obtain water environment parameters in different periods.
In some possible implementations, training and optimizing the established initial number of cultured life body estimation model according to water environment parameters of different periods to obtain a target number of cultured life body estimation model, including:
establishing a water environment database, and storing water environment parameters in different periods into the water environment database;
establishing an initial breeding life body quantity presumption model;
training and optimizing the initial culture life body quantity presumption model through the water environment database to obtain a transition culture life body quantity presumption model;
and evaluating the number estimation model of the transitional cultured living bodies to obtain a number estimation model of the target cultured living bodies meeting the preset precision requirement.
In some possible implementations, training and optimizing the initial number of cultured life bodies presumption model through the water environment database to obtain a transitional number of cultured life bodies presumption model includes:
extracting a sample set from a water environment database, and dividing the sample set into a training set and a testing set according to a preset proportion;
and repeatedly training the initial cultured life body number presumption model by using the training set and a preset training algorithm to obtain a transitional cultured life body number presumption model.
In some possible implementations, evaluating the transitional cultured life body number estimation model to obtain a target cultured life body number estimation model meeting a preset precision requirement includes:
inputting the test set into a transition breeding life body quantity presumption model for testing to obtain a test result;
evaluating the test result and the number of the actual cultured life bodies by using preset evaluation parameters, and determining the prediction precision;
when the prediction precision meets the preset precision requirement, the transitional cultured life body quantity presumption model is a target cultured life body quantity presumption model.
In a second aspect, the present application also provides a device for estimating the number of cultured living bodies, comprising:
the collection module is used for carrying out remote sensing data collection on the sample culture pond mouth in a preset collection period to obtain a plurality of remote sensing data in different periods;
the inversion module is used for preprocessing a plurality of remote sensing data in different periods and inverting the remote sensing water environment to obtain water environment parameters in different periods;
the training module is used for training and optimizing the established initial number of cultured life body presumption models according to the water environment parameters in different periods to obtain target number of cultured life body presumption models;
the presumption module is used for inputting water environment data of the pond mouth to be presumed into the number presumption model of the target cultured life bodies and presuming the number of the submerged cultured life bodies.
In a third aspect, the present application also provides a presumption device for the number of farmed living being, comprising a memory and a processor, wherein,
a memory for storing a program;
and a processor coupled to the memory for executing the program stored in the memory to implement the steps in the method of estimating the number of farmed life bodies in any of the implementations described above.
In a fourth aspect, the present application also provides a computer-readable storage medium storing a computer-readable program or instruction, which when executed by a processor, is capable of implementing the steps in the method for estimating the number of farmed life bodies in any of the above implementations.
The beneficial effects of adopting the embodiment are as follows: the application relates to a method, a device, equipment and a storage medium for presuming the number of cultured life bodies, wherein the method comprises the following steps: remote sensing data acquisition is carried out on the sample culture pond mouth with a preset acquisition period, so that a plurality of remote sensing data in different periods are obtained; preprocessing the remote sensing data in different periods and inverting the remote sensing water environment to obtain water environment parameters in different periods; training and optimizing the established initial number of cultured life bodies presumption model according to the water environment parameters in different periods to obtain a target number of cultured life bodies presumption model; and inputting water environment data of the mouth of the pond to be presumed into the target cultured life body quantity presumption model to presume the quantity of the cultured life bodies under water. According to the application, the water environment parameters are determined through remote sensing water environment inversion by collecting a plurality of remote sensing data of the sample culture pond mouth in different periods, then a model is established to analyze the relation between the water environment parameters and the number of the cultured life bodies, finally the number of the underwater cultured life bodies is speculated, the number of the underwater activity can be speculated only according to the collected plurality of remote sensing data of the sample culture pond mouth, the underwater environment is not required to be photographed, the cultured life bodies are not required to be tracked, the influence of the severe underwater imaging environment on the number speculation of the underwater cultured life bodies is eliminated, the speculation result is not influenced by the movement of the underwater cultured life bodies, and the accuracy of the speculation result is high.
Drawings
FIG. 1 is a schematic flow chart of an embodiment of a method for estimating the number of cultured living bodies according to the present application;
FIG. 2 is a flowchart illustrating an embodiment of removing interference data from remote sensing data according to the present application;
FIG. 3 is a schematic flow chart of an embodiment of the inversion of water environment provided by the present application;
FIG. 4 is a schematic flow chart of an embodiment of training and optimizing an initial number of cultured life models according to the present application;
FIG. 5 is a schematic flow chart of an embodiment of evaluating a model for estimating the number of transitional cultured life bodies according to the present application;
FIG. 6 is a schematic diagram of an embodiment of a device for estimating the number of cultured living things according to the present application;
fig. 7 is a schematic structural diagram of a device for estimating the number of cultured living bodies according to an embodiment of the present application.
Detailed Description
The following detailed description of preferred embodiments of the application is made in connection with the accompanying drawings, which form a part hereof, and together with the description of the embodiments of the application, are used to explain the principles of the application and are not intended to limit the scope of the application.
In the description of the present application, the meaning of "a plurality" is two or more, unless explicitly defined otherwise.
Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the application. The appearances of such phrases in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Those of skill in the art will explicitly and implicitly appreciate that the embodiments described herein may be combined with other embodiments.
The application provides a method, a device, equipment and a storage medium for estimating the number of cultured life bodies, which are respectively described below.
Referring to fig. 1, fig. 1 is a flow chart of an embodiment of a method for estimating the number of cultured living bodies according to the present application, and in one embodiment of the present application, a method for estimating the number of cultured living bodies is disclosed, including:
s101, carrying out remote sensing data acquisition on a sample culture pond mouth in a preset acquisition period to obtain a plurality of remote sensing data in different periods;
s102, preprocessing a plurality of remote sensing data in different periods and inverting the remote sensing water environment to obtain water environment parameters in different periods;
s103, training and optimizing the established initial number of cultured life body presumption models according to water environment parameters in different periods to obtain target number of cultured life body presumption models;
s104, inputting water environment data of the pond mouth to be presumed into a target cultured life body quantity presumption model, and presuming the quantity of the submerged cultured life bodies.
In the above embodiment, the sample culture pond mouths are selected firstly, the size and the range of the sample culture pond mouths are determined, the size, the number and the number of seedlings of the pond mouths need to be considered during selection, the sample culture pond mouths are selected for obtaining initial data, so that the water environment parameters in each sample culture pond mouths are known, and the relationship between the model simulated water environment parameters and the number of underwater culture life bodies is established.
The method for acquiring the remote sensing data can be satellite-borne remote sensing or unmanned plane remote sensing, can be specifically selected according to the size and the range of the sample culture pond mouth, has the advantages that the global range can be covered, the data with a large area can be acquired by one-time imaging, and has great advantages for large-scale and macroscopic application scenes; the unmanned aerial vehicle remote sensing coverage area is smaller, the imaging spatial resolution and the imaging time resolution are easier to control, and meanwhile, the imaging instrument is convenient to replace and install, so that the unmanned aerial vehicle remote sensing system has great advantages for small-scale and detailed application scenes. As a preferred embodiment, the data acquisition range of the sample culture pond mouth is relatively smaller, and the requirements on time and space resolution are higher, so that the unmanned aerial vehicle is more suitable for acquiring the data.
It should be noted that, the shorter the acquisition period, the larger the data volume obtained, and the higher the model accuracy established in theory, the preset acquisition period in the application is 24 hours, so that multiple remote sensing data of different periods of the sample culture pond mouth can be obtained, and the preset acquisition period can be set according to the actual situation, which is not limited by the application. According to the growth process of the underwater cultured life body, multispectral remote sensing images or hyperspectral remote sensing images in different periods are continuously acquired and used for recording the change of the water environment in the growth process of the underwater cultured life body.
The method is characterized in that a plurality of remote sensing data in different periods are initial data, interference exists, the remote sensing data are multispectral remote sensing images or hyperspectral remote sensing images, direct correlation with the number of underwater cultured life bodies is difficult, the interference needs to be removed through preprocessing and remote sensing water environment inversion, and the parameters of the water environment are determined by converting the interference into a unified reference system.
The initial aquaculture life body number presumption model established in the application is based on machine learning and deep learning technology, analyzes water environment data, and establishes water environment and aquaculture life body number change models of different numbers of underwater aquaculture life bodies based on remote sensing inversion according to actual data of underwater aquaculture life body fishing.
Finally, remote sensing data acquisition is carried out on the to-be-speculated pool opening, the acquired remote sensing data are preprocessed and the remote sensing water environment is inverted to obtain water environment parameters, the water environment parameters of the to-be-speculated pool opening are input into the obtained target aquaculture life body quantity speculated model, and the number of the speculated underwater aquaculture life bodies can be output.
Compared with the prior art, the method for estimating the number of the cultured living bodies provided by the embodiment comprises the following steps: remote sensing data acquisition is carried out on the sample culture pond mouth with a preset acquisition period, so that a plurality of remote sensing data in different periods are obtained; preprocessing the remote sensing data in different periods and inverting the remote sensing water environment to obtain water environment parameters in different periods; training and optimizing the established initial number of cultured life bodies presumption model according to the water environment parameters in different periods to obtain a target number of cultured life bodies presumption model; and inputting water environment data of the mouth of the pond to be presumed into the target cultured life body quantity presumption model to presume the quantity of the cultured life bodies under water. According to the application, the water environment parameters are determined through remote sensing water environment inversion by collecting a plurality of remote sensing data of the sample culture pond mouth in different periods, then a model is established to analyze the relation between the water environment parameters and the number of the cultured life bodies, finally the number of the underwater cultured life bodies is speculated, the number of the underwater activity can be speculated only according to the collected plurality of remote sensing data of the sample culture pond mouth, the underwater environment is not required to be photographed, the cultured life bodies are not required to be tracked, the influence of the severe underwater imaging environment on the number speculation of the underwater cultured life bodies is eliminated, the speculation result is not influenced by the movement of the underwater cultured life bodies, and the accuracy of the speculation result is high.
Because the plurality of remote sensing data acquired in step S101 includes other interferents such as water surface vegetation and inter-pond channels in addition to the water body data, in order to avoid the technical problem that the obtained water environment parameters are inaccurate due to the existence of the interferents, and thus the number of the estimated underwater cultured living bodies is inaccurate, in some embodiments of the present application, step S102 includes:
removing interference data in multiple remote sensing data in different periods to obtain multiple water body data in different periods;
determining a remote sensing water environment inversion formula, and carrying out water environment inversion on a plurality of pieces of water body data in different periods to obtain water environment parameters in different periods.
In the embodiment, the interference data in the remote sensing data are removed through the preset processing method, so that the influence of ground features irrelevant to inversion of the water environment, such as water surface vegetation, channels between ponds and other targets on the obtained water environment parameters is avoided, and the accuracy of the number of the estimated underwater cultured life bodies is improved.
It should be noted that: in order to avoid that interference data in the remote sensing data in different periods are removed, inversion inaccuracy is caused by too small data volume of the water body data in different periods, in some embodiments of the application, interpolation fitting can be performed on the notch after the data volume of the water body data in different periods is obtained, so that the water body data in different periods without relative interference is obtained.
The formula for calculating the water environment parameters by utilizing the relatively undisturbed multiple pieces of water body data in different periods is generally based on an empirical linear or nonlinear regression model, and the multiple pieces of water body data in different periods can be converted into the water environment parameters in the same dimension by carrying out water environment inversion on the multiple pieces of water body data in different periods so as to analyze the number of underwater culture life bodies.
Referring to fig. 2, fig. 2 is a flow chart of an embodiment of removing interference data from remote sensing data provided by the present application, because the remote sensing data is a multispectral remote sensing image or a hyperspectral remote sensing image, after removing the interference data, it is difficult to directly determine water environment parameters from the remote sensing data, so that the remote sensing data needs to be converted into a unified reference system, in some embodiments of the present application, removing the interference data from multiple remote sensing data in different periods to obtain multiple water body data in different periods, including:
s201, correcting the remote sensing data in different periods to obtain undistorted remote sensing data in different periods;
s202, carrying out radiation treatment on a plurality of remote sensing data without distortion in different periods to obtain a plurality of radiation calibrated data in different periods;
and S203, screening the data subjected to the radiation calibration in different periods to obtain the data of the water bodies in different periods.
In the above embodiment, the correction mode adopted in the present application is geometric correction, and the geometric correction is to eliminate or reduce the distortion of the hyperspectral image in the spatial position by using the control point or the reference image, so that the image is matched with the geographic coordinates or other images. Geometric correction methods typically use nearest neighbor interpolation, bilinear interpolation, cubic convolution interpolation, and the like. The formula for geometric correction is typically an affine transformation, which can be expressed as:
where (x, y) is the coordinates on the original image, (x ', y') is the coordinates on the corrected image, a, b, c, d, e, f are transformation parameters, which can be solved by the least squares method of the control points.
The radiation mode adopted by the application is radiometric calibration, and the radiometric calibration refers to converting the digital numerical value of the hyperspectral image into a physical quantity, such as radiance or radiation reflectivity (remote sensing reflectivity), by using the response function of a sensor or a radiation value measured in the field. The radiometric scaling method is usually relative scaling, absolute scaling, empirical scaling, the radiometric scaling formula is generally a linear transformation, which can be expressed as:
L＝G×DN+B；
where L is the radiance value, DN is the digital value, G is the gain factor, and B is the offset factor, these parameters being obtained by the response function of the sensor or the radiance value measured in the field.
Geometric distortion and radiation difference between different images can be eliminated through geometric correction and radiation calibration, so that comparability and fusibility of the images are improved, and a unified reference system is established.
When the obtained radiation parameters exceed the set values of the application, the radiation parameters are considered to be meaningless and should be removed, so that the screening of the radiation parameters in different periods is realized, and the data of the water bodies in different periods are obtained.
Referring to fig. 3, fig. 3 is a schematic flow chart of an embodiment of water environment inversion provided by the present application, in which although the obtained water body data is in a unified reference system, in order to accurately obtain water environment parameters from the water body data, the water environment inversion needs to be performed on the water body data, in some embodiments of the present application, a remote sensing water environment inversion formula is determined, and water environment inversion is performed on multiple pieces of water body data in different periods to obtain water environment parameters in different periods, including:
s301, establishing a mathematical relationship between water body data and water environment parameters;
s302, fitting the logarithmic relation to obtain a remote sensing water environment inversion formula;
s303, carrying out water environment inversion on a plurality of pieces of water body data in different periods according to a remote sensing water environment inversion formula to obtain water environment parameters in different periods.
In the above embodiment, the wavelength or the remote sensing reflectivity is used as an independent variable and the water environment parameter is used as an independent variable, a mathematical relationship is established, an approximate calculation formula is obtained by fitting, and then inversion of the water environment parameter is performed, as shown in table 1:
TABLE 1
Independent variable x is a plurality of pieces of water body data in different periods, and dependent variable y is a water environment parameter obtained by collecting water body data corresponding to a sample culture pond mouth and performing laboratory treatment.
It should be noted that the present application may also utilize deep learning or machine learning techniques, such as neural networks, support vector machines, random forests, etc., to establish a nonlinear mapping relationship between the water environment parameters and the remote sensing reflectivity, and perform estimation of the water environment parameters. This approach can deal with complex non-linearity problems, but requires a large amount of training data.
Referring to fig. 4, fig. 4 is a schematic flow chart of an embodiment of training and optimizing an initial number of cultured life bodies according to the present application, wherein in order to establish a relationship between a water environment parameter and a number of underwater cultured life bodies, a presumption model needs to be established, and a number of different periods of water environment parameters and an actual number of cultured life bodies are simulated, and in some embodiments of the present application, training and optimizing the established initial number of cultured life bodies presumption model according to different periods of water environment parameters to obtain a target number of cultured life bodies presumption model, including:
s401, establishing a water environment database, and storing water environment parameters in different periods into the water environment database;
s402, establishing an initial breeding life body quantity presumption model;
s403, training and optimizing the initial culture life body quantity estimation model through the water environment database to obtain a transition culture life body quantity estimation model;
s404, evaluating the number estimation model of the transitional cultured living bodies to obtain a target number estimation model of the cultured living bodies meeting the preset precision requirement.
In the above embodiment, a large number of water environment parameters in different periods are required for training the initial number of cultured life body estimation model, and the water environment database is established to store the water environment parameters in different periods of the plurality of sample cultured pond mouths, so that the water environment parameters can be directly called from the water environment database when the initial number of cultured life body estimation model is trained.
The preset regression method is a gradient lifting regression tree (Gradient Boosting Regression Trees, GBRT), the gradient lifting regression is a regression method with good generalization capability, the regression method consists of a plurality of decision trees, and the output results of all the trees are accumulated to be the final result. And establishing an initial breeding life body quantity presumption model through the gradient lifting regression tree.
The established initial number of cultured life bodies presumption model cannot meet the requirement on the number of the underwater cultured life bodies presumption, the presumption accuracy is improved through a large number of repeated training optimization, and the transitional cultured life body number presumption model is obtained at the moment.
The number of the life bodies in the transition culture can be estimated by the number estimation model of the life bodies in the underwater culture to a certain extent, but the accuracy of the estimation result is further required to be estimated, the estimation result is estimated by a preset evaluation function, whether the prediction accuracy meets the standard is judged according to the result, and whether the model is required to be optimized and improved is further determined.
The initial number of farmed life bodies presumption model is not accurate in accuracy of number of farmed life bodies presumption, a great amount of training is needed to improve the presumption accuracy, so that the number of underwater farmed life bodies is accurately presumed, in some embodiments of the application, the number of transitional farmed life bodies presumption model is obtained by training and optimizing the initial number of farmed life bodies presumption model through a water environment database, and the method comprises the following steps:
extracting a sample set from a water environment database, and dividing the sample set into a training set and a testing set according to a preset proportion;
and repeatedly training the initial cultured life body number presumption model by using the training set and a preset training algorithm to obtain a transitional cultured life body number presumption model.
In the embodiment, the water environment database is provided with recorded data of the change of the water environment according to different growth processes of the underwater cultured life body, and the water environment parameters such as the PH value, the temperature, the nitrogen, the ammonia content, the sediment content, the chlorophyll content, the oxygen saturation and the like in the water body are obtained after inversion of the remote sensing water environment; the number of aquatic breeding life bodies at each pond mouth is used as a dependent variable, the water parameter index value is used as an independent variable, and a basic sample set is provided for model training and testing.
It should be noted that the preset proportion may be set according to actual needs, which is not further limited by the present application. Machine learning is performed by using a training set and a GBRT algorithm, a plurality of regression trees are generated, a multidimensional and nonlinear relation between the water environment of remote sensing inversion and the number of cultured living bodies is constructed, GBRT integration can be initially trained by using Gradient Boosting Regressor in a python open-source scikit-learn machine learning library, the growth of the trees (such as max_depth, min_samples_leaf and the like) is controlled by parameters, and the number of subtrees for correcting errors (such as the number of base classifiers (n_optimizers) is set by parameters, so that a transient cultured living body number estimation model is obtained.
Referring to fig. 5, fig. 5 is a flow chart of an embodiment of evaluating a number of transitional cultured living bodies estimation model according to the present application, wherein the number of transitional cultured living bodies estimation model has a certain estimation accuracy, but the estimation accuracy is required to be evaluated to determine whether the actual use requirement can be met, and in some embodiments of the present application, the number of transitional cultured living bodies estimation model is evaluated to obtain a number of target cultured living bodies estimation model meeting a preset accuracy requirement, which includes:
s501, inputting a test set into a transition breeding life body quantity presumption model for testing to obtain a test result;
s502, evaluating the test result and the number of the actual cultured life bodies by using preset evaluation parameters, and determining the prediction precision;
s503, when the prediction precision meets the preset precision requirement, the transition breeding life body quantity estimation model is a target breeding life body quantity estimation model.
In the above embodiment, based on time requirements, the number of the underwater cultured life bodies in the planning period of the mouth of each sample culture pond where the test set is located is predicted by using the transient cultured life body number estimation model, the input data of the model are water environment parameter data in various different periods, and the output data, namely the test result of the number of the underwater cultured life bodies, is obtained through model calculation.
And comparing the test result of the number of the underwater cultured life bodies obtained by the test set data based on the model prediction with the number of the actual cultured life bodies, quantitatively evaluating the total deviation between the model prediction result and the true value by utilizing a decision coefficient R2, an average absolute error (mean absolute error, MAE) and an average absolute percentage error (mean absolute percentage error, MAPE), judging whether the prediction precision meets the standard according to the result, and further determining whether the model needs to be optimized and improved.
It is understood that the preset accuracy requirement can be set according to the actual requirement, the application does not limit this further, and when the prediction accuracy meets the preset accuracy requirement, the transitional cultured living body number estimation model is the target cultured living body number estimation model. And when the set maximum number of iterations is reached, training optimization of the model is stopped.
In order to better implement the method for estimating the number of cultured living things in the embodiment of the present application, referring to fig. 6, fig. 6 is a schematic structural diagram of an embodiment of an apparatus for estimating the number of cultured living things according to the present application, where the apparatus 600 for estimating the number of cultured living things includes:
the collection module 610 is configured to collect remote sensing data of a sample culture pond mouth in a preset collection period, so as to obtain multiple remote sensing data in different periods;
the inversion module 620 is configured to perform preprocessing and inversion of the remote sensing water environment on multiple remote sensing data in different periods, so as to obtain water environment parameters in different periods;
the training module 630 is configured to perform training optimization on the established initial number of cultured life body estimation model according to the water environment parameters in different periods, so as to obtain a target number of cultured life body estimation model;
the estimating module 640 is configured to input water environment data of the mouth to be estimated into the target number of cultured life bodies estimating model, and estimate the number of the cultured life bodies under water.
What needs to be explained here is: the apparatus 600 provided in the foregoing embodiments may implement the technical solutions described in the foregoing method embodiments, and the specific implementation principles of the foregoing modules or units may be referred to the corresponding content in the foregoing method embodiments, which is not repeated herein.
Referring to fig. 7, fig. 7 is a schematic structural diagram of a device for estimating the number of cultured living bodies according to an embodiment of the present application. Based on the above method for estimating the number of the cultured living bodies, the application also provides a corresponding device for estimating the number of the cultured living bodies, wherein the device for estimating the number of the cultured living bodies can be a mobile terminal, a desktop computer, a notebook computer, a palm computer, a server and other computing devices. The speculation means for the number of farmed life bodies includes a processor 710, a memory 720 and a display 730. Fig. 7 shows only some of the components of the speculation device for the number of farmed life bodies, but it should be understood that not all of the illustrated components need be implemented, and that more or fewer components may alternatively be implemented.
Memory 720 may be, in some embodiments, an internal storage unit of the speculation device for the number of farmed life, such as a hard disk or memory of the speculation device for the number of farmed life. The memory 720 may also be an external storage device of the estimation device of the number of farmed living being, such as a plug-in hard disk, smart Media Card (SMC), secure Digital (SD) Card, flash Card (Flash Card) or the like, which is provided on the estimation device of the number of farmed living being in other embodiments. Further, the memory 720 may also include both internal and external memory units of the presumption device for the number of farmed life bodies. The memory 720 is used for storing application software installed on the estimation device of the number of farmed living things and various kinds of data, such as program codes installed on the estimation device of the number of farmed living things. Memory 720 may also be used to temporarily store data that has been output or is to be output. In one embodiment, the memory 720 stores a number of cultured life bodies estimation program 740, and the number of cultured life bodies estimation program 740 is executed by the processor 710, so as to implement the number of cultured life bodies estimation method according to the embodiments of the present application.
The processor 710 may in some embodiments be a central processing unit (Central Processing Unit, CPU), microprocessor or other data processing chip for executing program code or processing data stored in the memory 720, such as performing methods for estimating the number of farmed life bodies, etc.
The display 730 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch, or the like in some embodiments. Display 730 is used to display information at the speculation device for the number of farmed life bodies and is used to display a visual user interface. The components 710-730 of the speculation device for cultivating the number of living things communicate with each other via a system bus.
In one embodiment, the steps in the method of speculating the number of farmed life as described above are implemented when the processor 710 executes the speculation program 740 for the number of farmed life in the memory 720.
The present embodiment also provides a computer-readable storage medium having stored thereon a program for estimating the number of farmed life bodies, which when executed by a processor, performs the steps of:
remote sensing data acquisition is carried out on the sample culture pond mouth with a preset acquisition period, so that a plurality of remote sensing data in different periods are obtained;
preprocessing a plurality of remote sensing data in different periods and inverting the remote sensing water environment to obtain water environment parameters in different periods;
training and optimizing the established initial number of cultured life body presumption models according to water environment parameters in different periods to obtain target number of cultured life body presumption models;
and inputting water environment data of the mouth of the pond to be presumed into a target cultured life body quantity presumption model to presume the quantity of the cultured life bodies under water.
In summary, the present embodiment provides a method, device, equipment and storage medium for estimating the number of cultured living bodies, where the method includes: remote sensing data acquisition is carried out on the sample culture pond mouth with a preset acquisition period, so that a plurality of remote sensing data in different periods are obtained; preprocessing the remote sensing data in different periods and inverting the remote sensing water environment to obtain water environment parameters in different periods; training and optimizing the established initial number of cultured life bodies presumption model according to the water environment parameters in different periods to obtain a target number of cultured life bodies presumption model; and inputting water environment data of the mouth of the pond to be presumed into the target cultured life body quantity presumption model to presume the quantity of the cultured life bodies under water. According to the application, the water environment parameters are determined through remote sensing water environment inversion by collecting a plurality of remote sensing data of the sample culture pond mouth in different periods, then a model is established to analyze the relation between the water environment parameters and the number of the cultured life bodies, finally the number of the underwater cultured life bodies is speculated, the number of the underwater activity can be speculated only according to the collected plurality of remote sensing data of the sample culture pond mouth, the underwater environment is not required to be photographed, the cultured life bodies are not required to be tracked, the influence of the severe underwater imaging environment on the number speculation of the underwater cultured life bodies is eliminated, the speculation result is not influenced by the movement of the underwater cultured life bodies, and the accuracy of the speculation result is high.
The present application is not limited to the above-mentioned embodiments, and any changes or substitutions that can be easily understood by those skilled in the art within the technical scope of the present application are intended to be included in the scope of the present application.

Claims (10)

1. A method for estimating the number of cultured living bodies, comprising:
remote sensing data acquisition is carried out on the sample culture pond mouth with a preset acquisition period, so that a plurality of remote sensing data in different periods are obtained;
preprocessing the remote sensing data in different periods and inverting the remote sensing water environment to obtain water environment parameters in different periods;
training and optimizing the established initial number of cultured life bodies presumption model according to the water environment parameters in different periods to obtain a target number of cultured life bodies presumption model;
and inputting water environment data of the mouth of the pond to be presumed into the target cultured life body quantity presumption model to presume the quantity of the cultured life bodies under water.
2. The method for estimating the number of cultured living bodies according to claim 1, wherein the preprocessing and the remote sensing water environment inversion of the plurality of remote sensing data in different periods to obtain water environment parameters in different periods comprises:
removing interference data in the remote sensing data in different periods to obtain water data in different periods;
determining a remote sensing water environment inversion formula, and carrying out water environment inversion on the multiple pieces of water body data in different periods to obtain water environment parameters in different periods.
3. The method for estimating the number of cultured living bodies according to claim 2, wherein the step of removing the interference data from the plurality of remote sensing data of different periods to obtain a plurality of pieces of water data of different periods comprises:
correcting the remote sensing data in different periods to obtain undistorted remote sensing data in different periods;
carrying out radiation treatment on the undistorted remote sensing data in different periods to obtain data calibrated by the radiation in different periods;
and screening the data subjected to the radiation calibration in different periods to obtain multiple water body data in different periods.
4. The method for estimating the number of cultured living bodies according to claim 2, wherein determining the remote sensing water environment inversion formula, performing water environment inversion on the plurality of pieces of water body data in different periods, and obtaining water environment parameters in different periods, comprises:
establishing a mathematical relationship between water body data and water environment parameters;
fitting the mathematical relationship to obtain a remote sensing water environment inversion formula;
and carrying out water environment inversion on the multiple water body data in different periods according to the remote sensing water environment inversion formula to obtain water environment parameters in different periods.
5. The method for estimating the number of cultured living things according to claim 1, wherein the training and optimizing the initial cultured living things number estimation model established according to the water environment parameters of the different periods to obtain the target cultured living things number estimation model comprises:
establishing a water environment database, and storing the water environment parameters in different periods into the water environment database;
establishing an initial breeding life body quantity presumption model;
training and optimizing the initial culture life body quantity presumption model through the water environment database to obtain a transition culture life body quantity presumption model;
and evaluating the transition breeding life body number estimation model to obtain a target breeding life body number estimation model meeting the preset precision requirement.
6. The method for estimating a number of cultured living things according to claim 5, wherein training and optimizing the initial cultured living things number estimation model by the water environment database to obtain a transitional cultured living things number estimation model comprises:
extracting a sample set from the water environment database, and dividing the sample set into a training set and a testing set according to a preset proportion;
and repeatedly training the initial cultured life body number estimation model by using the training set and a preset training algorithm to obtain a transitional cultured life body number estimation model.
7. The method for estimating a number of cultured living things according to claim 6, wherein evaluating the transitional cultured living things number estimation model to obtain a target cultured living things number estimation model satisfying a preset accuracy requirement comprises:
inputting the test set into the transition breeding life body quantity presumption model for testing to obtain a test result;
evaluating the test result and the number of the actual cultured life bodies by using preset evaluation parameters, and determining prediction accuracy;
when the prediction precision meets the preset precision requirement, the transition culture life body quantity estimation model is a target culture life body quantity estimation model.
8. A device for estimating the number of living things to be cultured, comprising:
the collection module is used for carrying out remote sensing data collection on the sample culture pond mouth in a preset collection period to obtain a plurality of remote sensing data in different periods;
the inversion module is used for preprocessing the remote sensing data in different periods and inverting the remote sensing water environment to obtain water environment parameters in different periods;
the training module is used for training and optimizing the established initial number of cultured life body presumption models according to the water environment parameters in different periods to obtain target number of cultured life body presumption models;
the presumption module is used for inputting water environment data of the pool mouth to be presumed into the target aquaculture life body quantity presumption model to presume the quantity of the underwater aquaculture life bodies.
9. A device for estimating the number of cultured living bodies is characterized by comprising a memory and a processor, wherein,
the memory is used for storing programs;
the processor, coupled to the memory, for executing the program stored in the memory to implement the steps in the method of speculating the number of farmed life bodies according to any one of the preceding claims 1 to 7.
10. A computer-readable storage medium storing a computer-readable program or instructions that, when executed by a processor, is capable of performing the steps in the method of estimating the number of farmed organisms according to any one of claims 1 to 7.
CN202311011575.9A 2023-08-10 2023-08-10 Method, device, equipment and storage medium for presuming number of cultured life bodies Pending CN117036921A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311011575.9A CN117036921A (en) 2023-08-10 2023-08-10 Method, device, equipment and storage medium for presuming number of cultured life bodies

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311011575.9A CN117036921A (en) 2023-08-10 2023-08-10 Method, device, equipment and storage medium for presuming number of cultured life bodies

Publications (1)

Publication Number Publication Date
CN117036921A true CN117036921A (en) 2023-11-10

Family

ID=88633154

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311011575.9A Pending CN117036921A (en) 2023-08-10 2023-08-10 Method, device, equipment and storage medium for presuming number of cultured life bodies

Country Status (1)

Country Link
CN (1) CN117036921A (en)

Similar Documents

Publication Publication Date Title
Li et al. A review of computer vision technologies for plant phenotyping
Müller-Linow et al. The leaf angle distribution of natural plant populations: assessing the canopy with a novel software tool
Bernotas et al. A photometric stereo-based 3D imaging system using computer vision and deep learning for tracking plant growth
GB2553631B (en) Data Processing of images of a crop
Andújar et al. Three-dimensional modeling of weed plants using low-cost photogrammetry
US20200020103A1 (en) Detecting crop related row from image
Wang et al. A rigorously-weighted spatiotemporal fusion model with uncertainty analysis
Liu et al. Estimating maize seedling number with UAV RGB images and advanced image processing methods
Magistri et al. Towards in-field phenotyping exploiting differentiable rendering with self-consistency loss
US11659783B2 (en) Calibrating crop row computer vision system
Liu et al. PocketMaize: An android-smartphone application for maize plant phenotyping
Barbole et al. GrapesNet: Indian RGB & RGB-D vineyard image datasets for deep learning applications
Dang et al. Vision-based white radish phenotypic trait measurement with smartphone imagery
Shi et al. Underwater fish mass estimation using pattern matching based on binocular system
Hossain et al. Machine learning techniques for estimating soil moisture from smartphone captured images
CN117036921A (en) Method, device, equipment and storage medium for presuming number of cultured life bodies
Patel et al. Deep Learning-Based Plant Organ Segmentation and Phenotyping of Sorghum Plants Using LiDAR Point Cloud
Dang et al. Vpbr: An automatic and low-cost vision-based biophysical properties recognition pipeline for pumpkin
Saeed et al. Cotton plant part 3D segmentation and architectural trait extraction using point voxel convolutional neural networks
Li et al. A technique system for the measurement, reconstruction and character extraction of rice plant architecture
Gonzalez et al. PhytoOracle: Scalable, modular phenomics data processing pipelines
Chen et al. Assessment of intertidal seaweed biomass based on RGB imagery
Wu et al. An approach to robot SLAM based on incremental appearance learning with omnidirectional vision
CN113408374A (en) Yield estimation method, device and equipment based on artificial intelligence and storage medium
Vuletić et al. Close-range multispectral imaging with Multispectral-Depth (MS-D) system

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