A method for quality inspection of eucalyptus branch biochar feed based on hyperspectral time series analysis

Abstract

The present invention belongs to the field of feed quality analysis and intelligent detection technology, and proposes a eucalyptus branch biochar feed quality inspection method based on hyperspectral time series analysis, which is specifically: obtaining multispectral data of aquaculture ponds, preprocessing the multispectral data to construct inversion data, identifying dissolution positive feedback areas from the inversion data, constructing a dissolution quality model according to the dissolution positive feedback areas to obtain sub-quality order values, establishing feed quality evaluation through time series analysis of sub-quality order values, and making an inductive evaluation on the suitability of eucalyptus branch biochar to aquaculture environment at the current addition amount through the dynamic quantitative determination, thereby providing a reference for the mixing amount of eucalyptus branch biochar in aquatic feed with targeting of aquaculture species and targeting of feed mixture ratio, thereby improving the adaptability of aquatic feed to application scenarios including water body direction and raw material direction, and effectively providing stability of feeding effect for improving weight gain rate or other feed functions of aquaculture animals.

Description

Technical Field

[0001] The invention belongs to the field of feed quality analysis and intelligent detection technology, and specifically relates to a eucalyptus branch biochar feed quality inspection method based on hyperspectral time series analysis. Background Art

[0002] Adding an appropriate amount of biochar to feed can increase the weight gain rate of aquatic animals, enhance the total antioxidant capacity and catalase activity of aquatic animals, and even improve the quality of aquaculture water, and can reduce the total nitrogen or total phosphorus in the water. Eucalyptus branch biochar is very beneficial to the preparation of aquatic feed due to its higher tolerance of pore structure and wide range of application for different feed additives; currently most aquatic feeds are designed as feed additives to improve the breeding effect of specific aquatic fish, including weight gain rate or meat crispness. However, when the feed is fed into the aquaculture water, the nutrients in the feed will be dissolved or shed to a certain extent before being swallowed by the fish, and the high tolerance of eucalyptus branch biochar allows the nutrients in the feed to be better retained in the feed before being swallowed by the fish. Since different mixing ratios of additives and raw materials will cause different dissolution rates of eucalyptus branch biochar aquatic feed in water, and the impact of the activity of aquatic species on the water environment will also lead to the stability of aquatic feed in water, there is an urgent need for a eucalyptus branch biochar feed quality inspection method based on hyperspectral time series analysis to determine the applicability of eucalyptus branch biochar addition to the aquaculture environment and effectively evaluate it. Summary of the invention

[0003] The purpose of the present invention is to propose a method for quality inspection of eucalyptus branch biochar feed based on hyperspectral time series analysis to solve one or more technical problems existing in the prior art and at least provide a beneficial choice or create conditions.

[0004] In order to achieve the above object, according to one aspect of the present invention, a method for quality inspection of eucalyptus branch biochar feed based on hyperspectral time series analysis is provided, and the method comprises the following steps:

[0005] Acquire multispectral data of aquaculture ponds, preprocess the multispectral data to construct inversion data, identify the dissolution positive feedback area from the inversion data, build a dissolution quality model based on the dissolution positive feedback area to obtain sub-quality order values, and establish feed quality evaluation through time series analysis of sub-quality order values.

[0006] Furthermore, the method for obtaining multispectral data of the breeding pond is: to obtain multispectral data by real-time data collection of the breeding pond through a multispectral camera, wherein the multispectral camera is carried by a drone.

[0007] Furthermore, the method of preprocessing the multispectral data to construct the inversion data is: performing geometric correction, radiation correction and atmospheric correction on the multispectral data, using the SDeCNN algorithm to denoise the hyperspectral image, enhancing the weak light characteristics through spectral differentiation, applying the inversion algorithm to obtain the corresponding content distribution image of each band, and the inversion data is constructed from the content distribution images obtained in each band.

[0008] The inversion algorithm includes any one of quantitative analysis method, photon diffusion analysis model, atomic fluorescence spectroscopy, atomic absorption spectroscopy or least squares support vector machine model.

[0009] Alternatively, the inverse calculation method may adopt the method of estimating the total suspended matter concentration in the nearshore area based on drone-borne hyperspectral images with patent number CN112051226B.

[0010] Further, the method for identifying the dissolution positive feedback area from the inversion data is: defining the inversion data obtained before the aquatic feed is put into the breeding pond as the benchmark inversion data, and the inversion data obtained after the aquatic feed is put into the breeding pond as the measured inversion data;

[0011] The measurement interval for determining the inversion data is once every 10-30 seconds, and the data collection time is 60-90 minutes;

[0012] Identify the dissolution positive feedback area for any measured inversion data obtained at one time: perform median filtering on the content distribution image of each band in the benchmark inversion data to form benchmark threshold data; for any pixel, if the inversion value of the pixel in the measured inversion data is greater than the inversion value in the benchmark threshold data in the same band, then the pixel is defined to meet the threshold requirement in this band, otherwise it does not meet the threshold requirement; when the number of bands where a pixel meets the threshold requirement exceeds half of the total number of bands, then the pixel is defined to meet the positive feedback requirement; all pixels that meet the positive feedback requirement constitute a dissolution positive feedback area.

[0013] Here, pixel refers to the smallest data unit of the content distribution image obtained in any band.

[0014] Furthermore, the method for constructing a dissolution quality model according to the dissolution positive feedback region to obtain the sub-quality order value is:

[0015] Set the size of the inversion data to Mp×Np×Lp, where Mp×Np is the pixel resolution size of the two-dimensional inversion data and Lp is the number of bands collected; set the pixel coordinates of the inversion data to (x, y, z), which means that the pixel coordinates of the two-dimensional inversion data with the serial number z in the band are (x, y); the range of x values ​​is [1 , Mp], the range of y values ​​is [1 , Np], and the range of z values ​​is [1 , Lp];

[0016] The dissolved positive feedback area is intercepted in the inversion data, and the inversion value of any pixel in the inversion data in different bands is calculated as the inversion mean of the pixel. The average spectral image E.IVC is constructed by the inversion mean of all pixels; the corresponding content dispersion threshold GDTh of any pixel in any band is defined as the ratio of the inversion mean of the pixel to the inversion value of the pixel in the corresponding band;

[0017] Or the content dispersion threshold is expressed as: GDTh(x,y,z)=E.IVC(x,y) / IVC(x,y,z); where GDTh(x,y,z) represents the content dispersion threshold with coordinates (x,y) in the two-dimensional inversion data with serial number z, E.IVC(x,y) represents the inversion value with coordinates (x,y) in the average spectrum image; IVC(x,y,z) represents the inversion value with serial number z and coordinates (x,y) in the inversion data;

[0018] The matrix formed by the content dispersion thresholds of each pixel in the same band is recorded as the content anisotropic matrix;

[0019] The average value and upper quartile value of all elements in the content anisotropic matrix are recorded as the content sensitization value Gvse and the second sensitization value respectively; the elements with values ​​greater than the content sensitization value and the elements greater than the second sensitization value in the content anisotropic matrix are recorded as the first sensitization point and the second sensitization point respectively;

[0020] Calculate the Euclidean distance between any first sensitizing point and all other second sensitizing points and write them into the sequence Ls.Sqp, obtain the median value of the values ​​in the sequence Ls.Sqp, and record the product of the median value and the content dispersion threshold corresponding to the first sensitizing point as the sensitization distance coefficient Sde of the first sensitizing point. The average value of each sensitization distance coefficient in the same band is recorded as e.Sde.

[0021] The maximum value of each sensitization distance coefficient obtained under any band is written into a sequence Ls.Sde; the sequence Ls.Sde includes the maximum value of the sensitization distance coefficient obtained under each band;

[0022] The sub-quality order value Ubva of the inversion data is calculated according to the sensitized distance coefficient, and its calculation method is:

[0023] ;

[0024] Where z1 is the cumulative variable, RMS{} represents the root mean square function, and the root mean square value of the calling sequence is returned by the root mean square function. z1 and e.Sde z1 are the maximum and average values ​​of the sensitization distance coefficients in the z1th band, ERG z1is the standard deviation of the content anisotropy matrix of the z1th band, and ln() is the logarithmic function with the natural number e as the base.

[0025] Furthermore, a method for establishing feed quality evaluation through time series analysis of sub-quality order values ​​is as follows: the ratio of the number of pixels in the positive feedback area at any time to the square of its maximum inner diameter is defined as the diffusion order value of the dissolution positive feedback area, and the difference between the diffusion order value obtained at any time and the previous time is the diffusion difference;

[0026] The diffusion difference and sub-mass order value obtained at any time are constructed into a binary group as a solubility diffusion feature group; the solubility diffusion feature group obtained during the monitoring period is constructed into a sequence and subjected to fast Fourier transform, and the corresponding time of the element with the maximum modulus length among the obtained elements is the diffusion critical point, and its corresponding modulus length is recorded as the high diffusion critical value;

[0027] The corresponding serial number value of the diffusion critical point in the solubility diffusion characteristic group is recorded as t_SSR, then the diffusion urgency is ln(t_SSR), and the diffusion critical characteristic is composed of the high diffusion critical value and the diffusion urgency;

[0028] The diffusion critical characteristics of aquatic feeds with different mixing ratios are read from the server and normalized. The high diffusion critical value and the root mean square value of the diffusion urgency after the normalization of the current diffusion critical characteristics are used as the feed quality evaluation. If the current feed quality evaluation is greater than the average value of each feed quality evaluation, it is judged to meet the high-quality aquatic feed, otherwise it is judged not to meet the high-quality aquatic feed conditions. It is believed that the corresponding eucalyptus branch biochar addition amount is good for the breeding environment when it meets the high-quality aquatic feed.

[0029] The maximum inner diameter of the positive feedback area refers to the distance between the two farthest pixels in the area; the monitoring period refers to the complete data collection period after the feed is fed.

[0030] Preferably, all undefined variables in the present invention, if not clearly defined, can be manually set thresholds.

[0031] The present invention also provides a eucalyptus branch biochar feed quality inspection system based on hyperspectral time series analysis, the eucalyptus branch biochar feed quality inspection system based on hyperspectral time series analysis includes: a processor, a memory, and a computer program stored in the memory and executable on the processor, the processor implements the steps in the eucalyptus branch biochar feed quality inspection method based on hyperspectral time series analysis when executing the computer program, the eucalyptus branch biochar feed quality inspection system based on hyperspectral time series analysis can be run in computing devices such as desktop computers, laptop computers, PDAs and cloud data centers, and the executable system may include, but is not limited to, processors, memories, and server clusters, and the processor executes the computer program to run in the following system units:

[0032] Spectral acquisition unit, used to obtain multi-spectral data of the aquaculture pond;

[0033] A data inversion unit, used for preprocessing multispectral data to construct inversion data;

[0034] A dissolution positive feedback region identification unit, used for identifying the dissolution positive feedback region from the inversion data;

[0035] A dissolution quality model analysis unit, used for constructing a dissolution quality model according to the dissolution positive feedback region to obtain a sub-quality order value;

[0036] The quality evaluation unit is used to establish feed quality evaluation through time series analysis of sub-quality order values.

[0037] The beneficial effects of the present invention are as follows: the present invention provides a eucalyptus branch biochar feed quality inspection method based on hyperspectral time series analysis, which obtains the distribution of scattered or dissolved matter in the feed by spectral analysis of the water in the breeding pond, timely identifies the dynamic distribution characteristics of the content of scattered or dissolved matter in the breeding pond and effectively quantifies it, and makes an inductive evaluation of the applicability of eucalyptus branch biochar to the breeding environment at the current addition amount through the dynamic quantification, thereby providing a reference for the mixing amount of eucalyptus branch biochar in aquatic feed with targeting of aquaculture species and feed mixture ratio, thereby improving the adaptability of aquatic feed to application scenarios including water body direction and raw material direction, and effectively providing the stability of feeding effect for improving the weight gain rate or other feed functions of aquaculture. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The above and other features of the present invention will become more obvious by describing in detail the embodiments shown in the accompanying drawings. The same reference numerals in the accompanying drawings of the present invention represent the same or similar elements. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other accompanying drawings can be obtained based on these accompanying drawings without creative work. In the accompanying drawings:

[0039] Figure 1 Shown is a flow chart of a method for quality inspection of eucalyptus branch biochar feed based on hyperspectral time series analysis;

[0040] Figure 2 Shown is a structural diagram of a eucalyptus branch biochar feed quality inspection system based on hyperspectral time series analysis. DETAILED DESCRIPTION

[0041] The following will be combined with the embodiments and drawings to clearly and completely describe the concept, specific structure and technical effects of the present invention, so as to fully understand the purpose, scheme and effect of the present invention. It should be noted that the embodiments and features in the embodiments of this application can be combined with each other without conflict.

[0042] like Figure 1 The figure shows a flow chart of a method for quality inspection of eucalyptus branch biochar feed based on hyperspectral time series analysis. Figure 1 To illustrate a method for quality inspection of eucalyptus branch biochar feed based on hyperspectral time series analysis according to an embodiment of the present invention, the method comprises the following steps:

[0043] Acquire multispectral data of aquaculture ponds, preprocess the multispectral data to construct inversion data, identify the dissolution positive feedback area from the inversion data, build a dissolution quality model based on the dissolution positive feedback area to obtain sub-quality order values, and establish feed quality evaluation through time series analysis of sub-quality order values.

[0044] Furthermore, the method for obtaining multispectral data of the breeding pond is: to obtain multispectral data by real-time data collection of the breeding pond through a multispectral camera, wherein the multispectral camera is carried by a drone.

[0045] Furthermore, the method for preprocessing the multispectral data to construct the inversion data is: geometric correction processing, radiation correction processing and atmospheric correction processing are performed on the multispectral data, the SDeCNN algorithm is used to denoise the hyperspectral image, the weak light feature is enhanced by spectral differentiation, and the inversion algorithm is applied to obtain the content distribution image corresponding to each band, and the content distribution image obtained in each band constitutes the inversion data. The band selection is selected according to the known components of the feed, or a hyperspectral processing method and a band selection method with invention patent number CN112067129B are referred to.

[0046] The inversion algorithm includes any one of quantitative analysis method, photon diffusion analysis model, atomic fluorescence spectroscopy, atomic absorption spectroscopy or least squares support vector machine model.

[0047] Alternatively, the inverse calculation method may adopt the method of estimating the total suspended matter concentration in the nearshore area based on drone-borne hyperspectral images with patent number CN112051226B.

[0048] Further, the method for identifying the dissolution positive feedback area from the inversion data is: defining the inversion data obtained before the aquatic feed is put into the breeding pond as the benchmark inversion data, and the inversion data obtained after the aquatic feed is put into the breeding pond as the measured inversion data;

[0049] The measurement interval for determining the inversion data is once every 10-30 seconds, and the data collection time is 60-90 minutes;

[0050] Identify the dissolution positive feedback area for any measured inversion data obtained at one time: perform median filtering on the content distribution image of each band in the benchmark inversion data to form benchmark threshold data; for any pixel, if the inversion value of the pixel in the measured inversion data is greater than the inversion value in the benchmark threshold data in the same band, then the pixel is defined to meet the threshold requirement in this band, otherwise it does not meet the threshold requirement; when the number of bands where a pixel meets the threshold requirement exceeds half of the total number of bands, then the pixel is defined to meet the positive feedback requirement; all pixels that meet the positive feedback requirement constitute a dissolution positive feedback area.

[0051] Here, pixel refers to the smallest data unit of the content distribution image obtained in any band.

[0052] Furthermore, the method for constructing a dissolution quality model according to the dissolution positive feedback region to obtain the sub-quality order value is:

[0053] Set the size of the inversion data to Mp×Np×Lp, where Mp×Np is the pixel resolution size of the two-dimensional inversion data and Lp is the number of bands collected; set the pixel coordinates of the inversion data to (x, y, z), which means that the pixel coordinates of the two-dimensional inversion data with the serial number z in the band are (x, y); the range of x values ​​is [1 , Mp], the range of y values ​​is [1 , Np], and the range of z values ​​is [1 , Lp];

[0054] The dissolved positive feedback area is intercepted in the inversion data, and the inversion value of any pixel in the inversion data in different bands is calculated as the inversion mean of the pixel. The average spectral image E.IVC is constructed by the inversion mean of all pixels; the corresponding content dispersion threshold GDTh of any pixel in any band is defined as the ratio of the inversion mean of the pixel to the inversion value of the pixel in the corresponding band;

[0055] Or the content dispersion threshold is expressed as: GDTh(x,y,z)=E.IVC(x,y) / IVC(x,y,z); where GDTh(x,y,z) represents the content dispersion threshold with coordinates (x,y) in the two-dimensional inversion data with serial number z, E.IVC(x,y) represents the inversion value with coordinates (x,y) in the average spectrum image; IVC(x,y,z) represents the inversion value with serial number z and coordinates (x,y) in the inversion data;

[0056] The matrix formed by the content dispersion thresholds of each pixel in the same band is recorded as the content anisotropic matrix;

[0057] The average value and upper quartile value of all elements in the content anisotropic matrix are recorded as the content sensitization value Gvse and the second sensitization value respectively; the elements with values ​​greater than the content sensitization value and the elements greater than the second sensitization value in the content anisotropic matrix are recorded as the first sensitization point and the second sensitization point respectively;

[0058] Calculate the Euclidean distance between any first sensitizing point and all other second sensitizing points and write them into the sequence Ls.Sqp, obtain the median value of the values ​​in the sequence Ls.Sqp, and record the product of the median value and the content dispersion threshold corresponding to the first sensitizing point as the sensitization distance coefficient Sde of the first sensitizing point. The average value of each sensitization distance coefficient in the same band is recorded as e.Sde.

[0059] The maximum value of each sensitization distance coefficient obtained under any band is written into a sequence Ls.Sde; the sequence Ls.Sde includes the maximum value of the sensitization distance coefficient obtained under each band;

[0060] The sub-quality order value Ubva of the inversion data is calculated according to the sensitized distance coefficient, and its calculation method is:

[0061] ;

[0062] Where z1 is the accumulated variable, RMS{} represents the root mean square function, and the root mean square value of the calling sequence is returned by the root mean square function. max.sdez1 and e.Sdez1 are the maximum and average values ​​of each sensitization distance coefficient in the z1th band, ERGz1 is the standard deviation of the content anisotropic matrix of the z1th band, and ln() is a logarithmic function with the natural number e as the base.

[0063] Furthermore, a method for establishing feed quality evaluation through time series analysis of sub-quality order values ​​is as follows: the ratio of the number of pixels in the positive feedback area at any time to the square of its maximum inner diameter is defined as the diffusion order value of the dissolution positive feedback area, and the difference between the diffusion order value obtained at any time and the previous time is the diffusion difference;

[0064] The diffusion difference and sub-mass order value obtained at any time are constructed into a binary group as a solubility diffusion feature group; the solubility diffusion feature group obtained during the monitoring period is constructed into a sequence and subjected to fast Fourier transform, and the corresponding time of the element with the maximum modulus length among the obtained elements is the diffusion critical point, and its corresponding modulus length is recorded as the high diffusion critical value;

[0065] The corresponding serial number value of the diffusion critical point in the solubility diffusion characteristic group is recorded as t_SSR, then the diffusion urgency is ln(t_SSR), and the diffusion critical characteristic is composed of the high diffusion critical value and the diffusion urgency;

[0066] The diffusion critical characteristics of aquatic feeds with different mixing ratios are read from the server and normalized. The high diffusion critical value and the root mean square value of the diffusion urgency after the normalization of the current diffusion critical characteristics are used as the feed quality evaluation. If the current feed quality evaluation is greater than the average value of each feed quality evaluation, it is judged to be in line with the high-quality aquatic feed, otherwise it is judged not to meet the high-quality aquatic feed conditions. It is believed that the current aquatic feed quality corresponding to the amount of eucalyptus branch biochar added is well applicable to the breeding environment.

[0067] The maximum inner diameter of the positive feedback area refers to the distance between the two farthest pixels in the area; the monitoring period refers to the complete data collection period after the feed is fed.

[0068] The embodiment of the present invention provides a eucalyptus branch biochar feed quality inspection system based on hyperspectral time series analysis, such as Figure 2 Shown is a structural diagram of a eucalyptus branch biochar feed quality inspection system based on hyperspectral time series analysis of the present invention. The eucalyptus branch biochar feed quality inspection system based on hyperspectral time series analysis of this embodiment includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps in the above-mentioned embodiment of the eucalyptus branch biochar feed quality inspection method based on hyperspectral time series analysis are implemented.

[0069] The system comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to run in the following units of the system:

[0070] Spectral acquisition unit, used to obtain multi-spectral data of the aquaculture pond;

[0071] A data inversion unit, used for preprocessing multispectral data to construct inversion data;

[0072] A dissolution positive feedback region identification unit, used for identifying the dissolution positive feedback region from the inversion data;

[0073] A dissolution quality model analysis unit, used for constructing a dissolution quality model according to the dissolution positive feedback region to obtain a sub-quality order value;

[0074] The quality evaluation unit is used to establish feed quality evaluation through time series analysis of sub-quality order values.

[0075] The eucalyptus branch biochar feed quality inspection system based on hyperspectral time series analysis can be run on computing devices such as desktop computers, laptops, PDAs and cloud servers. The eucalyptus branch biochar feed quality inspection system based on hyperspectral time series analysis can be operated on systems that may include, but are not limited to, processors and memories. Those skilled in the art will understand that the example is merely an example of a eucalyptus branch biochar feed quality inspection system based on hyperspectral time series analysis, and does not constitute a limitation on a eucalyptus branch biochar feed quality inspection system based on hyperspectral time series analysis. It may include more or fewer components than the example, or a combination of certain components, or different components. For example, the eucalyptus branch biochar feed quality inspection system based on hyperspectral time series analysis may also include input and output devices, network access devices, buses, etc.

[0076] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the operating system of the eucalyptus branch biochar feed quality inspection system based on hyperspectral time series analysis, and various interfaces and lines are used to connect the various parts of the entire eucalyptus branch biochar feed quality inspection system based on hyperspectral time series analysis.

[0077] The memory can be used to store the computer program and / or module, and the processor realizes various functions of the eucalyptus branch biochar feed quality inspection system based on hyperspectral time series analysis by running or executing the computer program and / or module stored in the memory and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0078] Although the description of the present invention has been quite detailed and has been described in particular with respect to several described embodiments, it is not intended to be limited to any of these details or embodiments or any particular embodiment, so as to effectively cover the intended scope of the present invention. In addition, the present invention is described above with the embodiments foreseeable by the inventors, and its purpose is to provide a useful description, and those non-substantial changes to the present invention that are not currently foreseen may still represent equivalent changes of the present invention.

Claims

1. A method for quality inspection of eucalyptus branch biochar feed based on hyperspectral time series analysis, characterized in that: The method comprises the following steps: obtaining multispectral data of a breeding pond, preprocessing the multispectral data to construct inversion data, identifying a dissolution positive feedback region from the inversion data, constructing a dissolution quality model according to the dissolution positive feedback region to obtain a sub-quality order value, and establishing a feed quality evaluation through a time series analysis of the sub-quality order value; The method for identifying the dissolution positive feedback area from the inversion data is: defining the inversion data obtained before the aquatic feed is put into the breeding pond as the benchmark inversion data, and the inversion data obtained after the aquatic feed is put into the breeding pond as the measured inversion data; The measurement interval for determining the inversion data is once every 10-30 seconds, and the data collection time is 60-90 minutes; Identify the dissolution positive feedback area for any measured inversion data obtained at one time: perform median filtering on the content distribution image of each band in the benchmark inversion data to form benchmark threshold data; for any pixel, if the inversion value of the pixel in the measured inversion data is greater than the inversion value in the benchmark threshold data in the same band, then the pixel is defined to meet the threshold requirement in this band, otherwise it does not meet the threshold requirement; when the number of bands where a pixel meets the threshold requirement exceeds half of the total number of bands, then the pixel is defined to meet the positive feedback requirement; all pixels that meet the positive feedback requirement constitute a dissolution positive feedback area; The method of constructing a dissolution quality model according to the dissolution positive feedback area to obtain the sub-quality order value is as follows: intercepting the dissolution positive feedback area in the inversion data, calculating the average value of the inversion value of any pixel in the inversion data in different bands as the inversion mean of the pixel, and constructing the average spectral image E.IVC from the inversion mean of all pixels; defining the corresponding content dispersion threshold GDTh of any pixel in any band as the ratio of the inversion mean of the pixel to the inversion value of the pixel in the corresponding band; The matrix formed by the content dispersion thresholds of each pixel in the same band is recorded as the content anisotropic matrix; The average value and upper quartile value of all elements in the content anisotropic matrix are recorded as the content sensitization value Gvse and the second sensitization value respectively; the elements with values ​​greater than the content sensitization value and the elements greater than the second sensitization value in the content anisotropic matrix are recorded as the first sensitization point and the second sensitization point respectively; Calculate the Euclidean distance between any first sensitizing point and all other second sensitizing points and write them into the sequence Ls.Sqp, obtain the median value of the values ​​in the sequence Ls.Sqp, and record the product of the median value and the content dispersion threshold corresponding to the first sensitizing point as the sensitization distance coefficient Sde of the first sensitizing point. The average value of each sensitization distance coefficient in the same band is recorded as e.Sde. The maximum value of each sensitization distance coefficient obtained in any band is written into the sequence Ls.Sde; the sub-quality order value Ubva of the inversion data is calculated according to the sensitization distance coefficient, and the calculation method is: ； Where z1 is the cumulative variable, Lp is the number of bands collected, RMS{} represents the root mean square function, max.sde z1 and e.Sde z1 are the maximum and average values ​​of the sensitization distance coefficients in the z1th band, ERG z1 is the standard deviation of the content anisotropy matrix of the z1th band; The method of establishing feed quality evaluation through time series analysis of sub-quality order values ​​is as follows: define the ratio of the number of pixels in the positive feedback area at any time to the square of its maximum inner diameter as the diffusion order value of the dissolution positive feedback area, and the difference between the diffusion order value obtained at any time and the previous time is the diffusion difference; The diffusion difference and sub-mass order value obtained at any time are constructed into a binary group as a solubility diffusion feature group; the solubility diffusion feature group obtained during the monitoring period is constructed into a sequence and subjected to fast Fourier transform. Among the obtained elements, the corresponding moment of the element with the maximum modulus length is the diffusion critical point, and its corresponding modulus length is recorded as the high diffusion critical value; The corresponding serial number value of the diffusion critical point in the solubility diffusion characteristic group is recorded as t_SSR, then the diffusion urgency is ln(t_SSR), and the diffusion critical characteristic is composed of the high diffusion critical value and the diffusion urgency; The diffusion critical characteristics of aquatic feeds with different mixing ratios are read from the server and normalized. The high diffusion critical value and the root mean square value of the diffusion urgency after the normalization of the current diffusion critical characteristics are used as the feed quality evaluation. If the current feed quality evaluation is greater than the average value of each feed quality evaluation, it is judged to be in line with the high-quality aquatic feed, otherwise it is judged to not meet the conditions of high-quality aquatic feed.

2. The method for quality inspection of eucalyptus branch biochar feed based on hyperspectral time series analysis according to claim 1, characterized in that: The method for obtaining multispectral data of the breeding pond is: to obtain multispectral data by real-time data collection of the breeding pond through a multispectral camera, wherein the multispectral camera is carried by a drone.

3. The method for quality inspection of eucalyptus branch biochar feed based on hyperspectral time series analysis according to claim 1, characterized in that: The method for preprocessing multispectral data to construct inversion data is: perform geometric correction, radiation correction and atmospheric correction on multispectral data, use SDeCNN algorithm to denoise hyperspectral images, enhance weak light features through spectral differentiation, apply inversion algorithm to obtain content distribution images corresponding to each band, and the content distribution images obtained in each band constitute inversion data.

4. A eucalyptus branch biochar feed quality inspection system based on hyperspectral time series analysis, characterized in that: The eucalyptus branch biochar feed quality inspection system based on hyperspectral time series analysis includes: a processor, a memory, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the steps of the eucalyptus branch biochar feed quality inspection method based on hyperspectral time series analysis described in any one of claims 1-3 are implemented. The eucalyptus branch biochar feed quality inspection system based on hyperspectral time series analysis runs on desktop computers, laptops, PDAs and computing devices in cloud data centers.

Images