Deep learning-based modified starch function performance prediction method and system

Abstract

The invention relates to the technical field of modified starch function performance prediction, in particular to a modified starch function performance prediction method and system based on deep learning. The method comprises the following steps: performing comparative analysis on modification process parameters and modification process parameter indexes to obtain a comparative analysis result, inputting the comparative analysis result into a modification process evaluation model, and outputting modification process parameter influence indexes; obtaining an infrared spectrogram of the original starch structure, calculating to obtain modification performance parameters based on the infrared spectrogram, and extracting starch structure characteristics based on the modification performance parameters; and inputting the modification process parameter influence index, the modification performance parameter and the starch structure characteristics as inputs into the CNN-LSTM performance index prediction model, and outputting the key performance index value of the modified starch, so that prediction of the key performance index value of the modified starch can be accurately realized, the development cost of the modified starch is reduced, and the development period is shortened.

Description

Technical Field

[0001] The present invention relates to the technical field of modified starch functional performance prediction, and in particular to a modified starch functional performance prediction method and system based on deep learning. Background Art

[0002] Modified starch, a functional material produced through physical, chemical, or enzymatic modification techniques, has been widely used in industries such as food, medicine, papermaking, and textiles. However, existing technologies for predicting the functional properties of modified starch require 50-100 kg of starch raw material for a single full-factor experiment, resulting in high costs, lengthy processing times, and low performance prediction efficiency and accuracy.

[0003] In response to the above problems, the present invention proposes a method and system for predicting the functional properties of modified starch based on deep learning. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for predicting the functional properties of modified starch based on deep learning: the existing technology has technical problems in the process of predicting the functional properties of modified starch, such as high cost, lengthy cycle and low performance prediction accuracy.

[0005] The purpose of the present invention can be achieved through the following technical solutions: On the one hand, a method for predicting the functional properties of modified starch based on deep learning, the method comprising: Obtaining the original starch type and modification process parameters, determining the modification process parameter index corresponding to the original starch based on the starch type, performing comparative analysis on the modification process parameters and the modification process parameter index to obtain comparative analysis results, and inputting the comparative analysis results into the modification process evaluation model to output the modification process parameter influence index; Obtaining an infrared spectrum of the original starch structure, calculating modified performance parameters based on the infrared spectrum, and extracting starch structural characteristics based on the modified performance parameters; The modification process parameter influence index, modification performance parameters and starch structural characteristics are input into the CNN-LSTM performance index prediction model, and the key performance index values ​​of the modified starch are output.

[0006] Furthermore, the modified process parameters and modified process parameter indicators are compared and analyzed to obtain the comparative analysis results, which specifically includes the following process: Obtaining a pH value in the modification process parameter, obtaining a preset pH value in the modification process parameter index, and calculating a difference A between the pH value in the modification process parameter and the preset pH value; Obtaining a modification reaction time in the modification process parameters, obtaining a preset modification reaction time in the modification process parameter index, and calculating a difference B between the modification reaction time in the modification process parameters and the preset modification reaction time; Obtaining a modification reaction temperature in the modification process parameters, obtaining a preset modification reaction temperature in the modification process parameter index, and calculating a difference C between the modification reaction temperature in the modification process parameters and the preset modification reaction temperature; Obtain the number of cross-linking agent types in the modification process parameters, and determine whether the number of cross-linking agent types exceeds the minimum cross-linking agent type threshold in the modification process parameter indicators. If so, the modification process qualification coefficient R is 1; if not, the modification process qualification coefficient R is 0; The difference A, difference B, difference C, and modification process qualification coefficient R are recorded as comparative analysis results.

[0007] Furthermore, the comparative analysis results are input into the modification process evaluation model, and the modification process parameter impact index is output. The specific process includes the following steps: The comparative analysis results are processed into a unified dimension and then input into the modified process evaluation formula: ; in, Indicates the influence index of modification process parameters, 、 is the weight coefficient, The pH value of the original starch is the difference between the pH value of the modified process parameters; is the change in crystallinity of the original starch; It is the difference between the single cross-linking agent addition amount and the minimum single cross-linking agent addition amount; the value of e is 2.72.

[0008] Furthermore, the modified performance parameters are calculated based on the infrared spectrum, specifically including the following process: Get the characteristic wavelength reflectivity increasing relationship diagram of the infrared spectrum, add the selected band to the characteristic wavelength reflectivity increasing relationship diagram, and select three eigenvectors for the selected band 、 、 , the wavelengths corresponding to the vectors are 、 、 , select the following formula to calculate the modified performance parameters : ; Among them, the three eigenvectors are The ratio of the reflectivity difference in the axial projection direction is equal.

[0009] Furthermore, extracting starch structural characteristics based on modified performance parameters specifically includes the following processes: Divide the infrared spectrum image into several parts and determine the Does it exceed the preset threshold? If so, the part is segmented and recorded as the starch structure feature part to be extracted. The third-order moment color feature extraction method is used to extract the starch structure feature part to obtain the starch structure feature: ; in, The structural characteristics of starch. represents the number of channels in the image, It means the Pixels on the The color value of each channel, Indicates the number of pixels. Indicates the The color mean of each channel.

[0010] Furthermore, the modification process parameter influence index, modification performance parameters, and starch structural characteristics are input into the CNN-LSTM performance index prediction model, and the key performance index values ​​of the modified starch are output, which specifically includes the following process: Based on starch structural characteristics A new vector is bound to the modification process parameter influence index and modification performance parameter ; The CNN-LSTM performance indicator prediction model consists of two parts: the encoding network f and the context network. The encoder network f consists of a five-layer convolutional network. In the structure of the five-layer convolutional network, the sizes of the convolution kernels are (10, 8, 4, 4, 4) respectively. The encoder outputs a low-frequency feature. , the hidden space after the encoder network mapping is represented by Z, where : ; Among them, m is the convolution receptive field; The context network is a convolutional neural network with the same structure as the encoding network f. It maps the hidden space into the context information space M. The output of the context network is ,in, : ; Among them, v is the convolution receptive field; Loss function based on CNN-LSTM performance indicator prediction model To train the model, the loss function It is expressed as follows: ; in, is the sequence length of the input data, is the loss of noise contrast estimation, is the true probability of the input sample, is the learning parameter; Output the key performance index values ​​of the modified starch, including paste strength, thermal stability and transparency.

[0011] On the other hand, a modified starch functional performance prediction system based on deep learning, the system comprising: A data acquisition module is used to obtain the original starch type and modification process parameters, determine the modification process parameter index corresponding to the original starch based on the starch type, compare and analyze the modification process parameters and the modification process parameter index to obtain a comparative analysis result, input the comparative analysis result into the modification process evaluation model, and output the modification process parameter influence index; A starch structure feature extraction module is used to obtain the infrared spectrum of the original starch structure, calculate the modified performance parameters based on the infrared spectrum, and extract the starch structure features based on the modified performance parameters; The starch functional performance prediction module is used to input the modification process parameter influence index, modification performance parameters and starch structural characteristics into the CNN-LSTM performance index prediction model and output the key performance index values ​​of the modified starch.

[0012] Compared with the existing solutions, the present invention achieves the following beneficial effects: The present invention obtains the original starch type and modification process parameters, compares and analyzes the modification process parameters and modification process parameter indicators to obtain a comparison analysis result, inputs the comparison analysis result into a modification process evaluation model, and outputs a modification process parameter influence index; obtains an infrared spectrum of the original starch structure, calculates modification performance parameters based on the infrared spectrum, and extracts starch structural characteristics based on the modification performance parameters; inputs the modification process parameter influence index, the modification performance parameters, and the starch structural characteristics into a CNN-LSTM performance indicator prediction model as input, and outputs a key performance indicator value of the modified starch. The present invention can accurately predict the key performance indicator value of the modified starch, thereby reducing the development cost of modified starch and shortening the development cycle. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0014] Figure 1 This is a workflow diagram of a method for predicting functional properties of modified starch based on deep learning according to an embodiment of the present invention; Figure 2 This is a system block diagram of a modified starch functional performance prediction system based on deep learning in an embodiment of the present invention. DETAILED DESCRIPTION

[0015] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0016] In addition, the described features, structures or characteristics can be combined in any suitable manner in one or more example embodiments. In the following description, many specific details are provided to provide a full understanding of the example embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure can be practiced while omitting one or more of the specific details, or other methods, components, steps, etc. can be adopted. In other cases, well-known structures, methods, implementations or operations are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0017] This embodiment provides a method for predicting the functional properties of modified starch based on deep learning. Figure 1 This is a workflow diagram of a method for predicting the functional properties of modified starch based on deep learning according to an embodiment of the present invention. Figure 1 As shown, the method includes the following steps: Step S101: obtaining the original starch type and modification process parameters, determining the modification process parameter index corresponding to the original starch based on the starch type, and performing comparative analysis on the modification process parameters and the modification process parameter index to obtain comparative analysis results; Step S102: inputting the comparative analysis results into the modification process evaluation model and outputting the modification process parameter impact index; Step S103: obtaining an infrared spectrum of the original starch structure, calculating modified performance parameters based on the infrared spectrum, and extracting starch structural characteristics based on the modified performance parameters; Step S104: The modification process parameter influence index, the modification performance parameter and the starch structure characteristics are input into the CNN-LSTM performance index prediction model, and the key performance index value of the modified starch is output.

[0018] In summary, the present invention obtains the original starch type and modification process parameters, compares and analyzes the modification process parameters and the modification process parameter indicators to obtain a comparative analysis result, inputs the comparative analysis result into a modification process evaluation model, and outputs a modification process parameter influence index; obtains an infrared spectrum of the original starch structure, calculates the modification performance parameters based on the infrared spectrum, and extracts the starch structure characteristics based on the modification performance parameters; inputs the modification process parameter influence index, the modification performance parameters and the starch structure characteristics as input into a CNN-LSTM performance indicator prediction model, and outputs the key performance indicator value of the modified starch, which can accurately realize the prediction of the key performance indicator value of the modified starch, thereby reducing the development cost of modified starch and shortening the development cycle.

[0019] In some embodiments, comparative analysis of the modified process parameters and the modified process parameter indicators to obtain comparative analysis results specifically includes the following process: Obtaining a pH value in the modification process parameter, obtaining a preset pH value in the modification process parameter index, and calculating a difference A between the pH value in the modification process parameter and the preset pH value; Obtaining a modification reaction time in the modification process parameters, obtaining a preset modification reaction time in the modification process parameter index, and calculating a difference B between the modification reaction time in the modification process parameters and the preset modification reaction time; Obtaining a modification reaction temperature in the modification process parameters, obtaining a preset modification reaction temperature in the modification process parameter index, and calculating a difference C between the modification reaction temperature in the modification process parameters and the preset modification reaction temperature; Obtain the number of cross-linking agent types in the modification process parameters, and determine whether the number of cross-linking agent types exceeds the minimum cross-linking agent type threshold in the modification process parameter indicators. If so, the modification process qualification coefficient R is 1; if not, the modification process qualification coefficient R is 0; The difference A, difference B, difference C, and modification process qualification coefficient R are recorded as comparative analysis results.

[0020] Furthermore, the comparative analysis results are input into the modification process evaluation model, and the modification process parameter impact index is output. The specific process includes the following steps: The comparative analysis results are processed into a unified dimension and then input into the modified process evaluation formula: ; in, Indicates the influence index of modification process parameters, 、 is the weight coefficient, The pH value of the original starch is the difference between the pH value of the modified process parameters; is the change in crystallinity of the original starch; It is the difference between the single cross-linking agent addition amount and the minimum single cross-linking agent addition amount; the value of e is 2.72.

[0021] It is worth noting that crosslinking agents can include: halides, such as phosphorus oxychloride, with a reaction pH of 8-12 and an addition level of 0.005%-0.25%. If the addition level is greater than 1%, crosslinked starch with gelatinization resistance can be obtained. Aldehydes, such as formaldehyde, glyoxal, and glutaraldehyde, are the earliest and most widely used crosslinking agents. Epoxides, such as epichlorohydrin, have active epoxy and chlorine groups in their molecules. Phosphates, such as sodium trimetaphosphate and sodium hexametaphosphate, are commonly used to prepare phosphate-crosslinked starch.

[0022] In some embodiments, calculating the modified performance parameters based on the infrared spectrum specifically includes the following process: Get the characteristic wavelength reflectivity increasing relationship diagram of the infrared spectrum, add the selected band to the characteristic wavelength reflectivity increasing relationship diagram, and select three eigenvectors for the selected band 、 、 , the wavelengths corresponding to the vectors are 、 、 , select the following formula to calculate the modified performance parameters : ; Among them, the three eigenvectors are The ratio of the reflectivity difference in the axial projection direction is equal.

[0023] Furthermore, extracting starch structural characteristics based on modified performance parameters specifically includes the following processes: Divide the infrared spectrum image into several parts and determine the Does it exceed the preset threshold? If so, the part is segmented and recorded as the starch structure feature part to be extracted. The third-order moment color feature extraction method is used to extract the starch structure feature part to obtain the starch structure feature: ; in, The structural characteristics of starch. represents the number of channels in the image, It means the Pixels on the The color value of each channel, Indicates the number of pixels. Indicates the The color mean of each channel.

[0024] In some embodiments, the modification process parameter influence index, the modification performance parameter, and the starch structural characteristics are input into the CNN-LSTM performance index prediction model, and the output of the modified starch key performance index value specifically includes the following process: Based on starch structural characteristics A new vector is bound to the modification process parameter influence index and modification performance parameter ; The CNN-LSTM performance indicator prediction model consists of two parts: the encoding network f and the context network. The encoder network f consists of a five-layer convolutional network. In the structure of the five-layer convolutional network, the sizes of the convolution kernels are (10, 8, 4, 4, 4) respectively. The encoder outputs a low-frequency feature. , the hidden space after the encoder network mapping is represented by Z, where : ; Among them, m is the convolution receptive field; The context network is a convolutional neural network with the same structure as the encoding network f. It maps the hidden space into the context information space M. The output of the context network is ,in, : ; Among them, v is the convolution receptive field; Loss function based on CNN-LSTM performance indicator prediction model To train the model, the loss function It is expressed as follows: ; in, is the sequence length of the input data, is the loss of noise contrast estimation, is the true probability of the input sample, is the learning parameter; Output the key performance index values ​​of the modified starch, including paste strength, thermal stability and transparency.

[0025] In some embodiments, Figure 2 This is a system block diagram of a modified starch functional performance prediction system based on deep learning according to an embodiment of the present invention. Figure 2 As shown, the system includes: A data acquisition module is used to obtain the original starch type and modification process parameters, determine the modification process parameter index corresponding to the original starch based on the starch type, compare and analyze the modification process parameters and the modification process parameter index to obtain a comparative analysis result, input the comparative analysis result into the modification process evaluation model, and output the modification process parameter influence index; A starch structure feature extraction module is used to obtain the infrared spectrum of the original starch structure, calculate the modified performance parameters based on the infrared spectrum, and extract the starch structure features based on the modified performance parameters; The starch functional performance prediction module is used to input the modification process parameter influence index, modification performance parameters and starch structural characteristics into the CNN-LSTM performance index prediction model and output the key performance index values ​​of the modified starch.

[0026] The above embodiments can be implemented in whole or in part via software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product comprises one or more computer instructions or computer programs. When loaded or executed on a computer, the processes or functions described in the embodiments of this application are fully or partially generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired means (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium accessible by a computer or a data storage device such as a server or data center that contains a collection of one or more available media. The available medium can be magnetic media (e.g., floppy disks, hard disks, tapes), optical media (e.g., DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0027] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0028] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0029] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is only for some logical functions. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0030] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0031] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A method for predicting the functional properties of modified starch based on deep learning, characterized in that the method include: Obtaining the original starch type and modification process parameters, determining the modification process parameter index corresponding to the original starch based on the starch type, performing comparative analysis on the modification process parameters and the modification process parameter index to obtain comparative analysis results, and inputting the comparative analysis results into the modification process evaluation model to output the modification process parameter influence index; Obtaining an infrared spectrum of the original starch structure, calculating modified performance parameters based on the infrared spectrum, and extracting starch structural characteristics based on the modified performance parameters; The modification process parameter influence index, modification performance parameters and starch structural characteristics are input into the CNN-LSTM performance index prediction model, and the key performance index values ​​of the modified starch are output.

2. A method for predicting the functional properties of modified starch based on deep learning according to claim 1, characterized in that: The modified process parameters and modified process parameter indicators were compared and analyzed to obtain the specific comparative analysis results. The following processes are included: Obtaining a pH value in the modification process parameter, obtaining a preset pH value in the modification process parameter index, and calculating a difference A between the pH value in the modification process parameter and the preset pH value; Obtaining a modification reaction time in the modification process parameters, obtaining a preset modification reaction time in the modification process parameter index, and calculating a difference B between the modification reaction time in the modification process parameters and the preset modification reaction time; Obtaining a modification reaction temperature in the modification process parameters, obtaining a preset modification reaction temperature in the modification process parameter index, and calculating a difference C between the modification reaction temperature in the modification process parameters and the preset modification reaction temperature; Obtain the number of cross-linking agent types in the modification process parameters, and determine whether the number of cross-linking agent types exceeds the minimum cross-linking agent type threshold in the modification process parameter indicators. If so, the modification process qualification coefficient R is 1; if not, the modification process qualification coefficient R is 0; The difference A, difference B, difference C, and modification process qualification coefficient R are recorded as comparative analysis results.

3. A method for predicting the functional properties of modified starch based on deep learning according to claim 2, characterized in that: Inputting the comparative analysis results into the modification process evaluation model and outputting the modification process parameter influence index specifically includes the following process: The comparative analysis results are processed into a unified dimension and then input into the modified process evaluation formula: ； in, Indicates the influence index of modification process parameters, 、 is the weight coefficient, The pH value of the original starch is the difference between the pH value of the modified process parameters; is the change in crystallinity of the original starch; It is the difference between the single cross-linking agent addition amount and the minimum single cross-linking agent addition amount; the value of e is 2.

72.

4. The method for predicting the functional properties of modified starch based on deep learning according to claim 1, wherein: The calculation of modified performance parameters based on infrared spectra specifically includes the following process: Get the characteristic wavelength reflectivity increasing relationship diagram of the infrared spectrum, add the selected band to the characteristic wavelength reflectivity increasing relationship diagram, and select three eigenvectors for the selected band 、 、 , the wavelengths corresponding to the vectors are 、 、 , select the following formula to calculate the modified performance parameters : ； Among them, the three eigenvectors are The ratio of the reflectivity difference in the axial projection direction is equal.

5. A method for predicting the functional properties of modified starch based on deep learning according to claim 4, characterized in that: Extraction of starch structural characteristics based on modified performance parameters The following processes are included: Divide the infrared spectrum image into several parts and determine the Does it exceed the preset threshold? If so, the part is segmented and recorded as the starch structure feature part to be extracted. The third-order moment color feature extraction method is used to extract the starch structure feature part to obtain the starch structure feature: ； in, The structural characteristics of starch. represents the number of channels in the image, It means the Pixels on the The color value of each channel, Indicates the number of pixels. Indicates the The color mean of each channel.

6. The method for predicting the functional properties of modified starch based on deep learning according to claim 1, wherein: The modification process parameter influence index, modification performance parameters, and starch structural characteristics are input into the CNN-LSTM performance index prediction model. The output of the key performance index values ​​of the modified starch specifically includes the following process: Based on starch structural characteristics A new vector is bound to the modification process parameter influence index and modification performance parameter ; The CNN-LSTM performance indicator prediction model consists of two parts: the encoding network f and the context network. The encoder network f consists of a five-layer convolutional network. In the structure of the five-layer convolutional network, the sizes of the convolution kernels are (10, 8, 4, 4, 4) respectively. The encoder outputs a low-frequency feature. , the hidden space after the encoder network mapping is represented by Z, where : ； Among them, m is the convolution receptive field; The context network is a convolutional neural network with the same structure as the encoding network f. It maps the hidden space into the context information space M. The output of the context network is ,in, : ； Among them, v is the convolution receptive field; Loss function based on CNN-LSTM performance indicator prediction model To train the model, the loss function It is expressed as follows: ； in, is the sequence length of the input data, is the loss of noise contrast estimation, is the true probability of the input sample, is the learning parameter; Output the key performance index values ​​of the modified starch, including paste strength, thermal stability and transparency.

7. A modified starch functional performance prediction system based on deep learning, characterized in that: A method for predicting the functional properties of modified starch based on deep learning applicable to any one of claims 1 to 6, the system comprising: A data acquisition module is used to obtain the original starch type and modification process parameters, determine the modification process parameter index corresponding to the original starch based on the starch type, compare and analyze the modification process parameters and the modification process parameter index to obtain a comparative analysis result, input the comparative analysis result into the modification process evaluation model, and output the modification process parameter influence index; A starch structure feature extraction module is used to obtain the infrared spectrum of the original starch structure, calculate the modified performance parameters based on the infrared spectrum, and extract the starch structure features based on the modified performance parameters; The starch functional performance prediction module is used to input the modification process parameter influence index, modification performance parameters and starch structural characteristics into the CNN-LSTM performance index prediction model and output the key performance index values ​​of the modified starch.

Images