Artificial intelligence-based method and system for identifying and analyzing saline-alkaline tolerant variety of mulberry

Abstract

The invention discloses a mulberry saline-alkaline tolerance variety identification and analysis method and system based on artificial intelligence, and the method comprises the steps: obtaining mulberry image data of different varieties of mulberries in each growth stage, carrying out the preprocessing, extracting attribute features, carrying out the classification, and constructing a first topological structure diagram; obtaining growth image data of each mulberry variety in different physiological environments, performing stress feature analysis and constructing a second topological structure chart; based on the first topological structure chart and the second topological structure chart, evaluating the saline-alkaline tolerance of each type of mulberries to obtain saline-alkaline tolerance evaluation information; and constructing a mulberry variety identification and analysis model, establishing a mulberry knowledge graph by using the first topological structure chart, the second topological structure chart and the saline-alkaline tolerance evaluation information, carrying out model training, identifying and analyzing the mulberry to be identified, and judging whether the mulberry is suitable for being planted in an expected planting area. The accuracy and the comprehensiveness of identifying the saline-alkaline-tolerant variety of the mulberry tree are improved; and the breeding identification period is shortened.

Description

Technical Field

[0001] The present invention relates to the technical field of mulberry variety identification, and in particular to an artificial intelligence-based salt-alkali tolerant mulberry variety identification and analysis method and system. Background Art

[0002] In the current agricultural development process, the improvement and utilization of saline-alkali land has become one of the important research directions in the world. However, due to the significant restrictions of saline-alkali stress on plant growth, the selection and cultivation of salt-alkali-tolerant crop varieties with strong adaptability is an important means to effectively improve the utilization efficiency of saline-alkali land. Under saline-alkali stress, the physiological characteristics and morphological features of plants will change significantly, such as changes in leaf color, thickening of surface texture, increase in dry areas, and shrinkage of plant morphology.

[0003] Traditionally, the identification of salt-alkali tolerant mulberry varieties is often carried out through expert experience and other methods, which has subjective problems and the problem of being unable to quantify salt-alkali tolerance, thus making it difficult to select and breed varieties for saline-alkali land. Therefore, how to improve the accuracy and comprehensiveness of the identification of salt-alkali tolerant mulberry varieties, thereby reducing the identification cycle of breeding work and providing certain guidance is an urgent problem to be solved. Summary of the invention

[0004] The present invention overcomes the defects of the prior art and provides an artificial intelligence-based salt-alkali tolerant mulberry variety identification and analysis method and system, the important purpose of which is to improve the accuracy and comprehensiveness of the identification of salt-alkali tolerant mulberry varieties.

[0005] To achieve the above-mentioned purpose, the first aspect of the present invention provides a method for identifying and analyzing salt-alkali tolerant mulberry varieties based on artificial intelligence, comprising: Acquire mulberry tree image data of different varieties of mulberry trees at various growth stages and perform preprocessing, extract attribute features of the preprocessed mulberry tree images and perform category classification, and construct a first topological structure diagram; Acquire growth image data of each mulberry tree variety in different physiological environments in the first topological structure diagram, analyze stress characteristics of the corresponding mulberry tree variety under each physiological environment, and construct a second topological structure diagram; Based on the first topological structure diagram and the second topological structure diagram, the salt-alkali tolerance of each type of mulberry tree under different physiological environments is evaluated to obtain salt-alkali tolerance evaluation information; Constructing a mulberry variety identification and analysis model, using the first topological structure diagram, the second topological structure diagram and the salt-alkali tolerance evaluation information to establish a mulberry knowledge graph and perform model training, identifying and analyzing the mulberry trees to be identified, and obtaining mulberry tree identification and analysis information; According to the mulberry tree identification and analysis information, it is determined whether the mulberry tree variety to be identified is suitable for planting in the desired planting area.

[0006] In this solution, the steps of obtaining image data of mulberry trees of different varieties at different growth stages and preprocessing them, extracting attribute features of the preprocessed mulberry tree images and classifying them into categories, and constructing a first topological structure diagram specifically include: Acquire mulberry tree image data of different varieties of mulberry trees at various growth stages, input the acquired mulberry tree image data into a Gaussian filter for smoothing filtering, and perform wavelet transformation on the filtered mulberry tree image data; The filtered image is decomposed into sub-bands of different scales and frequencies by wavelet transform, and threshold processing is performed by a preset processing threshold, and the wavelet coefficients smaller than the processing threshold are set to zero; After completing the threshold processing, high-frequency wavelet coefficients are extracted, and the extracted high-frequency wavelet coefficients are linearly enhanced by a linear enhancement method, and wavelet coefficients are merged based on the enhanced high-frequency wavelet coefficients; Acquire the merged wavelet coefficients and perform inverse wavelet transform, reconstruct subband information of different scales and frequencies to obtain a reconstructed mulberry tree image, and extract attribute features based on the reconstructed mulberry tree image to obtain attribute feature information; The attribute feature information is the variety attribute and growth stage attribute corresponding to each reconstructed mulberry tree image, and the attribute data label of each reconstructed image is generated according to the attribute feature information, and each reconstructed image is classified into categories in combination with a clustering algorithm to generate a first data set; Introducing the ORB image extraction algorithm to perform feature calculation and extract image features on each category of mulberry tree images in the first data set, and obtain feature vectors of each mulberry tree variety category at different growth stages; A directed description relationship is constructed with the mulberry tree type as the first node and the growth stage as the second node, the first node and the second node are connected to form a topological structure diagram, and the characteristic vectors of each mulberry tree variety category at different growth stages are associated to obtain a first topological structure diagram.

[0007] In this solution, the step of obtaining growth image data of each mulberry tree variety in different physiological environments in the first topological structure diagram, analyzing the stress characteristics of the corresponding mulberry tree variety in each physiological environment, and constructing a second topological structure diagram specifically includes: Based on big data retrieval, the growth image data of each mulberry tree variety in the first topological structure diagram in different physiological environments are obtained, and image preprocessing is performed on the obtained growth image data; The physiological environment characteristics of the mulberry tree in each growth image are obtained through the preprocessed growth image data, the Euclidean distance value between the physiological environment characteristics corresponding to each image is calculated, and the preprocessed growth image is classified by using the Euclidean distance value to obtain a number of physiological environment category subsets; The color moment method is used to calculate the color moment of each channel in the RGB three-color channels of each growth image in each category subset, including the first-order moment, the second-order moment and the third-order moment, and the calculated color moments are connected to form the color features of the corresponding growth image; Extracting morphological features of each growth image in each physiological environment category subset based on a machine vision algorithm, wherein the morphological features include leaf texture features, leaf edge features, and leaf area features; The extracted color features and morphological features constitute the stress features of each category subset, which are used to characterize the survival performance of the corresponding mulberry varieties in each physiological environment category subset and obtain stress feature information; The physiological environment characteristics of each physiological environment category subset are extracted, a second topological structure diagram is constructed with the mulberry tree species as the first node and the physiological environment characteristics as the second node, and the stress characteristic information is associated with the second topological structure diagram.

[0008] In this solution, the salt-alkali tolerance of each type of mulberry tree under different physiological environments is evaluated based on the first topological structure diagram and the second topological structure diagram to obtain salt-alkali tolerance evaluation information, specifically including: Obtaining a first topological structure diagram and a second topological structure diagram, extracting image features of mulberry trees of different categories at different growth stages based on the first topological structure diagram, and inputting the features into a generative adversarial network for adversarial training; By inputting the image features of mulberry trees of different categories at different growth stages, the generative adversarial network is made to learn the normal survival performance characteristics of mulberry trees of different categories, and the trained generative adversarial network is output; Obtaining survival performance characteristics of each category of mulberry trees under different physiological environments according to the second topological structure diagram, inputting them into the trained generative adversarial network for reconstruction analysis, and obtaining reconstructed survival performance characteristics of each category of mulberry trees under different physiological environments; Calculate the reconstruction error between the reconstructed survival performance characteristics of each category of mulberry trees under different physiological environments and the initial input survival performance characteristics of each category of mulberry trees under different physiological environments; Several reconstruction error intervals were set and associated with the salt-alkali tolerance levels. The salt-alkali tolerance of each category of mulberry trees under different physiological environments was evaluated through the calculated reconstruction errors, and the salt-alkali tolerance evaluation information was obtained.

[0009] In this solution, the construction of the mulberry variety identification and analysis model uses the first topological structure diagram, the second topological structure diagram and the salt-alkali tolerance evaluation information to establish a mulberry knowledge graph and perform model training, and the mulberry trees to be identified are identified and analyzed, specifically including: Obtaining a first topological structure diagram, a second topological structure diagram, and salt-alkali tolerance evaluation information, extracting survival performance characteristics of each mulberry tree variety at different growth stages based on the first topological structure diagram, and extracting survival performance characteristics of each mulberry tree variety under different physiological environments based on the second topological structure diagram; The survival performance characteristics of each mulberry tree variety at different growth stages are combined with the survival performance characteristics of each mulberry tree variety under different physiological environments to obtain the survival performance characteristics of each mulberry tree variety at each growth stage under different physiological environments; According to the survival performance characteristics of each mulberry variety at each growth stage under different physiological environments and the salt-alkali tolerance evaluation information, a mulberry knowledge graph with mulberry variety-physiological environment-salt-alkali tolerance as the association path is constructed; A mulberry variety identification and analysis model is constructed using a dual-channel convolutional neural network, and a training data set is constructed based on the mulberry knowledge graph to perform deep learning and training on the mulberry variety identification and analysis model to obtain a mulberry variety identification and analysis model that meets expectations; Obtaining information about the mulberry tree to be identified and information about the expected planting area, inputting the information into the mulberry tree variety identification and analysis model to identify and analyze the current unknown mulberry tree, extracting image features of the mulberry tree to be identified through a first channel, and extracting regional environmental features of the expected planting area through a second channel; The extracted image features and regional environmental features are input into the set parameter-sharing fully connected layer for weighted fusion. The variety of the mulberry trees to be identified and the salt-alkali tolerance are scored based on the fused features to obtain the mulberry tree identification analysis information.

[0010] In this solution, judging whether the mulberry tree variety to be identified is suitable for planting in the desired planting area according to the mulberry tree identification and analysis information specifically includes: Acquire mulberry tree identification and analysis information, extract the salt-alkali tolerance score of the target mulberry tree variety in the current expected planting area based on the mulberry tree identification and analysis information, and make a judgment with a preset threshold; If the salt-alkali tolerance score of the target mulberry variety in the current expected planting area is greater than the preset threshold, it means that the target mulberry variety is suitable for planting in the current expected planting area, and the recommended planting information is generated; If the salt-alkali tolerance score of the target mulberry variety in the current expected planting area is less than the preset threshold, it means that the target mulberry variety is not suitable for planting in the current expected planting area, and a non-planting recommendation information is generated.

[0011] The second aspect of the present invention provides a system for identifying and analyzing salt-alkali tolerant mulberry varieties based on artificial intelligence, the system comprising: a memory and a processor, the memory containing a program for identifying and analyzing salt-alkali tolerant mulberry varieties based on artificial intelligence, and the program for identifying and analyzing salt-alkali tolerant mulberry varieties based on artificial intelligence implementing the following steps when executed by the processor: Acquire mulberry tree image data of different varieties of mulberry trees at various growth stages and perform preprocessing, extract attribute features of the preprocessed mulberry tree images and perform category classification, and construct a first topological structure diagram; Acquire growth image data of each mulberry tree variety in different physiological environments in the first topological structure diagram, analyze stress characteristics of the corresponding mulberry tree variety under each physiological environment, and construct a second topological structure diagram; Based on the first topological structure diagram and the second topological structure diagram, the salt-alkali tolerance of each type of mulberry tree under different physiological environments is evaluated to obtain salt-alkali tolerance evaluation information; Constructing a mulberry variety identification and analysis model, using the first topological structure diagram, the second topological structure diagram and the salt-alkali tolerance evaluation information to establish a mulberry knowledge graph and perform model training, identifying and analyzing the mulberry trees to be identified, and obtaining mulberry tree identification and analysis information; According to the mulberry tree identification and analysis information, it is determined whether the mulberry tree variety to be identified is suitable for planting in the desired planting area.

[0012] The invention discloses a salt-alkali tolerant mulberry variety identification and analysis method and system based on artificial intelligence, including: obtaining mulberry image data of different varieties of mulberry trees at various growth stages and preprocessing, extracting attribute features and classifying, and constructing a first topological structure diagram; obtaining growth image data of each mulberry variety in different physiological environments, analyzing stress characteristics and constructing a second topological structure diagram; evaluating the salt-alkali tolerance of each category of mulberry trees based on the first topological structure diagram and the second topological structure diagram, and obtaining salt-alkali tolerance evaluation information; constructing a mulberry variety identification and analysis model, using the first topological structure diagram, the second topological structure diagram and the salt-alkali tolerance evaluation information to establish a mulberry knowledge map and perform model training, identifying and analyzing the mulberry trees to be identified, and judging whether they are suitable for planting in the desired planting area. The accuracy and comprehensiveness of the identification of salt-alkali tolerant mulberry varieties are improved and the breeding identification cycle is reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the technical solutions in the embodiments or exemplary embodiments of the present invention, the drawings required for use in the embodiments or exemplary descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained according to the drawings without paying creative work.

[0014] Figure 1 A flow chart of a method for identifying and analyzing salt-alkali tolerant mulberry varieties based on artificial intelligence provided by one embodiment of the present invention; Figure 2 A flow chart of a method for analyzing the planting of salt-alkali tolerant mulberry varieties provided in one embodiment of the present invention; Figure 3 A block diagram of an artificial intelligence-based identification and analysis system for salt-alkali tolerant mulberry varieties provided in one embodiment of the present invention; The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0015] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0016] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0017] Figure 1 A flow chart of a method for identifying and analyzing salt-alkali tolerant mulberry varieties based on artificial intelligence provided by one embodiment of the present invention; like Figure 1 As shown, the present invention provides a flow chart of a method for identifying and analyzing salt-alkali tolerant mulberry varieties based on artificial intelligence, comprising: S102, obtaining and preprocessing mulberry tree image data of different varieties of mulberry trees at various growth stages, extracting attribute features of the preprocessed mulberry tree images and classifying them into categories, and constructing a first topological structure diagram; S104, obtaining growth image data of each mulberry tree variety in different physiological environments in the first topological structure diagram, analyzing stress characteristics of the corresponding mulberry tree variety under each physiological environment, and constructing a second topological structure diagram; S106, evaluating the salt-alkali tolerance of each type of mulberry tree under different physiological environments based on the first topological structure diagram and the second topological structure diagram to obtain salt-alkali tolerance evaluation information; S108, constructing a mulberry variety identification and analysis model, using the first topological structure diagram, the second topological structure diagram and the salt-alkali tolerance evaluation information to establish a mulberry knowledge graph and perform model training, identifying and analyzing the mulberry trees to be identified, and obtaining mulberry tree identification and analysis information; S110, judging whether the mulberry tree variety to be identified is suitable for planting in the desired planting area according to the mulberry tree identification and analysis information.

[0018] It should be noted that the present invention provides a method for identifying and analyzing salt-alkali tolerant varieties of mulberry trees based on artificial intelligence. By obtaining leaf and plant images of mulberry trees under salt-alkali stress, image processing technology is used to extract multidimensional phenotypic data such as color, texture, morphology and edge features, and then a comprehensive analysis is performed in combination with multi-source data such as soil salinity and environmental conditions, thereby identifying and analyzing mulberry varieties. The stress characteristics represent the survival performance of this mulberry variety under different physiological environments. The salt-alkali tolerance is analyzed by the difference between the survival performance and the survival performance under a normal suitable living environment, thereby obtaining a salt-alkali tolerance evaluation result. Subsequently, a dual-channel convolutional neural network is used to establish a mulberry variety identification and analysis model, and a mulberry knowledge map is established using the first topological structure diagram, the second topological structure diagram and the salt-alkali tolerance evaluation information, and the mulberry trees to be identified are identified and analyzed to obtain mulberry identification and analysis information. According to the mulberry identification and analysis information, it is judged whether the mulberry variety to be identified is suitable for planting in the desired planting area. Improve the accuracy of the identification result.

[0019] Further, in a preferred embodiment of the present invention, the method of obtaining mulberry tree image data of different varieties of mulberry trees at various growth stages and preprocessing them, extracting attribute features of the preprocessed mulberry tree images and classifying them, and constructing a first topological structure diagram specifically includes: Acquire mulberry tree image data of different varieties of mulberry trees at various growth stages, input the acquired mulberry tree image data into a Gaussian filter for smoothing filtering, and perform wavelet transformation on the filtered mulberry tree image data; The filtered image is decomposed into sub-bands of different scales and frequencies by wavelet transform, and threshold processing is performed by a preset processing threshold, and the wavelet coefficients smaller than the processing threshold are set to zero; After completing the threshold processing, high-frequency wavelet coefficients are extracted, and the extracted high-frequency wavelet coefficients are linearly enhanced by a linear enhancement method, and wavelet coefficients are merged based on the enhanced high-frequency wavelet coefficients; Acquire the merged wavelet coefficients and perform inverse wavelet transform, reconstruct subband information of different scales and frequencies to obtain a reconstructed mulberry tree image, and extract attribute features based on the reconstructed mulberry tree image to obtain attribute feature information; The attribute feature information is the variety attribute and growth stage attribute corresponding to each reconstructed mulberry tree image, and the attribute data label of each reconstructed image is generated according to the attribute feature information, and each reconstructed image is classified into categories in combination with a clustering algorithm to generate a first data set; Introducing the ORB image extraction algorithm to perform feature calculation and extract image features on each category of mulberry tree images in the first data set, and obtain feature vectors of each mulberry tree variety category at different growth stages; A directed description relationship is constructed with the mulberry tree type as the first node and the growth stage as the second node, the first node and the second node are connected to form a topological structure diagram, and the characteristic vectors of each mulberry tree variety category at different growth stages are associated to obtain a first topological structure diagram.

[0020] It should be noted that high-resolution image data of different mulberry varieties at various growth stages are first obtained and input into a Gaussian filter for smoothing. Smoothing filtering can effectively reduce random noise in the image while retaining the main structure and feature information of the mulberry image. The filtered image data is subjected to wavelet transform to decompose the image into sub-bands of different scales and frequencies. The multi-scale analysis characteristics of wavelet transform can hierarchically analyze different information in the image, thereby separating detail information (high-frequency components) from global structures (low-frequency components). The wavelet coefficients are processed by a preset processing threshold, and the wavelet coefficients below the threshold are set to zero to remove weak signals or residual noise. After the threshold processing is completed, the high-frequency wavelet coefficients are extracted and enhanced by a linear enhancement method to improve the visibility and resolution of the details in the image. After the wavelet coefficients are merged, the sub-band information of different scales and frequencies is reconstructed by inverse wavelet transform to obtain a reconstructed mulberry image. These reconstructed images more clearly retain the important information of mulberry varieties and growth stages. Next, attribute feature information is extracted from the reconstructed image, including the variety attribute and growth stage attribute corresponding to the image, which is used to generate the attribute data label of the reconstructed image. Combined with the clustering algorithm, the reconstructed images with similar features are classified to form the first data set. The ORB (Oriented FAST and Rotated BRIEF) image extraction algorithm is introduced to calculate and extract the key features of each category of mulberry images in the first data set, and the feature vectors of different mulberry varieties at each growth stage are obtained to describe the growth morphology of different categories of mulberry trees at different stages. Subsequently, a topological structure diagram is constructed, with the mulberry variety category as the first node and the growth stage as the second node. The nodes are connected according to a directed relationship to form a topological structure that describes the association between mulberry varieties and growth stages. Finally, the above-extracted image feature vectors are associated with the topological graph structure to generate the first topological structure diagram, which provides support for subsequent variety identification and analysis.

[0021] Further, in a preferred embodiment of the present invention, the step of acquiring growth image data of each mulberry tree variety in different physiological environments in the first topological structure diagram, analyzing the stress characteristics of the corresponding mulberry tree variety under each physiological environment, and constructing the second topological structure diagram specifically includes: Based on big data retrieval, the growth image data of each mulberry tree variety in the first topological structure diagram in different physiological environments are obtained, and image preprocessing is performed on the obtained growth image data; The physiological environment characteristics of the mulberry tree in each growth image are obtained through the preprocessed growth image data, the Euclidean distance value between the physiological environment characteristics corresponding to each image is calculated, and the preprocessed growth image is classified by using the Euclidean distance value to obtain a number of physiological environment category subsets; The color moment method is used to calculate the color moment of each channel in the RGB three-color channels of each growth image in each category subset, including the first-order moment, the second-order moment and the third-order moment, and the calculated color moments are connected to form the color features of the corresponding growth image; Extracting morphological features of each growth image in each physiological environment category subset based on a machine vision algorithm, wherein the morphological features include leaf texture features, leaf edge features, and leaf area features; The extracted color features and morphological features constitute the stress features of each category subset, which are used to characterize the survival performance of the corresponding mulberry varieties in each physiological environment category subset and obtain stress feature information; The physiological environment characteristics of each physiological environment category subset are extracted, a second topological structure diagram is constructed with the mulberry tree species as the first node and the physiological environment characteristics as the second node, and the stress characteristic information is associated with the second topological structure diagram.

[0022] It should be noted that first, the growth image data of different mulberry varieties in the first topological structure diagram under various physiological environments are retrieved and obtained using big data technology. The features reflecting the performance of mulberry trees under different physiological environments are extracted from each image. These physiological environmental features include light, humidity, temperature, and salinity concentration. By calculating the Euclidean distance values ​​between the physiological environmental features corresponding to each image, their similarities or differences are quantified, and the growth images are automatically classified according to the Euclidean distance values ​​to form several physiological environment category subsets with similar environmental conditions. For each physiological environment category subset, the color moment method is used to analyze the color distribution characteristics of each growth image, and the first-order moment, second-order moment, and third-order moment of each image in the three RGB color channels are calculated respectively to quantify the color mean, standard deviation, and skewness of each channel. The calculated color moment feature values ​​are connected to form the overall color features of each image, which can fully reflect the color changes of leaves under different physiological environments. At the same time, morphological features are extracted from each growth image based on machine vision algorithms, including texture features, edge features, and area features of leaves. Among them, texture features capture the surface roughness of leaves and the distribution of pathological textures through grayscale co-occurrence matrix and other methods; edge features depict the curling or damage of leaf shapes through edge detection algorithms; leaf area features are used to quantify the impact of saline-alkali stress on mulberry growth. Then, stress feature information of each physiological environment category subset is generated through these color and morphological features, which is used to characterize the survival status and adaptive performance of the corresponding mulberry varieties in each category subset. Next, the physiological environment characteristics of each physiological environment category subset are associated with the mulberry species information, and the second topological structure diagram is constructed with the mulberry variety as the first node and the physiological environment characteristics as the second node. The stress feature information extracted from each category subset is associated with the second topological structure diagram to form a comprehensive analysis framework for quantifying and evaluating the performance of mulberry varieties under different environmental conditions.

[0023] Further, in a preferred embodiment of the present invention, the salt-alkali tolerance of each type of mulberry tree under different physiological environments is evaluated based on the first topological structure diagram and the second topological structure diagram to obtain salt-alkali tolerance evaluation information, specifically including: Obtaining a first topological structure diagram and a second topological structure diagram, extracting image features of mulberry trees of different categories at different growth stages based on the first topological structure diagram, and inputting the features into a generative adversarial network for adversarial training; By inputting the image features of mulberry trees of different categories at different growth stages, the generative adversarial network is made to learn the normal survival performance characteristics of mulberry trees of different categories, and the trained generative adversarial network is output; Obtaining survival performance characteristics of each category of mulberry trees under different physiological environments according to the second topological structure diagram, inputting them into the trained generative adversarial network for reconstruction analysis, and obtaining reconstructed survival performance characteristics of each category of mulberry trees under different physiological environments; Calculate the reconstruction error between the reconstructed survival performance characteristics of each category of mulberry trees under different physiological environments and the initial input survival performance characteristics of each category of mulberry trees under different physiological environments; Several reconstruction error intervals were set and associated with the salt-alkali tolerance levels. The salt-alkali tolerance of each category of mulberry trees under different physiological environments was evaluated through the calculated reconstruction errors, and the salt-alkali tolerance evaluation information was obtained.

[0024] It should be noted that, based on the first topological structure diagram, image features of mulberry trees of various categories at different growth stages are extracted from it. These features are specifically manifested as color features, texture features, morphological features, etc., which fully reflect the multidimensional properties of mulberry trees under their growth state. The extracted image features are then used as input and sent to the generative adversarial network (GAN) for adversarial training. In the generative adversarial network, the generator and the discriminator play a dynamic game. The generator tries to simulate and generate mulberry image features close to the real one, while the discriminator is responsible for distinguishing between real data and generated data. During the adversarial training process, the network gradually masters the normal survival performance characteristics of mulberry trees of various categories at different growth stages through continuous iterative learning. Finally, the trained generative adversarial network is output, and its generation ability can reflect the normal survival state of mulberry trees at different stages. Next, the survival performance characteristics of mulberry trees of various categories under different physiological environments recorded in the second topological structure diagram are sent to the trained generative adversarial network for reconstruction analysis. In this analysis process, the network attempts to generate the reconstructed data that the network considers based on the input features, that is, it tends to generate normal survival performance characteristics. Subsequently, the reconstruction error between the reconstructed survival performance characteristics of each category of mulberry trees under different physiological environments and their initial input characteristics was calculated. The reconstruction error reflects the difference between the growth state and the normal state of the corresponding mulberry tree variety under each physiological environment, that is, it represents the salt-alkali tolerance. Several reconstruction error intervals are set and the reconstruction error intervals are associated with the salt-alkali tolerance level. The salt-alkali tolerance of each category of mulberry trees under different physiological environments is evaluated through the calculated reconstruction error, and finally the salt-alkali tolerance evaluation information is output. This can efficiently and accurately quantify the salt-alkali tolerance level of different mulberry tree varieties, and also provide a scientific basis for variety screening and breeding strategies, which is helpful to meet the actual needs of agricultural development in saline-alkali land.

[0025] Further, in a preferred embodiment of the present invention, the construction of the mulberry variety identification and analysis model uses the first topological structure diagram, the second topological structure diagram and the salt-alkali tolerance evaluation information to establish a mulberry knowledge graph and perform model training, and the mulberry trees to be identified are identified and analyzed, specifically including: Obtaining a first topological structure diagram, a second topological structure diagram, and salt-alkali tolerance evaluation information, extracting survival performance characteristics of each mulberry tree variety at different growth stages based on the first topological structure diagram, and extracting survival performance characteristics of each mulberry tree variety under different physiological environments based on the second topological structure diagram; The survival performance characteristics of each mulberry tree variety at different growth stages are combined with the survival performance characteristics of each mulberry tree variety under different physiological environments to obtain the survival performance characteristics of each mulberry tree variety at each growth stage under different physiological environments; According to the survival performance characteristics of each mulberry variety at each growth stage under different physiological environments and the salt-alkali tolerance evaluation information, a mulberry knowledge graph with mulberry variety-physiological environment-salt-alkali tolerance as the association path is constructed; A mulberry variety identification and analysis model is constructed using a dual-channel convolutional neural network, and a training data set is constructed based on the mulberry knowledge graph to perform deep learning and training on the mulberry variety identification and analysis model to obtain a mulberry variety identification and analysis model that meets expectations; Obtaining information about the mulberry tree to be identified and information about the expected planting area, inputting the information into the mulberry tree variety identification and analysis model to identify and analyze the current unknown mulberry tree, extracting image features of the mulberry tree to be identified through a first channel, and extracting regional environmental features of the expected planting area through a second channel; The extracted image features and regional environmental features are input into the set parameter-sharing fully connected layer for weighted fusion. The variety of the mulberry trees to be identified and the salt-alkali tolerance are scored based on the fused features to obtain the mulberry tree identification analysis information.

[0026] It should be noted that based on the first topological structure diagram, the survival performance characteristics of each mulberry variety at different growth stages are extracted, and these characteristics reflect the performance of mulberry trees over time under normal conditions. At the same time, based on the second topological structure diagram, the survival performance characteristics of each mulberry variety under different physiological environments are extracted to reveal the impact of special conditions such as saline-alkali environment on its adaptability. By fusing the above two types of features, the characteristics of each mulberry variety at different growth stages are combined with its performance under different physiological environments to generate more comprehensive information. This fusion feature can comprehensively characterize the survival performance of each mulberry variety under diverse environments and growth conditions. Combined with the salt-alkali tolerance evaluation information, these multidimensional features are used to construct a knowledge graph with mulberry varieties, physiological environment and salt-alkali tolerance as the core association paths. This knowledge graph systematically integrates the growth characteristics, environmental adaptability and stress resistance of mulberry varieties, providing a structured knowledge basis for subsequent analysis. A dual-channel convolutional neural network (CNN) is used to design a mulberry variety identification and analysis model. The data extracted from the mulberry knowledge graph is used as a training data set to train the model for deep learning, ensuring that it can effectively identify and distinguish the characteristic performance of mulberry varieties and obtain a model with the expected performance. For the mulberry trees to be identified, their image information and environmental information of the target planting area are obtained and input into the trained identification and analysis model. The first channel of the model extracts the image features of the mulberry trees to be identified through convolution operations, such as leaf texture, edge shape and color change; the second channel extracts environmental characteristics of the target planting area, such as soil salinity, humidity and climatic conditions. These features are weighted fused through a preset parameter-sharing fully connected layer to associate the biological characteristics and environmental adaptability of mulberry trees. The fused features are further used for variety identification and salt-alkali tolerance scoring. Based on the weights of different features, the model accurately determines the specific varieties of mulberry trees to be identified, and scores their adaptability in the target area to generate the final mulberry identification and analysis information. The accurate identification of mulberry varieties and the scientific evaluation of salt-alkali tolerance are achieved, providing an efficient and intelligent solution for the selection of mulberry varieties for saline-alkali land.

[0027] Further, in a preferred embodiment of the present invention, judging whether the mulberry tree variety to be identified is suitable for planting in the desired planting area according to the mulberry tree identification analysis information specifically includes: Acquire mulberry tree identification and analysis information, extract the salt-alkali tolerance score of the target mulberry tree variety in the current expected planting area based on the mulberry tree identification and analysis information, and make a judgment with a preset threshold; If the salt-alkali tolerance score of the target mulberry variety in the current expected planting area is greater than the preset threshold, it means that the target mulberry variety is suitable for planting in the current expected planting area, and the recommended planting information is generated; If the salt-alkali tolerance score of the target mulberry variety in the current expected planting area is less than the preset threshold, it means that the target mulberry variety is not suitable for planting in the current expected planting area, and a non-planting recommendation information is generated.

[0028] It should be noted that, first, the salt-alkali tolerance score of the target mulberry variety in the current expected planting area is extracted by obtaining the mulberry identification analysis information. The score combines the characteristics of the mulberry variety and the environmental conditions of the expected area, reflecting the potential performance ability of the target variety under actual planting conditions. By comparing the score value with the preset threshold, it is judged whether the variety is suitable for planting in the target area. When the salt-alkali tolerance score of the target mulberry variety is higher than the preset threshold, it means that the variety has strong adaptability and resistance to the saline-alkali conditions in the target area and can grow well in the area. Therefore, recommended planting information is generated to provide guidance for growers and suggest that they choose the variety for planting in the target area. On the contrary, if the score is lower than the preset threshold, it indicates that the target mulberry variety may be difficult to adapt to salt-alkali stress or other environmental constraints in the current area and is not suitable for planting. At this time, information not recommended for planting will be generated to prompt growers that the variety is not adaptable enough in the target area, thereby avoiding waste of resources or failed planting attempts. Intelligently evaluate the environmental suitability of the planting area, provide a scientific basis for planting decisions, and improve the success rate and efficiency of mulberry planting projects.

[0029] Figure 2 A flow chart of a method for analyzing the planting of salt-alkali tolerant mulberry varieties provided in one embodiment of the present invention; like Figure 2 As shown, the present invention provides a flow chart of a planting analysis method for salt-alkali tolerant mulberry varieties, comprising: S202, obtaining information of the mulberry tree to be identified and information of the expected planting area, inputting the information into the mulberry tree variety identification and analysis model to identify and analyze the current unknown mulberry tree, and extracting image features of the mulberry tree to be identified and regional environmental features of the expected planting area; S204, inputting the extracted image features and regional environmental features into a set parameter-sharing fully connected layer for weighted fusion, and performing variety identification and salt-alkali tolerance scoring on the mulberry trees to be identified based on the fused features to obtain mulberry tree identification analysis information; S206, extracting the salt-alkali tolerance score of the target mulberry tree variety in the current expected planting area based on the mulberry tree identification analysis information, and comparing it with a preset threshold; S208, if the salt-alkali tolerance score of the target mulberry tree variety in the current expected planting area is greater than a preset threshold, it means that the target mulberry tree variety is suitable for planting in the current expected planting area, and then generating recommended planting information; S210, if the salt-alkali tolerance score of the target mulberry tree variety in the current expected planting area is less than a preset threshold, it means that the target mulberry tree variety is not suitable for planting in the current expected planting area, and a non-planting recommendation information is generated.

[0030] It should be noted that the information of the mulberry tree to be identified and the environmental information of the expected planting area are first input as input to the pre-trained mulberry variety identification and analysis model. The model extracts mulberry image features, such as leaf color, texture, edge characteristics, etc., through the first channel, and extracts environmental features of the planting area, including salinity, soil moisture, and climatic conditions, through the second channel. Next, the extracted image features and regional environmental features are input into the parameter-sharing fully connected layer, which uses a weighted fusion strategy to perform correlation analysis on multi-source data and generate a comprehensive feature representation. Based on the fused comprehensive features, the variety classification and salt-alkali tolerance score of the mulberry tree to be identified are output, and the mulberry identification and analysis information is output. After the identification and analysis are completed, the salt-alkali tolerance score of the target mulberry variety in the current expected planting area is extracted. The score is obtained by combining the variety characteristics and environmental adaptability output by the comprehensive model, and is the key basis for evaluating planting suitability. By comparing with the preset salt-alkali tolerance threshold, it is judged whether the target variety is suitable for planting in the target area. If the score is higher than the threshold, it means that the variety has good adaptability in the desired planting area and can grow normally in the target environment, thus generating information about recommended planting and reminding growers of the planting potential of the variety. On the contrary, if the score is lower than the threshold, it means that the target variety has weak adaptability to the desired area and may be difficult to grow due to adverse environmental conditions, thus generating information about not recommending planting and reminding growers to make cautious decisions to avoid possible planting failures or waste of resources.

[0031] Figure 3 An artificial intelligence-based salt-alkali tolerant mulberry variety identification and analysis system 3 is provided for one embodiment of the present invention. The system comprises: a memory 31 and a processor 32. The memory 31 contains an artificial intelligence-based salt-alkali tolerant mulberry variety identification and analysis method program. When the artificial intelligence-based salt-alkali tolerant mulberry variety identification and analysis method program is executed by the processor 32, the following steps are implemented: Acquire mulberry tree image data of different varieties of mulberry trees at various growth stages and perform preprocessing, extract attribute features of the preprocessed mulberry tree images and perform category classification, and construct a first topological structure diagram; Acquire growth image data of each mulberry tree variety in different physiological environments in the first topological structure diagram, analyze stress characteristics of the corresponding mulberry tree variety under each physiological environment, and construct a second topological structure diagram; Based on the first topological structure diagram and the second topological structure diagram, the salt-alkali tolerance of each type of mulberry tree under different physiological environments is evaluated to obtain salt-alkali tolerance evaluation information; Constructing a mulberry variety identification and analysis model, using the first topological structure diagram, the second topological structure diagram and the salt-alkali tolerance evaluation information to establish a mulberry knowledge graph and perform model training, identifying and analyzing the mulberry trees to be identified, and obtaining mulberry tree identification and analysis information; According to the mulberry tree identification and analysis information, it is determined whether the mulberry tree variety to be identified is suitable for planting in the desired planting area.

[0032] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0033] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0034] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0035] A person of ordinary skill in the art can understand that: all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: a mobile storage device, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0036] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0037] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for identifying and analyzing salt-alkali tolerant mulberry varieties based on artificial intelligence, characterized in that: include: Acquire mulberry tree image data of different varieties of mulberry trees at various growth stages and perform preprocessing, extract attribute features of the preprocessed mulberry tree images and perform category classification, and construct a first topological structure diagram; Acquire growth image data of each mulberry tree variety in different physiological environments in the first topological structure diagram, analyze stress characteristics of the corresponding mulberry tree variety under each physiological environment, and construct a second topological structure diagram; Based on the first topological structure diagram and the second topological structure diagram, the salt-alkali tolerance of each type of mulberry tree under different physiological environments is evaluated to obtain salt-alkali tolerance evaluation information; Constructing a mulberry variety identification and analysis model, using the first topological structure diagram, the second topological structure diagram and the salt-alkali tolerance evaluation information to establish a mulberry knowledge graph and perform model training, identifying and analyzing the mulberry trees to be identified, and obtaining mulberry tree identification and analysis information; According to the mulberry tree identification and analysis information, it is determined whether the mulberry tree variety to be identified is suitable for planting in the desired planting area.

2. The method for identifying and analyzing salt-alkali tolerant mulberry varieties based on artificial intelligence according to claim 1, characterized in that: The method of obtaining mulberry tree image data of different varieties of mulberry trees at various growth stages and preprocessing them, extracting attribute features of the preprocessed mulberry tree images and classifying them into categories, and constructing a first topological structure diagram specifically includes: Acquire mulberry tree image data of different varieties of mulberry trees at various growth stages, input the acquired mulberry tree image data into a Gaussian filter for smoothing filtering, and perform wavelet transformation on the filtered mulberry tree image data; The filtered image is decomposed into sub-bands of different scales and frequencies by wavelet transform, and threshold processing is performed by a preset processing threshold, and the wavelet coefficients smaller than the processing threshold are set to zero; After completing the threshold processing, high-frequency wavelet coefficients are extracted, and the extracted high-frequency wavelet coefficients are linearly enhanced by a linear enhancement method, and wavelet coefficients are merged based on the enhanced high-frequency wavelet coefficients; Acquire the merged wavelet coefficients and perform inverse wavelet transform, reconstruct subband information of different scales and frequencies to obtain a reconstructed mulberry tree image, and extract attribute features based on the reconstructed mulberry tree image to obtain attribute feature information; The attribute feature information is the variety attribute and growth stage attribute corresponding to each reconstructed mulberry tree image, and the attribute data label of each reconstructed image is generated according to the attribute feature information, and each reconstructed image is classified into categories in combination with a clustering algorithm to generate a first data set; Introducing the ORB image extraction algorithm to perform feature calculation and extract image features on each category of mulberry tree images in the first data set, and obtain feature vectors of each mulberry tree variety category at different growth stages; A directed description relationship is constructed with the mulberry tree type as the first node and the growth stage as the second node, the first node and the second node are connected to form a topological structure diagram, and the characteristic vectors of each mulberry tree variety category at different growth stages are associated to obtain a first topological structure diagram.

3. The method for identifying and analyzing salt-alkali tolerant mulberry varieties based on artificial intelligence according to claim 1, characterized in that: The step of obtaining growth image data of each mulberry tree variety in different physiological environments in the first topological structure diagram, analyzing stress characteristics of the corresponding mulberry tree variety in each physiological environment, and constructing a second topological structure diagram specifically includes: Based on big data retrieval, the growth image data of each mulberry tree variety in the first topological structure diagram in different physiological environments are obtained, and image preprocessing is performed on the obtained growth image data; The physiological environment characteristics of the mulberry tree in each growth image are obtained through the preprocessed growth image data, the Euclidean distance value between the physiological environment characteristics corresponding to each image is calculated, and the preprocessed growth image is classified by using the Euclidean distance value to obtain a number of physiological environment category subsets; The color moment method is used to calculate the color moment of each channel in the RGB three-color channels of each growth image in each category subset, including the first-order moment, the second-order moment and the third-order moment, and the calculated color moments are connected to form the color features of the corresponding growth image; Extracting morphological features of each growth image in each physiological environment category subset based on a machine vision algorithm, wherein the morphological features include leaf texture features, leaf edge features, and leaf area features; The extracted color features and morphological features constitute the stress features of each category subset, which are used to characterize the survival performance of the corresponding mulberry varieties in each physiological environment category subset and obtain stress feature information; The physiological environment characteristics of each physiological environment category subset are extracted, a second topological structure diagram is constructed with the mulberry tree species as the first node and the physiological environment characteristics as the second node, and the stress characteristic information is associated with the second topological structure diagram.

4. The method for identifying and analyzing salt-alkali tolerant mulberry varieties based on artificial intelligence according to claim 1, characterized in that: The step of evaluating the salt-alkali tolerance of each type of mulberry tree under different physiological environments based on the first topological structure diagram and the second topological structure diagram to obtain salt-alkali tolerance evaluation information specifically includes: Obtaining a first topological structure diagram and a second topological structure diagram, extracting image features of mulberry trees of different categories at different growth stages based on the first topological structure diagram, and inputting the features into a generative adversarial network for adversarial training; By inputting the image features of mulberry trees of different categories at different growth stages, the generative adversarial network is made to learn the normal survival performance characteristics of mulberry trees of different categories, and the trained generative adversarial network is output; Obtaining survival performance characteristics of each category of mulberry trees under different physiological environments according to the second topological structure diagram, inputting them into the trained generative adversarial network for reconstruction analysis, and obtaining reconstructed survival performance characteristics of each category of mulberry trees under different physiological environments; Calculate the reconstruction error between the reconstructed survival performance characteristics of each category of mulberry trees under different physiological environments and the initial input survival performance characteristics of each category of mulberry trees under different physiological environments; Several reconstruction error intervals were set and associated with the salt-alkali tolerance levels. The salt-alkali tolerance of each category of mulberry trees under different physiological environments was evaluated through the calculated reconstruction errors, and the salt-alkali tolerance evaluation information was obtained.

5. The method for identifying and analyzing salt-alkali tolerant mulberry varieties based on artificial intelligence according to claim 1, characterized in that: The construction of the mulberry variety identification and analysis model, using the first topological structure diagram, the second topological structure diagram and the salt-alkali tolerance evaluation information to establish a mulberry knowledge graph and perform model training, and identifying and analyzing the mulberry trees to be identified, specifically includes: Obtaining a first topological structure diagram, a second topological structure diagram, and salt-alkali tolerance evaluation information, extracting survival performance characteristics of each mulberry tree variety at different growth stages based on the first topological structure diagram, and extracting survival performance characteristics of each mulberry tree variety under different physiological environments based on the second topological structure diagram; The survival performance characteristics of each mulberry tree variety at different growth stages are combined with the survival performance characteristics of each mulberry tree variety under different physiological environments to obtain the survival performance characteristics of each mulberry tree variety at each growth stage under different physiological environments; According to the survival performance characteristics of each mulberry variety at each growth stage under different physiological environments and the salt-alkali tolerance evaluation information, a mulberry knowledge graph with mulberry variety-physiological environment-salt-alkali tolerance as the association path is constructed; A mulberry variety identification and analysis model is constructed using a dual-channel convolutional neural network, and a training data set is constructed based on the mulberry knowledge graph to perform deep learning and training on the mulberry variety identification and analysis model to obtain a mulberry variety identification and analysis model that meets expectations; Obtaining information about the mulberry tree to be identified and information about the expected planting area, inputting the information into the mulberry tree variety identification and analysis model to identify and analyze the current unknown mulberry tree, extracting image features of the mulberry tree to be identified through a first channel, and extracting regional environmental features of the expected planting area through a second channel; The extracted image features and regional environmental features are input into the set parameter-sharing fully connected layer for weighted fusion. The variety of the mulberry trees to be identified and the salt-alkali tolerance are scored based on the fused features to obtain the mulberry tree identification analysis information.

6. The method for identifying and analyzing salt-alkali tolerant mulberry varieties based on artificial intelligence according to claim 1, characterized in that: The step of judging whether the mulberry tree variety to be identified is suitable for planting in the desired planting area according to the mulberry tree identification and analysis information specifically includes: Acquire mulberry tree identification and analysis information, extract the salt-alkali tolerance score of the target mulberry tree variety in the current expected planting area based on the mulberry tree identification and analysis information, and make a judgment with a preset threshold; If the salt-alkali tolerance score of the target mulberry variety in the current expected planting area is greater than the preset threshold, it means that the target mulberry variety is suitable for planting in the current expected planting area, and the recommended planting information is generated; If the salt-alkali tolerance score of the target mulberry variety in the current expected planting area is less than the preset threshold, it means that the target mulberry variety is not suitable for planting in the current expected planting area, and a non-planting recommendation information is generated.

7. An artificial intelligence-based identification and analysis system for salt-alkali tolerant mulberry varieties, characterized in that: The system comprises: a memory and a processor, wherein the memory contains an artificial intelligence-based identification and analysis method program for salt-alkali tolerant mulberry varieties, and when the artificial intelligence-based identification and analysis method program for salt-alkali tolerant mulberry varieties is executed by the processor, the following steps are implemented: Acquire mulberry tree image data of different varieties of mulberry trees at various growth stages and perform preprocessing, extract attribute features of the preprocessed mulberry tree images and perform category classification, and construct a first topological structure diagram; Acquire growth image data of each mulberry tree variety in different physiological environments in the first topological structure diagram, analyze stress characteristics of the corresponding mulberry tree variety under each physiological environment, and construct a second topological structure diagram; Based on the first topological structure diagram and the second topological structure diagram, the salt-alkali tolerance of each type of mulberry tree under different physiological environments is evaluated to obtain salt-alkali tolerance evaluation information; Constructing a mulberry variety identification and analysis model, using the first topological structure diagram, the second topological structure diagram and the salt-alkali tolerance evaluation information to establish a mulberry knowledge graph and perform model training, identifying and analyzing the mulberry trees to be identified, and obtaining mulberry tree identification and analysis information; According to the mulberry tree identification and analysis information, it is determined whether the mulberry tree variety to be identified is suitable for planting in the desired planting area.

Images