Method and system for determining moisture content of Psoralea corylifolia seeds based on spectral technology

Abstract

The present invention relates to the technical field of water content determination, and in particular to a method and system for determining the moisture content of Psoralea corylifolia seeds based on spectral technology. The present invention obtains the water content characteristic band and material characteristic band under each stage batch according to the overall intensity of the spectral response of all seeds in different bands and the water content data under each stage batch; obtains the material similarity of the newly added seeds relative to the seeds under each stage batch according to the similarity of the overall intensity of the spectral response between the newly added seeds and the different seeds under each stage batch in the material characteristic band; constructs a water content regression equation according to the corresponding spectral response intensity and water content data of the different seeds under each stage batch in the water content characteristic band; and then obtains the water content prediction value of the newly added seeds. The present invention takes into account the similarity of the spectral data between the seeds, selects the characteristic band with larger contribution to the moisture content for regression analysis, and improves the accuracy of the determination of the moisture content of the seeds.

Description

Technical Field

[0001] The present invention relates to the technical field of water content determination, and in particular to a method and system for determining the water content of Psoralea corylifolia seeds based on spectroscopy technology. Background Art

[0002] As a species-specific seed, Peach Seven seeds have important application value in agricultural production and ecological research. With the increasing demands of modern agriculture for seed quality and planting efficiency, seed moisture content has become a key factor affecting seed storage, germination, and growth and development. Both high and low seed moisture content can affect long-term seed storage.

[0003] In the existing technology, the measurement method based on spectral technology uses SPA to select characteristic bands related to water content changes. However, because the biological characteristics of seeds at different growth stages and in different regions and the differences in water distribution are not taken into account, the same characteristic bands are selected for seeds at different periods, resulting in poor accuracy in moisture content measurement. Summary of the Invention

[0004] In order to solve the technical problem that the accuracy of moisture content measurement is poor when the same characteristic band is selected for analysis without considering the differences between seeds in different growth stages and different regions, the object of the present invention is to provide a method and system for measuring the moisture content of Psoralea corylifolia seeds based on spectral technology. The technical solutions adopted are as follows:

[0005] The present invention proposes a method for determining the moisture content of Psoralea corylifolia seeds based on spectroscopy technology, the method comprising:

[0006] Obtaining hyperspectral data and water content data of Psoralea corylifolia seeds at different stages and batches, wherein the hyperspectral data includes spectral response intensities at different positions in different bands;

[0007] For any batch of seeds at any stage, the position cluster of each seed is obtained based on the relative distances of all spectral response intensities between different positions; the overall spectral response intensity of each seed in each band is obtained based on the distribution of spectral response intensities of all positions in each seed position cluster corresponding to different bands;

[0008] Based on the overall intensity of the spectral response and water content data of all seeds in each batch at different bands, the water content characteristic band and material characteristic band of each batch are obtained; based on the similarity of the overall intensity of the spectral response of the newly added seeds and different seeds in each batch at different bands in material characteristic bands, the material similarity of the newly added seeds relative to the seeds in each batch is obtained;

[0009] A water content regression equation was constructed based on the spectral response intensity and water content data of different seeds in each stage batch in the water content characteristic band; the water content prediction value of the new seeds was obtained based on the material similarity of the new seeds relative to the seeds in different stage batches, the overall intensity of the spectral response of the water content characteristic band in the corresponding stage batches and the water content regression equation.

[0010] Furthermore, the method for obtaining the position cluster of each seed includes:

[0011] The spectral response intensities of each position and the corresponding positions are combined into a vector; the Euclidean distance between the vectors of different positions is obtained as the relative distance between the different positions;

[0012] Based on the relative distances between different locations, K-means clustering is performed on all locations to obtain the location clusters of each seed.

[0013] Furthermore, the method for obtaining the overall intensity of the spectral response includes:

[0014] The spectral response intensity of all positions in the position cluster of each seed corresponding to each same band is averaged to obtain the overall spectral response intensity of each seed in each band.

[0015] Furthermore, the method for obtaining the water content characteristic band includes:

[0016] Obtain the correlation coefficient between the sequence composed of the overall intensity of the spectral response of all seeds in each band in each batch and the sequence composed of the water content data as the first correlation coefficient; select the band corresponding to the maximum first correlation coefficient as the first water content characteristic band;

[0017] The cumulative sum of the correlation coefficients of the sequence composed of the overall intensity of the spectral response between each band and all water content characteristic bands for all seeds in each batch at each stage is obtained as the second correlation coefficient; the ratio of the first correlation coefficient and the second correlation coefficient corresponding to each band is obtained as the water content characteristic possibility of each band;

[0018] The band with the largest water content feature possibility value among all bands is selected, and the corresponding band is used as the next water content feature band; the water content feature possibility of each band is continuously analyzed to obtain a preset first number of water content feature bands.

[0019] Furthermore, the method for obtaining the material characteristic band includes:

[0020] The band corresponding to the minimum first correlation coefficient is selected as the first material characteristic band;

[0021] The sum of the first correlation coefficient and the second correlation coefficient corresponding to each band is obtained as the insignificant material feature of each band;

[0022] The band with the smallest material characteristic insignificance value among all the bands is selected, and the corresponding band is used as the next material characteristic band; the material characteristic insignificance of each band is continuously analyzed to obtain a preset second number of material characteristic bands.

[0023] Furthermore, the method for obtaining the substance similarity includes:

[0024] Obtain the relative distance between the newly added seeds and each seed in each batch in the sequence composed of the overall intensity of the spectral response in the material characteristic band, and perform negative correlation mapping as the local similarity between the newly added seeds and each seed in each batch;

[0025] The local similarity accumulation value between the newly added seeds and different seeds in each stage batch is obtained as the material similarity of the newly added seeds relative to the seeds in each stage batch.

[0026] Furthermore, the method for obtaining the water content regression equation includes:

[0027] A linear regression equation was constructed with the overall intensity of the spectral response corresponding to the water content characteristic band of the seeds as the independent variable and the corresponding water content data of the seeds as the dependent variable;

[0028] The spectral response intensity and water content data corresponding to the water content characteristic band of different seeds in each batch are substituted into the linear regression equation, and the least squares method is used to obtain the parameters of the linear regression equation to form the water content regression equation.

[0029] Furthermore, the method for obtaining the water content prediction value includes:

[0030] The material similarity of the newly added seeds relative to the seeds in each batch is normalized as the water content weight of each batch;

[0031] According to the water content weights of batches at different stages, the water content regression equations of the corresponding batches are weighted summed to obtain a water content prediction model for the newly added seeds;

[0032] The overall intensity of the spectral response of the newly added seeds in each water content characteristic band is substituted into the water content prediction model, and the model results are used as the predicted value of the water content of the newly added seeds.

[0033] Furthermore, a reciprocal method is used to perform negative correlation mapping.

[0034] The present invention also proposes a system for determining the moisture content of Psoralea corylifolia seeds based on spectral technology, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any one of the steps of the method for determining the moisture content of Psoralea corylifolia seeds based on spectral technology.

[0035] The present invention has the following beneficial effects:

[0036] The present invention obtains the position cluster of each seed according to the relative distance of all spectral response intensities between different positions for seeds in any stage batch, and performs an overall analysis of the spectral characteristics within a single seed area; obtains the overall spectral response intensity of each seed in each band according to the spectral response intensity distribution of all positions in the position cluster of each seed in different bands, and more comprehensively reflects the overall intensity of each seed; obtains the water content characteristic band and material characteristic band of each stage batch according to the overall spectral response intensity and water content data of all seeds in different bands under each stage batch, identifies the band with greater water and material contribution characteristics, and is more helpful for evaluating the subsequent physiological state; and according to the newly added The similarity of the overall intensity of the spectral response of the material characteristic band between the different seeds under the seed and each stage batch, obtain the material similarity of the newly added seed relative to the seed under each stage batch, reflect the closest stage batch of the newly added seed, the more helpful as a reference analysis; According to the corresponding spectral response intensity and water content data of the different seeds in the water content characteristic band under each stage batch, construct a water content regression equation, evaluate the water content regression equation of each stage batch, and more accurately analyze the water content change situation; According to the material similarity of the newly added seed relative to the seed under different stage batches, the overall intensity of the spectral response of the water content characteristic band under the corresponding stage batch and the water content regression equation, obtain the water content predicted value of the newly added seed. The present invention considers the similarity of spectral data between seeds, selects the larger characteristic band of moisture content contribution to carry out regression analysis, and improves the accuracy of the determination of seed moisture content. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0038] Figure 1 A flow chart of a method for determining the moisture content of Psoralea corylifolia seeds based on spectroscopy technology provided by one embodiment of the present invention;

[0039] Figure 2A flow chart of a method for obtaining a characteristic band of water content provided by one embodiment of the present invention;

[0040] Figure 3 A flow chart of a method for obtaining a water content prediction value provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail a method and system for determining the moisture content of Psoralea corylifolia seeds based on spectral technology proposed by the present invention, as well as its specific implementation methods, structures, features and effects. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0042] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0043] The following describes in detail a method and system for determining the moisture content of Psoralea corylifolia seeds based on spectroscopy technology provided by the present invention in conjunction with the accompanying drawings.

[0044] See also Figure 1 , which shows a method flow chart of a method for determining the moisture content of Psoralea corylifolia seeds based on spectroscopy technology provided by one embodiment of the present invention, specifically comprising:

[0045] Step S1: Acquire hyperspectral data and water content data of seeds in different batches at different stages, wherein the hyperspectral data includes spectral response intensities at different positions in different bands.

[0046] In an embodiment of the present invention, in order to accurately measure the water content of seeds and avoid ignoring the differences in water distribution at different growth stages, the hyperspectral data and water content data of the seeds are processed.

[0047] First, seeds have different growth stages, such as the water absorption and swelling stage, the seed germination stage, and the germination stage. Hyperspectral imaging instruments were used to obtain hyperspectral data of seeds. The water content of batches of seeds at different stages can be measured by drying them in a 130°C constant-temperature drying oven. The seeds were then removed, cooled, and weighed. The drying method was repeated for another 30 minutes, followed by cooling and weighing, until the error between the two drying results was less than or equal to 0.002g. The average of these two weights was taken as the final dried seed weight. The difference between the pre-drying and post-drying seed weights was calculated and then divided by the pre-drying seed weight to obtain the water content data for each seed. Therefore, hyperspectral data and water content data of Taoerqi seeds were obtained for batches at different stages. The hyperspectral data includes the spectral response intensity at different locations in different bands.

[0048] It should be noted that, in one embodiment of the present invention, a batch of seeds is obtained for analysis every 7 days, and 200 seeds are obtained for analysis in each batch; in other embodiments of the present invention, the implementers can make specific settings according to the specific circumstances, which will not be limited or elaborated here.

[0049] Step S2: For any batch of seeds at any stage, obtain the position cluster of each seed based on the relative distances of all spectral response intensities between different positions; obtain the overall spectral response intensity of each seed in each band based on the spectral response intensity distribution of all positions in the position cluster of each seed corresponding to different bands.

[0050] As the band changes, the spectral response degrees in the non-seed area and the seed corresponding area change differently, and the spectral characteristics of the non-seed area, i.e. the background area, are consistent. By analyzing the relative distances of all spectral response intensities between different positions, positions with similar characteristics and close distances can be clustered together, which helps to analyze each seed in a targeted manner; for seeds in any batch at any stage, the position cluster of each seed is obtained based on the relative distances of all spectral response intensities between different positions.

[0051] Preferably, in one embodiment of the present invention, the method for obtaining the location cluster of each seed includes:

[0052] The spectral response intensities of each position and the corresponding positions are combined into a vector; the Euclidean distance between the vectors of different positions is obtained as the relative distance between the different positions;

[0053] Based on the relative distances between different locations, K-means clustering is performed on all locations to obtain the location clusters of each seed.

[0054] It should be noted that the K-means clustering algorithm clusters multiple clustering objects into specified K clusters based on the similarity between the clustering objects. Each clustering object belongs to and only belongs to one cluster with the smallest distance to the center of the position cluster. In one embodiment of the present invention, in order to divide each seed into a separate cluster for analysis, the number of clusters is set to the sum of the number of seeds and the number of non-seed areas, and the position of each seed is clustered into one cluster, and the positions of the non-seed areas are all clustered into one cluster. The specific K-means clustering algorithm is a numerical technical means of those skilled in the art and will not be elaborated here.

[0055] The spectral response intensity of a single seed at each wavelength at different positions may be different. In order to conduct a holistic analysis of the seeds and reflect the overall spectral characteristics of the seeds, the overall spectral response intensity of each seed in each band is obtained based on the distribution of spectral response intensities at different bands corresponding to all positions in the position cluster of each seed.

[0056] Preferably, in one embodiment of the present invention, the method for obtaining the overall intensity of the spectral response includes:

[0057] The spectral response intensity of all positions in the position cluster of each seed corresponding to each same band is averaged to obtain the overall spectral response intensity of each seed in each band.

[0058] Step S3: Based on the overall intensity of the spectral response and water content data of all seeds in each batch at different bands, obtain the water content characteristic band and material characteristic band of each batch; based on the similarity of the overall intensity of the spectral response in the material characteristic band between the newly added seeds and different seeds in each batch, obtain the material similarity of the newly added seeds relative to the seeds in each batch.

[0059] Due to the different growth states of seeds at different stages, there are many bands of substances unrelated to water content. By analyzing the overall intensity of the spectral response and water content data of different bands, we can select the characteristic bands with greater water content contribution, as well as the bands of material characteristics of seeds in each stage and batch, reflecting the correlation between the overall intensity of the spectral response of different bands and the change of water content data, as well as the germination of seeds; according to the overall intensity of the spectral response and water content data of all seeds in different bands in each stage and batch, the water content characteristic bands and material characteristic bands of each stage and batch are obtained.

[0060] Preferably, in one embodiment of the present invention, the method for obtaining the characteristic band of water content can be found in Figure 2 , which shows a flow chart of a method for obtaining a characteristic band of water content, including:

[0061] Step S201: Obtain the correlation coefficient between the sequence composed of the overall intensity of the spectral response of all seeds in each band in each stage batch and the sequence composed of the water content data as the first correlation coefficient; select the band corresponding to the maximum first correlation coefficient as the first water content characteristic band.

[0062] In order to subsequently determine the water content, the correlation coefficient between the sequence composed of the overall intensity of the spectral response and the sequence composed of the water content data is analyzed. The larger the correlation coefficient, the more likely it is that the characteristic band has a greater contribution to the water content.

[0063] As an example, wavelength The sequence composed of the overall intensity of the spectral response is , the sequence of water content data is ,in, Indicates the Seeds at wavelength The overall intensity of the spectral response under Indicates the The water content of the seeds.

[0064] It should be noted that, in the embodiment of the present invention, the correlation coefficient is the Pearson correlation coefficient. The larger the correlation coefficient, the more correlated the changes in the sequence composed of the overall intensity of the spectral response of each band and the sequence composed of the water content data, and the greater the influence of the overall intensity of the spectral response on the water content data. The specific Pearson correlation coefficient is a technical means well known to those skilled in the art and will not be elaborated here.

[0065] Step S202: Obtain the cumulative sum of the correlation coefficients of the sequence composed of the overall intensity of the spectral response between each band and all water content characteristic bands for all seeds in each stage batch as the second correlation coefficient; obtain the ratio of the first correlation coefficient and the second correlation coefficient corresponding to each band as the water content characteristic possibility of each band.

[0066] In order to avoid excessive correlation with the spectral response degree under the obtained water content characteristic band, and the correlation between the spectral response degree of the band and the water content cannot be too low, to ensure that it has a certain contribution to the change in water content; by analyzing the correlation coefficient of the sequence composed of the overall intensity of the spectral response between each band and all water content characteristic bands, the smaller the correlation coefficient, the more inconsistent the performance characteristics of the band and the water content characteristic band, and the larger the correlation coefficient between the spectral response degree of the band and the water content, the more it can express the characteristics of the water content, the greater the possibility of the water content characteristics of the corresponding band, and the more helpful it is for the determination and analysis of the water content.

[0067] Step S203: Select the band with the largest water content feature possibility value among all bands, and use the corresponding band as the next water content feature band; continuously analyze the water content feature possibility of each band to obtain a preset first number of water content feature bands.

[0068] In one embodiment of the present invention, the water content characteristic probability is expressed as:

[0069] ;

[0070] in, Indicates band Possibility of water content characteristics; Indicates band The correlation coefficient between the sequence composed of the overall intensity of the spectral response and the sequence composed of the water content data, that is, the first correlation coefficient; Indicates band The sequence composed of the overall intensity of the spectral response; Indicates the Water content characteristic band The sequence composed of the overall degree of spectral response; Indicates band Hedi Water content characteristic band The correlation coefficient of the sequence composed of the overall intensity of the spectral response between them; Indicates the number of characteristic bands of water content.

[0071] In the formula for the water content characteristic probability, Add 0.01 to avoid the denominator of the formula being 0, which would make the formula meaningless; Indicates band and characteristic bands of different water contents The sum of the correlation coefficients of the sequence composed of the overall intensity of the spectral response between the two bands is the second correlation coefficient. The smaller the sum, the stronger the band. and characteristic bands of different water contents The smaller the correlation between the spectral characteristics of the band The greater the correlation coefficient between the sequence composed of the overall intensity of the spectral response and the sequence composed of the water content data, the greater the impact on the water content. The more new bands contribute to the change in water content, the greater the possibility of water content characteristics.

[0072] It should be noted that since there are many bands in the seeds of Peach Seven that are not related to the water content, in order to more accurately reflect the characteristics of the water content, in one embodiment of the present invention, the preset first number is set to 5, that is, after selecting each water content characteristic band in turn, the water content characteristics of each band are analyzed and the next one is selected until the number of selected water content characteristic bands is the preset first number; in other embodiments of the present invention, the size of the preset first number can be set according to the specific situation, and is not limited or elaborated here.

[0073] Preferably, in one embodiment of the present invention, the method for acquiring the material characteristic band includes:

[0074] The band corresponding to the minimum first correlation coefficient is selected as the first material characteristic band;

[0075] The sum of the first correlation coefficient and the second correlation coefficient corresponding to each band is obtained as the insignificant material feature of each band;

[0076] The band with the smallest material characteristic insignificance value among all the bands is selected, and the corresponding band is used as the next material characteristic band; the material characteristic insignificance of each band is continuously analyzed to obtain a preset second number of material characteristic bands.

[0077] In one embodiment of the present invention, the formula for the insignificant substance characteristic is expressed as:

[0078] ;

[0079] in, Indicates wavelength The material characteristics of Indicates band The correlation coefficient between the sequence composed of the overall intensity of the spectral response and the sequence composed of the water content data, that is, the first correlation coefficient; Indicates band The sequence composed of the overall intensity of the spectral response; Indicates the Material characteristic bands The sequence composed of the overall degree of spectral response; Indicates band Hedi Material characteristic bands The correlation coefficient of the sequence composed of the overall intensity of the spectral response between them; Indicates the number of material characteristic bands.

[0080] In formulas where the material characteristics are not explicit, Indicates band Hedi Material characteristic bands The sum of the correlation coefficients of the sequence composed of the overall intensity of the spectral response between the two bands is the second correlation coefficient. The smaller the sum, the stronger the band. Hedi Material characteristic bands The smaller the correlation between the spectral characteristics, the more likely it is to show the characteristics of different substances; The smaller the correlation coefficient between the sequence composed of the overall intensity of the spectral response and the sequence composed of the water content data, the smaller the correlation contribution to the change in water content, the more likely it is to be the manifestation of other material components, and the less obvious the material characteristics are.

[0081] It should be noted that in order to analyze the seed germination characteristics of each batch in each stage, attention is paid to the changes in the remaining material components in the seeds. The more material components are expressed, the more helpful it is to analyze the seed germination characteristics. In one embodiment of the present invention, the preset second number is 9; in other embodiments of the present invention, the preset second number can be set according to the specific circumstances, and no limitation or elaboration is made here.

[0082] The material characteristic bands reflect the germination status of seeds within each batch. By analyzing the similarity in the overall intensity of the spectral response of the material characteristic bands between the newly added seeds and different seeds within each batch, the material similarity of the newly added seeds relative to the seeds within each batch is assessed. This facilitates analysis based on the status of seeds within the corresponding batches. Based on the similarity in the overall intensity of the spectral response of the material characteristic bands between the newly added seeds and different seeds within each batch, the material similarity of the newly added seeds relative to the seeds within each batch is obtained.

[0083] Preferably, in one embodiment of the present invention, the method for obtaining material similarity includes:

[0084] Obtain the relative distance between the newly added seeds and each seed in each batch in the sequence composed of the overall intensity of the spectral response in the material characteristic band, and perform negative correlation mapping as the local similarity between the newly added seeds and each seed in each batch;

[0085] The local similarity accumulation value between the newly added seeds and different seeds in each stage batch is obtained as the material similarity of the newly added seeds relative to the seeds in each stage batch.

[0086] It should be noted that the overall intensity of the spectral response of the newly added seeds in each band is analyzed according to the same method and steps for obtaining the overall intensity of the spectral response of each seed in each band in step S2; in an embodiment of the present invention, the relative distance can be calculated by an existing distance algorithm such as Euclidean distance or Manhattan distance. The specific means are technical means well known to those skilled in the art and will not be elaborated here.

[0087] It should be noted that, in one embodiment of the present invention, a negative correlation mapping is performed on the relative distance of the spectral response intensity vector by taking the reciprocal. The larger the relative distance, the smaller the material similarity, and the less likely the seeds are in the same germination state. When taking the reciprocal, a threshold is artificially added to avoid the denominator of the formula being 0, which makes the formula meaningless. In other embodiments of the present invention, the negative correlation mapping can also be performed by taking the reciprocal. The specific means for performing negative correlation mapping are well known to those skilled in the art and will not be elaborated here.

[0088] Step S4: Construct a water content regression equation based on the spectral response intensity and water content data of different seeds in each stage batch in the water content characteristic band; obtain the water content prediction value of the new seeds based on the material similarity of the new seeds relative to the seeds in different stage batches, the overall intensity of the spectral response of the water content characteristic band in the corresponding stage batch, and the water content regression equation.

[0089] Seeds from batches at different stages have different moisture spectral relationships. By establishing a moisture content regression equation corresponding to the characteristic band of moisture content, it is more conducive to evaluating the oil content of seeds in each batch and reflecting the germination status.

[0090] A water content regression equation was constructed based on the spectral response intensity and water content data corresponding to the water content characteristic band of different seeds in each batch at each stage.

[0091] Preferably, in one embodiment of the present invention, the method for obtaining the water content regression equation includes:

[0092] A linear regression equation was constructed with the overall intensity of the spectral response corresponding to the water content characteristic band of the seeds as the independent variable and the water content data of the corresponding seeds as the dependent variable; the linear regression equation is: ,in, Indicates water content data; represents the intercept term; Characteristic band representing water content Variable parameters of Indicates the first water content characteristic band The overall intensity of the spectral response; Characteristic band representing water content Variable parameters of Indicates the Water content characteristic band The overall intensity of the spectral response; represents the random component;

[0093] The spectral response intensity and water content data corresponding to the water content characteristic band of different seeds in each batch are substituted into the linear regression equation, and the least squares method is used to obtain the parameters of the linear regression equation to form the water content regression equation.

[0094] It should be noted that the specific least squares method for solving parameters is a technical means well known to those skilled in the art and will not be described in detail here.

[0095] Material similarity reflects the similarity of growth status between the newly added seeds and batches at different stages, which helps to match the most similar stage batches of the newly added seeds. The larger the proportion of the corresponding regression equation, the more comprehensive and accurate the analysis of the water content characteristics of the newly added seeds; based on the material similarity of the newly added seeds relative to seeds at different stage batches, the overall intensity of the spectral response of the water content characteristic bands at the corresponding stage batches, and the water content regression equation, the predicted value of the water content of the newly added seeds is obtained.

[0096] Preferably, in one embodiment of the present invention, the method for obtaining the water content prediction value can be found in Figure 3 , which shows a flow chart of a method for obtaining a water content prediction value, including:

[0097] Step S301: normalize the material similarity of the newly added seeds relative to the seeds in each stage batch as the water content weight of each stage batch.

[0098] Material similarity reflects the degree of similarity between the characteristics of the newly added seeds and the seeds of which stage and batch. The greater the similarity, the more helpful it is for analyzing the water content of the newly added seeds, and the greater the weight of the water content.

[0099] Step S302: performing weighted summation on the water content regression equations of the batches at the corresponding stages according to the water content weights of the batches at different stages, to obtain a water content prediction model for the newly added seeds.

[0100] In one embodiment of the present invention, the water content prediction model is expressed as follows:

[0101] ;

[0102] in, represents the predicted water content of newly added seeds; Indicates the Material similarity between phase batches; Indicates the Regression equation for water content of phase batches; represents the normalization function; Indicates the number of batches in the stage.

[0103] In the formula of the water content prediction model, Indicates the The material similarity of the stage batches is normalized, that is, the water content weight. The greater the material similarity, the more consistent the germination of the newly added seeds is with the seeds in the stage batches, and the more conducive it is to the prediction of water content.

[0104] Step S303: Substitute the overall intensity of the spectral response of the newly added seed in each water content characteristic band into the water content prediction model, and use the model result as the water content prediction value of the newly added seed.

[0105] Based on this, by assigning greater weights to stage batches with greater material similarity with the newly added seeds, we can avoid prediction bias caused by differences in germination stages, which will help to accurately analyze the water content of the newly added seeds and obtain moisture information more quickly and non-destructively.

[0106] In summary, the present invention obtains the water content characteristic band and material characteristic band under each stage batch according to the overall intensity of the spectral response of all seeds in different bands and the water content data under each stage batch; According to the similarity of the overall intensity of the spectral response in the material characteristic band between the newly added seeds and the different seeds under each stage batch, the material similarity of the newly added seeds relative to the seeds under each stage batch is obtained; According to the corresponding spectral response intensity and water content data of the different seeds in the water content characteristic band under each stage batch, a water content regression equation is constructed; And then the water content prediction value of the newly added seeds is obtained. The present invention takes into account the similarity of the spectral data between seeds, selects the characteristic band with larger moisture content contribution for regression analysis, and improves the accuracy of seed moisture content determination.

[0107] The present invention also proposes a system for determining the moisture content of Psoralea corylifolia seeds based on spectral technology, which includes a memory, a processor, and a computer program stored in the memory and runnable on the processor. When the processor executes the computer program, it implements any step of a method for determining the moisture content of Psoralea corylifolia seeds based on spectral technology.

[0108] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0109] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A method for determining the moisture content of Psoralea corylifolia seeds based on spectroscopy, characterized in that: The method comprises: Obtaining hyperspectral data and water content data of Psoralea corylifolia seeds at different stages and batches, wherein the hyperspectral data includes spectral response intensities at different positions in different bands; For any batch of seeds at any stage, the position cluster of each seed is obtained based on the relative distances of all spectral response intensities between different positions; the overall spectral response intensity of each seed in each band is obtained based on the distribution of spectral response intensities of all positions in each seed position cluster corresponding to different bands; Based on the overall intensity of the spectral response and water content data of all seeds in each batch at different bands, the water content characteristic band and material characteristic band of each batch are obtained; based on the similarity of the overall intensity of the spectral response of the newly added seeds and different seeds in each batch at different bands in material characteristic bands, the material similarity of the newly added seeds relative to the seeds in each batch is obtained; A water content regression equation was constructed based on the spectral response intensity and water content data of different seeds in each batch at each stage. The water content prediction value of the newly added seeds was obtained based on the material similarity of the newly added seeds relative to the seeds in different batches at different stages, the overall intensity of the spectral response of the water content characteristic band in the corresponding batches at different stages, and the water content regression equation. The method for obtaining the water content prediction value includes: The material similarity of the newly added seeds relative to the seeds in each batch is normalized as the water content weight of each batch; According to the water content weights of batches at different stages, the water content regression equations of the corresponding batches are weighted summed to obtain a water content prediction model for the newly added seeds; The overall intensity of the spectral response of the newly added seeds in each water content characteristic band is substituted into the water content prediction model, and the model results are used as the predicted value of the water content of the newly added seeds.

2. A method for determining the moisture content of Psoralea corylifolia seeds based on spectroscopy according to claim 1, characterized in that: The method for obtaining the position cluster of each seed includes: The spectral response intensities of each position and the corresponding positions are combined into a vector; the Euclidean distance between the vectors of different positions is obtained as the relative distance between the different positions; Based on the relative distances between different locations, K-means clustering is performed on all locations to obtain the location clusters of each seed.

3. A method for determining the moisture content of Psoralea corylifolia seeds based on spectroscopy according to claim 1, wherein The method for obtaining the overall intensity of the spectral response includes: The spectral response intensity of all positions in the position cluster of each seed corresponding to each same band is averaged to obtain the overall spectral response intensity of each seed in each band.

4. A method for determining the moisture content of Psoralea corylifolia seeds based on spectroscopy according to claim 1, characterized in that, The method for obtaining the water content characteristic band includes: Obtain the correlation coefficient between the sequence composed of the overall intensity of the spectral response of all seeds in each band in each batch and the sequence composed of the water content data as the first correlation coefficient; select the band corresponding to the maximum first correlation coefficient as the first water content characteristic band; The cumulative sum of the correlation coefficients of the sequence composed of the overall intensity of the spectral response between each band and all water content characteristic bands for all seeds in each batch at each stage is obtained as the second correlation coefficient; the ratio of the first correlation coefficient and the second correlation coefficient corresponding to each band is obtained as the water content characteristic possibility of each band; The band with the largest water content feature possibility value among all bands is selected, and the corresponding band is used as the next water content feature band; the water content feature possibility of each band is continuously analyzed to obtain a preset first number of water content feature bands.

5. A method for determining the moisture content of Psoralea corylifolia seeds based on spectroscopy according to claim 4, characterized in that, The method for obtaining the material characteristic band includes: The band corresponding to the minimum first correlation coefficient is selected as the first material characteristic band; The sum of the first correlation coefficient and the second correlation coefficient corresponding to each band is obtained as the insignificant material feature of each band; The band with the smallest material characteristic insignificance value among all the bands is selected, and the corresponding band is used as the next material characteristic band; the material characteristic insignificance of each band is continuously analyzed to obtain a preset second number of material characteristic bands.

6. A method for determining the moisture content of Psoralea corylifolia seeds based on spectroscopy according to claim 1, characterized in that, The method for obtaining the substance similarity includes: Obtain the relative distance between the newly added seeds and each seed in each batch in the sequence composed of the overall intensity of the spectral response in the material characteristic band, and perform negative correlation mapping as the local similarity between the newly added seeds and each seed in each batch; The local similarity accumulation value between the newly added seeds and different seeds in each stage batch is obtained as the material similarity of the newly added seeds relative to the seeds in each stage batch.

7. A method for determining the moisture content of Psoralea corylifolia seeds based on spectroscopy according to claim 1, characterized in that, The method for obtaining the water content regression equation includes: A linear regression equation was constructed with the overall intensity of the spectral response corresponding to the water content characteristic band of the seeds as the independent variable and the corresponding water content data of the seeds as the dependent variable; The spectral response intensity and water content data corresponding to the water content characteristic band of different seeds in each batch are substituted into the linear regression equation, and the least squares method is used to obtain the parameters of the linear regression equation to form the water content regression equation.

8. A method for determining the moisture content of Psoralea corylifolia seeds based on spectroscopy according to claim 6, characterized in that: The negative correlation mapping is performed using the reciprocal method.

9. A system for measuring the moisture content of Psoralea corylifolia seeds based on spectroscopy, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of a method for determining the moisture content of Psoralea corylifolia seeds based on spectral technology as described in any one of claims 1 to 8 are implemented.

Images