Intelligent determination method, system and medium for winter wheat development period based on multi-source parameters

Abstract

The present application discloses a method, system and medium for intelligently determining the development period of winter wheat based on multi-source parameters, which mainly relates to the technical field of determining the development period of winter wheat, and is used to solve the problem that the existing green crop image cutting method may produce misjudgment, thereby causing insufficient background shadow separation and overfitting in the scheme of data training classification network. It includes: determining the pixel points to be processed whose color gamut value in HSV communication is within the preset adjustment range and / or whose saturation is less than the preset threshold; adjusting the image transparency of the pixel points to be processed to 0 to obtain the adjusted image; converting the adjusted image into a grayscale image, obtaining the position of the black area from the grayscale image, determining the position of the black area, removing it from the initial collected image, and obtaining the final image; obtaining the specific numerical value corresponding to the final training parameter in the final image and obtaining the annotated development period corresponding to the final image, and then using them as training data for the preset classification algorithm to obtain a trained preset classification algorithm.

Description

Technical Field

[0001] The present application relates to the technical field of winter wheat development period determination, and in particular to a method, system and medium for intelligent determination of winter wheat development period based on multi-source parameters. Background Art

[0002] The development period of winter wheat refers to the various stages from sowing to maturity. These stages mainly include three leaves, tillering, greening, jointing, heading, heading, flowering, milky maturity and maturity. By accurately judging the development stage of winter wheat, agricultural experts and farmers can more accurately predict its growth and final yield. This is crucial for formulating reasonable planting plans, adjusting management measures and optimizing resource allocation.

[0003] The existing intelligent judgment process of winter wheat development period based on digital images requires image segmentation. The traditional green crop image segmentation method is the super green (GR) segmentation algorithm, which is based on the channel difference method, that is, the digital image is subtracted using a pre-set channel difference combination (such as RB, RG, R+BG, etc.), and the combination with the largest image contrast is obtained through multiple experiments, and background separation is performed based on this. However, due to the involvement of exogenous debris in the image of winter wheat growing in the natural environment of the field, including pure green weeds and non-green debris, and the change of canopy color from green to yellow after winter wheat enters heading, the super green (GR) segmentation algorithm will cause misjudgment, which in turn leads to insufficient separation of background shadows.

[0004] Therefore, there is an urgent need for an intelligent determination method, system and medium for the development period of winter wheat based on multi-source parameters to solve the problem that the existing green crop image cutting method will produce misjudgment, which in turn leads to insufficient separation of background shadows. In addition, the existing intelligent determination scheme for the development period of winter wheat mainly imports all parameter data into the classification network and then trains it into a classification model. However, the above scheme of training the classification model with all parameter data has the problem of overfitting, and the generalization ability is relatively weak, and it cannot be widely applied to the intelligent determination of the development period of winter wheat in different ecological zones. Summary of the invention

[0005] In view of the shortcomings of the above-mentioned prior art, the present application provides a method, system and medium for intelligent determination of the development period of winter wheat based on multi-source parameters, so as to solve the problems that the existing green crop image cutting method may produce misjudgment, thereby resulting in insufficient background shadow separation, and the existing parameter data training classification model has overfitting and weak generalization ability.

[0006] In a first aspect, the present application provides a method for intelligently determining the development period of winter wheat based on multi-source parameters, the method comprising: obtaining an original digital image corresponding to winter wheat in a preset winter wheat development period in a preset area; wherein the format of the original digital image is RGB format; after removing images with light interference, foreign matter interference and lens attachment interference from the original digital image, an initial acquired image is obtained; the initial acquired image in RGB format is converted into an HSV image, and pixels to be processed whose color gamut values ​​in HSV communication are within a preset adjustment range and / or whose saturation is less than a preset threshold are determined; the image transparency of the pixels to be processed is adjusted to 0 to obtain an adjusted image; the adjusted image is converted into a grayscale image, and the position of a black area is obtained from the grayscale image; the initial acquired image corresponding to the grayscale image is obtained, and the position of the black area is in the regional position corresponding to the initial acquired image; the regional position is removed from the initial acquired image, and the obtained The retained part of the initially collected image is used as the final image and saved in a preset format; a number of preset RGB channel parameters are extracted from the final image, and a number of preset meteorological parameters are extracted from a meteorological database; the number of preset RGB channel parameters and the number of preset meteorological parameters are used as a group of data to obtain each preset meteorological parameter and the Kaiser-Meyer-Olkin value and Bartlett variance value corresponding to the preset meteorological parameter; the accuracy and weight corresponding to each preset meteorological parameter and the preset meteorological parameter are obtained; based on the Kaiser-Meyer-Olkin value, Bartlett variance value, accuracy and weight, two preset RGB channel parameters and two preset meteorological parameters are extracted as final training parameters; the specific values ​​corresponding to the final training parameters in the final image and the annotated development period corresponding to the final image are obtained, which are then used as training data for a preset classification network to obtain a trained preset classification model.

[0007] Furthermore, the preset winter wheat development stages include the three-leaf stage, tillering stage, greening stage, jointing stage, booting stage, heading stage, flowering stage, milky stage and maturity stage; the preset RGB channel parameters include at least the R channel skewness and the G channel kurtosis; the preset meteorological parameters include at least: the cumulative value of the daily average temperature and the cumulative amount of photosynthetically active radiation; the preset format is jpg format.

[0008] Furthermore, the original digital image corresponding to the winter wheat in the preset winter wheat development period in the preset area is obtained, specifically including: in the preset collection time period, the original digital image is collected at a preset collection frequency by a monitoring camera device located in the winter wheat field; wherein, one first preset focal length picture and one second preset focal length picture are collected at each frequency, and the installation height is a preset height.

[0009] Furthermore, images with light interference, foreign matter interference and lens attachment interference are eliminated from the original digital images to obtain the initial collected image, specifically including: determining the original digital images with foreign matter interference or lens attachment interference through an edge detection algorithm, and then removing the original digital images with foreign matter interference or lens attachment interference; determining the original digital images with local highlight areas through an image brightness analysis algorithm, and then eliminating the original digital images with local highlight areas; wherein the local highlight area is an area with a local brightness greater than a preset brightness value.

[0010] Furthermore, removing the region position from the initial acquired image specifically includes: using an image segmentation algorithm to remove the region position from the initial acquired image.

[0011] Further, based on the Kaiser-Meyer-Olkin value, the Bartlett variance value, the accuracy and the weight, two preset RGB channel parameters and two preset meteorological parameters are extracted as final training parameters, specifically including: selecting, from the preset RGB channel parameters and the preset meteorological parameters, the preset RGB channel parameters and the preset meteorological parameters whose Kaiser-Meyer-Olkin value is greater than the first preset value and whose Bartlett variance value is greater than the second preset value; by the formula: , calculate the impact factor H corresponding to each preset RGB channel parameter and preset meteorological parameter; where X represents the accuracy, Represents weight; from all preset RGB channel parameters, select the first two with the largest impact factors as the final training parameters; from all preset meteorological parameters, select the first two with the largest impact factors as the final training parameters.

[0012] In a second aspect, the present application provides an intelligent determination system for the development period of winter wheat based on multi-source parameters, the system comprising: an acquisition module, used to acquire the original digital image corresponding to the winter wheat in the preset development period of winter wheat in a preset area; wherein the format of the original digital image is RGB format; the acquisition module, used to remove images with light interference, foreign matter interference and lens attachment interference from the original digital image, to obtain an initial acquired image; an area acquisition module, used to convert the initial acquired image in RGB format into an HSV image, determine the pixel points to be processed whose color gamut value in the HSV communication is within a preset adjustment range and / or whose saturation is less than a preset threshold; adjust the image transparency of the pixel points to be processed to 0, to obtain an adjusted image; convert the adjusted image into a grayscale image, and obtain the position of the black area from the grayscale image; an image acquisition module, used to acquire the initial acquired image corresponding to the grayscale image, determine the position of the black area in the area corresponding to the initial acquired image; convert the area position from the initial acquired image The module is used to remove the preset RGB channel parameters from the final image, obtain the retained part of the initial collected image as the final image, and save it in a preset format; a parameter determination module is used to extract a number of preset RGB channel parameters from the final image, and extract a number of preset meteorological parameters from the meteorological database; the number of preset RGB channel parameters and the number of preset meteorological parameters are used as a group of data to obtain each preset meteorological parameter and the Kaiser-Meyer-Olkin value and Bartlett variance value corresponding to the preset meteorological parameter; the number of preset meteorological parameters and the accuracy and weight corresponding to the preset meteorological parameter are obtained; based on the Kaiser-Meyer-Olkin value, Bartlett variance value, accuracy and weight, two preset RGB channel parameters and two preset meteorological parameters are extracted as final training parameters; a training module is used to obtain the specific values ​​corresponding to the final training parameters in the final image and the annotated development period corresponding to the final image, and then use them as training data for the preset classification network to obtain a trained preset classification model.

[0013] Furthermore, the acquisition module includes a collection unit, which is used to collect original digital images at a preset collection frequency through a monitoring camera device located in the winter wheat field during a preset collection time period; wherein, one first preset focal length image and one second preset focal length image are collected each frequency, and the installation height is a preset height.

[0014] Furthermore, the image acquisition module includes a segmentation unit for removing the region position from the original digital image using an image segmentation algorithm.

[0015] In a third aspect, the present application provides a non-volatile computer storage medium having computer instructions stored thereon, which when executed implement an intelligent method for determining the development period of winter wheat based on multi-source parameters as described in any one of the above items.

[0016] Those skilled in the art can understand that the present application has at least the following beneficial effects:

[0017] The present application obtains the position of the black area from the grayscale image; obtains the original digital image corresponding to the grayscale image, determines the position of the black area in the area corresponding to the original digital image; removes the area position from the original digital image, and obtains the retained part of the original digital image as the final image. The present application first separates foreign matter using the preset adjustment range corresponding to the color gamut value, and then separates the winter wheat in the late growth period from the soil using the preset threshold corresponding to the saturation. It does not involve the green extraction (super green cutting algorithm) in the existing green crop image cutting process, and avoids the canopy color change from green to yellow after the winter wheat enters the heading stage, which leads to misjudgment when the super green cutting algorithm is used, and then causes the problem of insufficient background shadow separation.

[0018] In addition, the present application extracts two preset RGB channel parameters and two preset meteorological parameters from several preset RGB channel parameters and several preset meteorological parameters as final training parameters through Kaiser-Meyer-Olkin value and Bartlett variance value to train the network, thereby improving the generalization ability of the preset classification algorithm and avoiding the problem of overfitting. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Some embodiments of the present disclosure are described below with reference to the accompanying drawings, in which:

[0020] Figure 1 This is a flow chart of a method for intelligently determining the development period of winter wheat based on multi-source parameters provided in an embodiment of the present application.

[0021] Figure 2 It is a schematic diagram of the internal structure of a winter wheat development period intelligent determination system based on multi-source parameters provided in an embodiment of the present application. DETAILED DESCRIPTION

[0022] It should be understood by those skilled in the art that the embodiments described below are only preferred embodiments of the present disclosure, and do not mean that the present disclosure can only be implemented through the preferred embodiments. The preferred embodiments are only used to explain the technical principles of the present disclosure, and are not used to limit the protection scope of the present disclosure. Based on the preferred embodiments provided by the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work should still fall within the protection scope of the present disclosure.

[0023] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0024] The technical solution proposed in the embodiments of the present application is described in detail below with reference to the accompanying drawings.

[0025] The present application embodiment provides a method for intelligently determining the winter wheat development period based on multi-source parameters, such as Figure 1 As shown, the method provided in the embodiment of the present application mainly includes the following steps:

[0026] Step 110: Acquire an original digital image corresponding to winter wheat in a preset winter wheat growth period in a preset area.

[0027] It should be noted that the format of the original digital image is RGB format. The preset winter wheat development period includes three-leaf stage, tillering stage, greening stage, jointing stage, booting stage, heading stage, flowering stage, milk stage and maturity stage. Several preset RGB channel parameters include at least R channel skewness and G channel kurtosis, and several preset meteorological parameters include at least: the cumulative value of daily average temperature and the cumulative amount of photosynthetically active radiation.

[0028] This step may be specifically as follows: during a preset collection time period, original digital images are collected at a preset collection frequency by a monitoring camera device located in the winter wheat field; wherein one first preset focal length image and one second preset focal length image are collected at each frequency, and the installation height is a preset height.

[0029] It should be noted that the preset collection time period may be 8:00-16:00; the preset collection frequency may be once per hour; and the preset height may be 280 cm. The first preset focal length may be the focal length corresponding to the close-up image, and the specific value may be determined by those skilled in the art according to actual conditions. The second preset focal length may be the focal length corresponding to the long-range image, and the specific value may be determined by those skilled in the art according to actual conditions.

[0030] Step 120: Eliminate images with light interference, foreign matter interference, and lens attachment interference from the original digital image to obtain an initial captured image.

[0031] This step can be specifically as follows:

[0032] Through the edge detection algorithm, the original digital image with interference from foreign objects or lens attachments is determined, and then the original digital image with interference from foreign objects or lens attachments is removed; through the image brightness analysis algorithm, the original digital image with local highlight areas is determined, and then the original digital image with local highlight areas is removed; wherein the local highlight area is an area where the local brightness is greater than a preset brightness value.

[0033] Step 130: convert the initial collected image in RGB format into an HSV image, determine the pixels to be processed whose color gamut values ​​in the HSV communication are within a preset adjustment range and / or whose saturation is less than a preset threshold; adjust the image transparency of the pixels to be processed to 0 to obtain an adjusted image; convert the adjusted image into a grayscale image, and obtain the position of the black area from the grayscale image.

[0034] It should be noted that the specific values ​​of the preset adjustment range and preset threshold of the HSV image (Hue, Saturation, Value, color gamut value, saturation, image transparency) can be determined by those skilled in the art according to actual conditions. For example, the preset adjustment range can be [0.1, 0.5], and the preset threshold can be 0.3.

[0035] Step 140, obtaining an initial acquisition image corresponding to the grayscale image, determining the black area position at the area position corresponding to the initial acquisition image; removing the area position from the initial acquisition image, obtaining the retained portion of the initial acquisition image as the final image, and saving it in a preset format.

[0036] It should be noted that the default format is jpg.

[0037] The region position is removed from the initial collected image, which may specifically be: using an image segmentation algorithm to remove the region position from the initial collected image.

[0038] Step 150, extracting a number of preset RGB channel parameters from the final image, and extracting a number of preset meteorological parameters from the meteorological database; taking the number of preset RGB channel parameters and the number of preset meteorological parameters as a group of data, obtaining each preset meteorological parameter, and the Kaiser-Meyer-Olkin value and Bartlett variance value corresponding to the preset meteorological parameter; obtaining each preset meteorological parameter, and the accuracy and weight corresponding to the preset meteorological parameter; based on the Kaiser-Meyer-Olkin value, Bartlett variance value, accuracy and weight, extracting two preset RGB channel parameters and two preset meteorological parameters as final training parameters.

[0039] It should be noted that the data in the meteorological database can be obtained through the climate center corresponding to the preset area. The specific method for obtaining each preset meteorological parameter, the accuracy and weight corresponding to the preset meteorological parameter can be obtained through the preset interface.

[0040] Among them, based on the Kaiser-Meyer-Olkin value, Bartlett variance value, accuracy and weight, two preset RGB channel parameters and two preset meteorological parameters are extracted as the final training parameters, which can be specifically:

[0041] From the preset RGB channel parameters and the preset meteorological parameters, the preset RGB channel parameters and the preset meteorological parameters are selected, whose Kaiser-Meyer-Olkin value is greater than the first preset value and whose Bartlett variance value is greater than the second preset value.

[0042] By formula: , calculate the impact factor H corresponding to each preset RGB channel parameter and preset meteorological parameter; where X represents the accuracy, Represents weight.

[0043] From all preset RGB channel parameters, the first two with the largest impact factors are selected as the final training parameters; from all preset meteorological parameters, the first two with the largest impact factors are selected as the final training parameters.

[0044] It should be noted that the first preset value can be any value in the range of (0, 1], and the second preset value can be any value in the range of (0, 1]. Those skilled in the art can determine the specific values ​​according to actual needs.

[0045] Step 160 , obtaining specific values ​​corresponding to the final training parameters in the final image and obtaining the annotated developmental period corresponding to the final image, which are then used as training data for the preset classification algorithm to obtain a trained preset classification algorithm.

[0046] As an example, when the final training parameters are R channel skewness, G channel kurtosis, cumulative value of daily average temperature, cumulative amount of photosynthetically active radiation, and the preset classification algorithm is the Bayesian algorithm: extract the R channel skewness and G channel kurtosis from the final image, extract the cumulative value of daily average temperature and the cumulative amount of photosynthetically active radiation corresponding to the final image from the meteorological database, and obtain the annotated developmental period corresponding to the final image; extract the R channel skewness, G channel kurtosis, cumulative value of daily average temperature, cumulative amount of photosynthetically active radiation and annotated developmental period as training data for the Bayesian algorithm to obtain a trained Bayesian algorithm.

[0047] besides, Figure 2 The present invention provides an intelligent winter wheat development period determination system based on multi-source parameters. Figure 2As shown, the system provided in the embodiment of the present application mainly includes:

[0048] The acquisition module 210 is used to acquire the original digital image corresponding to the winter wheat in the preset winter wheat growth period in the preset area.

[0049] It should be noted that the format of the original digital image is RGB format.

[0050] The acquisition module 210 includes a collection unit, which is used to collect original digital images at a preset collection frequency through a monitoring camera device located in the winter wheat field during a preset collection time period; wherein one first preset focal length image and one second preset focal length image are collected each frequency, and the installation height is a preset height.

[0051] The acquisition module 220 is used to remove images with light interference, foreign matter interference and lens attachment interference from the original digital image to obtain the initial collected image.

[0052] The acquisition module 220 includes a rejection unit, which is used to determine the original digital image with interference from lens attachments through an edge detection algorithm, and then remove the original digital image with interference from lens attachments; determine the original digital image with local highlight areas through an image brightness analysis algorithm, and then reject the original digital image with local highlight areas; wherein the local highlight area is an area with local brightness greater than a preset brightness value.

[0053] The area acquisition module 230 is used to convert the initial collected image in RGB format into an HSV image, determine the pixel points to be processed whose color gamut values ​​in the HSV communication are within a preset adjustment range and / or whose saturation is less than a preset threshold; adjust the image transparency of the pixel points to be processed to 0 to obtain an adjusted image; convert the adjusted image into a grayscale image, and obtain the position of the black area from the grayscale image.

[0054] The image acquisition module 240 is used to acquire the initial acquisition image corresponding to the grayscale image, determine the black area position at the area position corresponding to the initial acquisition image; remove the area position from the initial acquisition image, obtain the retained part of the initial acquisition image as the final image, and save it in a preset format.

[0055] The image acquisition module 240 includes a segmentation unit for removing the region position from the initial acquired image using an image segmentation algorithm.

[0056] The parameter determination module 250 is used to extract a number of preset RGB channel parameters from the final image and a number of preset meteorological parameters from the meteorological database; take the number of preset RGB channel parameters and the number of preset meteorological parameters as a group of data to obtain each preset meteorological parameter and the Kaiser-Meyer-Olkin value and Bartlett variance value corresponding to the preset meteorological parameter; obtain each preset meteorological parameter and the accuracy and weight corresponding to the preset meteorological parameter; based on the Kaiser-Meyer-Olkin value, Bartlett variance value, accuracy and weight, extract two preset RGB channel parameters and two preset meteorological parameters as final training parameters.

[0057] The training module 260 is used to obtain the specific values ​​corresponding to the final training parameters in the final image and the annotated developmental stages corresponding to the final image, which are then used as training data for the preset classification algorithm to obtain a trained preset classification algorithm.

[0058] In addition, an embodiment of the present application further provides a non-volatile computer storage medium on which executable instructions are stored. When the executable instructions are executed, an intelligent method for determining the development period of winter wheat based on multi-source parameters as described above is implemented.

[0059] So far, the technical solutions of the present disclosure have been described in combination with the above multiple embodiments, but it is easy for those skilled in the art to understand that the protection scope of the present disclosure is not limited to these specific embodiments. Without departing from the technical principles of the present disclosure, those skilled in the art can split and combine the technical solutions in the above-mentioned various embodiments, and can also make equivalent changes or replacements to the relevant technical features. Any changes, equivalent replacements, improvements, etc. made within the technical concept and / or technical principle of the present disclosure will fall within the protection scope of the present disclosure.

Claims

1. A method for intelligently determining the development period of winter wheat based on multi-source parameters, characterized in that: The method comprises: Acquire an original digital image corresponding to winter wheat in a preset winter wheat development period in a preset area; wherein the format of the original digital image is RGB format; Eliminate images with light interference, foreign matter interference and lens attachment interference from the original digital images to obtain the initial acquisition image; Convert the initially acquired image in RGB format into an HSV image, determine the pixels to be processed whose color gamut values ​​in the HSV communication are within a preset adjustment range and / or whose saturation is less than a preset threshold; adjust the image transparency of the pixels to be processed to 0 to obtain an adjusted image; convert the adjusted image into a grayscale image, and obtain the position of the black area from the grayscale image; Obtain an initial acquisition image corresponding to the grayscale image, determine the black area position at the area position corresponding to the initial acquisition image; remove the area position from the initial acquisition image, obtain the retained part of the initial acquisition image as the final image, and save it in a preset format; Extracting a number of preset RGB channel parameters from the final image, and extracting a number of preset meteorological parameters from the meteorological database; Taking a number of preset RGB channel parameters and a number of preset meteorological parameters as a set of data, obtaining each preset meteorological parameter, Kaiser-Meyer-Olkin value and Bartlett variance value corresponding to the preset meteorological parameter; Obtain each preset meteorological parameter, and the accuracy and weight corresponding to the preset meteorological parameter; Based on the Kaiser-Meyer-Olkin value, Bartlett variance value, accuracy and weight, two preset RGB channel parameters and two preset meteorological parameters are extracted as the final training parameters; specifically including: From the preset RGB channel parameters and the preset meteorological parameters, select the preset RGB channel parameters and the preset meteorological parameters whose Kaiser-Meyer-Olkin value is greater than the first preset value and whose Bartlett variance value is greater than the second preset value; by the formula: , calculate the impact factor H corresponding to each preset RGB channel parameter and preset meteorological parameter; where X represents the accuracy, Indicates weight; from all preset RGB channel parameters, select the first two with the largest impact factors as the final training parameters; from all preset meteorological parameters, select the first two with the largest impact factors as the final training parameters; The specific values ​​corresponding to the final training parameters in the final image and the annotated developmental period corresponding to the final image are obtained, which are then used as training data for the preset classification algorithm to obtain a trained preset classification model.

2. The method for intelligently determining the winter wheat growth period based on multi-source parameters according to claim 1, characterized in that: The preset winter wheat development stages include the three-leaf stage, tillering stage, greening stage, jointing stage, booting stage, heading stage, flowering stage, milky stage and maturity stage; The preset RGB channel parameters include at least the skewness of the R channel and the kurtosis of the G channel, and the preset meteorological parameters include at least: the cumulative value of the daily average temperature and the cumulative amount of photosynthetically active radiation; The default format is jpg.

3. The method for intelligently determining the winter wheat growth period based on multi-source parameters according to claim 1, characterized in that: Acquiring the original digital image corresponding to the winter wheat in the preset winter wheat growth period in the preset area, specifically including: During a preset collection time period, original digital images are collected at a preset collection frequency by a monitoring camera device located in the winter wheat field; wherein, one first preset focal length image and one second preset focal length image are collected at each frequency, and the installation height is a preset height.

4. The method for intelligently determining the winter wheat growth period based on multi-source parameters according to claim 1, characterized in that: Eliminate the images with light interference, foreign matter interference and lens attachment interference from the original digital images to obtain the initial acquisition image, including: By using edge detection algorithm, the original digital image with interference from foreign objects or lens attachments is determined, and then the original digital image with interference from foreign objects or lens attachments is removed; Through the brightness analysis algorithm of the image, the original digital image with local highlight areas is determined, and then the original digital image with local highlight areas is eliminated; wherein the local highlight area is an area with local brightness greater than a preset brightness value.

5. The method for intelligently determining the winter wheat growth period based on multi-source parameters according to claim 1, characterized in that: Remove the region from the initial acquisition image, including: Using an image segmentation algorithm, the region locations are removed from the initial acquired image.

6. An intelligent determination system for winter wheat development period based on multi-source parameters, characterized in that: The system comprises: The acquisition module is used to acquire the original digital image corresponding to the winter wheat in the preset winter wheat development period in the preset area; wherein the format of the original digital image is RGB format; An acquisition module is used to remove images with light interference, foreign matter interference and lens attachment interference from the original digital image to obtain an initial acquisition image; The region acquisition module is used to convert the initial collected image in RGB format into an HSV image, determine the pixels to be processed whose color gamut values ​​in the HSV communication are within a preset adjustment range and / or whose saturation is less than a preset threshold; adjust the image transparency of the pixels to be processed to 0 to obtain an adjusted image; convert the adjusted image into a grayscale image, and obtain the position of the black area from the grayscale image; The image acquisition module is used to acquire the initial acquisition image corresponding to the grayscale image, determine the black area position at the area position corresponding to the initial acquisition image; remove the area position from the initial acquisition image, obtain the retained part of the initial acquisition image as the final image, and save it in a preset format; The parameter determination module is used to extract several preset RGB channel parameters from the final image and several preset meteorological parameters from the meteorological database; take the several preset RGB channel parameters and the several preset meteorological parameters as a group of data to obtain each preset meteorological parameter and the Kaiser-Meyer-Olkin value and Bartlett variance value corresponding to the preset meteorological parameter; obtain each preset meteorological parameter and the accuracy and weight corresponding to the preset meteorological parameter; based on the Kaiser-Meyer-Olkin value, Bartlett variance value, accuracy and weight, extract two preset RGB channel parameters and two preset meteorological parameters as the final training parameters; specifically including: From the preset RGB channel parameters and the preset meteorological parameters, select the preset RGB channel parameters and the preset meteorological parameters whose Kaiser-Meyer-Olkin value is greater than the first preset value and whose Bartlett variance value is greater than the second preset value; by the formula: , calculate the impact factor H corresponding to each preset RGB channel parameter and preset meteorological parameter; where X represents the accuracy, Indicates weight; from all preset RGB channel parameters, select the first two with the largest impact factors as the final training parameters; from all preset meteorological parameters, select the first two with the largest impact factors as the final training parameters; The training module is used to obtain the specific values ​​corresponding to the final training parameters in the final image and the annotated developmental period corresponding to the final image, which are then used as training data for the preset classification algorithm to obtain a trained preset classification model.

7. The intelligent determination system for winter wheat development period based on multi-source parameters according to claim 6 is characterized in that: The acquisition module includes a collection unit, Used to collect original digital images at a preset collection frequency during a preset collection time period through a monitoring camera device located in a winter wheat field; wherein one first preset focal length image and one second preset focal length image are collected each time, and the installation height is a preset height.

8. The winter wheat growth stage intelligent determination system based on multi-source parameters according to claim 6 is characterized in that: The image acquisition module includes a segmentation unit, Used to remove area locations from the original digital image using image segmentation algorithms.

9. A non-volatile computer storage medium, characterized in that: Computer instructions are stored thereon, and when the computer instructions are executed, the method for intelligently determining the development period of winter wheat based on multi-source parameters as described in any one of claims 1 to 5 is implemented.

Images