Crop growth water demand prediction system and method based on big data

Abstract

The invention discloses a crop growth water demand prediction system and method based on big data, relates to the technical field of crop water demand prediction, and solves the problems that the evaluation accuracy of a corresponding operation partition cannot be fully guaranteed, and the predicted water demand accuracy needs to be improved. According to the invention, based on deep analysis of a soil humidity change curve in a monitoring period, multiple factors such as current soil humidity, a humidity change trend and duration from next irrigation are comprehensively considered, and required water volume intervals of different image areas are accurately locked; an adaptive water demand range can be given according to a unique soil humidity dynamic change rule for a standard image area or a substandard image area, so that a quantitative basis is provided for precise irrigation; compared with traditional empirical irrigation, the prediction method based on data driving greatly reduces adverse effects on crops due to insufficient or excessive irrigation.

Description

Technical Field

[0001] The present invention relates to the technical field of crop water quantity prediction, and in particular to a system and method for predicting crop growth water quantity demand based on big data. Background Art

[0002] When predicting the amount of water required for crop growth, the radial basis neural network algorithm is generally introduced. As an efficient feedforward neural network, the radial basis neural network has excellent nonlinear approximation ability and fast learning convergence characteristics. In the process of crop growth water demand prediction based on big data, it can make full use of the massive crop image data, soil moisture data, meteorological data and other multi-source information that have been collected. Taking crop image data as an example, the neural network accurately captures the potential mapping relationship between different image areas and crop water demand by deep learning the pixel features and contour features of image partitions. For soil moisture data, the dynamic change pattern of soil moisture and its influence mechanism on water demand can be accurately summarized according to the shape and fluctuation frequency of the moisture change curve. With the powerful modeling ability of the radial basis neural network, the complex laws hidden behind the multi-source data are presented in the form of mathematical models, further optimizing the prediction accuracy of the water demand range of different image areas, making the entire prediction method more robust and reliable, and promoting agricultural production in all directions towards intelligence and refinement, opening a new chapter for the development of modern agriculture.

[0003] The application with publication number CN111915062B discloses a method for regulating water demand of greenhouse crops by coordinating water utilization rate and photosynthetic rate, obtaining net photosynthetic rate and WUE data under different temperature, photon flux density, CO2 concentration and soil moisture nesting conditions, and constructing photosynthetic rate prediction model and WUE prediction model based on radial basis neural network; according to the photosynthetic rate prediction model, the response curve of photosynthetic rate to soil moisture under different temperature, photon flux density and CO2 concentration nesting is obtained, and its discrete curvature is calculated and the regulation interval is constructed; within the interval, the soil moisture value corresponding to the maximum WUE point is obtained based on the particle swarm optimization algorithm, and this is used as the regulation target value; the water demand model integrating WUE-photosynthetic rate is constructed by using the SVR algorithm, and the water demand of greenhouse crops is regulated based on the model. The invention can take into account both crop demand and economic benefits, and provides a theoretical basis for dynamic and efficient soil moisture regulation of facility crops;

[0004] When predicting the water demand for crop growth, the green performance of the corresponding crop zone is generally used to identify whether the water demand in the corresponding crop production area is sufficient. However, this method is not comprehensive, because there are green areas that meet the standards and green areas that do not meet the standards in the corresponding operation zones. The original identification method is relatively one-sided and cannot fully guarantee the assessment accuracy of the corresponding operation zones. The accuracy of the predicted water demand needs to be improved. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides a system and method for predicting the water demand for crop growth based on big data, which solves the problem that the assessment accuracy of the corresponding operation zones cannot be fully guaranteed and the accuracy of the predicted water demand needs to be improved.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: A method for predicting the water demand for crop growth based on big data, comprising the following steps:

[0007] Step 1: Acquire crop images of different crops. The crop images are acquired by monitoring probes set at a designated height. Pixel analysis is performed on the crop images of different crops, and different image partitions are determined from the crop images of the corresponding crops. The specific sub-steps are as follows:

[0008] S11. Perform image analysis on crop images of different crops, identify the pixel values ​​associated with different pixel points in the crop images and mark them as X i , where i represents different pixel points, and the pixel values ​​associated with multiple groups of pixel points adjacent to this pixel point are determined and marked as SZ1, SZ2, ..., SZ8;

[0009] S12: Based on the determined pixel point and the associated pixel value X i , confirm the gradient features of this pixel, which include the horizontal and vertical gradients Gx i And the vertical gradient Gy i :

[0010] Where Gx i =(-1)×SZ1+0×SZ2+1×SZ3+(-2)×SZ4+0×X i +2×SZ5+(-1)×SZ6+0×SZ7+1×SZ8;

[0011] Gy i =(-1)×SZ1+(-2)×SZ2+(-1)×SZ3+0×SZ4+0×X i +0×SZ5+1×SZ6+2×SZ7+1×SZ8;

[0012] use Determine the gradient feature G associated with this undetermined point i ;

[0013] S13, the determined gradient feature G i Compare with the preset value Y1, where Y1 is the set image area segmentation standard gradient, and G i Pixels ≥Y1 are marked as contour points;

[0014] S14, based on the several groups of contour points determined in the crop image, adjacent contour points are connected to confirm the image contour, the area covered by the image contour is calibrated, and different image partitions are determined from the crop image;

[0015] Step 2: Based on the different groups of image partitions determined in the crop image, the different image partitions are divided into qualified image areas or non-qualified image areas according to different image features in the different image partitions. The specific sub-steps are as follows:

[0016] S21, based on different pixel values ​​associated with different pixel points in different image partitions, performing average processing on several groups of pixel values ​​associated with several pixel points, confirming the pixel average, and using the confirmed pixel average as the image feature of this image partition;

[0017] S22, confirming the pixel standard interval associated with the crop image corresponding to the crop, wherein the pixel standard interval is a preset interval, and calibrating the image partition with the image feature∈the pixel standard interval as the qualified image area, wherein different crop images have different pixel standard intervals, otherwise no calibration is performed;

[0018] Step 3: Based on the confirmed qualified or unqualified image areas in the crop image, monitor the soil moisture in different image areas, predict the water demand of different image areas based on the soil moisture changes within the cycle, and display the predicted water demand of the image areas. The specific sub-steps are as follows:

[0019] S31, defining a set of monitoring periods, confirming the soil moisture change data of the corresponding image area within the monitoring period, wherein the monitoring period is a preset period, and generating a soil moisture change curve of the corresponding image area in this monitoring period based on different soil moisture change data associated with different moments in this monitoring period;

[0020] S32, confirming the change trend from the confirmed soil moisture change curve: identifying the soil moisture change values ​​associated with adjacent moments, where the soil moisture change value = the soil moisture associated with the previous moment - the soil moisture associated with the next moment, selecting the maximum soil moisture change value and the minimum soil moisture change value from the corresponding soil moisture change curve, and confirming the soil moisture change interval belonging to this image area;

[0021] S33, confirm the specific time from the current moment to the next irrigation moment, the irrigation moment is the preset moment, and then based on the soil moisture associated with the current moment and the soil moisture change interval, extend the curve, and determine the soil moisture data associated with the extension end moment: calibrate the maximum soil moisture change value as Rmax, calibrate the minimum soil moisture change value as Rmin, calibrate the confirmed specific time as Sc, and calibrate the soil moisture associated with the current moment as SD, and use: DRmin = SD-(Rmax×Sc) to confirm the minimum soil moisture data DRmin associated with the end moment, and then use DRmax = SD-(Rmin×Sc) to confirm the maximum soil moisture data DRmax associated with the end moment;

[0022] S34, based on the DRmin and DRmax confirmed in the corresponding image area, lock the humidity data interval [DRmin, DRmax]:

[0023] If this image area belongs to the standard image area, determine the soil moisture demand value X1, identify the difference Cz between [DRmin, DRmax] and X1, where Czmax=X1-DRmin, Czmin=X1-DRmax, use Czmax×C1=XQmax to determine the maximum required water volume XQmax, and then use Czmin×C1=XQmin to determine the minimum required water volume XQmin, where C1 is a preset fixed coefficient factor to lock the required water volume interval associated with this image area;

[0024] If this image area belongs to the non-standard image area, determine the soil moisture demand value X2, identify the difference Cz between [DRmin, DRmax] and X2, where Czmax=X2-DRmin, Czmin=X2-DRmax, use Czmax×C1=XQmax to determine the maximum required water volume XQmax, and then use Czmin×C1=XQmin to determine the minimum required water volume XQmin, where C1 is a preset fixed coefficient factor, and lock the required water volume interval associated with this image area, where X1 and X2 are both preset values, and X1<X2;

[0025] Step 4: Based on the different water demand intervals predicted for different image areas in the corresponding crop image, the specific water demand associated with the corresponding image area is selected, and the selected specific water demand is displayed. The specific sub-steps are as follows:

[0026] Based on the different water demand intervals associated with different image areas in the crop image, identify whether there are cross-demand water demand segments in several groups of different water demand intervals. These cross-demand water demand segments belong to each group of water demand intervals:

[0027] If it exists, the maximum water demand of this cross-water demand segment is selected, and this maximum water demand is used as the predicted water demand of the corresponding crop image and displayed;

[0028] If it does not exist, confirm the maximum values ​​of several different water demand intervals and mark the confirmed maximum value as QJ k , where k represents different water demand intervals, locking QJ k min, and lock the QJ k min is used as the predicted water requirement for the corresponding crop image and displayed.

[0029] Preferably, a crop growth water demand prediction system based on big data comprises:

[0030] The image partition processing end obtains crop images of different crops, which are obtained by monitoring probes set at a designated height, performs pixel analysis on the crop images of different crops, and determines different image partitions from the crop images of the corresponding crops;

[0031] The image area calibration processing end, based on a number of different image areas determined in the crop image, divides the different image areas into qualified image areas or unqualified image areas according to different image features in the different image areas;

[0032] The water demand processing end monitors the soil moisture in different image areas based on the image areas that meet or do not meet the standards confirmed in the crop image, predicts the water demand in different image areas based on the soil moisture changes within the cycle, and displays the predicted water demand for the image areas;

[0033] The demand water selection end selects the specific demand water associated with the corresponding image area based on the different demand water ranges predicted in different image areas within the corresponding crop image, and displays the selected specific demand water.

[0034] The present invention provides a system and method for predicting the amount of water required for crop growth based on big data. Compared with the prior art, it has the following beneficial effects:

[0035] Based on the in-depth analysis of the soil moisture change curve within the monitoring period, the present invention comprehensively considers multiple factors such as the current soil moisture, the moisture change trend, and the time until the next irrigation, and accurately locks the required water volume range of different image areas; whether it is an image area that meets the standard or an image area that does not meet the standard, it can give an adapted water demand range according to its unique soil moisture dynamic change law, providing a quantitative basis for precision irrigation; compared with traditional empirical irrigation, this data-driven prediction method greatly reduces the adverse effects of insufficient or excessive irrigation on crops, ensuring that every drop of water is used on the "cutting edge", saving water while ensuring a stable supply of water required for crop growth;

[0036] According to different crop types, exclusive image standards, pixel standard ranges, soil moisture requirement values ​​and other parameters are set, which is suitable for a variety of crop planting scenarios. No matter whether it is food crops, cash crops or fruit and vegetable crops, as long as the corresponding parameters are adjusted in a targeted manner, it can run accurately. Moreover, with the continuous accumulation of agricultural big data, the system can continuously optimize and learn itself, further improve the accuracy and reliability of the prediction, and easily cope with the problem of predicting crop water demand in different regions, different climatic conditions and different planting patterns, providing a solid technical guarantee for the widespread promotion of smart agriculture. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a schematic diagram of the process of the present invention;

[0038] Figure 2 It is a schematic diagram of the principle framework of the present invention. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] First embodiment

[0041] See also Figure 1 The present application provides a method for predicting the water demand for crop growth based on big data, comprising the following steps:

[0042] Step 1: Acquire crop images of different crops. The crop images are acquired by monitoring probes set at a designated height. The acquisition time interval is a preset time, which is prepared in advance by the operator. Generally, the value is 48 hours or more. Pixel analysis is performed on the crop images of different crops, and different image partitions are determined from the crop images of the corresponding crops. Specifically, there are different image standards for crop images of different crops. According to the preset image standards, the qualified areas of the corresponding crop images can be effectively confirmed, and the area confirmation of the corresponding crop images is completed;

[0043] The specific sub-steps of pixel analysis of crop images of different crops are as follows:

[0044] S11. Perform image analysis on crop images of different crops, identify the pixel values ​​associated with different pixel points in the crop images and mark them as X i , where i represents different pixel points, the pixel values ​​associated with multiple groups of pixel points adjacent to this pixel point are determined and marked as SZ1, SZ2, ..., SZ8. The pixel points are arranged in an array in the image, so there are only eight groups of pixel values ​​around each group of pixel points;

[0045] S12: Based on the determined pixel point and the associated pixel value X i , confirm the gradient features of this pixel, which include the horizontal and vertical gradients Gx i And the vertical gradient Gy i :

[0046] Where Gx i =(-1)×SZ1+0×SZ2+1×SZ3+(-2)×SZ4+0×X i +2×SZ5+(-1)×SZ6+0×SZ7+1×SZ8;

[0047] Among them, Gy i =(-1)×SZ1+(-2)×SZ2+(-1)×SZ3+0×SZ4+0×X i +0×SZ5+1×SZ6+2×SZ7+1×SZ8;

[0048] use Determine the gradient feature G associated with this undetermined point i ;

[0049] S13, the determined gradient feature G i Compare with the preset value Y1, where Y1 is the set image area segmentation standard gradient, which is prepared in advance by relevant operators based on experience. i Pixels with Y1 or above are calibrated as contour points, otherwise no calibration is performed;

[0050] S14, based on the several groups of contour points determined in the crop image, adjacent contour points are connected to confirm the image contour, the area covered by the image contour is calibrated, and different image partitions are determined from the crop image;

[0051] Specifically, in the corresponding crop images, the areas covered by different green plants have different contour features, and the differences between pixel points are relatively large. Therefore, the contour can be confirmed based on the corresponding gradient features, so that the corresponding crop image can be divided into multiple different image partitions, and the edge contours associated with each different image partition are different.

[0052] Step 2: Based on the different groups of image partitions determined in the crop image, the different image partitions are divided into qualified image areas or non-qualified image areas according to different image features in the different image partitions, wherein the specific sub-steps of the division are:

[0053] S21, based on different pixel values ​​associated with different pixel points in different image partitions, performing average processing on several groups of pixel values ​​associated with several pixel points, confirming the pixel average, and using the confirmed pixel average as the image feature of this image partition;

[0054] S22, confirm the pixel standard interval associated with the crop image corresponding to the crop, the pixel standard interval is a preset interval, which is prepared by the operator in advance based on experience, and the image partition with the image feature ∈ the pixel standard interval is marked as the qualified image area, otherwise, the corresponding image partition is marked as the non-qualified image area, wherein different crop images have different pixel standard intervals, and in the corresponding crop image, different image values ​​represent different image signs of the corresponding area, so that the qualified image can be specifically confirmed based on different image signs, and then the qualified image is confirmed from the confirmed image, and the humidity values ​​of different lands in the qualified image and the non-qualified image need to be distinguished and analyzed, and based on the actually analyzed humidity values, the relevant prediction of the required water volume of the crop partition associated with the crop image is performed;

[0055] Step 3: Based on the confirmed qualified image areas or non-qualified image areas in the crop image, the soil moisture of different image areas is monitored, and the required water volume of different image areas is predicted based on the soil moisture changes within the cycle, and the predicted required water volume of the image areas is displayed. The specific sub-steps of the prediction are as follows:

[0056] S31, defining a set of monitoring periods, confirming the soil moisture change data of the corresponding image area within the monitoring period, wherein the monitoring period is a preset period, the specific value of which is determined by the operator based on experience, and generating a soil moisture change curve of the corresponding image area in this monitoring period based on different soil moisture change data associated with different moments within this monitoring period;

[0057] S32, confirming the change trend from the confirmed soil moisture change curve: identifying the soil moisture change values ​​associated with adjacent moments, where the soil moisture change value = the soil moisture associated with the previous moment - the soil moisture associated with the next moment, selecting the maximum soil moisture change value and the minimum soil moisture change value from the corresponding soil moisture change curve, and confirming the soil moisture change interval belonging to this image area;

[0058] S33, confirm the specific time from the current moment to the next irrigation moment, the irrigation moment is a preset moment, which is prepared in advance by the relevant operators, and then based on the soil moisture associated with the current moment and the soil moisture change range, extend the curve, and determine the soil moisture data associated with the extension end moment: calibrate the maximum soil moisture change value as Rmax, the minimum soil moisture change value as Rmin, the confirmed specific time as Sc, and the soil moisture associated with the current moment as SD, and use: DRmin = SD-(Rmax×Sc) to confirm the minimum soil moisture data DRmin associated with the end moment, and then use DRmax = SD-(Rmin×Sc) to confirm the maximum soil moisture data DRmax associated with the end moment;

[0059] S34, based on the DRmin and DRmax confirmed in the corresponding image area, lock the humidity data interval [DRmin, DRmax]:

[0060] If this image area belongs to the standard image area, determine the soil moisture demand value X1, identify the difference Cz between [DRmin, DRmax] and X1, where Czmax=X1-DRmin, Czmin=X1-DRmax, use Czmax×C1=XQmax to determine the maximum required water volume XQmax, and then use Czmin×C1=XQmin to determine the minimum required water volume XQmin, where C1 is a preset fixed coefficient factor, and its specific value is determined by the operator based on experience to lock the required water volume interval associated with this image area;

[0061] If this image area belongs to the non-standard image area, determine the soil moisture demand value X2, identify the difference Cz between [DRmin, DRmax] and X2, where Czmax=X2-DRmin, Czmin=X2-DRmax, use Czmax×C1=XQmax to determine the maximum required water volume XQmax, and then use Czmin×C1=XQmin to determine the minimum required water volume XQmin, where C1 is a preset fixed coefficient factor, and its specific value is determined by the operator based on experience, and the required water volume interval associated with this image area is locked, where X1 and X2 are both preset values, which are determined by relevant operators based on experience, and X1<X2;

[0062] Specifically, different types of image areas have different humidity data changes. Based on the current time and the specific time of the next irrigation, the relevant humidity data change duration can be identified. Based on the corresponding change duration, the subsequent associated humidity data can be specifically confirmed, and the required water volume of the corresponding area can be determined based on the confirmed specific humidity data. Therefore, the specific water volume data of this image area can be confirmed to lock the specific required water volume range.

[0063] Step 4: Based on the different water demand intervals predicted for different image areas in the corresponding crop image, the specific water demand associated with the corresponding image area is selected, and the selected specific water demand is displayed for external relevant personnel to view. The specific sub-steps for selecting the specific water demand are:

[0064] Based on the different water demand intervals associated with different image areas in the crop image, identify whether there are cross-demand water segments in several groups of different water demand intervals. These cross-demand water segments belong to each group of water demand intervals (that is, these water segments exist in each group of water demand intervals):

[0065] If it exists, the maximum water demand of this cross-demand water demand segment is selected, and this maximum water demand is used as the predicted water demand of the corresponding crop image and displayed;

[0066] If it does not exist, confirm the maximum values ​​of several different water demand intervals and mark the confirmed maximum value as QJ k , where k represents different water demand intervals, locking QJ k min, and lock the QJ k min is used as the predicted water requirement for the corresponding crop image and displayed.

[0067] Specifically, in the actual processing process, if there are identical water volume intervals between multiple intervals, the maximum value is directly locked from the water volume interval, and the locked maximum value is used as the water demand of this area. This water demand characteristic has better performance and will not affect the water demand of other areas, so as to achieve better water volume prediction effect. If there is no overlapping water volume segment, the minimum water demand is confirmed. If the maximum water volume is selected, it will cause excessive water for crops in some areas, which will easily cause crops to be flooded. Therefore, the minimum demand is selected to achieve better water volume confirmation effect.

[0068] Second embodiment

[0069] The crop growth water demand prediction system based on big data includes:

[0070] The image partition processing end obtains crop images of different crops, which are obtained by monitoring probes set at a designated height, performs pixel analysis on the crop images of different crops, and determines different image partitions from the crop images of the corresponding crops;

[0071] The image area calibration processing end, based on a number of different image areas determined in the crop image, divides the different image areas into qualified image areas or unqualified image areas according to different image features in the different image areas;

[0072] The water demand processing end monitors the soil moisture in different image areas based on the image areas that meet or do not meet the standards confirmed in the crop image, predicts the water demand in different image areas based on the soil moisture changes within the cycle, and displays the predicted water demand for the image areas;

[0073] The demand water selection end selects the specific demand water associated with the corresponding image area based on the different demand water ranges predicted in different image areas within the corresponding crop image, and displays the selected specific demand water.

[0074] Some of the data in the above formulas are dimensionless and numerically calculated. Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0075] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A method for predicting water demand for crop growth based on big data, characterized in that: The following steps are involved: Step 1: Acquire crop images of different crops, wherein the crop images are acquired by a monitoring probe set at a designated height, and pixel analysis is performed on the crop images of different crops to determine different image partitions from the crop images of corresponding crops; Step 2: based on the different groups of image partitions determined in the crop image, the different image partitions are divided into qualified image areas or unqualified image areas according to different image features in the different image partitions; Step 3: Based on the confirmed qualified image areas or non-qualified image areas in the crop image, the soil moisture of different image areas is monitored, and the required water volume of different image areas is predicted based on the soil moisture changes within the cycle, and the predicted required water volume of the image areas is displayed; Step 4: Based on the different water demand intervals predicted for different image areas in the corresponding crop image, a specific water demand associated with the corresponding image area is selected, and the selected specific water demand is displayed.

2. The method for predicting crop growth water demand based on big data according to claim 1, characterized in that: In step 1, the specific sub-steps of performing pixel analysis on crop images of different crops are: S11. Perform image analysis on crop images of different crops, identify the pixel values ​​associated with different pixel points in the crop images and mark them as X i , where i represents different pixel points, and the pixel values ​​associated with multiple groups of pixel points adjacent to this pixel point are determined and marked as SZ1, SZ2, ..., SZ8; S12: Based on the determined pixel point and the associated pixel value X i , confirm the gradient features of this pixel point, its gradient features include horizontal and vertical gradients Gx i And the vertical gradient Gy i : Where Gx i =(-1)×SZ1+0×SZ2+1×SZ3+(-2)×SZ4+0×X i +2×SZ5+(-1)×SZ6+0×SZ7+1×SZ8; Among them, Gy i =(-1)×SZ1+(-2)×SZ2+(-1)×SZ3+0×SZ4+0×X i +0×SZ5+1×SZ6+2×SZ7+1×SZ8; use Determine the gradient feature G associated with this undetermined point i ; S13, the determined gradient feature G i Compare with the preset value Y1, where Y1 is the set image area segmentation standard gradient, and G i Pixels ≥Y1 are marked as contour points; S14. Based on the several groups of contour points determined in the crop image, adjacent contour points are connected to confirm the image contour, the area covered by the image contour is calibrated, and different image partitions are determined from the crop image.

3. The method for predicting crop growth water demand based on big data according to claim 2, characterized in that: In step S13, for G i No calibration is performed on pixels < Y1.

4. The method for predicting crop growth water demand based on big data according to claim 1, characterized in that: In the step 2, the specific sub-steps of dividing different image areas are: S21, based on different pixel values ​​associated with different pixel points in different image partitions, performing average processing on several groups of pixel values ​​associated with several pixel points, confirming the pixel average, and using the confirmed pixel average as the image feature of this image partition; S22, confirming the pixel standard interval associated with the crop image corresponding to the crop, wherein the pixel standard interval is a preset interval, and marking the image partition with the image feature∈the pixel standard interval as the qualified image area, wherein different crop images have different pixel standard intervals.

5. The method for predicting crop growth water demand based on big data according to claim 4, characterized in that: In step S22, for No calibration is performed on the image partitions.

6. The method for predicting crop growth water demand based on big data according to claim 1, characterized in that: In step 3, the specific sub-steps for predicting the required water volume of different image areas are: S31, defining a set of monitoring periods, confirming the soil moisture change data of the corresponding image area within the monitoring period, wherein the monitoring period is a preset period, and generating a soil moisture change curve of the corresponding image area in this monitoring period based on different soil moisture change data associated with different moments in this monitoring period; S32, confirming the change trend from the confirmed soil moisture change curve: identifying the soil moisture change values ​​associated with adjacent moments, where the soil moisture change value = the soil moisture associated with the previous moment - the soil moisture associated with the next moment, selecting the maximum soil moisture change value and the minimum soil moisture change value from the corresponding soil moisture change curve, and confirming the soil moisture change interval belonging to this image area; S33, confirm the specific time from the current moment to the next irrigation moment, the irrigation moment is the preset moment, and then based on the soil moisture associated with the current moment and the soil moisture change interval, extend the curve, and determine the soil moisture data associated with the extension end moment: calibrate the maximum soil moisture change value as Rmax, calibrate the minimum soil moisture change value as Rmin, calibrate the confirmed specific time as Sc, and calibrate the soil moisture associated with the current moment as SD, and use: DRmin = SD-(Rmax×Sc) to confirm the minimum soil moisture data DRmin associated with the end moment, and then use DRmax = SD-(Rmin×Sc) to confirm the maximum soil moisture data DRmax associated with the end moment; S34, based on the DRmin and DRmax confirmed in the corresponding image area, lock the humidity data interval [DRmin, DRmax]: If this image area belongs to the standard image area, determine the soil moisture demand value X1, identify the difference Cz between [DRmin, DRmax] and X1, where Czmax=X1-DRmin, Czmin=X1-DRmax, use Czmax×C1=XQmax to determine the maximum required water volume XQmax, and then use Czmin×C1=XQmin to determine the minimum required water volume XQmin, where C1 is a preset fixed coefficient factor to lock the required water volume interval associated with this image area; If this image area belongs to the non-standard image area, determine the soil moisture demand value X2, identify the difference Cz between [DRmin, DRmax] and X2, where Czmax = X2-DRmin, Czmin = X2-DRmax, use Czmax×C1=XQmax to determine the maximum required water volume XQmax, and then use Czmin×C1=XQmin to determine the minimum required water volume XQmin, where C1 is a preset fixed coefficient factor, locking the required water volume interval associated with this image area, where X1 and X2 are both preset values, and X1<X2.

7. The method for predicting crop growth water demand based on big data according to claim 6, characterized in that: In step 4, the specific sub-steps for selecting the specific required water volume are: Based on the different water demand intervals associated with different image areas in the crop image, identify whether there are cross-demand water demand segments in several groups of different water demand intervals. These cross-demand water demand segments belong to each group of water demand intervals: If it exists, the maximum water demand of this cross-water demand segment is selected, and this maximum water demand is used as the predicted water demand of the corresponding crop image and displayed.

8. The method for predicting crop growth water demand based on big data according to claim 7, characterized in that: If it does not exist, confirm the maximum values ​​of several different water demand intervals and mark the confirmed maximum value as QJ k , where k represents different water demand intervals, locking QJ k min, and lock the QJ k min is used as the predicted water requirement for the corresponding crop image and displayed.

9. A crop growth water demand prediction system based on big data, the prediction system is operated according to a crop growth water demand prediction method based on big data according to any one of claims 1 to 8, characterized in that: include: The image partition processing end obtains crop images of different crops, which are obtained by monitoring probes set at a designated height, performs pixel analysis on the crop images of different crops, and determines different image partitions from the crop images of the corresponding crops; The image area calibration processing end, based on a number of different image areas determined in the crop image, divides the different image areas into qualified image areas or unqualified image areas according to different image features in the different image areas; The water demand processing end monitors the soil moisture in different image areas based on the image areas that meet or do not meet the standards confirmed in the crop image, predicts the water demand in different image areas based on the soil moisture changes within the cycle, and displays the predicted water demand for the image areas; The demand water selection end selects the specific demand water associated with the corresponding image area based on the different demand water ranges predicted in different image areas within the corresponding crop image, and displays the selected specific demand water.

Images