Soil moisture inversion model generation method and device

Abstract

The application relates to the technical field of soil humidity detection, in particular to a soil humidity inversion model generation method and device, a soil humidity inversion method and device, an irrigation method and device, an irrigation anomaly determination method and device, a computer readable storage medium and an electronic device, which solve the problems that obtaining soil humidity is large in workload, low in efficiency and accuracy, cannot intuitively understand the soil humidity condition, and cannot detect the humidity of water-soaked soil. The soil humidity inversion model generation method provided in the application embodiment trains an initial network model by using rich training samples to generate a soil humidity inversion model, thereby improving the accuracy of the soil humidity inversion model. The soil humidity inversion model can be used to determine the humidity image corresponding to the to-be-detected soil region, without needing to perform a large amount of soil sampling, thereby reducing the workload and improving the efficiency of soil humidity detection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of neural networks, in particular to a soil moisture inversion model generation method and device, a soil moisture inversion method and device, an irrigation method and device, an irrigation anomaly determination method and device, and a computer readable storage medium and an electronic device. BACKGROUND

[0002] At present, the method for obtaining soil moisture is the drying method. The drying method mainly calculates soil moisture based on the initial collected soil weight and the soil weight after drying.

[0003] However, the drying method of manually collecting soil samples requires a large amount of manpower, resulting in a large amount of work and low efficiency in obtaining soil moisture. In addition, factors affecting soil moisture include soil clumps caused by flatness defects, loose surface caused by cultivation, soil type differences, footprints, etc. Various factors affecting soil moisture can make it difficult for the drying method to take all types of soil during sampling, thereby reducing the accuracy of the drying method. Furthermore, the soil moisture obtained by the drying method is some discrete humidity values, which cannot intuitively understand the overall soil moisture situation. In addition, the drying method cannot detect the humidity of water-soaked soil. SUMMARY

[0004] Therefore, the embodiments of the present application provide a soil moisture inversion model generation method and device, a soil moisture inversion method and device, an irrigation method and device, an irrigation anomaly determination method and device, and a computer readable storage medium and an electronic device to solve the problems of large amount of work, low efficiency, low accuracy, and inability to intuitively understand the soil moisture situation, and inability to detect the humidity of water-soaked soil.

[0005] In a first aspect, an embodiment of the present application provides a soil moisture inversion model generation method, comprising: determining a plurality of remote sensing images corresponding to a soil area sample, wherein the plurality of remote sensing images correspond one-to-one to a plurality of preset collection time points; determining a plurality of soil moisture data sets corresponding to the soil area sample, wherein the plurality of soil moisture data sets correspond one-to-one to the plurality of preset collection time points; generating a humidity labeled image corresponding to each of the plurality of remote sensing images based on the plurality of soil moisture data sets and the plurality of remote sensing images; and training an initial network model based on the plurality of remote sensing images and the humidity labeled image corresponding to each of the plurality of remote sensing images to generate a soil moisture inversion model.

[0006] In some implementations of the first aspect, the soil region sample includes a plurality of to-be-tested regions, and generating the humidity labeled image corresponding to each of the plurality of remote sensing images based on the soil humidity data set corresponding to each of the plurality of remote sensing images includes: labeling the plurality of to-be-tested regions for each of the plurality of remote sensing images to obtain a region labeled image corresponding to the remote sensing image, where the region labeled image includes a plurality of labeled regions, and the plurality of labeled regions correspond one-to-one to the plurality of to-be-tested regions; and generating the humidity labeled image corresponding to the remote sensing image based on the region labeled image and the soil humidity data set corresponding to the remote sensing image.

[0007] In some implementations of the first aspect, generating the humidity labeled image corresponding to each of the plurality of remote sensing images based on the plurality of soil humidity data sets and the plurality of remote sensing images includes: determining a soil humidity data set corresponding to each of the plurality of remote sensing images based on a preset collection time point corresponding to each of the plurality of remote sensing images; and generating the humidity labeled image corresponding to each of the plurality of remote sensing images based on the plurality of remote sensing images and the soil humidity data set corresponding to each of the plurality of remote sensing images.

[0008] In some implementations of the first aspect, the soil region sample includes a plurality of to-be-tested regions, and generating the humidity labeled image corresponding to each of the plurality of remote sensing images based on the soil humidity data set corresponding to each of the plurality of remote sensing images includes: labeling the plurality of to-be-tested regions for each of the plurality of remote sensing images to obtain a region labeled image corresponding to the remote sensing image, where the region labeled image includes a plurality of labeled regions, and the plurality of labeled regions correspond one-to-one to the plurality of to-be-tested regions; and generating the humidity labeled image corresponding to the remote sensing image based on the region labeled image and the soil humidity data set corresponding to the remote sensing image.

[0009] In some implementations of the first aspect, generating the humidity labeled image corresponding to each of the plurality of remote sensing images based on the region labeled image and the soil humidity data set corresponding to the remote sensing image includes: generating humidity labeled data corresponding to each of the plurality of labeled regions based on the plurality of labeled regions included in the region labeled image and the soil humidity data set corresponding to the remote sensing image; and generating the humidity labeled image corresponding to the remote sensing image based on the region labeled image and the humidity labeled data corresponding to each of the plurality of labeled regions.

[0010] In some implementations of the first aspect, generating the humidity labeled data corresponding to each of the plurality of labeled regions based on the plurality of labeled regions included in the region labeled image and the soil humidity data set corresponding to the remote sensing image includes: determining, for each of the plurality of labeled regions included in the region labeled image, a to-be-tested region to which the labeled region belongs and a relative position relationship between the to-be-tested region to which the labeled region belongs and a humidity sensor corresponding to the to-be-tested region; and determining the humidity labeled data corresponding to the labeled region based on the relative position relationship and the soil humidity data set corresponding to the remote sensing image.

[0011] In a second aspect, an embodiment of the present application provides a soil moisture inversion method, including: determining a soil moisture inversion model, wherein the soil moisture inversion model is generated based on the soil moisture inversion model generation method mentioned in the first aspect; and determining a humidity image corresponding to a to-be-detected soil region based on a remote sensing image corresponding to the to-be-detected soil region by using the soil moisture inversion model.

[0012] In a third aspect, an embodiment of the present application provides an irrigation method, including: obtaining farmland soil moisture, wherein the farmland soil moisture is obtained based on the soil moisture inversion method mentioned in the second aspect; determining different humidity regions in the farmland according to the obtained farmland soil moisture; and determining an irrigation scheme according to the humidity regions.

[0013] In a fourth aspect, an embodiment of the present application provides an irrigation anomaly determination method, including: obtaining farmland soil moisture, wherein the farmland soil moisture is obtained based on the soil moisture inversion method mentioned in the second aspect; determining different humidity regions in the farmland according to the obtained farmland soil moisture; comparing and analyzing the humidity of the different humidity regions to determine an abnormal humidity region; and determining an irrigation anomaly point according to the abnormal humidity region.

[0014] In a fifth aspect, an embodiment of the present application provides a soil moisture inversion model generation device, including: a first confirmation module configured to determine a plurality of remote sensing images corresponding to a soil region sample, wherein the plurality of remote sensing images correspond to a plurality of preset collection time points in a one-to-one manner; a second confirmation module configured to determine a plurality of soil moisture data sets corresponding to the soil region sample, wherein the plurality of soil moisture data sets correspond to the plurality of preset collection time points in a one-to-one manner; a generation module configured to generate a humidity labeled image corresponding to each of the plurality of remote sensing images based on the plurality of soil moisture data sets and the plurality of remote sensing images; and a model training module configured to train an initial network model based on the plurality of remote sensing images and the humidity labeled image corresponding to each of the plurality of remote sensing images to generate a soil moisture inversion model.

[0015] In a sixth aspect, an embodiment of the present application provides a soil moisture inversion device, including: a model determination module configured to determine a soil moisture inversion model, wherein the soil moisture inversion model is generated based on the soil moisture inversion model generation method mentioned in the first aspect; and a soil moisture analysis module configured to determine a humidity image corresponding to a to-be-detected soil region based on a remote sensing image corresponding to the to-be-detected soil region by using the soil moisture inversion model.

[0016] In a seventh aspect, an embodiment of the present application provides an irrigation device, comprising: a first farmland humidity acquisition module configured to acquire farmland soil humidity, wherein the farmland soil humidity is obtained based on the soil humidity inversion method mentioned in the second aspect; a first humidity region confirmation module configured to determine different humidity regions in the farmland according to the acquired farmland soil humidity; and an irrigation scheme determination module configured to determine an irrigation scheme according to the humidity regions.

[0017] In an eighth aspect, an embodiment of the present application provides an irrigation anomaly determination device, comprising: a second farmland humidity acquisition module configured to acquire farmland soil humidity, wherein the farmland soil humidity is obtained based on the soil humidity inversion method mentioned in the second aspect; a second humidity region confirmation module configured to determine different humidity regions in the farmland according to the acquired farmland soil humidity; an anomaly analysis module configured to compare and analyze the humidity of the different humidity regions to determine an abnormal humidity region; and an anomaly point confirmation module configured to determine an irrigation anomaly point according to the abnormal humidity region.

[0018] In a ninth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores instructions, and when the instructions are executed by a processor of an electronic device, the electronic device can execute the soil humidity inversion model generation method mentioned in the first aspect and / or the soil humidity inversion method mentioned in the second aspect and / or the irrigation method mentioned in the third aspect and / or the irrigation anomaly determination method mentioned in the fourth aspect.

[0019] In a tenth aspect, an embodiment of the present application provides an electronic device, which comprises: a processor and a memory for storing computer executable instructions; and the processor is configured to execute the computer executable instructions to implement the soil humidity inversion model generation method mentioned in the first aspect and / or the soil humidity inversion method mentioned in the second aspect and / or the irrigation method mentioned in the third aspect and / or the irrigation anomaly determination method mentioned in the fourth aspect.

[0020] The soil humidity inversion model generation method provided by the embodiments of the present application generates a plurality of humidity labeled images based on a plurality of soil humidity data sets corresponding to a plurality of preset acquisition time points and a plurality of remote sensing images, so as to form rich training samples corresponding to the plurality of preset acquisition time points. The initial network model is trained by using the rich training samples to generate the soil humidity inversion model, thereby improving the accuracy of the soil humidity inversion model.

[0021] The soil humidity inversion method uses the soil humidity inversion model, can directly output the humidity image, that is, directly shows the soil humidity condition, and improves the user experience. In addition, the soil humidity inversion method uses the soil humidity inversion model to determine the humidity image corresponding to the to-be-measured soil region based on the remote sensing image corresponding to the to-be-measured soil region, without the need for a large amount of soil sampling, thereby reducing the workload and improving the efficiency of soil humidity detection. The soil humidity inversion method does not need soil sampling, can directly output the humidity labeled image of a complex farmland scene, is not limited by complex and diverse farmland scenes, and improves the user experience. BRIEF DESCRIPTION OF DRAWINGS

[0022] FIG. 1 Fig. 1 shows an application scenario of a soil humidity inversion model generation method according to an embodiment of the present application.

[0023] FIG. 2 Fig. 2 shows an application scenario of a soil humidity inversion model generation method according to another embodiment of the present application.

[0024] FIG. 3 Fig. 3 shows a flowchart of a soil humidity inversion model generation method according to an embodiment of the present application.

[0025] FIG. 4 Fig. 4 shows a flowchart of a soil humidity inversion model generation method according to another embodiment of the present application.

[0026] FIG. 5 Fig. 5 shows a flowchart of a soil humidity inversion model generation method according to another embodiment of the present application.

[0027] FIG. 6 Fig. 6 shows a flowchart of a soil humidity inversion model generation method according to another embodiment of the present application.

[0028] FIG. 7 Fig. 7 shows a flowchart of a soil humidity inversion model generation method according to another embodiment of the present application.

[0029] FIG. 8 Fig. 8 shows a flowchart of a soil humidity inversion model generation method according to another embodiment of the present application.

[0030] FIG. 9 Fig. 9 shows a flowchart of a soil humidity inversion method according to an embodiment of the present application.

[0031] FIG. 10 Fig. 10 shows a flowchart of an irrigation method according to an embodiment of the present application.

[0032] FIG. 11 Fig. 11 shows a flowchart of an irrigation anomaly determination method according to an embodiment of the present application.

[0033] FIG. 12 The diagram shown is a schematic diagram of the soil moisture inversion model generation device provided in an embodiment of this application.

[0034] FIG. 13 The diagram shown is a structural schematic of the second determining module provided in an embodiment of this application.

[0035] FIG. 14 The diagram shown is a structural schematic of a generation module provided in an embodiment of this application.

[0036] FIG. 15 The diagram shown is a schematic diagram of the structure of an annotation image generation unit provided in an embodiment of this application.

[0037] FIG. 16 The diagram shown is a structural schematic of an annotation image generation subunit provided in an embodiment of this application.

[0038] FIG. 17 The diagram shown is a structural schematic of a subunit for determining annotation data according to an embodiment of this application.

[0039] FIG. 18 The diagram shown is a structural schematic of a soil moisture inversion device provided in an embodiment of this application.

[0040] FIG. 19 The diagram shown is a structural schematic of an irrigation device provided in an embodiment of this application.

[0041] FIG. 20 The diagram shown is a structural schematic of an irrigation anomaly determination device provided in an embodiment of this application.

[0042] FIG. 21 The diagram shown is a structural schematic of an electronic device provided in an embodiment of this application. Detailed Implementation

[0043] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0044] Exemplary scenarios

[0045] FIG. 1 The diagram shown is an application scenario diagram of a soil moisture inversion model generation method provided in an embodiment of this application. FIG. 1The illustrated scene includes a server 110 and an image shooting device 120 connected in communication with the server 110. Specifically, the image shooting device 120 is configured to shoot remote sensing images corresponding to a soil region sample and send the remote sensing images to the server 110. The server 110 is configured to receive the remote sensing images corresponding to the soil region sample shot by the image shooting device 120, and then determine a plurality of remote sensing images corresponding to the soil region sample, wherein the plurality of remote sensing images correspond to a plurality of preset collection time points one by one; determine a plurality of soil humidity data sets corresponding to the soil region sample, wherein the plurality of soil humidity data sets correspond to the plurality of preset collection time points one by one; generate a humidity labeled image corresponding to each of the plurality of remote sensing images based on the plurality of soil humidity data sets and the plurality of remote sensing images; and train an initial network model based on the plurality of remote sensing images and the humidity labeled image corresponding to each of the plurality of remote sensing images, to generate a soil humidity inversion model.

[0046] FIG. 2 The illustrated is an application scene schematic diagram of a soil humidity inversion model generation method provided by another embodiment of the present application. FIG. 2 The illustrated scene includes a server 110 and an image shooting device 120, a humidity sensor 130 connected in communication with the server 110. Specifically, the image shooting device 120 is configured to shoot remote sensing images corresponding to a soil region sample and send the remote sensing images to the server 110. The humidity sensor 130 is configured to collect soil humidity data sets corresponding to the soil region sample and send the soil humidity data sets to the server 110. The server 110 is configured to receive the remote sensing images corresponding to the soil region sample shot by the image shooting device 120, receive the soil humidity data sets corresponding to the soil region sample collected by the humidity sensor 130, and then determine a plurality of remote sensing images corresponding to the soil region sample, wherein the plurality of remote sensing images correspond to a plurality of preset collection time points one by one; determine a plurality of soil humidity data sets corresponding to the soil region sample, wherein the plurality of soil humidity data sets correspond to the plurality of preset collection time points one by one; generate a humidity labeled image corresponding to each of the plurality of remote sensing images based on the plurality of soil humidity data sets and the plurality of remote sensing images; and train an initial network model based on the plurality of remote sensing images and the humidity labeled image corresponding to each of the plurality of remote sensing images, to generate a soil humidity inversion model.

[0047] It should be noted that the server 110 can be replaced by other types of processing devices, such as mobile phones, tablet computers or personal computers, etc. electronic devices with processing capabilities, but not limited to this. For example, in one device, it can have the function of executing the above soil humidity inversion model generation method, and also have the function of the image shooting device 120. The present application does not limit this.

[0048] The following is an example of the above soil humidity inversion model generation method executed in the server.

[0049] Exemplary methods

[0050] FIG. 3 Fig. 1 shows a flowchart of a method for generating a soil moisture inversion model according to an embodiment of the present application. As shown in Fig. 1, the method for generating a soil moisture inversion model according to an embodiment of the present application includes the following steps. FIG. 3

[0051] In step 310, a plurality of remote sensing images corresponding to the soil region sample are determined.

[0052] For example, the plurality of remote sensing images correspond to a plurality of preset collection time points one by one. It can be understood that each remote sensing image corresponds to a preset collection time point. The soil region sample can be one or more farmlands. For example, a cotton field. The plurality of remote sensing images can be high-definition remote sensing images taken by a drone. The preset collection time points can be 8 o'clock, 10 o'clock and 12 o'clock every day, or 2 o'clock, 5 o'clock and 7 o'clock every day, or other time points. The user can select according to actual needs, and the present application does not make specific limitations.

[0053] In step 320, a plurality of soil moisture data sets corresponding to the soil region sample are determined.

[0054] For example, the plurality of soil moisture data sets correspond to a plurality of preset collection time points one by one. It can be understood that each soil moisture data set corresponds to a preset collection time point. The plurality of soil moisture data sets can be humidity data detected by a humidity sensor. The humidity sensor can be a soil monitor. The humidity sensor can continuously collect soil moisture data sets of the soil region sample. Then the server can extract a plurality of soil moisture data sets corresponding to a plurality of preset collection time points one by one from the continuously collected soil moisture data sets of the soil region sample.

[0055] In step 330, a plurality of humidity labeled images corresponding to the plurality of remote sensing images are generated based on the plurality of soil moisture data sets and the plurality of remote sensing images.

[0056] ​Exemplarily, the plurality of soil humidity data sets and the plurality of remote sensing images are in one-to-one correspondence. The remote sensing image corresponding to a preset collection time point can be labeled according to the soil humidity data set corresponding to the preset collection time point. For example, the remote sensing image at 8:00 am on August 15, 2021 can be labeled according to the soil humidity data set at 8:00 am on August 15, 2021, so as to obtain a humidity labeled image. In this way, the time-synchronous soil humidity data and remote sensing images can be associated, and a plurality of humidity labeled images rich in time can be formed. The humidity labeled image can be represented by a matrix. In addition, the humidity sensor can be numbered. Since the humidity sensor has high-precision positioning information, the server can determine the positioning information of the position to be labeled according to the remote sensing image, then determine the humidity sensor number closest to the position to be labeled according to the positioning information, and finally extract a plurality of soil humidity data sets corresponding to a plurality of preset collection time points from the soil humidity data set according to the number and the preset collection time point.

[0057] In step 340, an initial network model is trained based on the plurality of remote sensing images and the humidity labeled images corresponding to the plurality of remote sensing images, so as to generate a soil humidity inversion model.

[0058] Exemplarily, the soil humidity inversion model is used to generate a humidity image corresponding to a to-be-measured soil region based on a remote sensing image corresponding to the to-be-measured soil region. That is, the remote sensing image corresponding to the to-be-measured soil region is input into the soil humidity inversion model, and the soil humidity inversion model can output the humidity image corresponding to the to-be-measured soil region. The humidity image can be represented by a data matrix. The initial network model can be a neural network model mainly composed of a semantic segmentation network and a regression network. The training of the initial network model can be regression training of the initial network model, and whether the initial network model is trained completely can be determined based on a result of a loss function. The loss function can include a mean square error (MSE), and can also include a least absolute error (LAE) and a least squared error (LSE), but is not limited to the MSE, the LAE, and the LSE. The person skilled in the art can select according to actual needs.

[0059] This application provides a method for generating a soil moisture inversion model. Based on multiple soil moisture datasets and multiple remote sensing images corresponding to multiple preset collection time points, it generates multiple moisture-labeled images, thus forming rich training samples corresponding to multiple preset collection time points. The rich training samples are used to train an initial network model to generate the soil moisture inversion model, improving the accuracy of the model. Using the soil moisture inversion model, moisture-labeled images can be directly output, directly displaying the soil moisture status, improving the user experience. Furthermore, the soil moisture inversion model can determine the moisture image corresponding to the soil area to be measured based on the remote sensing image corresponding to the soil area, eliminating the need for extensive soil sampling, reducing workload, and improving the efficiency of soil moisture detection.

[0060] FIG. 4 The diagram shown is a flowchart illustrating a method for generating a soil moisture inversion model according to another embodiment of this application. FIG. 3 This application extends from the embodiments shown. FIG. 4 The illustrated embodiment will be described in detail below. FIG. 4 The illustrated embodiments and FIG. 3 The differences between the embodiments shown are not repeated here, and the similarities are not repeated here.

[0061] like FIG. 4 As shown in the embodiments of this application, the steps for determining multiple soil moisture datasets corresponding to soil area samples include the following steps.

[0062] Step 410: Determine the subset of soil moisture data for each area to be tested at each preset collection time point.

[0063] For example, a soil region sample may include multiple test regions. A soil moisture dataset includes a subset of soil moisture data from each test region of the soil region sample at a preset collection time point. The soil moisture data subset of a test region includes moisture data at different detection depths at a preset sampling time point. Therefore, each test region can correspond to moisture data at different detection depths, thus forming multiple spatially rich moisture-annotated images. For example, each test region can be equipped with a soil monitoring instrument, and each soil monitoring instrument has five detection depths. The five detection depths can be 0cm, -10cm, -20cm, -30cm, and -40cm. A detection depth of 0cm represents the soil surface. A detection depth of -10cm represents 10cm below the soil surface, and so on. The multiple test regions can be understood as different sub-regions of the same plot, or as regions of different plots or even different latitudes and longitudes, thus obtaining spatially richer sample data.

[0064] In an embodiment, when the plurality of to-be-tested regions are different sub-regions of the same plot, the plot can be divided into the plurality of to-be-tested regions in the following manner:

[0065] The first manner: a circular region formed with the soil monitor as the center and the effective monitoring distance of the soil monitor as the radius can be taken as the to-be-tested region. In this way, when a plurality of soil monitors are provided, a plurality of to-be-tested regions can be formed, but the present application is not limited thereto. In short, the to-be-tested regions can be obtained by dividing the plot based on the position information of the soil monitor. In this way, the soil moisture data subsets corresponding to each to-be-tested region can be obtained based on the correspondence between the to-be-tested region and the soil monitor, and these subsets can constitute the soil moisture data set corresponding to the soil region sample.

[0066] The second manner: the plurality of to-be-tested regions can be obtained based on the environmental information of the soil region, instead of the position information of the soil monitor. The environmental information can include at least one of the following, but is not limited thereto: soil type and soil surface condition. The soil type can include, but is not limited to, sandy soil, clay soil and loam soil; the soil surface condition can include, but is not limited to, a first condition indicating that the soil surface is paved with a drip irrigation belt, a second condition indicating that the soil surface is paved with a mulch film, and a third condition indicating that the soil surface is not paved with a mulch film and a drip irrigation belt. For the third condition, it can be understood that the region corresponding to the soil surface in this condition is used for planting plants. In this way, the plot can be divided into a plurality of to-be-tested regions according to the purpose of the region corresponding to the soil surface in the plot or the soil type, and the soil type and / or the indicated soil surface condition included in the same to-be-tested region are the same. Accordingly, to obtain the soil moisture data subset of the to-be-tested region, the corresponding soil moisture data subset can also be obtained from the soil moisture data set based on the position information of the to-be-tested region and the position information of the soil monitor. By setting a plurality of preset collection time points and making each preset collection time point correspond to one soil moisture data set, the correspondence between the preset collection time points and the soil moisture data set can be determined, the soil moisture data set is enriched in time, and thus an accurate and rich data basis is provided for generating a plurality of humidity labeled images. In addition, by using the plurality of remote sensing images corresponding to the soil region sample including the plurality of to-be-tested regions and the plurality of soil moisture data sets, and making the soil moisture data subset of the to-be-tested region include humidity data of different detection depths, the soil moisture data set is further enriched in space, and thus a more rich data basis is provided for generating a plurality of humidity labeled images. In this way, the accuracy and robustness of the soil moisture inversion model trained based on the humidity labeled images enriched in time and space are further improved.

[0067] FIG. 5 The soil moisture inversion model generation method provided by another embodiment of the present application is shown in the flowchart. Based on the embodiment of the present application FIG. 3 The embodiment of the present applicationFIG. 5 In the embodiments shown, the following is described in detail FIG. 5 The embodiments shown are different from FIG. 3 The same parts of the embodiments shown will not be described again.

[0068] As FIG. 5 As shown in the embodiments of the present application, the step of generating the humidity labeled image corresponding to each of the plurality of remote sensing images based on the plurality of soil humidity data sets and the plurality of remote sensing images includes the following steps.

[0069] Step 510, based on the preset acquisition time point corresponding to each of the plurality of remote sensing images, determine the soil humidity data set corresponding to each of the plurality of remote sensing images.

[0070] Exemplarily, each remote sensing image corresponds to a preset acquisition time point. According to the one preset acquisition time point, the soil humidity data set at the one preset acquisition time point is determined.

[0071] Step 520, based on the plurality of remote sensing images and the soil humidity data set corresponding to each of the plurality of remote sensing images, generate the humidity labeled image corresponding to each of the plurality of remote sensing images.

[0072] In actual application, according to the soil humidity data set corresponding to each of the plurality of remote sensing images, the plurality of remote sensing images are labeled respectively, thereby generating the humidity labeled image corresponding to each of the plurality of remote sensing images.

[0073] By determining the soil humidity data set corresponding to each of the plurality of remote sensing images based on the preset acquisition time point corresponding to each of the plurality of remote sensing images, and labeling the plurality of remote sensing images respectively according to the soil humidity data set corresponding to each of the plurality of remote sensing images, the humidity labeled image corresponding to each of the plurality of remote sensing images is generated, so that the generated humidity labeled image fuses the remote sensing image and the soil humidity data set at the same preset acquisition time point, ensuring the accuracy of the humidity labeled image, thereby providing an accurate data basis for training the initial network model.

[0074] FIG. 6 The flowchart shown is a soil humidity inversion model generation method provided by another embodiment of the present application. In the present application FIG. 5 The embodiments shown extend the embodiments of the present application FIG. 6 In the embodiments shown, the following is described in detail FIG. 6 The embodiments shown are different from FIG. 5 The same parts of the embodiments shown will not be described again.

[0075] As FIG. 6 As shown in the embodiments of the present application, the step of generating the humidity labeled image corresponding to each of the plurality of remote sensing images based on the plurality of soil humidity data sets and the plurality of remote sensing images includes the following steps.

[0076] In step 610, for each of the plurality of remote sensing images, based on the pixel information of the remote sensing image or the position information of the plurality of to-be-tested regions contained in the soil region sample corresponding to the remote sensing image, the remote sensing image is labeled to form a region-labeled image containing a plurality of labeled regions.

[0077] For example, since the soil region sample can include a plurality of to-be-tested regions, which can be different sub-regions of the same plot, or regions of different plots or even different latitude and longitude areas, when the plurality of to-be-tested regions are different sub-regions of the same plot, the soil detectors corresponding to the plurality of to-be-tested regions or the soil types and / or soil surface conditions characterized by the soil detectors are different. Therefore, in this case, in order to accurately generate the humidity-labeled image corresponding to each remote sensing image, the region detection can be performed on each remote sensing image to form a region-labeled image containing a plurality of labeled regions.

[0078] On the one hand, in the case of labeling the remote sensing image to form a region-labeled image containing a plurality of labeled regions based on the pixel information of the remote sensing image, as an implementation manner, the remote sensing image can be directly clustered based on the pixel information of the remote sensing image, such as pixel value difference, to generate a region-labeled image containing a plurality of labeled regions. Through clustering, the regions with different soil types and / or different soil surface conditions contained in the plot corresponding to the remote sensing image can be divided to form a plurality of labeled regions. As another implementation manner, the plurality of labeled regions can be formed based on the region division operation input by the user. For example, the remote sensing image can be displayed to the user, and the user can input a division operation under the condition of distinguishing the regions with different soil types and / or different soil surface conditions, and then the regions with different soil types and / or different soil surface conditions can be divided based on the division operation to form a plurality of labeled regions.

[0079] On the other hand, in the case of labeling the remote sensing image to form a region-labeled image containing a plurality of labeled regions based on the position information of the plurality of to-be-tested regions contained in the soil region sample corresponding to the remote sensing image, the labeled region at the corresponding position can be directly determined from the remote sensing image based on the position information of each to-be-tested region, and thus the region-labeled image containing a plurality of labeled regions can be obtained.

[0080] Exemplarily, the region-labeled image can include a label mask and a remote sensing image arranged in a stack. The height dimension and the width dimension of the label mask and the remote sensing image are the same. The unlabeled region can be uniformly set to 0 by the label mask, or can be uniformly set to -1, or can be uniformly set to other numerical values, and the designer can select according to actual needs, which is not limited in the present application. The unlabeled region in the region-labeled image can also be set to a uniform gray value, for example, the gray value can be set to 0.

[0081] In step 620, a humidity-labeled image corresponding to the remote sensing image is generated based on the region-labeled image and the soil humidity data set corresponding to the remote sensing image.

[0082] Exemplarily, in one aspect, in the case of labeling the remote sensing image to form a region-labeled image containing multiple labeled regions based on the pixel information of the remote sensing image, the soil humidity data corresponding to each labeled region can be determined based on the position information of the multiple labeled regions and the multiple to-be-measured regions. For example, for each labeled region, if the labeled region is within a to-be-measured region or the overlapping area of the labeled region and a to-be-measured region is greater than or equal to a preset threshold, the soil humidity data corresponding to the to-be-measured region can be taken as the humidity data of the labeled region; if the labeled region overlaps with at least two to-be-measured regions or overlaps with only one to-be-measured region and the overlapping area is less than the preset threshold, the humidity data corresponding to the overlapping to-be-measured regions can be corrected, and the corrected humidity data is taken as the humidity data of the to-be-measured region. The correction method can be the product of the humidity data and a set coefficient, or the average or weighted value of the humidity data; and the set coefficient can be set as needed or determined according to the overlapping area size. In another aspect, in the case of labeling the remote sensing image to form a region-labeled image containing multiple labeled regions based on the position information of the multiple to-be-measured regions contained in the soil region sample corresponding to the remote sensing image, the multiple labeled regions correspond one-to-one to the multiple to-be-measured regions. Therefore, the multiple labeled regions can be labeled according to the soil humidity data set of the multiple to-be-measured regions, so as to obtain the soil humidity data of each labeled region, i.e., the soil humidity data of the multiple labeled regions included in each region-labeled image.

[0083] By labeling the multiple to-be-measured regions to obtain a region-labeled image corresponding to the remote sensing image, and then generating a humidity-labeled image corresponding to the remote sensing image based on the region-labeled image and the soil humidity data set corresponding to the remote sensing image, the remote sensing image is labeled in zones, so that a more accurate humidity-labeled image is obtained, and a more accurate data basis is provided for training the initial network model.

[0084] FIG. 7 The figure is a flowchart of a soil humidity inversion model generation method provided by another embodiment of the present application. In the present applicationFIG. 6 The application extends from the basis of the embodiments shown FIG. 7 In the embodiments shown, the following is focused on FIG. 7 In the embodiments shown, the differences and the same are not repeated. FIG. 6

[0085] As shown FIG. 7 In the embodiments of the present application, based on the soil moisture data set corresponding to the region annotation image and the remote sensing image, the step of generating the humidity annotation image corresponding to the remote sensing image includes the following steps.

[0086] Step 710, based on the multiple annotation regions included in the region annotation image, the soil moisture data set corresponding to the remote sensing image is used to generate the humidity annotation data corresponding to each of the multiple annotation regions.

[0087] Illustratively, by setting the humidity sensor in multiple areas to be measured, the soil moisture data set corresponding to the remote sensing image can be obtained according to the values measured by the humidity sensor. That is, the soil moisture data set can be the values measured directly by the humidity sensor. By processing the soil moisture data set corresponding to the remote sensing image, the processed data can be obtained, and then the processed data is annotated in multiple annotation regions, so as to obtain the humidity annotation data corresponding to each of the multiple annotation regions. The values measured directly by the humidity sensor can be multiple values of different depths at the same position point, so the average, maximum or minimum of the multiple values of different depths can be obtained to obtain the humidity data of the position point. That is, the processing of the soil moisture data set corresponding to the remote sensing image can be the processing of the multiple values of different depths.

[0088] Step 720, based on the region annotation image and the humidity annotation data corresponding to each of the multiple annotation regions, the humidity annotation image corresponding to the remote sensing image is generated.

[0089] Illustratively, a region annotation image includes multiple annotation regions, and each region annotation image includes multiple annotation regions, and each region annotation image includes multiple annotation regions. According to the humidity annotation data corresponding to each of the multiple annotation regions, the multiple annotation regions included in each region annotation image are annotated, so as to obtain the humidity annotation image corresponding to each remote sensing image.

[0090] By generating the humidity annotation data corresponding to each of the multiple annotation regions based on the multiple annotation regions included in the region annotation image and using the soil moisture data set corresponding to the remote sensing image, the soil moisture data set corresponding to the remote sensing image can be processed and annotated, so as to enable the user to select the processing method of the soil moisture data set according to the actual demand, thereby improving the user experience.

[0091] FIG. 8 The soil moisture inversion model generation method provided by another embodiment of the present application is shown in the flowchart. In the present application FIG. 7 ​This application extends from the embodiments shown. FIG. 8 The illustrated embodiment will be described in detail below. FIG. 8 The illustrated embodiments and FIG. 7 The differences between the embodiments shown are not repeated here, and the similarities are not repeated here.

[0092] like FIG. 8 As shown in the embodiments of this application, the step of generating humidity labeling data corresponding to each of the multiple labeled regions based on the multiple labeled regions included in the region-labeled image and using the soil moisture dataset corresponding to the remote sensing image includes the following steps.

[0093] Step 810: For each of the multiple labeled regions included in the region labeled image, determine the relative positional relationship between the test area to which the labeled region belongs and the humidity sensor corresponding to the test area to which the labeled region belongs.

[0094] For example, there are two possible relative positions between the area to be measured and the humidity sensor: one is that the humidity sensor is located outside the area to be measured, and the other is that the humidity sensor is located within the area to be measured.

[0095] Step 820: Based on the relative positional relationship and the soil moisture dataset corresponding to the remote sensing image, determine the moisture labeling data corresponding to the labeled area.

[0096] For example, when the humidity sensor is located outside the area to be measured, the humidity labeling data corresponding to the area to be measured can be determined by the humidity sensor reading and the soil type coefficient. That is, the humidity labeling data corresponding to the area to be measured can be equal to the product of the humidity sensor reading and the soil type coefficient. The soil type coefficient can be an empirical coefficient determined based on the soil type of the area to be measured. Conversely, when the humidity sensor is located within the area to be measured, the humidity labeling data corresponding to the area to be measured can be determined by the humidity sensor reading.

[0097] When a humidity sensor is within the measurement area, the humidity it detects can be interpreted as the humidity of that area. However, when the sensor is outside the measurement area, the detected humidity may not be the same as the humidity in the measurement area, as they may be located in different humidity zones. Therefore, analyzing the relative positional relationship between the measurement area and the humidity sensor to determine the calculation method for humidity labeling data, and considering the differences in humidity detection by the sensor outside the measurement area, helps improve the accuracy of humidity labeling data.

[0098] FIG. 9 The diagram shown is a schematic flowchart of a soil moisture inversion method provided in an embodiment of this application. FIG. 9 As shown in the embodiments of this application, the soil moisture inversion method includes the following steps.

[0099] At step 910, a soil moisture inversion model is determined.

[0100] Specifically, the soil moisture inversion model is generated based on the soil moisture inversion model generation method in the above embodiment.

[0101] At step 920, a soil moisture inversion model is used to generate a humidity image corresponding to a to-be-detected soil region based on a remote sensing image corresponding to the to-be-detected soil region.

[0102] Illustratively, the humidity labeled image can label different humidities as different colors, so that the user can intuitively see the humidity condition of the to-be-detected soil region.

[0103] The soil moisture inversion method provided by the embodiment of the present application generates a plurality of humidity labeled images based on a plurality of soil moisture data sets corresponding to a plurality of preset collection time points and a plurality of remote sensing images, thereby forming rich training samples corresponding to the plurality of preset collection time points. The initial network model is trained using the rich training samples to generate a soil moisture inversion model, thereby improving the accuracy of the soil moisture inversion model.

[0104] The soil moisture inversion method can directly output the humidity labeled image using the soil moisture inversion model, that is, directly display the soil moisture condition, thereby improving the user experience. Moreover, the soil moisture inversion method can determine the humidity image corresponding to the to-be-detected soil region based on the remote sensing image corresponding to the to-be-detected soil region using the soil moisture inversion model, without the need for a large amount of soil sampling, thereby reducing the workload and improving the efficiency of soil moisture detection. The soil moisture inversion method does not need soil sampling and can directly output the humidity labeled image of a complex farmland scene, which is not limited by complex and diverse farmland scenes, thereby improving the user experience.

[0105] In an embodiment of the present application, the soil moisture inversion model can determine the humidity image corresponding to the to-be-detected soil region based on the remote sensing image corresponding to the to-be-detected soil region. The humidity labeled image can be used to guide irrigation operations or determine irrigation failure points based on the humidity condition of the to-be-detected soil region. For example, in the scenario of guiding irrigation operations, for a to-be-detected soil region with high humidity, the irrigation water amount can be reduced or irrigation can be not performed, and for a to-be-detected soil region with low humidity, irrigation can be performed or the irrigation water amount can be increased. In the scenario of determining irrigation failure points, a to-be-detected soil region with continuously high humidity or a to-be-detected soil region with continuously low humidity can be caused by a drip irrigation belt leakage or clogging, so that the irrigation failure point can be determined. After the irrigation failure point is determined, the failure point region or the failed drip irrigation belt can be displayed on a map, so that the user can intuitively understand and conveniently perform on-site maintenance.

[0106] In an embodiment of the present application, the bad point area or the damaged drip irrigation belt can be displayed by using some markers. For example, the bad point area or the damaged drip irrigation belt can be marked by using colored lines. In addition, a walking path can be automatically generated based on the location of the user and the location of the bad point, so as to provide navigation for the user and improve the efficiency of the user to the bad point.

[0107] In an embodiment of the present application, the soil area with high humidity or the soil area with low humidity can be determined from the output humidity annotation image. For example, the greater the value of the humidity data displayed on the humidity annotation image, the greater the soil humidity, and vice versa. In actual application, one or more humidity thresholds can be set to determine normal humidity and abnormal humidity. For example, a first humidity threshold and a second humidity threshold can be set. If the humidity data is greater than the first humidity threshold and less than the second humidity threshold, it can be determined as normal humidity. If the humidity data is less than the first humidity threshold or greater than the second humidity threshold, it can be determined as abnormal humidity. The specific humidity threshold can be set according to actual demand or experiment, which is not limited in the present application.

[0108] FIG. 10 Fig. 1 shows a flowchart of an irrigation method provided by an embodiment of the present application. As shown in Fig. 1, the irrigation method provided by the embodiment of the present application includes the following steps. FIG. 10

[0109] Step 1010, obtaining the soil humidity of the farmland.

[0110] Specifically, the soil humidity of the farmland can be obtained by using the soil humidity inversion method in the above embodiment.

[0111] Step 1020, determining different humidity areas in the farmland according to the obtained soil humidity of the farmland.

[0112] Specifically, according to the obtained soil humidity of the farmland, the humidity distribution in the farmland can be determined, and different humidity areas can be determined (for example, demarcated) according to the humidity distribution. For example, the same humidity areas with the same humidity or the humidity within a set range can be divided according to the humidity distribution.

[0113] Step 1030, determining an irrigation scheme according to the humidity areas.

[0114] Specifically, a target humidity value can be determined in advance. If the humidity of the humidity area is lower than the target humidity value, the irrigation water amount can be increased in the next irrigation. If the humidity of the humidity area is higher than the target humidity value, the irrigation water amount can be reduced or no irrigation can be performed in the next irrigation.

[0115] ​In an embodiment of the present application, the target humidity value can be set as the target humidity value at or before the next irrigation (e.g., the day before or several hours before). The target humidity value can be set according to actual needs, which is not limited in the present application. The same or different target humidity values can be set for different areas.

[0116] FIG. 11 Fig. 1 shows a flowchart of an irrigation anomaly determination method provided by an embodiment of the present application. As shown in Fig. 1, the irrigation anomaly determination method provided by the embodiment of the present application includes the following steps. FIG. 11

[0117] Step 1110, obtaining the soil humidity of the farmland.

[0118] Specifically, the soil humidity of the farmland can be obtained by using the soil humidity inversion method in the above embodiments.

[0119] Step 1120, determining different humidity areas in the farmland according to the obtained soil humidity of the farmland.

[0120] Specifically, according to the obtained soil humidity of the farmland, the humidity distribution in the farmland can be determined, and different humidity areas can be determined (e.g., demarcated) according to the humidity distribution. For example, the same humidity areas can be divided according to the humidity distribution, which have the same humidity or the humidity within a set range.

[0121] Step 1130, comparing and analyzing the humidity of different humidity areas to determine an abnormal humidity area.

[0122] In an embodiment of the present application, the humidity of different humidity areas can be compared, and the humidity area with the maximum or minimum humidity can be determined as the abnormal humidity area. In another embodiment of the present application, the average value of the humidity of different humidity areas can be calculated, and the humidity of each humidity area can be compared with the average value. If the difference between the humidity and the average value exceeds a set threshold, it can be determined that the humidity area corresponding to the humidity is an abnormal humidity area. In another embodiment of the present application, a target humidity value can be set. The target humidity value can be set as the target humidity value at or before the next irrigation (e.g., the day before or several hours before). The humidity of the humidity area can be compared with the target humidity value, and if the difference between the humidity and the target humidity value exceeds a set threshold, it can be determined that the humidity area corresponding to the humidity is an abnormal humidity area.

[0123] Step 1140, determining an irrigation anomaly point according to the abnormal humidity area.

[0124] ​Specifically, for areas that are consistently humid or consistently dry, the problem might be leaks or blockages in the irrigation system's irrigation belts. Irrigation anomalies (damaged points) within the abnormal humidity area can be identified. In one embodiment of this application, the location of each irrigation point in the irrigation system can be known. After identifying the abnormal humidity area, irrigation points associated with it (e.g., located within or adjacent to the abnormal humidity area) can be identified. These irrigation points can be considered as irrigation anomalies. In another embodiment of this application, each irrigation point in the irrigation system can be equipped with a positioning device, such as a GPS (Global Positioning System) positioning device or a BeiDou positioning device. The positioning device can transmit the location information of the irrigation point.

[0125] In one embodiment of this application, after identifying the irrigation anomaly point, the abnormal area or broken irrigation strip can be displayed on a map in a graphical form, so that the user can intuitively determine the location and facilitate on-site repair.

[0126] In one embodiment of this application, markers can be used to distinguish and display irrigation anomalies.

[0127] In one embodiment of this application, positioning technology can also be incorporated. Specifically, after identifying the irrigation anomaly point, the user's location can be obtained. For example, the user can carry a positioning device, such as a GPS positioning device or a mobile terminal with positioning capabilities (e.g., a mobile phone or remote control). The user's mobile terminal can receive the location information of the abnormal irrigation point, and the positioning function of the mobile terminal or the positioning device can obtain the user's current location. The mobile terminal can generate a travel path based on the location of the irrigation anomaly point and the user's current location, and provide navigation functionality to the user based on the travel path.

[0128] The above text combined FIG. 3 to FIG. 11 The method embodiments of this application are described in detail below, in conjunction with... FIG. 12 to FIG. 20 The present application provides a detailed description of the apparatus embodiments. It should be understood that the descriptions of the method embodiments correspond to the descriptions of the apparatus embodiments; therefore, any parts not described in detail can be found in the foregoing method embodiments.

[0129] Exemplary apparatuses

[0130] FIG. 12 The diagram shown is a schematic representation of a soil moisture inversion model generation device provided in an embodiment of this application. FIG. 12 As shown, the soil moisture inversion model generation device 1200 includes a first confirmation module 1210, a second confirmation module 1220, a generation module 1230, and a model training module 1240.

[0131] Specifically, the first confirmation module 1210 is configured to determine a plurality of remote sensing images corresponding to the soil region sample, wherein the plurality of remote sensing images correspond to a plurality of preset collection time points one by one. The second confirmation module 1220 is configured to determine a plurality of soil humidity data sets corresponding to the soil region sample, wherein the plurality of soil humidity data sets correspond to the plurality of preset collection time points one by one. The generation module 1230 is configured to generate a humidity labeled image corresponding to each of the plurality of remote sensing images based on the plurality of soil humidity data sets and the plurality of remote sensing images. The model training module 1240 is configured to train an initial network model based on the plurality of remote sensing images and the humidity labeled images corresponding to the plurality of remote sensing images to generate a soil humidity inversion model, wherein the soil humidity inversion model is used to generate a humidity image corresponding to a to-be-measured soil region based on a remote sensing image corresponding to the to-be-measured soil region.

[0132] FIG. 13 Fig. 2 shows a structural schematic diagram of the second determination module according to an embodiment of the present application. FIG. 12 Based on the embodiment shown in Fig. 2, the second determination module according to another embodiment of the present application is extended. FIG. 13 Based on the embodiment shown in Fig. 2, the second determination module according to another embodiment of the present application is extended. FIG. 13 Based on the embodiment shown in Fig. 2, the second determination module according to another embodiment of the present application is extended. FIG. 12 Based on the embodiment shown in Fig. 2, the second determination module according to another embodiment of the present application is extended.

[0133] As shown in Fig. 3, in the embodiment of the present application, the second confirmation module 1220 includes a humidity set determination unit 1221. FIG. 13 Specifically, the humidity set determination unit 1221 is configured to determine a soil humidity data set of each to-be-measured region at each preset collection time point, wherein the soil humidity data set of the to-be-measured region includes humidity data of different detection depths.

[0134]

[0135] Fig. 4 shows a structural schematic diagram of the generation module according to an embodiment of the present application. FIG. 14 Based on the embodiment shown in Fig. 4, the generation module according to another embodiment of the present application is extended. FIG. 12 Based on the embodiment shown in Fig. 4, the generation module according to another embodiment of the present application is extended. FIG. 14 Based on the embodiment shown in Fig. 4, the generation module according to another embodiment of the present application is extended. FIG. 14 Based on the embodiment shown in Fig. 4, the generation module according to another embodiment of the present application is extended. FIG. 12 Based on the embodiment shown in Fig. 4, the generation module according to another embodiment of the present application is extended.

[0136] As shown in Fig. 5, in the embodiment of the present application, the generation module 1230 includes an image data determination unit 1231 and a labeled image generation unit 1232. FIG. 14

[0137] ​Specifically, the image data determination unit 1231 is configured to determine the soil moisture dataset corresponding to each of the multiple remote sensing images based on the preset acquisition time points corresponding to each of the multiple remote sensing images. The annotation image generation unit 1232 is configured to generate moisture annotation images corresponding to each of the multiple remote sensing images based on the multiple remote sensing images and their corresponding soil moisture datasets.

[0138] FIG. 15 The diagram shown is a schematic representation of the structure of an annotated image generation unit provided in an embodiment of this application. FIG. 14 Extending from the illustrated embodiment FIG. 15 The illustrated embodiment will be described in detail below. FIG. 15 The illustrated embodiments and FIG. 14 The differences between the embodiments shown are not repeated here, and the similarities are not repeated here.

[0139] like FIG. 15 As shown in the embodiment of this application, the annotation image generation unit 1232 includes an annotation region determination subunit 1510 and an annotation image generation subunit 1520.

[0140] Specifically, the annotation region determination subunit 1510 is configured to annotate multiple test regions for each of the multiple remote sensing images, obtaining an annotation image of the region corresponding to the remote sensing image. The annotation image includes multiple annotated regions, and each annotated region corresponds one-to-one with a test region. The annotation image generation subunit 1520 is configured to generate a humidity annotation image corresponding to the remote sensing image based on the annotation image of the region and the soil moisture dataset corresponding to the remote sensing image.

[0141] FIG. 16 The diagram shown is a structural schematic of an annotation image generation subunit provided in an embodiment of this application. FIG. 15 Extending from the illustrated embodiment FIG. 16 The illustrated embodiment will be described in detail below. FIG. 16 The illustrated embodiments and FIG. 15 The differences between the embodiments shown are not repeated here, and the similarities are not repeated here.

[0142] like FIG. 16 As shown in the embodiment of this application, the labeled image generation subunit 1520 includes a labeled data determination subunit 1521 and a humidity labeled image generation subunit 1522.

[0143] Specifically, the annotation data determination subunit 1521 is configured to generate, based on the multiple annotation regions included in the region annotation image, the multiple humidity annotation data corresponding to the multiple annotation regions respectively, by using the soil moisture data set corresponding to the remote sensing image. The humidity annotation image generation subunit 1522 is configured to generate, based on the region annotation image and the multiple humidity annotation data corresponding to the multiple annotation regions respectively, the humidity annotation image corresponding to the remote sensing image.

[0144] FIG. 17 Fig. 1 shows a structural schematic diagram of an annotation data determination subunit provided by an embodiment of the present application. In the embodiment shown, FIG. 16 Based on the embodiment shown, FIG. 17 In the embodiment shown, the differences between the embodiment shown and FIG. 17 In the embodiment shown, the differences between the embodiment shown and FIG. 16 The same parts of the embodiment shown will not be described again.

[0145] As shown in the embodiment, FIG. 17 In the embodiment of the present application, the annotation data determination subunit 1521 includes a relative position determination subunit 1710 and a humidity annotation data determination subunit 1720.

[0146] Specifically, the relative position determination subunit 1710 is configured to determine, for each of the multiple annotation regions included in the region annotation image, a relative position relationship between the humidity sensor corresponding to the to-be-measured region to which the annotation region belongs. The humidity annotation data determination subunit 1720 is configured to determine, based on the relative position relationship and the soil moisture data set corresponding to the remote sensing image, the humidity annotation data corresponding to the annotation region.

[0147] FIG. 18 Fig. 2 shows a structural schematic diagram of a soil moisture inversion device provided by an embodiment of the present application. As shown in the embodiment, FIG. 18 The soil moisture inversion device 1800 includes a model determination module 1810 and a soil moisture analysis module 1820.

[0148] Specifically, the model determination module 1810 is configured to determine a soil moisture inversion model, wherein the soil moisture inversion model is obtained based on the soil moisture inversion model generation method of the above-mentioned embodiment. The soil moisture analysis module 1820 is configured to generate, by using the soil moisture inversion model, a humidity image corresponding to a to-be-measured soil region based on a remote sensing image corresponding to the to-be-measured soil region.

[0149] FIG. 19 Fig. 3 shows a structural schematic diagram of an irrigation device provided by an embodiment of the present application. As shown in the embodiment, FIG. 19 The irrigation device 1900 includes a first farmland humidity acquisition module 1910, a first humidity region confirmation module 1920, and an irrigation scheme determination module 1930.

[0150] Specifically, the first farmland humidity obtaining module 1910 is configured to obtain farmland soil humidity, wherein the farmland soil humidity is obtained based on the soil humidity inversion method of claim 7. The first humidity region confirming module 1920 is configured to determine different humidity regions in the farmland according to the obtained farmland soil humidity. The irrigation scheme determining module 1930 is configured to determine an irrigation scheme according to the humidity regions.

[0151] FIG. 20 Fig. 2 shows a structural schematic diagram of an irrigation anomaly determining apparatus provided by an embodiment of the present application. As shown in FIG. 20 The irrigation anomaly determining apparatus 2000 includes a second farmland humidity obtaining module 2010, a second humidity region confirming module 2020, an anomaly analyzing module 2030 and an anomaly point confirming module 2040.

[0152] Specifically, the second farmland humidity obtaining module 2010 is configured to obtain farmland soil humidity, wherein the farmland soil humidity is obtained based on the soil humidity inversion method of claim 7. The second humidity region confirming module 2020 is configured to determine different humidity regions in the farmland according to the obtained farmland soil humidity. The anomaly analyzing module 2030 is configured to compare and analyze the humidity of different humidity regions to determine an abnormal humidity region. The anomaly point confirming module 2040 is configured to determine an irrigation anomaly point according to the abnormal humidity region.

[0153] Exemplary electronic devices

[0154] Next, an electronic device according to an embodiment of the present application will be described with reference to FIG. 21 Fig. 2. FIG. 21 Fig. 2 shows a structural schematic diagram of an electronic device provided by an embodiment of the present application.

[0155] As shown in FIG. 21 The electronic device 2100 includes one or more processors 2101 and a memory 2102, and computer program instructions stored in the memory 2102, which, when executed by the processor 2101, cause the processor 2101 to perform the soil humidity inversion model generation method and / or the soil humidity inversion method and / or the irrigation method and / or the irrigation anomaly determining method of any of the above embodiments.

[0156] The processor 2101 can be a central processing unit (CPU) or other forms of processing unit having data processing capability and / or instruction execution capability, and can control other components in the electronic device to perform desired functions.

[0157] The memory 2102 can include one or more computer program products that can include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory, for example, can include random access memory (RAM), and / or a cache, etc. The non-volatile memory, for example, can include read-only memory (ROM), hard disks, flash memory, etc. One or more computer program instructions can be stored on the computer-readable storage media, and the processor 2101 can run the program instructions to implement the steps in the soil moisture inversion model generation method and / or the soil moisture inversion method and / or the irrigation method and / or the irrigation anomaly determination method of various embodiments of the present application described above and / or other desired functions.

[0158] In one example, the electronic device 2100 can further include an input device 2103 and an output device 2104, which are interconnected through a bus system and / or other forms of connection mechanisms (not shown in the figure). FIG. 21

[0159] In addition, the input device 2103 can also include, for example, a keyboard, a mouse, a microphone, etc.

[0160] The output device 2104 can output various information to the outside. The output device 2104 can include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, etc.

[0161] Of course, for the sake of simplicity, FIG. 21 In the figure, only some of the components in the electronic device 2100 related to the present application are shown, and components such as buses, input / output interfaces, etc. are omitted. In addition, according to specific application circumstances, the electronic device 2100 can also include any other appropriate components.

[0162] In addition to the above methods and devices, the embodiments of the present application can also be computer program products including computer program instructions, which, when run by a processor, cause the processor to perform the steps in the soil moisture inversion model generation method and / or the soil moisture inversion method and / or the irrigation method and / or the irrigation anomaly determination method of any one of the above embodiments.

[0163] ​The computer program product can be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C++ or the like, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code can execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device or entirely on the remote computing device or server.

[0164] Furthermore, embodiments of the present application can also be a computer readable storage medium, having stored thereon computer program instructions which, when executed by a processor, cause the processor to perform the steps described in the above "Exemplary Methods" section of the present specification for generating a soil moisture inversion model and / or for performing a soil moisture inversion method and / or for performing an irrigation method and / or for determining an irrigation anomaly according to various embodiments of the present application.

[0165] The computer readable storage medium can be a combination of one or more computer readable media. The computer readable media can be a computer readable signal medium, or a computer readable storage medium, or a combination of computer readable signal medium and computer readable storage medium.

[0166] The basic principles of the present application are described above in conjunction with specific embodiments, however, it should be noted that the advantages, benefits, effects and the like mentioned in the present application are only examples and are not limiting, and these advantages, benefits, effects and the like cannot be considered as necessary for each embodiment of the present application. In addition, the above-mentioned specific details are only for the purpose of example and understanding, and are not limiting, and the above-mentioned details do not limit the present application to be necessarily implemented with the above-mentioned specific details.

[0167] In addition, the present application also provides a movable platform, which comprises:

[0168] An image capturing device for acquiring a remote sensing image of the land to be measured;

[0169] The processor is configured to generate a soil moisture inversion model based on the remote sensing image by using the soil moisture inversion model method described in the above embodiments; and / or determine the soil moisture of the to-be-tested land plot based on the remote sensing image by using the soil moisture inversion method described in the above embodiments; and / or determine the irrigation scheme of the to-be-tested land plot based on the remote sensing image by using the irrigation method described in the above embodiments; and / or determine the irrigation abnormal point of the to-be-tested land plot based on the remote sensing image by using the irrigation abnormality determination method described in the above embodiments.

[0170] In the above embodiments, the movable platform can be an unmanned aerial vehicle, or can be a movable ground device, such as an unmanned vehicle.

[0171] The block diagrams of the devices, apparatuses, equipment, and systems involved in the present application are only illustrative examples and are not intended to require or imply that the connections, arrangements, configurations must be as shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "include", "contain", "have", and the like are open-ended words that mean "including but not limited to", and can be used interchangeably. The words "or" and "and" used herein mean the word "and / or", and can be used interchangeably unless the context clearly indicates otherwise. The word "such as" used herein means the phrase "such as but not limited to", and can be used interchangeably.

[0172] It should also be noted that in the devices, apparatuses, and methods of the present application, each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be considered as equivalent solutions of the present application.

[0173] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein can be applied to other aspects without departing from the scope of the present application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0174] The above description has been given for the purpose of illustration and description. Furthermore, this description does not intend to limit the embodiments of the present application to the forms disclosed herein. Although a number of example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions, and sub-combinations thereof.

[0175] The above description is only the preferred embodiments of the present application and is not intended to limit the present application. Any modification, equivalent replacement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A soil moisture inversion model generation method characterized by comprising: The method comprises the following steps: determining a plurality of remote sensing images corresponding to a soil region sample, wherein the plurality of remote sensing images correspond to a plurality of preset collection time points one by one, the soil region sample comprises a plurality of to-be-measured regions, any to-be-measured region of the plurality of to-be-measured regions corresponds to humidity data of different detection depths, the plurality of to-be-measured regions comprise different sub-regions of a same land plot or regions of different land plots, and the determination of the plurality of to-be-measured regions comprises division based on environmental information of the soil region, wherein the environmental information comprises a soil type and / or a soil surface condition; determining a plurality of soil humidity data sets corresponding to the soil region sample, wherein the plurality of soil humidity data sets correspond to the plurality of preset collection time points one by one; generating a humidity labeled image corresponding to each of the plurality of remote sensing images based on the plurality of soil humidity data sets and the plurality of remote sensing images; training an initial network model based on the plurality of remote sensing images and the humidity labeled image corresponding to each of the plurality of remote sensing images to generate a soil humidity inversion model, wherein the initial network model is a neural network model; wherein the determination of the plurality of soil humidity data sets corresponding to the soil region sample comprises: determining a soil humidity data subset of each to-be-measured region at each preset collection time point, wherein the soil humidity data subset of the to-be-measured region comprises humidity data of different detection depths.

2. The soil moisture inversion model generation method according to claim 1, characterized in that, The generation of the humidity labeled image corresponding to each of the plurality of remote sensing images based on the plurality of soil humidity data sets and the plurality of remote sensing images comprises: determining a soil humidity data set corresponding to each of the plurality of remote sensing images based on a preset collection time point corresponding to each of the plurality of remote sensing images; generating the humidity labeled image corresponding to each of the plurality of remote sensing images based on the plurality of remote sensing images and the soil humidity data set corresponding to each of the plurality of remote sensing images.

3. The soil moisture inversion model generation method according to claim 2, characterized in that, The generation of the humidity labeled image corresponding to each of the plurality of remote sensing images based on the plurality of remote sensing images and the soil humidity data set corresponding to each of the plurality of remote sensing images comprises: for each of the plurality of remote sensing images, labeling the remote sensing image to form a region labeled image comprising a plurality of labeled regions based on pixel information of the remote sensing image or position information of a plurality of to-be-measured regions included in a soil region sample corresponding to the remote sensing image; generating the humidity labeled image corresponding to the remote sensing image based on the region labeled image and the soil humidity data set corresponding to the remote sensing image.

4. The soil moisture inversion model generation method according to claim 3, characterized in that, The generation of the humidity labeled image corresponding to the remote sensing image based on the region labeled image and the soil humidity data set corresponding to the remote sensing image comprises: generating humidity labeled data corresponding to each of the plurality of labeled regions based on the plurality of labeled regions included in the region labeled image and the soil humidity data set corresponding to the remote sensing image; generating the humidity labeled image corresponding to the remote sensing image based on the region labeled image and the humidity labeled data corresponding to each of the plurality of labeled regions.

5. The soil moisture inversion model generation method according to claim 4, characterized in that, The generation of the humidity labeled data corresponding to each of the plurality of labeled regions based on the plurality of labeled regions included in the region labeled image and the soil humidity data set corresponding to the remote sensing image comprises: For each of the plurality of labeled regions included in the labeled image, determine a to-be-detected region to which the labeled region belongs, and a relative position relationship between the labeled region and a humidity sensor corresponding to the to-be-detected region to which the labeled region belongs; Based on the relative position relationship and the soil humidity data set corresponding to the remote sensing image, determine the humidity label data corresponding to the labeled region.

6. A soil moisture inversion method characterized by, Comprise: Determine a soil humidity inversion model, wherein the soil humidity inversion model is obtained based on the soil humidity inversion model generation method of any one of claims 1 to 5; Using the soil humidity inversion model, generate a humidity image corresponding to a to-be-detected soil region based on a remote sensing image corresponding to the to-be-detected soil region.

7. An irrigation method characterized by, Comprise: Obtain farmland soil humidity, wherein the farmland soil humidity is obtained based on the soil humidity inversion method of claim 6; Determine different humidity regions in the farmland according to the obtained farmland soil humidity; Determine an irrigation scheme according to the humidity regions.

8. An irrigation anomaly determination method characterized by comprising: Comprise: Obtain farmland soil humidity, wherein the farmland soil humidity is obtained based on the soil humidity inversion method of claim 6; Determine different humidity regions in the farmland according to the obtained farmland soil humidity; Compare and analyze the humidity of different humidity regions to determine an abnormal humidity region; Determine an irrigation abnormal point according to the abnormal humidity region.

9. A soil moisture inversion model generation apparatus characterized by comprising: Comprise: A first confirmation module configured to determine a plurality of remote sensing images corresponding to a soil region sample, wherein the plurality of remote sensing images correspond one-to-one to a plurality of preset collection time points, the soil region sample includes a plurality of to-be-detected regions, any to-be-detected region of the plurality of to-be-detected regions corresponds to humidity data of different detection depths, the soil region sample includes a plurality of to-be-detected regions, any to-be-detected region of the plurality of to-be-detected regions corresponds to humidity data of different detection depths, the plurality of to-be-detected regions include different sub-regions of a same plot or regions of different plots, and determination of the plurality of to-be-detected regions includes division based on environmental information of the soil region, the environmental information including soil type and / or soil surface condition; A second confirmation module configured to determine a plurality of soil humidity data sets corresponding to the soil region sample, wherein the plurality of soil humidity data sets correspond one-to-one to the plurality of preset collection time points; A generation module configured to generate a humidity labeled image corresponding to each of the plurality of remote sensing images based on the plurality of soil humidity data sets and the plurality of remote sensing images; A model training module configured to train an initial network model based on the plurality of remote sensing images and the humidity labeled images corresponding to the plurality of remote sensing images to generate a soil humidity inversion model, the initial network model being a neural network model; The second confirmation module comprises a humidity set determination unit. The humidity set determination unit is configured to determine a soil humidity data set of each to-be-detected region at each preset collection time point, wherein the soil humidity data set of a to-be-detected region includes humidity data of different detection depths.

10. A soil moisture inversion apparatus, characterized by, Comprise: a model determining module configured to determine a soil moisture inversion model, wherein the soil moisture inversion model is generated based on the soil moisture inversion model generation method in any one of claims 1 to 5; a soil moisture analysis module configured to determine a moisture image corresponding to a to-be-detected soil region based on a remote sensing image corresponding to the to-be-detected soil region by using the soil moisture inversion model.

11. An irrigation device, characterized in that comprising: a first farmland moisture acquisition module configured to acquire farmland soil moisture, wherein the farmland soil moisture is obtained based on the soil moisture inversion method in claim 6; a first moisture region confirmation module configured to determine different moisture regions in a farmland according to the acquired farmland soil moisture; an irrigation scheme determination module configured to determine an irrigation scheme according to the moisture regions.

12. An irrigation anomaly determination apparatus characterized by comprising: comprising: a second farmland moisture acquisition module configured to acquire farmland soil moisture, wherein the farmland soil moisture is obtained based on the soil moisture inversion method in claim 6; a second moisture region confirmation module configured to determine different moisture regions in a farmland according to the acquired farmland soil moisture; an anomaly analysis module configured to compare and analyze the moisture of different moisture regions to determine an abnormal moisture region; an abnormal point confirmation module configured to determine an irrigation abnormal point according to the abnormal moisture region.

13. A computer-readable storage medium, the storage medium storing instructions, when executed by a processor of an electronic device, enable the electronic device to perform the soil moisture inversion model generation method in any one of claims 1 to 5 and / or the soil moisture inversion method in claim 6 and / or the irrigation method in claim 7 and / or the irrigation abnormality determination method in claim 8.

14. An electronic device, the electronic device comprising: a processor; a memory for storing computer executable instructions; the processor, for executing the computer executable instructions to implement the soil moisture inversion model generation method in any one of claims 1 to 5 and / or the soil moisture inversion method in claim 6 and / or the irrigation method in claim 7 and / or the irrigation abnormality determination method in claim 8.

15. A movable platform, characterized by comprising: an image capturing device for acquiring a remote sensing image of a to-be-detected land plot; a processor for generating a soil moisture inversion model based on the remote sensing image by the soil moisture inversion model method in any one of claims 1 to 5; and / or, determining soil moisture of the to-be-detected land plot based on the remote sensing image by the soil moisture inversion method in claim 6; and / or, determining an irrigation scheme of the to-be-detected land plot based on the remote sensing image by the irrigation method in claim 7; and / or, determining an irrigation abnormal point of the to-be-detected land plot based on the remote sensing image by the irrigation abnormality determination method in claim 8.

Images