Crop root system image generation method and device, crop health determination method and device

Abstract

The present application provides a crop root image generation method and device, and a crop health determination method and device, which relate to the field of smart agriculture. The crop root image generation method includes: determining an underground cross-sectional image of the crop to be observed; using a root contour image generation model to determine a root contour image of the crop to be observed based on the underground cross-sectional image; and determining a crop root image of the crop to be observed based on the underground cross-sectional image and the root contour image. The present application does not require cumbersome collection, has less manual participation, and has a high reusability of the root contour image generation model. It can output the corresponding crop root map of the crop to be observed in a timely and efficient manner based on the input underground cross-sectional image of the crop to be observed, and is suitable for scenes with larger areas of land (such as farmland).

Description

Technical Field

[0001] The present application relates to the field of intelligent agricultural technology, and specifically to a method and device for generating crop root system images, and a method and device for determining crop health. Background Art

[0002] In recent years, smart agriculture has gradually emerged, and convolutional neural network feature extraction technology has been applied in all aspects of agriculture, such as judging the growth status of crops by analyzing crop leaf and stem images by computer, judging the growth of crops by analyzing crop images by computer, etc. However, the research on using crop root images in the soil to analyze the health status of crops is not perfect enough, and there are still some difficulties in obtaining crop images in underground soil. Summary of the invention

[0003] In order to solve the above technical problems, the present application is proposed. The embodiments of the present application provide a method and device for generating a crop root system image, and a method and device for determining crop health.

[0004] In a first aspect, an embodiment of the present application provides a method for generating a crop root image, the method comprising: determining an underground cross-section image of the crop to be observed; using a root contour image generation model to determine a root contour image of the crop to be observed based on the underground cross-section image; and determining a crop root image of the crop to be observed based on the underground cross-section image and the root contour image.

[0005] In combination with the first aspect, in certain implementations of the first aspect, the loss function of the root contour image generation model includes a root area pixel value prediction function and an overall image label function, wherein, for root section image samples, the function value of the root area pixel value prediction function is determined based on the classification label data corresponding to the root section image samples and the model output results, and the function value of the overall image label function is determined based on the classification label data corresponding to the root section image samples.

[0006] In combination with the first aspect, in some implementations of the first aspect, the loss function includes:

[0007]

[0008] represents the root area pixel value prediction function, represents the overall image label function, for the root section image sample, n represents the total number of pixels of the root section image sample, y i Represents the root label data of the i-th pixel, p iRepresents the model prediction probability of the i-th pixel point, wherein the classification label data corresponding to the root section image sample includes the root label data corresponding to each pixel point contained in the root section image sample, and the model prediction probability of the i-th pixel point is determined based on the model output result corresponding to the root section image sample.

[0009] In combination with the first aspect, in certain implementations of the first aspect, before using the root contour image generation model to determine the root contour image of the crop to be observed based on the underground section image, it also includes: determining a training data set, the training data set including M root section image samples and classification label data corresponding to each of the M root section image samples, the classification label data corresponding to the root section image samples including root label data corresponding to each pixel point contained in the root section image samples; training the initial network model based on the training data set to obtain a root contour image generation model.

[0010] In combination with the first aspect, in certain implementations of the first aspect, the M root section image samples include at least one of root section image samples of different types of soil samples, root section image samples of different types of crops, root section image samples of different growth conditions of the same type of crops, and root section image samples of the same type of crops under different light intensities.

[0011] In combination with the first aspect, in certain implementations of the first aspect, the underground section image includes N section image pixels, the root contour image includes N contour image pixels, the crop root image includes N root image pixels, and the N section image pixels and the N contour image pixels are in a one-to-one correspondence, and the N section image pixels and the N root image pixels are in a one-to-one correspondence. Based on the underground section image and the root contour image, the crop root image of the crop to be observed is determined, including: for each of the N section image pixels, multiplying the pixel value of the section image pixel by the pixel value of the contour image pixel corresponding to the section image pixel, to obtain the pixel value of the root image pixel corresponding to the section image pixel; based on the pixel values ​​of the root image pixels corresponding to each of the N section image pixels, the crop root image is determined.

[0012] In a second aspect, an embodiment of the present application provides a method for determining crop health, the method comprising: determining a crop root image of a crop to be observed, the crop root image being determined based on the method described in the first aspect; analyzing the crop root image to obtain growth status data of the crop to be observed.

[0013] In combination with the second aspect, in certain implementations of the second aspect, the crop root image is analyzed to obtain the growth status data of the crop to be observed, including: converting the crop root image into the HSV space, analyzing the color distribution ratio data of the crop root image; and determining the growth status data of the crop to be observed based on the color distribution ratio data.

[0014] In a third aspect, an embodiment of the present application provides a method for generating a pesticide formula, the method comprising: determining health status data of crops in a current plot, the health status data being determined based on the method described in the second aspect; and determining the pesticide formula of the current plot based on the health status data of crops in the current plot.

[0015] In a fourth aspect, an embodiment of the present application provides a crop control method, which includes: determining the pesticide formula of the current plot based on the health status data of the crops in the current plot, wherein the pesticide formula is determined based on the method described in the third aspect; based on the pesticide formula, generating a spraying operation route corresponding to the current plot, so as to perform the pesticide spraying task based on the spraying operation route.

[0016] In a fifth aspect, an embodiment of the present application provides a crop root image generating device, which includes: a first determination module, used to determine an underground cross-section image of the crop to be observed; a second determination module, used to use a root contour image generation model to determine the root contour image of the crop to be observed based on the underground cross-section image; a third determination module, used to determine the crop root image of the crop to be observed based on the underground cross-section image and the root contour image.

[0017] In the sixth aspect, an embodiment of the present application provides a crop health observation device, which includes: a crop root image determination module, used to determine the crop root image of the crop to be observed, the crop root image is determined based on the method described in the first aspect; a crop root image analysis module, used to analyze the crop root image to obtain health status data of the crop to be observed.

[0018] In the seventh aspect, an embodiment of the present application provides a pesticide formula generation device, which includes: a health status data determination module, used to determine the health status data of crops in the current plot, the health status data is determined based on the method described in the second aspect; a pesticide formula generation module, used to determine the pesticide formula of the current plot based on the health status data of crops in the current plot.

[0019] In the eighth aspect, an embodiment of the present application provides a crop control device, which includes a pesticide formula determination module for determining the pesticide formula of the current plot based on the health status data of the crops in the current plot, wherein the pesticide formula is determined based on the method described in the third aspect; a spraying operation generation module for generating a spraying operation route corresponding to the current plot based on the pesticide formula, so as to perform the pesticide spraying task based on the spraying operation route.

[0020] In a ninth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program, and the computer program is used to execute any of the methods described in the first to fourth aspects mentioned in any of the above embodiments.

[0021] In the tenth aspect, an embodiment of the present application provides an electronic device, which includes: a processor; a memory for storing processor executable instructions; the processor is used to execute any method described in any of the first to fourth aspects mentioned in any of the above embodiments.

[0022] In the eleventh aspect, an embodiment of the present application provides a crop root image generation system, the system comprising: an image acquisition system, the image acquisition system is used to capture underground cross-section images of the crop to be observed; and an electronic device as described in the tenth aspect, the electronic device is connected to the image acquisition system.

[0023] In combination with the eleventh aspect, in certain implementations of the eleventh aspect, the image acquisition system includes: a soil contact mechanism, the soil contact mechanism is provided with a transparent window for the soil; and an image acquisition mechanism, an optical path is formed between a camera of the image acquisition mechanism and the transparent window.

[0024] The method for generating a crop root image provided in an embodiment of the present application inputs an underground cross-section image to be observed, uses a root contour image generation model, and outputs a corresponding crop root image of the crop to be observed. This method has less manual participation and the root contour image generation model has high reusability, and is suitable for large-area land (such as farmland) scenes. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0026] Figure 1 Shown is a schematic diagram of a scenario applicable to an embodiment of the present application.

[0027] Figure 2 Shown is a schematic flow chart of a crop root system image generation method provided in one embodiment of the present application.

[0028] Figure 3 Shown is a flow chart of a crop root system image generation method provided in another embodiment of the present application.

[0029] Figure 4 The figure is a schematic diagram of a process of determining a crop root system image of a crop to be observed provided by an embodiment of the present application.

[0030] Figure 5 Shown is a schematic flow chart of a crop health determination method provided in an embodiment of the present application.

[0031] Figure 6 The figure is a schematic diagram of a process for analyzing crop root images and obtaining health status data of the crop to be observed provided by an embodiment of the present application.

[0032] Figure 7 Shown is a flow chart of a method for generating a pesticide formulation according to an embodiment of the present application.

[0033] Figure 8 The figure is a flow chart of a job control method provided in one embodiment of the present application.

[0034] Fig. 9 Shown is a schematic structural diagram of a crop root system image generating device provided in one embodiment of the present application.

[0035] Fig.10 Shown is a schematic structural diagram of a crop health observation device provided in one embodiment of the present application.

[0036] Fig.11 Shown is a schematic diagram of the structure of a pesticide formulation generating device provided in one embodiment of the present application.

[0037] Fig.12 Shown is a schematic structural diagram of a job control device provided in one embodiment of the present application.

[0038] Fig.13 Shown is a schematic structural diagram of an electronic device provided in one embodiment of the present application.

[0039] Fig.14 Shown is a schematic diagram of the structure of an image acquisition system provided in one embodiment of the present application.

[0040] Fig.15 Shown is a schematic diagram of the structure of an image acquisition system provided in one embodiment of the present application.

[0041] Fig.16 Shown is a schematic diagram of the structure of an image acquisition system provided in one embodiment of the present application. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0043] Traditional farmland management generally requires farmland managers to regularly visit the fields to observe the growth of crops, judge the health of crops based on the growth of crops, and then determine whether it is necessary to treat the problem of poor crop health based on the experience of farmland managers. However, this method is limited by time and space, and there is a risk of being untimely or missing, thereby missing the best time to take plant protection measures.

[0044] The method of analyzing crop health by real-time monitoring of the status of crop leaves and stems through smart agricultural technology is mainly applicable to situations where lesions have already manifested in the crop phenotype. However, when lesions manifest in the crop phenotype, it usually causes certain economic losses.

[0045] The root system is an important organ for crops to obtain water and nutrients from the soil and other media environments. A basic summary is "white roots are strong, yellow roots are life-saving, black roots are diseased, and gray roots are fatal." It can be seen that the different states of the root system reflect the health status of the crop to a certain extent, and the early lesions of the roots are not reflected in the crop phenotype, which can play a role in early detection and early control.

[0046] However, there are still some difficulties in acquiring crop root images in the soil, and the processing and analysis of crop root images are not perfect enough.

[0047] Figure 1 The figure shows a schematic diagram of an application scenario provided by an embodiment of the present application. Figure 1 As shown, the application scenario provided by the embodiment of the present application is a farmland application scenario. Specifically, the scenario includes a farmland 110, a server 120, and an image acquisition system 130 connected to the server 120. Specifically, the farmland 110 can be regarded as a plot area to be detected, and the server 120 is deployed with a root contour image generation model that can determine the root contour image of the crop to be observed.

[0048] For example, in actual application, the image acquisition system 130 captures an underground cross-sectional image of the crop to be observed in the farmland 110, and uploads the captured underground cross-sectional image of the crop to be observed to the server 120. After acquiring the underground cross-sectional image of the crop to be observed, the server 120 generates a model using the root system contour image, analyzes the underground cross-sectional image of the crop to be observed, and obtains a crop root system image of the crop to be observed in the farmland 110.

[0049] In the above application scenarios, human participation is less, and after determining the underground cross-section image of the crop to be observed, the root system image of the crop to be observed is obtained by using the root system contour image generation model analysis, which is not only timely and efficient, but also fully applicable to larger farmland scenes.

[0050] It should be noted that the server can be replaced by other types of processing devices, such as electronic devices with processing capabilities such as tablet computers or personal computers, but is not limited to this. For example, in one device, it can have the function of executing the above-mentioned crop root image generation method and can also have the function of an image acquisition system. This application does not limit this.

[0051] Figure 2 The figure is a flow chart of a crop root system image generation method provided in an embodiment of the present application. Exemplarily, the crop root system image generation method can be executed in a processor of a hardware device such as a farming machine, or in a related server.

[0052] like Figure 2 As shown, a crop root system image generation method provided by an exemplary embodiment of the present application includes the following steps.

[0053] Step 230: determine the underground cross-sectional image of the crop to be observed.

[0054] Exemplarily, the crops to be observed refer to crops for which root images need to be obtained to facilitate other detection and analysis operations, such as crops in a farmland for which root images are required to analyze the growth status of the crops.

[0055] Exemplarily, the underground cross-sectional image refers to an underground cross-sectional image of the crop to be observed including the crop root system, for example, an underground cross-sectional image taken by placing a photographing device underground near the crop root system.

[0056] Step 240 , using the root system contour image generation model, based on the underground section image, determines the root system contour image of the crop to be observed.

[0057] In some embodiments, the root contour image generation model refers to a deep neural network model, the deep neural network model input is an underground section image, and the output is a root contour image. The number of pixels contained in the width and height directions of the underground section image and the root contour image is the same.

[0058] For example, the pixels of the underground section image are 500×700, which means that the underground section image has 500 pixels in the width direction and 700 pixels in the height direction, and the pixels of the root system contour image are also 500×700.

[0059] Exemplarily, in addition to the root system of the crop to be detected, the underground section image also includes other non-root parts, such as soil, impurities in the soil, and underground fruits of the crop. The root contour image output by the root contour image generation model only includes the root part. The pixel value of the corresponding position of the root part on the root contour image is 1, and the pixel value of the corresponding position of the non-root part on the root contour image is 0.

[0060] In actual application, the root contour image generation model can be deployed to the edge computing device on the ground for real-time calculation, or it can be deployed to the cloud server. The embodiments of the present application do not make a unified limitation on this.

[0061] Step 250, determining a crop root system image of the crop to be observed based on the underground section image and the root system contour image.

[0062] For example, the number of pixels in the width and height directions of the crop root system image is the same as that of the underground section image and the root system contour image. The crop root system image only contains the crop root system, and the pixel value of the crop root system at the corresponding position on the crop root system image is the same as the pixel value of the crop root system at the corresponding position on the underground section image, and the pixel value of the non-root area on the crop root system image is 0.

[0063] The crop root system image generation method provided in the embodiment of the present application inputs the underground section image of the crop to be observed into the root system contour image generation model to obtain the root system contour image, and multiplies the pixel values ​​at the corresponding positions of the underground section image and the root system contour image to obtain the crop root system image. The method can quickly, timely and accurately obtain the crop root system image, so that people can obtain information about the crop growth status from the crop root system image in a timely manner.

[0064] Figure 3 FIG. 1 is a flow chart of a crop root system image generation method provided by another embodiment of the present application. Figure 2 Based on the embodiment shown, Figure 3 The embodiment shown is described below in detail. Figure 3 The embodiment shown and Figure 2The differences and similarities of the illustrated embodiments are not described in detail.

[0065] like Figure 3 As shown, before using the root contour image generation model to determine the root contour image of the crop to be observed based on the underground section image, the following steps are also included.

[0066] Step 210: determine a training data set.

[0067] Specifically, the training data set includes M root section image samples and label data corresponding to each of the M root section image samples. The label data corresponding to the root section image samples include root label data corresponding to each pixel point included in the root section image samples.

[0068] Exemplarily, the M root section image samples include at least one of root section image samples of different types of soil samples, root section image samples of different types of crops, root section image samples of the same type of crops under different growth conditions, and root section image samples of the same type of crops under different light intensities.

[0069] In actual application, the image capture device is buried underground for a long time. In order to obtain a clear image, it is generally necessary to take photos with fill light, that is, the image capture device and the fill light module are used together to shoot the root system of the crop. The fill light module will wear out over time, affecting the fill light intensity, which will result in different fill light intensities even if the fill light control is performed with the same power. Therefore, before training the root contour image generation model, the embodiment of the present application also needs to collect root section image samples of the same type of crops under different light intensities.

[0070] Exemplarily, the pixel values ​​of the pixels at the regional positions containing roots on the root section image sample are marked as 1, and the pixel values ​​of the pixels at the regional positions containing non-roots on the root section image sample are marked as 0. The label data corresponding to the root section image sample includes the pixel values ​​1 of the regional positions containing roots on all root section image samples and the pixel values ​​0 of all regional positions containing non-roots.

[0071] In some embodiments, the root section image sample is an image in RGB three-channel image format. The RGB three-channel image format is a color standard in the industry. Various colors are obtained by changing the three color channels of red (R), green (G), and blue and superimposing them on each other. This standard includes almost all colors that can be perceived by human vision and is one of the most widely used color systems.

[0072] In some embodiments, root section image samples are preprocessed, and the preprocessed root section image samples are included in a training set. The preprocessing includes rotation and / or cropping operations. Specifically, rotation and / or cropping operations are performed on the same root section image sample to obtain new root section image samples, and all the obtained new root section image samples are included in the training data set, thereby increasing the diversity of samples to improve the accuracy of the obtained root contour image generation model.

[0073] Step 220, training an initial network model based on the training data set to obtain a root system contour image generation model.

[0074] It can be understood that the initial network model is a pre-established neural network model. The difference between the initial network model and the root contour image generation model lies in the difference in model parameters, that is, the model parameters of the initial network model are adjusted by training with the training data set to obtain a converged image difference detection model.

[0075] Specifically, the loss function of the root contour image generation model includes a root area pixel value prediction function and an overall image label function, wherein, for a root section image sample, the function value of the root area pixel value prediction function is determined based on the classification label data corresponding to the root section image sample and the model output result, and the function value of the overall image label function is determined based on the classification label data corresponding to the root section image sample.

[0076] Furthermore, the formula of the loss function is:

[0077]

[0078] represents the root area pixel value prediction function, represents the overall image label function, for the root section image sample, n represents the total number of pixels of the root section image sample, y i Represents the root label data of the i-th pixel, p i Represents the model prediction probability of the i-th pixel point, wherein the classification label data corresponding to the root section image sample includes the root label data corresponding to each pixel point contained in the root section image sample, and the model prediction probability of the i-th pixel point is determined based on the model output result corresponding to the root section image sample.

[0079] For example, for a root section image sample with a pixel size of 300×400, if the pixel at the position (200, 100) has a root part, the root label data y of the pixel at the position (200, 100) is marked as 1; if the pixel at the position (260, 200) does not contain a root part, the root label data y of the pixel at the position (260, 200) is marked as 0.

[0080] The crop root image generation method in the embodiment of the present application trains the initial network model through a large and rich training data set, thereby improving the output accuracy of the root contour image generation model obtained. In addition, the denominator of the loss function of the root contour image generation model is the total number of label data of the entire image, which prevents the loss function from being too small when the root area is relatively small and cannot effectively reflect the loss of the root area. The output accuracy of the root contour image generation model is further improved by the selected loss function suitable for this model.

[0081] Figure 4 The figure shows a flow chart of determining the crop root system image of the crop to be observed provided by an embodiment of the present application. Figure 2 Based on the embodiment shown, Figure 4 The embodiment shown is described below in detail. Figure 4 The embodiment shown and Figure 2 The differences and similarities of the illustrated embodiments are not described in detail.

[0082] like Figure 4 As shown, the step of determining the crop root system image of the crop to be observed based on the underground section image and the root system contour image includes the following steps.

[0083] Step 251, for each of the N cross-sectional image pixels, multiply the pixel value of the cross-sectional image pixel by the pixel value of the contour image pixel corresponding to the cross-sectional image pixel to obtain the pixel value of the root system image pixel corresponding to the cross-sectional image pixel.

[0084] Specifically, the underground section image includes N section image pixels, the root contour image includes N contour image pixels, the crop root image includes N root image pixels, and the N section image pixels and the N contour image pixels are in a one-to-one correspondence, and the N section image pixels and the N root image pixels are in a one-to-one correspondence.

[0085] As an example, the underground section image, the root system outline image, and the crop root system image are all 500×600 pixels. The pixel point at the position (200, 300) on the underground section image contains the crop root system, and the pixel value of the pixel point at this position is 163. The pixel value of the corresponding pixel point at the position (200, 300) on the root system outline image is 1, and the pixel value of the corresponding pixel point at the position (200, 300) on the crop root system image is 163×1=163. The pixel point at the position (150, 400) on the underground section image does not contain the root system, and the pixel value of the pixel point at this position is 36. The pixel value of the corresponding pixel point at the position (150, 400) on the root system outline image is 0, and the pixel value of the corresponding pixel point at the position (150, 400) on the crop root system image is 36×0=0.

[0086] Step 252, determining the crop root system image based on the pixel values ​​of the root system image pixels corresponding to the N slice image pixels.

[0087] Continuing with the example in step 251, by multiplying the pixel values ​​of 500×600 pixels in the underground section image with the pixel values ​​of 1 or 0 of 500×600 pixels in the corresponding root contour image, the resulting new pixel values ​​of 500×600 pixels are assigned to the crop root image.

[0088] The crop root image generation method in the embodiment of the present application obtains the crop root image by multiplying the pixel values ​​of the pixel points in the underground section image with the pixel values ​​of the pixel points in the root contour image. This method can convert the pixel values ​​of the non-root areas in the underground section image to 0, and only retain the pixel values ​​of the root area, thereby avoiding the influence of the non-root area and facilitating accurate analysis of the crop root area.

[0089] Figure 5 FIG. 1 is a flow chart of a crop health determination method provided in an embodiment of the present application. Figure 5 As shown, the crop health determination method provided in the embodiment of the present application includes the following steps.

[0090] Step 510: determine the crop root system image of the crop to be observed.

[0091] Exemplarily, the crop root system image of the crop to be observed is determined based on the crop root system image generation method mentioned in any of the above embodiments.

[0092] Step 520: Analyze the crop root system image to obtain health status data of the crop to be observed.

[0093] The color distribution ratio in the crop root system image is analyzed, and the health status data of the crop to be observed is obtained according to the color distribution ratio.

[0094] Figure 6 The figure shows a flow chart of analyzing crop root system images and obtaining health status data of the crop to be observed according to an embodiment of the present application. Figure 5 Based on the embodiment shown, Figure 6 The embodiment shown is described below in detail. Figure 6 The embodiment shown and Figure 5 The differences and similarities of the illustrated embodiments are not described in detail.

[0095] like Figure 6 As shown, in the crop health determination method provided in the embodiment of the present application, the step of analyzing the crop root system image to obtain the health status data of the crop to be observed includes the following steps.

[0096] Step 521, converting the crop root system image into the HSV (Hue, Saturation, Value) space, and analyzing the color distribution ratio data of the crop root system image.

[0097] Specifically, the RGB image of the crop root system image is converted to the HSV space, and the color distribution ratio of the crop root system image is analyzed according to Table 1.

[0098] Table 1 Hue, saturation and brightness range values ​​of different colors

[0099] black Ash yellow white Minimum hue value 0 0 26 0 Maximum hue value 180 180 34 180 Minimum saturation value 0 0 43 0 Maximum saturation value 255 43 255 30 Minimum brightness value 0 46 46 221 Maximum brightness value 46 220 255 255

[0100] In some embodiments, the hue value, saturation value, and brightness value of the pixel points in the crop root image are analyzed one by one to see whether they are all within the range of the hue value, saturation value, and brightness value of a certain color. If they are all within the range, the pixel point is marked as the corresponding color.

[0101] For example, if the hue value of pixel a in the crop root system image is 60, the saturation value is 100, and the lightness value is 30, the color of pixel a is black.

[0102] Step 522, determining the health status data of the crop to be observed based on the color distribution ratio data.

[0103] Specifically, a first ratio threshold, a second ratio threshold, a third ratio threshold, and a fourth ratio threshold are preset, and the values ​​of the first ratio threshold, the second ratio threshold, the third ratio threshold, and the fourth ratio threshold are respectively determined according to historical growth status data of the corresponding crop.

[0104] For example, for the crop peanut, the first ratio threshold is determined to be 70%, the second ratio threshold is 50%, the third ratio threshold is 40%, and the fourth ratio threshold is 40% according to its corresponding historical growth status data.

[0105] Exemplarily, if the proportion of white in the crop root image is greater than or equal to the first proportion threshold of 70%, the crop is in good health; if the proportion of yellow in the crop root image is greater than or equal to the second proportion threshold of 50%, the crop is in acceptable health; if the proportion of black in the crop root image is greater than or equal to the third proportion threshold of 40%, the crop is sick; if the proportion of gray in the crop root image is greater than or equal to the fourth proportion threshold of 40%, the crop will die.

[0106] The crop health status observation method in this embodiment converts the crop root system image into the HSV space and analyzes the proportion of various colors in the crop root system image to judge the health status of the crop. This method is simple, scientific, and eliminates the need for manual observation. It can detect early lesions of crops and facilitate people to take corresponding treatment measures based on the crop health status data obtained in real time to avoid large economic losses in the later stage.

[0107] Figure 7 FIG. 1 is a flow chart of a method for generating a pesticide formulation according to an embodiment of the present application. Figure 7 As shown, the pesticide formulation generation method provided in the embodiment of the present application includes the following steps.

[0108] Step 710, determining health status data of crops in the current plot.

[0109] Exemplarily, the health status data of the crops in the current plot is determined based on the crop health determination method mentioned in the above embodiment.

[0110] Step 720, determining the pesticide formula for the current plot based on the health status data of the crops in the current plot.

[0111] For example, if it is detected that the crops in the current plot are diseased, a corresponding pesticide formula can be prepared or selected for treatment.

[0112] For example, for crops that are in good health, other remedial measures such as fertilization or watering can be taken to help the crops overcome the immediate growth bottleneck. For dying crops, they can be pulled out in time.

[0113] Figure 8 FIG. 1 is a flow chart of a method for controlling a job according to an embodiment of the present application. Figure 8 As shown, the job control method provided in the embodiment of the present application includes the following steps.

[0114] Step 810, determining the pesticide formula for the current plot based on the health status data of the crops in the current plot.

[0115] Exemplarily, the pesticide formula of the current plot is determined based on the pesticide formula generation method mentioned in the above embodiment.

[0116] Step 820, based on the pesticide formula, generate a spraying operation route corresponding to the current plot.

[0117] Combination of the above Figures 2 to 8 , describes the method embodiment of the present application in detail, and the following is combined with Figures 9 to 16 , describing the device embodiment of the present application. It should be understood that the description of the method embodiment corresponds to the description of the device embodiment, so the parts not described in detail can refer to the previous method embodiment.

[0118] Fig. 9 FIG. 1 is a schematic diagram of the structure of a crop root system image generation device provided by an embodiment of the present application. Fig. 9 As shown, the crop root image generation device provided in the embodiment of the present application includes a first determination module 910, which is used to determine the underground cross-sectional image of the crop to be observed; a second determination module 920, which is used to use the root contour image generation model to determine the root contour image of the crop to be observed based on the underground cross-sectional image; and a third determination module 930, which is used to determine the crop root image of the crop to be observed based on the underground cross-sectional image and the root contour image.

[0119] In some embodiments, the loss function of the root contour image generation model includes a root area pixel value prediction function and an overall image label function, wherein, for a root section image sample, the function value of the root area pixel value prediction function is determined based on the classification label data corresponding to the root section image sample and the model output result, and the function value of the overall image label function is determined based on the classification label data corresponding to the root section image sample.

[0120] In some embodiments, the loss function includes:

[0121]

[0122] represents the root area pixel value prediction function, represents the overall image label function, for the root section image sample, n represents the total number of pixels of the root section image sample, y i Represents the root label data of the i-th pixel, p i Represents the model prediction probability of the i-th pixel point, wherein the classification label data corresponding to the root section image sample includes the root label data corresponding to each pixel point contained in the root section image sample, and the model prediction probability of the i-th pixel point is determined based on the model output result corresponding to the root section image sample.

[0123] In some embodiments, the second determination module 920 is also used to determine a training data set, which includes M root section image samples and classification label data corresponding to each of the M root section image samples, and the classification label data corresponding to the root section image samples include root label data corresponding to each pixel point contained in the root section image samples; the initial network model is trained based on the training data set to obtain a root contour image generation model.

[0124] In some embodiments, the M root section image samples include at least one of root section image samples of different types of soil samples, root section image samples of different types of crops, root section image samples of the same type of crops under different growth conditions, and root section image samples of the same type of crops under different light intensities.

[0125] In some embodiments, the underground cross-sectional image includes N cross-sectional image pixels, the root contour image includes N contour image pixels, the crop root image includes N root image pixels, and the N cross-sectional image pixels are in one-to-one correspondence with the N contour image pixels, and the N cross-sectional image pixels are in one-to-one correspondence with the N root image pixels. The third determination module 930 is also used to multiply the pixel value of the cross-sectional image pixel with the pixel value of the contour image pixel corresponding to the cross-sectional image pixel for each of the N cross-sectional image pixels, to obtain the pixel value of the root image pixel corresponding to the cross-sectional image pixel; based on the pixel values ​​of the root image pixels corresponding to each of the N cross-sectional image pixels, determine the crop root image.

[0126] Fig.10 The figure is a schematic diagram of the structure of a crop health observation device provided by an embodiment of the present application. Fig.10 As shown, the crop health observation device provided in the embodiment of the present application includes a crop root system image determination module 1010, which is used to determine the crop root system image of the crop to be observed. A crop root system image analysis module 1020, which is used to analyze the crop root system image to obtain health status data of the crop to be observed.

[0127] In some embodiments, the crop root image analysis module 1020 is also used to convert the crop root image into the HSV space, analyze the color distribution ratio data of the crop root image; and determine the health status data of the crop to be observed based on the color distribution ratio data.

[0128] Fig.11 FIG. 1 is a schematic diagram of the structure of a pesticide formulation generating device provided in an embodiment of the present application. Fig.11As shown, the pesticide formula generation device of the embodiment of the present application includes a health data determination module 1110, which is used to determine the health data of the crops in the current plot. The pesticide formula generation module 1120 is used to determine the pesticide formula of the current plot based on the health data of the crops in the current plot.

[0129] Fig.12 FIG. 1 is a schematic diagram of the structure of a job control device provided by an embodiment of the present application. Fig.12 As shown, the operation control device provided in the embodiment of the present application includes a pesticide formula determination module 1210, which is used to determine the pesticide formula of the current plot based on the health status data of the crops in the current plot. A spraying crop generation module is used to generate a spraying operation route corresponding to the current plot based on the pesticide formula, so as to perform the pesticide spraying task based on the spraying operation route.

[0130] Fig.13 The figure is a schematic diagram of the structure of an electronic device provided by an embodiment of the present application. Fig.13 To describe an electronic device according to an embodiment of the present application.

[0131] like Fig.13 As shown, the electronic device 70 includes one or more processors 701 and a memory 702 .

[0132] The processor 701 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 70 to perform desired functions.

[0133] The memory 702 may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 701 may run the program instructions to implement the relevant methods of the various embodiments of the present application described above and / or other desired functions. Various contents such as underground section images of crops to be observed, root contour images of crops to be observed, and crop root images of crops to be observed may also be stored in the computer-readable storage medium.

[0134] In one example, the electronic device 70 may further include: an input device 703 and an output device 704, and these components are interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0135] The input device 703 may include, for example, a keyboard, a mouse, etc.

[0136] The output device 704 can output various information to the outside, including root system outline images of the crops to be observed, crop root system images of the crops to be observed, etc. The output device 704 can include, for example, a display, a speaker, a printer, a communication network and its connected remote output device, etc.

[0137] Of course, to simplify, Fig.13 Only some of the components related to the present application in the electronic device 70 are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, according to specific application conditions, the electronic device 70 may also include any other appropriate components.

[0138] In addition to the above-mentioned methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps in the related methods described above in this specification according to various embodiments of the present application.

[0139] The computer program product may be written in any combination of one or more programming languages ​​to write program codes for performing the operations of the embodiments of the present application, including object-oriented programming languages, such as Java, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as an independent software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0140] In addition, an embodiment of the present application may also be a computer-readable storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, the processor executes the steps of the related methods described above in this specification according to various embodiments of the present application.

[0141] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0142] In some embodiments, a crop root system image generation system is also provided, which includes an image acquisition system and an electronic device as mentioned in the above embodiments, and the image acquisition system and the electronic device have a communication connection relationship. The image acquisition system is used to capture the underground section image of the crop to be observed.

[0143] In some embodiments, the image acquisition system includes: a soil contact mechanism, the soil contact mechanism is provided with a transparent window for use in the soil; and an image acquisition mechanism, wherein an optical path is formed between a camera of the image acquisition mechanism and the transparent window.

[0144] Combine the following Figures 14 to 16 An example is given to illustrate the image acquisition system.

[0145] See also Fig.14 , Fig.14 This is a schematic diagram of the structure of a first optional method of an image acquisition device provided in an embodiment of the present application. Fig.14 As shown, the soil contact mechanism 1 can be completely buried in the soil 5 and has a container with a accommodating cavity; the soil contact mechanism 1 includes a protective cover; a transparent window 2 is provided on one side of the protective cover; the image acquisition mechanism 3 is installed in the protective cover, and the camera of the image acquisition mechanism 3 faces the transparent window 2, and the cable 4 is used to transmit the underground soil photos taken by the image acquisition mechanism 3.

[0146] See also Fig.15 , Fig.15 This is a schematic diagram of the structure of a second optional method of an image acquisition device provided in another embodiment of the present application. Fig.15As shown, the soil surface forming component 1 can be a periscope having a first periscope port and a second periscope port, the first periscope port of the periscope is a soil observation port, and the second periscope port of the periscope is a camera observation port; the transparent window 2 is formed by the soil observation port or the transparent window 2 is sealed and installed on the soil observation port; the camera of the image acquisition mechanism 3 faces the camera observation port. In addition, during the operation of the image acquisition device, the camera observation port and the image acquisition mechanism 3 are located in a space outside the soil, that is, above the ground, so as to facilitate their operation and maintenance.

[0147] Optionally, see Fig.16 , Fig.16 This is a schematic diagram of the structure of a third optional method of an image acquisition device provided in another embodiment of the present application. Fig.16 As shown, the inner diameter of the periscope channel located between the two plane mirrors in the periscope increases from the camera observation port to the soil observation port. That is to say, the periscope tube located between the first periscope port and the second periscope port of the periscope is a trumpet tube with a gradually increasing inner diameter, and the soil observation port is arranged at the end of the trumpet tube with a larger inner diameter, and the end with a larger tube diameter is buried in the soil 5 to increase the shooting field of view.

[0148] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0149] The basic principles of the present application are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present application are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present application. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, not for limitation, and the above details do not limit the present application to being implemented by adopting the above specific details.

[0150] The block diagrams of the devices, apparatuses, equipment, and systems involved in this application are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagram. As will be appreciated by those skilled in the art, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open words, referring to "including but not limited to", and can be used interchangeably with them. The words "or" and "and" used here refer to the words "and / or" and can be used interchangeably with them, unless the context clearly indicates otherwise. The words "such as" used here refer to the phrase "such as but not limited to", and can be used interchangeably with them.

[0151] It should also be noted that in the apparatus, device and method of the present application, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0152] 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 general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0153] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A method for generating a crop root system image, characterized in that: include: Determine the underground cross-sectional image of the crop to be observed; Using a root system contour image generation model, based on the underground section image, a root system contour image of the crop to be observed is determined, wherein the root system contour image only includes a root system portion, a pixel value of a corresponding position of the root system portion on the root system contour image is 1, and a pixel value of a corresponding position of a non-root system portion on the root system contour image is 0; Determine a crop root system image of the crop to be observed based on the underground section image and the root system contour image, wherein the crop root system image only includes a crop root system portion, a pixel value of a corresponding position of the crop root system portion on the crop root system image is the same as a pixel value of a corresponding position of the crop root system portion on the underground section image, and a pixel value of a non-root system region on the crop root system image is 0; Analyzing the crop root system image to obtain health status data of the crop to be observed; The loss function of the root contour image generation model includes a root region pixel value prediction function and an overall image label function, wherein, for a root section image sample, the function value of the root region pixel value prediction function is determined based on the classification label data corresponding to the root section image sample and the model output result, and the function value of the overall image label function is determined based on the classification label data corresponding to the root section image sample; The loss function includes: Said represents the root area pixel value prediction function, the represents the overall image label function, for the root section image sample, n represents the total number of pixels of the root section image sample, y i Represents the root label data of the i-th pixel, p i Represents the model prediction probability of the i-th pixel point, wherein the classification label data corresponding to the root section image sample includes the root label data corresponding to each pixel point contained in the root section image sample, and the model prediction probability of the i-th pixel point is determined based on the model output result corresponding to the root section image sample.

2. The crop root system image generation method according to claim 1, characterized in that: Before using the root system contour image generation model to determine the root system contour image of the crop to be observed based on the underground section image, the method further includes: Determine a training data set, the training data set comprising M root section image samples and classification label data corresponding to each of the M root section image samples, the classification label data corresponding to the root section image samples comprising root label data corresponding to each pixel point included in the root section image samples; An initial network model is trained based on the training data set to obtain the root system outline image generation model.

3. The crop root system image generation method according to claim 2, characterized in that: The M root section image samples include at least one of root section image samples of different types of soil samples, root section image samples of different types of crops, root section image samples of the same type of crops under different growth conditions, and root section image samples of the same type of crops under different light intensities.

4. The crop root system image generation method according to any one of claims 1 to 3, characterized in that: The underground section image includes N section image pixels, the root system contour image includes N contour image pixels, the crop root system image includes N root system image pixels, and the N section image pixels and the N contour image pixels are in a one-to-one correspondence, and the N section image pixels and the N root system image pixels are in a one-to-one correspondence; The step of determining the crop root system image of the crop to be observed based on the underground section image and the root system contour image comprises: For each of the N cross-sectional image pixel points, multiply the pixel value of the cross-sectional image pixel point by the pixel value of the contour image pixel point corresponding to the cross-sectional image pixel point to obtain the pixel value of the root system image pixel point corresponding to the cross-sectional image pixel point; The crop root system image is determined based on the pixel values ​​of the root system image pixel points corresponding to each of the N section image pixel points.

5. The crop root system image generation method according to claim 1, characterized in that: The step of analyzing the crop root system image to obtain health status data of the crop to be observed includes: Convert the crop root system image into HSV space, and analyze the color distribution ratio data of the crop root system image; Based on the color distribution ratio data, health status data of the crop to be observed is determined.

6. A method for generating a pesticide formulation, characterized in that: include: Determining health status data of crops in a current plot, wherein the health status data is determined based on the method according to any one of claims 1 to 5; Based on the health status data of the crops in the current plot, a pesticide formula for the current plot is determined.

7. A method for controlling a job, characterized in that: include: Determining a pesticide formula for the current plot based on the health status data of the crops in the current plot, wherein the pesticide formula is determined based on the method of claim 6; Based on the pesticide formula, a spraying operation route corresponding to the current plot is generated, so as to perform the pesticide spraying task based on the spraying operation route.

8. A crop root system image generation device, characterized in that: include: A first determination module is used to determine an underground section image of the crop to be observed; A second determination module is used to determine the root system contour image of the crop to be observed based on the underground section image by using a root system contour image generation model, wherein the root system contour image only includes a root system portion, and a pixel value of a corresponding position of the root system portion on the root system contour image is 1, and a pixel value of a corresponding position of a non-root system portion on the root system contour image is 0; A third determination module is used to determine a crop root system image of the crop to be observed based on the underground section image and the root system contour image, wherein the crop root system image only includes a crop root system portion, a pixel value of a corresponding position of the crop root system portion on the crop root system image is the same as a pixel value of a corresponding position of the crop root system portion on the underground section image, and a pixel value of a non-root system area on the crop root system image is 0; An analysis module, used for analyzing the crop root system image to obtain health status data of the crop to be observed; The loss function of the root contour image generation model includes a root region pixel value prediction function and an overall image label function, wherein, for a root section image sample, the function value of the root region pixel value prediction function is determined based on the classification label data corresponding to the root section image sample and the model output result, and the function value of the overall image label function is determined based on the classification label data corresponding to the root section image sample; The loss function includes: Said represents the root area pixel value prediction function, the represents the overall image label function, for the root section image sample, n represents the total number of pixels of the root section image sample, y i Represents the root label data of the i-th pixel, p i Represents the model prediction probability of the i-th pixel point, wherein the classification label data corresponding to the root section image sample includes the root label data corresponding to each pixel point contained in the root section image sample, and the model prediction probability of the i-th pixel point is determined based on the model output result corresponding to the root section image sample.

9. A pesticide formulation generating device, characterized in that: include: A health data determination module, used to determine health data of crops in a current plot, wherein the health data is determined based on the method according to any one of claims 1 to 5; The pesticide formula generation module is used to determine the pesticide formula of the current plot based on the health status data of the crops in the current plot.

10. A work control device, characterized in that: include: A pesticide formula determination module, used to determine the pesticide formula of the current plot based on the health status data of the crops in the current plot, wherein the pesticide formula is determined based on the method of claim 6; The spraying operation generation module is used to generate a spraying operation route corresponding to the current plot based on the pesticide formula, so as to perform the pesticide spraying task based on the spraying operation route.

11. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and the computer program is used to execute the method according to any one of claims 1 to 7.

12. An electronic device, characterized in that: The electronic device comprises: processor; a memory for storing instructions executable by the processor; The processor is used to execute the method according to any one of claims 1 to 7.

13. A crop root system image generation system, characterized in that: include: An image acquisition system, wherein the image acquisition system is used to capture an underground cross-sectional image of the crop to be observed; as well as The electronic device of claim 12, connected to the image acquisition system.

14. The crop root system image generation system according to claim 13, characterized in that: The image acquisition system comprises: a soil contact mechanism having a transparent viewing window for use in the soil; An image acquisition mechanism, wherein an optical path is formed between a camera of the image acquisition mechanism and the transparent window.

Images