Prescription homework method, plant protection equipment and storage medium

Abstract

The invention relates to the technical field of unmanned aerial vehicles, and discloses a prescription work method, plant protection equipment and a storage medium, and the method comprises the steps: extracting a pixel parameter distribution diagram of each channel of a global image of a target land parcel to generate a vegetation index distribution diagram, and obtaining a growth grading diagram of green plants based on the vegetation index distribution diagram; and performing prescription work on the target plot according to the growth grading graph. In this way, when the green plants are operated, the corresponding prescription can be determined according to the growth vigor, namely the amount of the operation is determined, so that real-time matching of the prescription is achieved, the situation that the amount of the operation is too small or too large is avoided, and the problem that an undifferentiated operation scheme in the prior art cannot be matched with the growth vigor of the crops is solved.

Description

Technical Field

[0001] The present application relates to the technical field of unmanned aerial vehicles, and in particular to a prescription operation method, plant protection equipment and storage medium. Background Art

[0002] In the modern agricultural planting process, most of the intelligent equipment such as drones are used to replace manual operations, especially in sowing, fertilizing, spraying, etc. At present, when drones are used for fertilization and spraying, they are basically operated uniformly in the whole field in an indiscriminate manner. This operation method does not take into account the growth distribution of crops in the field and the actual distribution of pests and diseases. It often causes excessive application of pesticides and fertilizers, which easily leads to pesticide residues and waste of agricultural materials, increasing the cost of agricultural operations. Summary of the invention

[0003] The present application provides a prescription operation method, plant protection equipment and storage medium, which are used to solve the problem that the indiscriminate operation scheme of the prior art cannot match the growth of crops.

[0004] The first aspect of the present application provides a prescription operation method, which is applied to plant protection equipment, and the method includes: acquiring a global image of a target plot; extracting pixel parameters of each channel in the global image, and generating a pixel parameter distribution map of each channel, wherein the pixel parameter is the brightness value of a spectral channel or the reflectivity of a spectral channel; generating a vegetation index distribution map of the target plot based on the pixel parameter distribution map of each channel; analyzing the vegetation index distribution map to obtain a growth grading map of green plants in the target plot; and performing prescription operations on the target plot based on the growth grading map.

[0005] Optionally, the global image only includes a multispectral image; and extracting pixel parameters of each channel in the global image to generate a pixel parameter distribution map of each channel includes: extracting the reflectivity of each spectral channel in the multispectral image to generate a reflectivity distribution map of the target plot.

[0006] Optionally, the global image includes an RGB image; and extracting pixel parameters of each channel in the global image to generate a pixel parameter distribution map of each channel includes: extracting the brightness values ​​of R, G, and B channels of each pixel point in the RGB image to generate a brightness distribution map of the three channels of the target plot.

[0007] Optionally, generating the vegetation index distribution map of the target plot based on the pixel parameter distribution map of each channel includes: combining and calculating the brightness values ​​of the same pixel point in the brightness distribution map of each channel to obtain the vegetation index distribution map of the target plot.

[0008] Optionally, the global image also includes a multispectral image; before analyzing the vegetation index distribution map to obtain the growth classification map of green plants in the target plot, it also includes: calculating the green plant coverage of the target plot based on the vegetation index distribution map; determining the type of the multispectral vegetation index distribution map based on the green plant coverage, and generating a multispectral vegetation index distribution map corresponding to the type based on the multispectral image.

[0009] Optionally, the calculating the green plant coverage of the target plot based on the vegetation index distribution map includes: calculating the ratio between the number of pixels in the green plant area and the number of all pixels in the vegetation index distribution map to obtain the green plant coverage of the target plot; or

[0010] Determine multiple local areas in the global map; calculate the regional green plant coverage corresponding to the local area according to the ratio between the number of green plant area pixels corresponding to each local area and the number of all pixels in the area in the vegetation index distribution map, and obtain multiple regional green plant coverages; determine the average value of the green plant coverages of the multiple areas, and use the average value as the green plant coverage of the target plot.

[0011] Optionally, determining the type of the multispectral vegetation index distribution map based on the green plant coverage, and generating a multispectral vegetation index distribution map corresponding to the type based on the multispectral image, includes: judging whether the green plant coverage is less than a preset coverage threshold; if so, determining that the type of the multispectral vegetation index distribution map is a first type, and generating a first multispectral vegetation index distribution map based on the multispectral image; if not, determining that the type of the multispectral vegetation index distribution map is a second type, and generating a second multispectral vegetation index distribution map based on the multispectral image.

[0012] Optionally, the analyzing the vegetation index distribution map to obtain a growth classification map of green plants in the target plot includes: analyzing the first multispectral vegetation index distribution map or the second multispectral vegetation index distribution map using a natural breakpoint algorithm to obtain a growth classification map of green plants in the target plot.

[0013] Optionally, before analyzing the vegetation index distribution map to obtain the growth classification map of green plants in the target plot, the method further includes: binarizing the vegetation index distribution map to obtain a vegetation index distribution map without background.

[0014] A second aspect of the present application provides a plant protection device, which includes a mobile platform, a spreading device, a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory to enable the plant protection device to execute the above-mentioned prescription operation method.

[0015] A third aspect of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the above-mentioned prescription operation method.

[0016] In the technical solution provided by the present application, a vegetation index distribution map is generated by extracting the pixel parameter layout of each channel of the global image of the target plot, and then a growth classification map of green plants is obtained based on the vegetation index distribution map, and prescription operations are performed on the target plot according to the growth classification map. This method enables prescription operations to be performed according to the growth of green plants, that is, the amount of operations is determined, thereby achieving real-time matching of prescriptions, avoiding excessive or insufficient operations, and solving the problem that the indiscriminate operation scheme of the prior art cannot match the growth of crops. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic diagram of a first embodiment of a prescription operation method provided in an embodiment of the present application;

[0018] Figure 2 A schematic diagram of a second embodiment of the prescription operation method provided in the embodiment of the present application;

[0019] Figure 3 A schematic diagram of a third embodiment of the prescription operation method provided in the embodiment of the present application;

[0020] Figure 4 This is a schematic diagram of an embodiment of the plant protection equipment in this application. DETAILED DESCRIPTION

[0021] The present application provides a prescription operation method, plant protection equipment and storage medium, which are used to generate a vegetation index distribution map of green plants in the target plot through the brightness distribution map of each channel of the image data of the target plot when fertilizing, spraying and other operations are performed on the target plot, and then the generated vegetation index distribution map is graded using the algorithm of natural break points to generate a growth classification map, and the prescription operation is performed based on the growth classification map. This method not only realizes the identification of the growth of green plants in the target plot, but also enables the matching of the corresponding fertilizer and spraying amounts based on the growth stage of the green plants corresponding to the growth during fertilization, spraying and other operations, so as to achieve the matching of the operation with the growth stage of the green plants, avoiding the problem of excessive application of pesticides and fertilizers in the indiscriminate operation of the prior art, which easily leads to pesticide residues and waste of agricultural materials, etc., and increases the cost of agricultural operations.

[0022] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0023] It is understandable that the present invention is mainly applied to unmanned equipment, such as drones, unmanned vehicles, etc., and is not specifically limited here. Any equipment that can realize unmanned operation on green plants in the target plot can be used. The operation may include fertilization, spraying, etc. In the embodiment of the present invention, the drone is the execution subject, and the drone is equipped with an image acquisition device. The image acquisition device can collect green plant information in the target plot in real time for analysis of subsequent prescription operations.

[0024] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In this embodiment, an embodiment of the prescription operation method includes:

[0025] 110. Obtain a global image of the target plot.

[0026] In this embodiment, the global image is an image including green plants and background, and the target plot can be a complete farmland or a part of the farmland. When obtaining the global image of the target plot, the drone can be used to fly to a certain height above the target plot and then take the photo, or the drone can be used to cruise the target plot based on a pre-planned cruise route and then stitch them together in sequence.

[0027] In practical applications, it is also possible to plan the acquisition of fixed-point images of the target plot in advance, then control the drone or image acquisition device to move to each fixed point to take images, and then stitch the images taken at each fixed point to obtain a global image of the target plot.

[0028] It should be noted that the global image includes at least one type of image, such as an RGB image and a multispectral image. Of course, it can also be a video, which is not limited here. When collecting RGB images and multispectral images, different cameras are used. The two cameras here are integrated on one device and are realized through different lenses. When collecting the two images, the camera controls different lenses at the same time to collect them.

[0029] 120. Extract pixel parameters of each channel in the global image and generate a pixel parameter distribution map of each channel.

[0030] In this embodiment, the pixel parameter is the brightness value of the spectral channel or the reflectivity of the spectral channel. Similarly, the generated pixel parameter distribution map can be a brightness distribution map or a reflectivity distribution map. Whether to generate a brightness distribution map or a reflectivity distribution map depends on the specific type of the global image.

[0031] If the global image is an RGB image, the channel images in the global image are separated by an image analysis algorithm, and the brightness value of each pixel in each channel image is extracted, thereby converting the channel image into a brightness distribution map.

[0032] It is understandable that each channel here can be a channel defined from the perspective of color, such as the R, G, and B channels of the three primary colors. The channel images of the R, G, and B channels are separated from the global image by color processing and then saved separately.

[0033] In practical applications, we can use OpenCV's cv2.split() function to decompose the global image into three independent color channels: R, G, and B. This function returns a tuple containing three NumPy arrays, corresponding to the R, G, and B channels. Further, we use OpenCV's cv2.imshow() function to display the data of the three channels, thereby extracting the corresponding channel image.

[0034] In this embodiment, when generating a brightness distribution map, after obtaining the channel images, the brightness values ​​of the pixels in each channel image are extracted, and a corresponding brightness distribution map is drawn according to the positions of the pixels and the corresponding brightness values.

[0035] It is understandable that the brightness distribution map can be expressed in the form of a histogram. After the brightness value of each pixel is extracted, the brightness values ​​of adjacent pixels are merged according to a preset brightness level to obtain a brightness level distribution, and then a brightness distribution map is drawn based on the level distribution. It should be noted that a brightness distribution map is drawn based on the image of each channel.

[0036] If the global image is a multispectral image, the reflectivity corresponding to each spectral channel in the multispectral image is extracted to obtain the reflectivity distribution map of each spectral channel of the entire image. It should be noted that if the reflectivity distribution map is generated, the global image only contains the multispectral image, and the reflectivity of each spectral channel in the multispectral image is extracted to generate the reflectivity distribution map of the target plot. The spectral channel can be three spectral channels divided by the three primary colors of R, G, and B.

[0037] 130. Generate a vegetation index distribution map of the target plot based on the pixel parameter distribution map of each channel.

[0038] In this embodiment, the vegetation index distribution map includes two types, one is an EXGR (extra red and ultra green difference index) index distribution map calculated based on brightness values, and the other is a multispectral vegetation index distribution map obtained based on multispectral images.

[0039] The choice of vegetation index distribution map is determined by the acquired global image. When the global image only contains RGB images, the EXGR index distribution map is selected as the vegetation index distribution map. If the global image includes both RGB images and multispectral images, any one of the two vegetation index distribution maps can be selected. Of course, it is preferred to generate a multispectral vegetation index distribution map.

[0040] It should be noted that when an EXGR index distribution map is generated, the brightness value of the same pixel in the brightness distribution map of each channel is extracted for combined calculation, and then the combined values ​​of all the calculated pixels are used to draw the EXGR index distribution map. Of course, the EXGR index distribution map can be one for each image. For example, when the global image is formed by stitching multiple sampling point images, it can be calculated for the brightness distribution map of each channel of each sampling point image to calculate an EXGR index distribution map, and then the EXGR index distribution maps are merged into one.

[0041] If a multispectral vegetation index distribution map is generated, the vegetation coverage is calculated through the brightness distribution map, and the multispectral vegetation index distribution map is generated based on the vegetation coverage.

[0042] 104. By analyzing the vegetation index distribution map, a growth classification map of green plants in the target plot can be obtained.

[0043] It should be noted that when analyzing the vegetation index distribution map, the vegetation index is graded based on the growth stages of various green plants in the target plot, and then the vegetation index of each pixel position in the vegetation index distribution map is extracted using the natural break point algorithm, and the index is classified and merged according to the classification to obtain a graded distribution, thereby outputting a growth classification map of green plants in the target plot.

[0044] 105. Prescribe treatment for the target plot based on the growth classification map.

[0045] After obtaining the growth grading map, the corresponding growth stage is determined using the level display in the growth grading map, and the corresponding operation parameters are configured based on the growth stage to control the unmanned equipment to operate on the target plot.

[0046] In this embodiment, it also includes pre-setting the correspondence between data such as the amount of fertilizer and the amount of spraying for different growth stages of different green plants, and then determining the correspondence based on the growth classification diagram corresponding to the vegetation index distribution diagram and the type of green plants, and controlling the operating device of the unmanned equipment to perform operations based on the data such as the amount of fertilizer and the amount of spraying recorded in the correspondence.

[0047] The technical solution provided in this embodiment generates a growth grading map of green plants through images of the target plots, and matches the corresponding fertilizer and spraying amounts based on the growth stages of the green plants corresponding to the growth in the growth grading map, so as to achieve matching of operations with the growth stages of the green plants, thereby avoiding the problems of indiscriminate operations in the prior art, such as excessive application of pesticides and fertilizers, which easily lead to pesticide residues and waste of agricultural materials, and increase the cost of agricultural operations.

[0048] In another feasible embodiment, if the global image only includes a multispectral image, it generates a pixel parameter distribution map of each channel, specifically extracting the reflectivity of the R, G, and B channels of each pixel point in the multispectral image to generate a reflectivity distribution map of the three channels of the target plot.

[0049] After the reflectivity distribution map is generated, the emissivity of the corresponding spectral channel at the same pixel point in the reflectivity distribution map of each channel is combined and calculated to obtain the vegetation index distribution map of the target plot, and finally the growth classification analysis is performed based on the vegetation index distribution map to guide the prescription operation. By calculating the growth of vegetation based on the vegetation index distribution map of the target plot, the vegetation is prescribed based on the growth, so that the operation corresponds to the vegetation and the growth, which not only improves the pertinence of the operation, but also reduces the unnecessary waste of resources during the operation.

[0050] In another feasible embodiment, if the global image includes at least an RGB image, the step of generating a pixel parameter distribution map of each channel is to generate a brightness distribution map of the three channels of the target plot by extracting the brightness values ​​of the R, G, and B channels of each pixel point in the RGB image.

[0051] It should be noted that a brightness distribution map is generated for each channel, and then the brightness values ​​of the same pixel point in the brightness distribution map of each channel are combined and calculated to obtain the vegetation index distribution map of the target plot. Finally, the growth classification analysis is performed based on the vegetation index distribution map. After the growth classification is completed, according to the actual operation scene, differentiated operation parameters are set for different grade areas in the growth classification map to perform prescription operations. This operation control method greatly reduces the resources required for operation and the operating cost of operation control.

[0052] In another embodiment of the present application, when the global image includes both an RGB image and a multispectral image, the vegetation index distribution map of the target plot is first calculated using the RGB image; then the green plant coverage of the target plot is calculated based on the vegetation index distribution map; the type of the multispectral vegetation index distribution map is determined based on the green plant coverage, and a multispectral vegetation index distribution map corresponding to the type is generated based on the multispectral image.

[0053] It should be noted that the coverage of vegetation will be different under different growth conditions. The corresponding type of multispectral vegetation index distribution map is selected according to the green plant coverage for the subsequent generation of growth grading maps to improve the division of growth grading, which is more conducive to the subsequent prescription operations based on the growth grading maps, and further improves the accuracy of the operations and the matching degree with the growth condition.

[0054] In this embodiment, the green plant coverage of the target plot can be calculated from two directions. One direction is to use the global map as an object for calculation, and the other direction is to use the local map as an object for calculation.

[0055] If the global map is used as the object, the green coverage of the target plot is obtained by calculating the ratio between the number of pixels in the green plant area and the number of all pixels in the vegetation index distribution map.

[0056] If a local map is taken as the object, first determine multiple local areas in the global map; then calculate the regional green plant coverage corresponding to the local area according to the ratio between the number of green plant area pixels corresponding to each local area and the number of all pixels in the area in the vegetation index distribution map, and obtain multiple regional green plant coverages; finally, determine the average value of the green plant coverages of the multiple areas, and use the average value as the green plant coverage of the target plot.

[0057] In another embodiment, a multispectral vegetation index distribution map is generated based on green plant coverage and a multispectral image by judging whether the green plant coverage is less than a preset coverage threshold; if so, determining that the type of the multispectral vegetation index distribution map is a first type, and generating a first multispectral vegetation index distribution map based on the multispectral image; if not, determining that the type of the multispectral vegetation index distribution map is a second type, and generating a second multispectral vegetation index distribution map based on the multispectral image.

[0058] Among them, the first type may be an NDVI (Normalized Difference Ratio Vegetation Index) index distribution map, and the second type may be an NDRE (Normalized Difference Ratio Vegetation Index Red Edge) index distribution map.

[0059] By using the calculated green plant coverage to generate a vegetation index distribution map that is more suitable for the current vegetation growth stage, the constructed vegetation index distribution map can more accurately present the growth of vegetation, thereby ensuring the accuracy of subsequent growth classification and improving the accuracy of the growth classification map in guiding prescription operations.

[0060] In another embodiment, the analyzing the vegetation index distribution map to obtain the growth classification map of the green plants in the target plot includes: analyzing the first multispectral vegetation index distribution map or the second multispectral vegetation index distribution map using a natural breakpoint algorithm to obtain the growth classification map of the green plants in the target plot.

[0061] Analyzing the vegetation index distribution map by means of natural break points can not only achieve hierarchical extraction of information in the map, but also reduce the factors caused by human intervention that affect the accuracy of growth classification.

[0062] In another embodiment, before analyzing the vegetation index distribution map to obtain the growth classification map of the green plants in the target plot, it also includes: binarizing the vegetation index distribution map to obtain a vegetation index distribution map without background. This removes the influence of the background on the growth classification analysis, improves the analysis accuracy of vegetation growth, and provides accurate guidance for subsequent differential operations.

[0063] Reference Figure 2 This embodiment takes the global image as an RGB image as an example to propose another embodiment of the prescription operation method, and the method includes the following steps:

[0064] 210. Obtain an RGB image of the target plot.

[0065] It should be noted that the RGB image is a global RGB image that includes both green plants and background. It is obtained by remotely controlling a remote sensing drone equipped with an RGB camera to fly to a certain altitude above the target plot or to fly to various sampling positions on the target plot in sequence.

[0066] When the drone is remotely controlled to take images over the target plot, the drone is first controlled to fly to the target plot at the default altitude, and then the target plot image is sampled and the sampled image is identified to see whether it meets the requirements for calculating the vegetation index distribution map. The requirement here is to be able to clearly identify the leaves of the green plants in the image. If the sampled image does not meet the requirements, the drone's flight altitude is controlled until it meets the requirements. After the requirements are met, the image of the target plot is collected through the RGB camera.

[0067] When controlling the RGB camera to capture the RGB image of the target plot, a panoramic image can be captured by continuous shooting or by setting multiple points of interest. When acquiring by setting multiple points of interest, the points of interest of the target plot are first set according to the image stitching logic and the maximum shooting range of the RGB image, and then the drone is controlled to fly to each point of interest to capture one or more images, and then the images of the multiple points of interest captured are stitched according to the positions of the points of interest to obtain a global RGB image of the target plot.

[0068] 220. Extract the brightness distribution map of each channel of R, G, and B in the RGB image.

[0069] In this embodiment, by extracting the brightness values ​​of the R, G, and B channels of each pixel in the RGB image, a brightness distribution map of the three channels of the target land parcel is generated.

[0070] 230. Generate an EXGR index distribution map based on the brightness distribution map of each RGB channel.

[0071] The EXGR index distribution map is a collection of EXGR indices of all pixels in the image of the target plot. Specifically, the vegetation index distribution map of the target plot (i.e., the EXGR index distribution map) is obtained by combining and calculating the brightness values ​​of the same pixel in the brightness distribution maps of each channel.

[0072] The EXGR value of each pixel in the distribution map is calculated by combining the brightness values ​​of the R, G, and B channels. The formula for the combined calculation is as follows:

[0073]

[0074] Among them, R, G, and B in the formula correspond to the brightness values ​​of the R, G, and B channels respectively.

[0075] 240. Remove the non-vegetation background in the EXGR index distribution map to obtain the EXGR index distribution map of green plants.

[0076] In this embodiment, the background in the EXGR index distribution map is removed by binarization processing to obtain the EXGR index distribution map of the green plants.

[0077] It is understandable that a threshold is set to filter out the non-vegetation background in the EXGR index distribution map, and the EXGR index distribution map of green plants is output. Here, the threshold is set to 0.05. The part with an index greater than the threshold is green plants, and the part not greater than the threshold is the background.

[0078] 250. The natural breakpoint algorithm is used to classify the EXGR index distribution map after filtering out the non-vegetation background to generate a growth classification map.

[0079] 260. After grading, the operator manually sets differentiated operating parameters for different grade areas in the growth grading diagram according to the actual operating scenario and performs prescription operations.

[0080] The implementation principles of steps 250 and 260 may be implemented in the manner of steps 140 and 150 described above, and will not be repeated here.

[0081] In this embodiment, when analyzing the growth of green plants in the target plot, the growth parameters of the green plants are determined by analyzing the vegetation index in the index distribution map, thereby determining the growth stage of the corresponding green plants.

[0082] The specific growth parameters include plant height, green plant coverage, vegetation index, etc. The growth parameters are used to indicate the growth of the plants. The plants are graded according to the growth parameters to obtain a growth classification diagram.

[0083] Taking rice as an example, the growth parameters of rice include plant height, green plant coverage, and vegetation index. Among them, green plant coverage is used to characterize leaf area, vegetation index characterizes leaf color, and plant height layer is used to characterize plant height. By analyzing the vegetation index of each pixel point, the leaf area, leaf color, plant height, etc. of the green plants are evaluated based on the vegetation index to determine the growth of the green plants. After obtaining the growth of the green plants at each pixel point, the growth is classified and merged according to the level, so as to obtain the growth classification map of the green plants in the target plot.

[0084] In another embodiment, when a prescription operation is performed based on a reflectance distribution map generated based on a global image, the current growth stage of the plant is determined by analyzing the reflectance of light in each spectral channel in the reflectance distribution map and the distribution of light in the corresponding spectral channel in the image, and a prescription is determined based on the growth stage to achieve subsequent spraying or spreading operations.

[0085] The technical solution provided in this embodiment generates an EXGR index distribution map by extracting the brightness distribution map of each channel in the RGB image, and then performs a growth grading analysis based on the EXGR index distribution map to obtain a growth grading map of green plants in the target plot. Prescription operations are performed based on the growth grading map, thereby achieving a match between the amount of fertilizer or spraying during the operation and the current state of the green plants, solving the problem of matching the operation control of the prior art with the growth of crops, resulting in poor operation results and waste of resources.

[0086] Reference Figure 3 In this embodiment, a global image including both an RGB image and a multispectral image is taken as an example to explain in detail the prescription operation method proposed in this application. The method includes:

[0087] 310. Obtain an RGB image and a multispectral image of the target land parcel.

[0088] 320. Extract the brightness value of each channel in the RGB image and generate a brightness distribution map of each channel.

[0089] 330. Generate a vegetation index distribution map of the target plot based on the brightness distribution map of each channel.

[0090] The implementation principles of steps 310 - 330 in this embodiment are the same as the implementation principles of steps 110 - 130 in the above embodiment, and are not repeated here.

[0091] In practical applications, the remote sensing drone is controlled to fly to various points of interest on the target plot, and the RGB camera on the remote sensing drone is used to randomly and evenly collect multiple single-point low-altitude images to obtain images of the points of interest. Specifically, the RGB camera is controlled to capture the RGB image of the target plot at a collection height of 5 to 10 meters close to the crop. Of course, the collection height is specifically set based on the standard of being able to clearly identify the crop leaves through the image. While capturing the RGB image, the multispectral camera is also controlled to capture the multispectral image of the target plot, thereby obtaining multiple RGB images and multiple multispectral images.

[0092] The brightness distribution maps of the R, G, and B channels in the images of each point of interest are extracted and spliced ​​to obtain complete brightness distribution maps of each channel.

[0093] Furthermore, the collected multiple multispectral images are also spliced ​​to generate a global reflectance map of each multispectral channel.

[0094] It should be noted that the vegetation index distribution map in step 330 is an EXGR (super red super green difference index) index distribution map, which is specifically generated based on the brightness distribution map of each RGB channel. The EXGR value of each pixel is calculated by combining the brightness values ​​of the R, G, and B channels. The calculation formula is as follows:

[0095]

[0096] 340. Calculate the green plant coverage of the target plot based on the vegetation index distribution map.

[0097] Specifically, the green coverage of the target plot is obtained by calculating the ratio between the number of pixels in the green plant area and the number of all pixels in the vegetation index distribution map (EXGR index distribution map).

[0098] It should be noted that the number of pixels in the green plant area refers to the total number of pixels belonging to green plants in the RGB image of the target plot, which can be obtained by filtering by setting a threshold of the vegetation index.

[0099] For example, when the threshold is set to 0.05, each EXGR index distribution map is binarized with 0.05 as the threshold to obtain the EXGR index distribution map of green plants, and then the number of valid pixels after binarization is extracted to obtain the number of pixels in the green plant area. During the binarization process, those greater than 0.05 are green plants, and those not greater than 0.05 are directly eliminated.

[0100] After obtaining the number of pixels in the green plant area and the number of all pixels, the green plant coverage is calculated. The calculation formula is as follows:

[0101] Green plant coverage = number of pixels in green plant area / total number of pixels.

[0102] In this embodiment, since the image of the target plot is obtained by setting multiple points of interest, multiple images will be obtained, and correspondingly there will be multiple exponential distribution maps. When calculating the vegetation coverage, it is necessary to average all the green plant coverage calculated based on the interest point images, and the average value is used as the green plant coverage of the target plot.

[0103] 350. Determine the type of the multispectral vegetation index distribution map based on the green plant coverage, and generate a multispectral vegetation index distribution map corresponding to the type based on the multispectral image.

[0104] It should be noted that there are two types of multispectral vegetation index distribution maps, which are specifically distinguished by setting the boundary of green plant coverage, that is, by judging whether the green plant coverage is less than a preset coverage threshold; if so, the type of the multispectral vegetation index distribution map is determined to be the first type, and a first multispectral vegetation index distribution map is generated based on the multispectral image; if not, the type of the multispectral vegetation index distribution map is determined to be the second type, and a second multispectral vegetation index distribution map is generated based on the multispectral image.

[0105] In practical applications, the first multispectral vegetation index distribution map is an NDVI index distribution map, and the second multispectral vegetation index distribution map is an NDRE index distribution map. Select and generate a suitable multispectral vegetation index map according to the green plant coverage value. When the green plant coverage is less than 80%, select to generate an NDVI (normalized ratio vegetation index) index distribution map; when the green plant coverage is ≥80%, select to generate an NDRE (red edge normalized ratio vegetation index) index distribution map.

[0106] Among them, the formulas for generating the NDVI index distribution map and the NDRE index distribution map are:

[0107]

[0108] Where NRI is the red light normalization value,

[0109] The NDVI index distribution map is closely related to parameters such as leaf area index, plant canopy background, vegetation coverage, biomass, and crop growth, and can reflect the health of vegetation and the growth of vegetation; it is linearly related to the spatial distribution density of crops. For crops that grow normally, it is usually positive, and increases with the increase of vegetation coverage. The crop area with good growth has a larger value. At the same time, this index can eliminate most of the irradiance changes related to remote sensing instrument calibration, sensor observation angle, solar altitude angle, cloud influence and atmospheric conditions, enhance the ability to respond to vegetation, and is widely used in resource environment and agricultural remote sensing. However, the NDVI index distribution map has low sensitivity to high vegetation areas. The index value is distributed in the range of -1 to 1. If NDVI <0, it means that the ground is covered with clouds, water, snow, etc., and it is highly reflective of visible light; if NDVI = 0, it means rock and bare soil, and nir is approximately equal to red; if NDVI > 0, it means that there is vegetation coverage and it is proportional.

[0110] Meanwhile, the index values ​​in the NDRE index distribution graph are distributed in the range of -1 to 1, where -1 to 0.2 indicates bare soil or developing crops; 0.2 to 0.6 can be interpreted as unhealthy plants or immature crops; 0.6 to 1 is a good value indicating healthy, mature, and mature crops; and in the middle and late stages of crop growth, values ​​below 0.6 almost certainly indicate some form of crop damage.

[0111] 360. The multispectral vegetation index distribution map is analyzed using the natural breakpoint algorithm to obtain a growth classification map of green plants in the target plot.

[0112] Specifically, the first multispectral vegetation index distribution map or the second multispectral vegetation index distribution map is analyzed using a natural breakpoint algorithm to obtain a growth classification map of green plants in the target plot.

[0113] 370. Prescribe treatment for the target plot based on the growth classification map.

[0114] Based on the growth grading map and the actual operation scenarios of the target plots, differentiated operation parameters are set for different grade areas in the growth grading map to carry out prescription operations.

[0115] In the technical solution provided in this embodiment, RGB images are combined with multispectral images to generate an EXGR index distribution map, and the green plant coverage is calculated based on the EXGR index distribution map, and a multispectral vegetation index map more suitable for the current stage is selected to generate a growth potential classification map, and the prescription operation is guided based on the growth potential classification map. The combination of RGB images and multispectral images is more reasonable for the selection of vegetation index, and also solves the problem that the indiscriminate operation scheme of the prior art cannot match the growth potential of crops.

[0116] Figure 4 4 is a schematic diagram of the structure of a plant protection device provided in an embodiment of the present application. The plant protection device 400 includes a memory 420 and a processor 410. Instructions are stored in the memory 420. The plant protection device 400 may have relatively large differences due to different configurations or performances. It may include one or more processors (central processing units, CPU) 410 (for example, one or more processors) and a memory 420, and one or more storage media 430 (for example, one or more mass storage devices) storing application programs 433 or data 432. Among them, the memory 420 and the storage medium 430 can be short-term storage or permanent storage. The program stored in the storage medium 430 may include one or more modules (not shown in the figure), and each module may include a series of computer-readable instruction operations in the plant protection device 400. Furthermore, the processor 410 may be configured to communicate with the storage medium 430, and execute a series of computer-readable instruction operations in the storage medium 430 on the plant protection device 400. When the computer-readable instructions are executed by the processor, the processor executes the steps of the prescription operation method described in the above embodiments.

[0117] The plant protection device 400 may further include one or more power supplies 440, one or more wired or wireless network interfaces 450, one or more input and output interfaces 460, and / or one or more operating systems 431, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be appreciated by those skilled in the art that Figure 4 The structure of the plant protection device shown does not constitute a limitation of the plant protection device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0118] The present application also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium, and when the instructions are executed on a computer, the computer executes the steps of the prescription operation method.

[0119] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0120] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program code.

[0121] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A prescription operation method, applied to plant protection equipment, characterized in that: The method comprises: Get a global image of the target plot; Extracting pixel parameters of each channel in the global image and generating a pixel parameter distribution map of each channel, wherein the pixel parameter is a brightness value of a spectral channel or a reflectivity of a spectral channel; Generate a vegetation index distribution map of the target plot based on the pixel parameter distribution map of each channel; Analyze the vegetation index distribution map to obtain a growth classification map of green plants in the target plot; Prescription operations are carried out on the target plot based on the growth classification map.

2. The prescription operation method according to claim 1, characterized in that: The global image only includes the multispectral image; extracting the pixel parameters of each channel in the global image to generate a pixel parameter distribution map of each channel includes: The reflectivity of each spectral channel in the multispectral image is extracted to generate a reflectivity distribution map of the target land parcel.

3. The prescription operation method according to claim 1, characterized in that: The global image includes an RGB image; extracting pixel parameters of each channel in the global image to generate a pixel parameter distribution map of each channel includes: The brightness values ​​of the R, G, and B channels of each pixel in the RGB image are extracted to generate a brightness distribution map of the three channels of the target plot.

4. The prescription operation method according to claim 3, characterized in that: The generating of the vegetation index distribution map of the target plot based on the pixel parameter distribution map of each channel includes: The brightness value of the same pixel point in the brightness distribution map of each channel is combined and calculated to obtain the vegetation index distribution map of the target plot.

5. The prescription operation method according to claim 3, characterized in that: The global image also includes a multispectral image; before analyzing the vegetation index distribution map to obtain the growth classification map of the green plants in the target plot, it also includes: Calculating the green plant coverage of the target plot based on the vegetation index distribution map; The type of the multispectral vegetation index distribution map is determined based on the green plant coverage, and a multispectral vegetation index distribution map corresponding to the type is generated based on the multispectral image.

6. The prescription operation method according to claim 5, characterized in that: The calculating the green plant coverage of the target plot based on the vegetation index distribution map includes: Calculate the ratio between the number of pixels in the green area and the number of all pixels in the vegetation index distribution map to obtain the green coverage of the target plot; or Determine multiple local areas in the global map; calculate the regional green plant coverage corresponding to the local area according to the ratio between the number of green plant area pixels corresponding to each local area and the number of all pixels in the area in the vegetation index distribution map, and obtain multiple regional green plant coverages; determine the average value of the green plant coverages of the multiple areas, and use the average value as the green plant coverage of the target plot.

7. The prescription operation method according to claim 5, characterized in that: The determining the type of the multispectral vegetation index distribution map based on the green plant coverage, and generating a multispectral vegetation index distribution map corresponding to the type based on the multispectral image, comprises: Determine whether the green plant coverage is less than a preset coverage threshold; If yes, determining that the type of the multispectral vegetation index distribution map is a first type, and generating a first multispectral vegetation index distribution map based on the multispectral image; If not, it is determined that the type of the multispectral vegetation index distribution map is the second type, and a second multispectral vegetation index distribution map is generated based on the multispectral image.

8. The prescription operation method according to claim 7, characterized in that: The step of analyzing the vegetation index distribution map to obtain a growth classification map of green plants in the target plot includes: The first multispectral vegetation index distribution map or the second multispectral vegetation index distribution map is analyzed by using a natural breakpoint algorithm to obtain a growth classification map of green plants in the target plot.

9. The prescription operation method according to any one of claims 1 to 8, characterized in that: Before analyzing the vegetation index distribution map to obtain the growth classification map of green plants in the target plot, the method further includes: The vegetation index distribution map is binarized to obtain a vegetation index distribution map without background.

10. A plant protection device, characterized in that: The plant protection device comprises a mobile platform, a spreading device, a memory and at least one processor, wherein instructions are stored in the memory; The at least one processor calls the instructions in the memory to enable the plant protection device to execute the prescription operation method according to any one of claims 1 to 9.

11. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are read and executed, the prescription operation method according to any one of claims 1 to 9 is executed.

Images