Method, storage medium and device for monitoring crop moisture status based on drone

Abstract

The present application provides a method, storage medium and equipment for monitoring crop moisture status based on drones, which are used to accurately monitor the moisture status of field crops. On the one hand, it can reliably describe the growth status of crops, and on the other hand, it can provide effective support for field moisture management and the formulation of irrigation systems. Traditional field sensor measurement methods require a lot of manpower and material resources and the accuracy is affected by subjective factors. The method of high-throughput monitoring using drones can not only liberate labor, but also effectively improve monitoring efficiency. Agricultural water conservation is of great significance to my country's current agricultural production. Accurately and efficiently monitoring crop moisture status is part of achieving precision irrigation and agricultural water conservation. This method will play a significant role in monitoring crop moisture status. The purpose of the present invention is to provide a vegetation index that accurately describes the moisture status of crops, which can quickly evaluate the moisture status of field crops at low cost and high throughput.

Description

Technical Field

[0001] The present invention relates to the technical field of remote sensing monitoring of crop drought, and in particular to a method, storage medium and device for monitoring crop moisture status based on an unmanned aerial vehicle. Background Art

[0002] Crop water status (CWS) is a key indicator during the crop growth period and is usually characterized by vegetation water content (VMC) or soil water content (SMC). CWS can effectively describe the water supply and demand in the field and can be further used to formulate irrigation plans. Accurate monitoring of CWS is of great significance to improving agricultural water use efficiency and agricultural production. In the past few decades, scholars have proposed many methods to evaluate CWS, such as time domain, frequency domain reflectometers, neutron measurement, negative pressure gauges, etc. However, these field measurement methods are usually sensitive to the method and location of field installation. When it comes to monitoring large areas, these measurement methods are inefficient and it is difficult to capture the spatial heterogeneity of CWS in the field.

[0003] With the development of various sensors in recent years, scholars have proposed hundreds of drought indices based on these sensor observations to characterize the moisture status of regional or field crops. These indices can generally be divided into station-based and remote sensing-based indices. Station-based indices are usually constructed using information such as temperature and precipitation observed at meteorological stations, such as the Standardized Precipitation Index (SPI), the Palmer Drought Severity Index (PDSI), and the Standardized Precipitation Evaporation Index (SPEI). Remote sensing-based indices usually use vegetation indices calculated by sensors equipped in satellites, aircraft, or unmanned aerial vehicles (UAVs) to indirectly characterize field CWS. These sensors mainly include RGB, multispectral, hyperspectral, thermal infrared, and radar sensors. Typical vegetation indices include the vegetation condition index (VCI) calculated using the red band and near infrared band of a multispectral sensor, the normalized difference water index (NDWI) calculated using the near infrared band and shortwave infrared band of a multispectral remote sensor, and the normalized multiband drought index (NMWI) calculated using the 1640nm and 2130nm bands of a hyperspectral sensor. Both site-based and remote sensing-based indices are reliable and widely used for characterizing CWS and monitoring drought. In particular, the remote sensing-based vegetation index can provide the spatial distribution of field moisture conditions and has higher efficiency and lower cost than on-site measurement methods.

[0004] However, these indices all have certain limitations, which are mainly reflected in the following aspects:

[0005] (1) The accuracy of the site-based index depends mainly on the density of observation sites and the method of spatial interpolation. For field scales with small areas and small climate differences, the site-based index cannot capture the spatial heterogeneity of field CWS.

[0006] (2) Remote sensing-based indices have poor ability to characterize the temporal characteristics of CWS and are easily affected by spectral saturation effects. Summary of the invention

[0007] In view of the deficiencies in the prior art, the present invention provides a method, storage medium and device for monitoring crop moisture status based on unmanned aerial vehicles, which can accurately characterize the crop moisture status in the field and its temporal and spatial distribution characteristics.

[0008] To achieve the above object, the present invention adopts the following technical solutions:

[0009] A method for monitoring crop moisture status based on drones, comprising the following steps:

[0010] S1. Use the multispectral and thermal infrared sensors carried by the drone to obtain long-term multispectral images and thermal infrared images of the target field; at the same time, obtain the temperature Ta of the target field observed by the meteorological station;

[0011] S2. Use the georeferencing tool of ArcGIS software to georeference the multispectral image and thermal infrared image acquired by the drone, that is, correct the position offset of the two images; then use the gray card and ASD to perform radiometric calibration on the georeferenced multispectral image, so that the DN value in the multispectral image is converted into reflectivity;

[0012] S3, using the resampling tool in ArcGIS software to downsample the multispectral image and thermal infrared image to the same spatial resolution;

[0013] S4, using multispectral images to calculate the normalized vegetation index NDVI of the target field for a long time series to obtain an NDVI image, and using thermal infrared images to calculate the surface temperature Ts of the target field for a long time series to obtain a Ts image;

[0014] S5, putting the NDVI value, Ts value and the temperature value Ta of the corresponding period into a three-dimensional rectangular coordinate system for each period of NDVI image and Ts image pixel by pixel to construct a three-dimensional feature space of Ts-NDVI-Ta;

[0015] S6. In the three-dimensional feature space of Ts-NDVI-Ta, select a certain NDVI value and a certain Ta value, and extract the maximum surface temperature Ts value under the corresponding conditions, that is, Ts maxAt the same time, in the three-dimensional feature space of Ts-NDVI-Ta, a certain NDVI value and a certain Ta value are selected to extract the minimum surface temperature Ts value under the corresponding conditions, that is, Ts min ;

[0016] S7, select the NDVI value, Ta value and Ts extracted under the corresponding conditions max The selected NDVI value, Ta value and Ts under the corresponding conditions are fitted accordingly to obtain the dry surface. min The values ​​are fitted accordingly to obtain the wet surface; the fitting method is based on a random forest regression algorithm and is implemented through the sklearn library of Python language;

[0017] S8. Determine the three-dimensional drought index TDDI using the Ts image and the corresponding dry surface and wet surface, and judge the moisture status of the crop based on the three-dimensional drought index TDDI.

[0018] To optimize the above technical solutions, the specific measures taken also include:

[0019] Furthermore, in step S2, the calculation formula for radiometric calibration of the geo-referenced multispectral image using the gray plate and ASD is as follows:

[0020]

[0021] In the formula, R i is the reflectance of the i-th band of the multispectral image; DN i is the gray value of the i-th band of the multispectral image; DN i_Board , R i_Board They are the grayscale value of the i-th band of the multispectral image measured by the gray plate and the actual reflectivity of the i-th band of the multispectral image measured by ASD.

[0022] Furthermore, in step S4, the normalized vegetation index NDVI of the target field long time series is calculated using multispectral images. The specific calculation formula is:

[0023]

[0024] Where NIR is the near infrared band in the multispectral image; R is the reflectance of the red band in the multispectral image, which is obtained from the data R i Obtained in.

[0025] Furthermore, in step S4, the specific calculation formula for calculating the long-term surface temperature Ts of the target field using thermal infrared images is:

[0026]

[0027] Where DN′ is the gray value of the thermal infrared image; DN Board Ts is the grayscale value of the thermal infrared image of the blackboard arranged on the ground of the target field; Board A blackboard is arranged for the temperature sensor to observe the temperature of the target field ground; and the maximum 5% and minimum 5% pixels of each thermal infrared image are removed as abnormal values.

[0028] Further, in step S8, the specific calculation formula for determining the three-dimensional drought index TDDI using the Ts image and the corresponding dry surface and wet surface is:

[0029]

[0030] Where, Ts i is the surface temperature value of pixel i in Ts image; Ts maxi is the maximum surface temperature value corresponding to the NDVI value and Ta value of Ts image pixel i; mini is the minimum surface temperature value corresponding to the NDVI value and Ta value of Ts image pixel i.

[0031] Furthermore, in step S8, the specific content of judging the moisture status of crops based on the three-dimensional drought index TDDI is:

[0032] The value of the three-dimensional drought index TDDI is between 0 and 1. The closer the value is to 1, the higher the moisture content of the crop and the better the moisture status; conversely, the worse the moisture status.

[0033] A computer-readable storage medium stores a computer program, wherein the computer program enables a computer to execute any of the above methods for monitoring crop moisture conditions.

[0034] An electronic device comprises: a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for monitoring the moisture status of crops as described in any one of the above items is implemented.

[0035] The beneficial effects of the present invention are:

[0036] The method of the present invention mainly constructs a method for monitoring crop moisture status based on drones, which mainly provides a new vegetation index based on drone observation, which can accurately characterize the crop moisture status in the field and its spatiotemporal distribution characteristics. Compared with traditional ground observations, the method of the present invention greatly improves the monitoring efficiency at the field scale; compared with traditional remote sensing vegetation indices and site-based indices, the method of the present invention can more accurately describe the spatiotemporal distribution characteristics of field moisture status. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a schematic diagram of the overall solution flow of the present invention.

[0038] Figure 2 It is a schematic diagram of a blackboard for thermal infrared calibration of the target field ground layout of the present invention.

[0039] Figure 3 It is a schematic diagram of the surface temperature image and a schematic diagram of outlier removal of the present invention.

[0040] Figure 4 is a schematic diagram of the three-dimensional space of Ts-NDVI-Ta of the present invention, wherein Figure 4 (a) is a three-dimensional scatter plot of Ts-NDVI-Ta; Figure 4 (b) is a schematic diagram of the Ts-NDVI-Ta feature space based on a three-dimensional scatter plot, i.e., a schematic diagram of a dry surface and a wet surface; Figure 4 (c) is the two-dimensional scatter plot of Ts-NDVI, which is also the side projection in (a); Figure 4 (d) is the two-dimensional scatter plot of Ts-Ta, which is the orthographic projection in (a); Figure 4 (e) is the Ta-NDVI two-dimensional scatter plot, which is the top-down projection in (a).

[0041] Figure 5 It is a schematic diagram of the calculation results of the three-dimensional drought index TDDI of the target field at different periods of the present invention.

[0042] Figure 6 It is a schematic diagram of the correlation between TDDI and vegetation water content of the present invention. DETAILED DESCRIPTION

[0043] The present invention will now be described in further detail with reference to the accompanying drawings.

[0044] refer to Figure 1 , the overall scheme flow diagram of the present invention comprises the following steps:

[0045] A. Use the multispectral and thermal infrared sensors carried by drones to obtain long-term multispectral and thermal infrared multi-temporal images of the target fields.

[0046] B. Use the georeferencing tool of ArcGIS software to georeference the images acquired by the drone, that is, correct the offset of the image position.

[0047] C. Use gray plates and ASD to calibrate the multispectral images acquired by the drone (Formula 1), that is, convert the DN value of the original image into actual reflectivity.

[0048]

[0049] Where: R i DNi is the reflectance and DN value of the i-th band of the image; DN i_Borad , R i_Borad is the DN value of the gray card and the actual reflectivity of the gray card measured by ASD.

[0050] D. Obtain the temperature Ta of the target field observed by the weather station.

[0051] E. Use the resampling tool in ArcGIS to downsample the multispectral and thermal infrared images to the same spatial resolution, and use the nearest neighbor method for downsampling.

[0052] F. Use multispectral images to calculate the normalized vegetation index NDVI image of the target field for a long time series, and use the GDAL library in Python language to perform batch calculations on multiple images. The calculation method is shown in Formula 2:

[0053]

[0054] Where: NIR and R are the reflectances of the near-infrared band and the red band in the multispectral image, respectively.

[0055] G. Use thermal infrared images to calculate the long-term surface temperature image Ts of the target field, and use the GDAL library in Python language to perform batch calculations on multiple images. The calculation method is shown in Formula 3:

[0056]

[0057] Where: DN is the gray value of the thermal infrared image; DN Board Arrange a blackboard for the floor (reference Figure 2 ) thermal infrared image gray value; Ts Board The temperature of the blackboard arranged on the ground observed by the temperature sensor. The characteristics of the blackboard are: metal material, black color, and a temperature sensor to monitor the temperature of the blackboard in real time. The maximum 5% and minimum 5% of pixels in each image are removed as abnormal values ​​(reference Figure 3 ).

[0058] H. Put the long-term NDVI, Ts, pixel by pixel and the corresponding date Ta into a three-dimensional rectangular coordinate system to construct the three-dimensional feature space of Ts-NDVI-Ta, which is implemented by the GDAL library in Python (refer to Figure 4 In this coordinate system, the horizontal axis can be set to Ta, the vertical axis can be set to NDVI, and the vertical axis can be set to Ts).

[0059] I. Extract different NDVI, Ta and the corresponding maximum Ts and minimum Ts according to the three-dimensional feature space of Ts-NDVI-Ta, that is, Ts max and Ts min(refer to Figure 4 ,in Figure 4 In (a), a point value is selected on the horizontal axis to draw a vertical line, and a point value is selected on the vertical axis to draw a vertical line, and then an intersection is formed on the NDVI-Ta plane. Based on the intersection, a straight line is drawn in the vertical direction. The straight line corresponds to multiple point values. Figure 4 (b), where the maximum value falls on the dry surface and the minimum value falls on the wet surface. Similarly, multiple maximum and minimum values ​​are obtained and fitted into dry and wet surfaces)

[0060] J. Fitting Ts max The relationship between NDVI and Ta, i.e. dry surface; fitting Ts min and NDVI, a function of Ta, i.e., wet surface; the fitting method uses the random forest regression algorithm, implemented by the sklearn library in Python, with the parameter n_estimator set to 15 and the rest of the parameters as default (refer to Figure 4 ).

[0061] K. The three-dimensional drought index TDDI is calculated using the Ts image and the corresponding dry and wet surfaces (ref. Figure 5 ), which is calculated as Equation 4:

[0062]

[0063] Where, T si is the surface temperature Ts of pixel i; Ts maxi and Ts mini are the maximum and minimum surface temperatures corresponding to the NDVI and Ta of pixel i, i.e., Ts max and Ts min , calculated by the random forest model established in the previous step. The value of TDDI is between 0 and 1. The closer the value is to 1, the higher the water content and the better the water status; conversely, the worse the water status (refer to Figure 6 ).

[0064] Furthermore, the present application also provides: a computer-readable storage medium storing a computer program, wherein the computer program enables a computer to execute any of the above methods for monitoring crop moisture conditions. An electronic device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for monitoring crop moisture conditions as described in any of the above is implemented.

[0065] It should be noted that the terms such as "upper", "lower", "left", "right", "front", "back", etc. cited in the invention are only for the convenience of description and are not used to limit the scope of implementation of the present invention. Changes or adjustments in their relative relationships should be regarded as the scope of implementation of the present invention without substantially changing the technical content.

[0066] The above are only preferred embodiments of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should be regarded as the protection scope of the present invention.

Claims

1. A method for monitoring crop moisture status based on drones, characterized in that: The following steps are involved: S1. Use the multispectral and thermal infrared sensors carried by the drone to obtain long-term multispectral images and thermal infrared images of the target field; at the same time, obtain the temperature Ta of the target field observed by the meteorological station; S2. Use the georeferencing tool of ArcGIS software to georeference the multispectral image and thermal infrared image acquired by the drone, that is, correct the position offset of the two images; then use the gray card and ASD to perform radiometric calibration on the georeferenced multispectral image, so that the DN value in the multispectral image is converted into reflectivity; S3, using the resampling tool in ArcGIS software to downsample the multispectral image and thermal infrared image to the same spatial resolution; S4, using multispectral images to calculate the normalized vegetation index NDVI of the target field for a long time series to obtain an NDVI image, and using thermal infrared images to calculate the surface temperature Ts of the target field for a long time series to obtain a Ts image; S5, putting the NDVI value, Ts value and the temperature value Ta of the corresponding period into a three-dimensional rectangular coordinate system for each period of NDVI image and Ts image pixel by pixel to construct a three-dimensional feature space of Ts-NDVI-Ta; S6. In the three-dimensional feature space of Ts-NDVI-Ta, select a certain NDVI value and a certain Ta value, and extract the maximum surface temperature Ts value under the corresponding conditions, that is, Ts max At the same time, in the three-dimensional feature space of Ts-NDVI-Ta, a certain NDVI value and a certain Ta value are selected to extract the minimum surface temperature Ts value under the corresponding conditions, that is, Ts min ; S7, select the NDVI value, Ta value and Ts extracted under the corresponding conditions max The values ​​are fitted accordingly to obtain the dry surface; The selected NDVI value, Ta value and Ts under the corresponding conditions min The values ​​are fitted accordingly to obtain the wet surface; the fitting method is based on a random forest regression algorithm and is implemented through the sklearn library of Python language; S8. Determine the three-dimensional drought index TDDI using the Ts image and the corresponding dry surface and wet surface, and judge the moisture status of the crop based on the three-dimensional drought index TDDI.

2. The method for monitoring crop moisture status based on an unmanned aerial vehicle according to claim 1, characterized in that: In step S2, the calculation formula for radiometric calibration of the geo-referenced multispectral image using the gray plate and ASD is as follows: In the formula, R i is the reflectance of the i-th band of the multispectral image; DN i is the gray value of the i-th band of the multispectral image; DN i_Board , R i_Board They are the grayscale value of the i-th band of the multispectral image measured by the gray plate and the actual reflectivity of the i-th band of the multispectral image measured by ASD.

3. The method for monitoring crop moisture status based on drone according to claim 2, characterized in that: In step S4, the normalized vegetation index NDVI of the target field long time series is calculated using multispectral images. The specific calculation formula is: Where NIR is the near infrared band in the multispectral image; R is the reflectance of the red band in the multispectral image, which is obtained from the data R i Obtained in.

4. The method for monitoring crop moisture status based on an unmanned aerial vehicle according to claim 3, characterized in that: In step S4, the thermal infrared image is used to calculate the surface temperature Ts of the target field in a long time series. The specific calculation formula is: Where DN′ is the gray value of the thermal infrared image; DN Board Ts is the grayscale value of the thermal infrared image of the blackboard arranged on the ground of the target field; Board A blackboard is arranged for the temperature sensor to observe the temperature of the target field ground; and the maximum 5% and minimum 5% pixels of each thermal infrared image are removed as abnormal values.

5. The method for monitoring crop moisture status based on an unmanned aerial vehicle according to claim 4, characterized in that: In step S8, the specific calculation formula for determining the three-dimensional drought index TDDI using the Ts image and the corresponding dry surface and wet surface is: In the formula, Ts i is the surface temperature value of pixel i in Ts image; Ts maxi is the maximum surface temperature value corresponding to the NDVI value and Ta value of pixel i in Ts image; Ts mini is the minimum surface temperature value corresponding to the NDVI value and Ta value of Ts image pixel i.

6. The method for monitoring crop moisture status based on drone according to claim 1, characterized in that: In step S8, the specific content of judging the crop moisture status based on the three-dimensional drought index TDDI is: The value of the three-dimensional drought index TDDI is between 0 and 1. The closer the value is to 1, the higher the moisture content of the crop and the better the moisture status; conversely, the worse the moisture status.

7. A computer-readable storage medium storing a computer program, characterized in that: The computer program enables the computer to execute the method for monitoring crop moisture conditions as described in any one of claims 1 to 6.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for monitoring crop moisture conditions according to any one of claims 1 to 6 is implemented.

Images