Cotton field early identification method and device based on crop planting management difference

Abstract

This invention provides a method and apparatus for early identification of cotton fields based on differences in crop planting and management, relating to the field of smart agriculture technology. The method involves acquiring surface reflectance data of a target area during historical years' mulching periods to determine a surface reflectance image sequence. Based on the image sequence, a mulching difference index and an average texture feature index are determined to characterize the difference in mulched area ratio between cotton and non-cotton fields. A classifier is trained using the mulching difference index, the average texture feature index, and the location information of cotton and non-cotton fields from historical years to obtain an early cotton field identification model. The surface reflectance data of the target area during the current year's mulching period is input into the model to obtain the early cotton field identification results. This invention enables early mapping and meets the early management decision-making needs of cotton in areas such as irrigation water allocation and agricultural insurance, without waiting for cotton canopy development or canopy closure.

Description

Technical Field

[0001] This invention relates to the field of smart agriculture technology, and in particular to a method and device for early identification of cotton fields based on differences in crop planting and management. Background Technology

[0002] Cotton is an important economic crop. Timely and accurate information on cotton distribution is helpful for early crop monitoring, disaster warning, and agricultural insurance. In addition, obtaining information on cotton planting conditions in the early stages of growth is particularly crucial for agricultural water resource management in cotton-growing areas.

[0003] Early mapping refers to crop identification using remote sensing data from the early or mid-growing season to obtain distribution information throughout the crop growing season. Existing early crop identification methods mainly rely on differences in physiological and phenological characteristics among crops. These differences can usually only be captured when the crop canopy develops or closes in. Due to the low farmland cover and significant influence from soil background in the early stages of growth, the spectral differences in vegetation among crops are relatively limited. Currently, the earliest time for crop identification generally ranges from a few weeks to several months before harvest, which is insufficient to meet the needs of early management decisions such as irrigation water allocation and agricultural insurance.

[0004] Therefore, it is necessary to further explore the differences in crop management practices in the early growing season in order to provide a technical solution that can achieve early mapping of cotton, thereby meeting the needs of early management decision-making. Summary of the Invention

[0005] This invention provides a method and apparatus for early identification of cotton fields based on differences in crop planting and management, which addresses the shortcomings of existing technologies that rely on differences in physiological and phenological characteristics among cotton crops for identification, making it difficult to meet the needs of early management decisions such as irrigation water allocation and agricultural insurance for cotton.

[0006] This invention provides a method for early identification of cotton fields based on differences in crop planting and management, comprising: Obtain surface reflectance data of the target area during historical annual mulching periods to determine the surface reflectance image sequence; Based on the surface reflectance image sequence, the film covering difference index and the average texture feature index are determined. The film covering difference index and the average texture feature index are used to characterize the difference in the proportion of film covering area between cotton fields and non-cotton fields. Based on the mulch difference index, the average texture feature index, and the location information of cotton fields and non-cotton fields in historical years, the classifier is trained to obtain an early identification model for cotton fields. Based on the surface reflectance data of the target area during the current year's mulching period and the cotton field early identification model, the cotton field early identification results output by the cotton field early identification model are obtained.

[0007] As one embodiment, determining the overlay difference index and the average texture feature index based on the surface reflectance image sequence includes: Based on the first band data of the surface reflectance image sequence, the film covering difference index is determined; Based on the second band data of the surface reflectance image sequence, the texture feature index is determined and averaged. The first band data includes surface reflectance data with center wavelengths of 490nm, 560nm, 842nm and 865nm, and the second band data includes surface reflectance data with a center wavelength of approximately 490nm.

[0008] As an example, the formula for calculating the coating difference index is as follows: ; PMI represents the coating difference index. The data are surface reflectance data with center wavelengths of 490nm, 560nm, 842nm, and 865nm, respectively.

[0009] As an example, the formula for calculating the sum and average texture feature index is as follows: ; Wherein, SAVG represents the average texture feature index, pixel pairs in the gray-level co-occurrence matrix calculated for surface reflectance data with a center wavelength of 490 nm The sum of the corresponding gray values, For pixel pairs in the gray-level co-occurrence matrix The probability of its occurrence, The grayscale level is denoted by .

[0010] As an example, the proportion of cotton fields covered with plastic film is greater than the proportion of non-cotton fields covered with plastic film.

[0011] As an example, the proportion of the cotton field covered with plastic film ranges from 67–73.6% or 62–69.6%; the proportion of the non-cotton field covered with plastic film ranges from 59.97–65.35% or 38.23–41.18%.

[0012] The present invention also provides a cotton distribution identification device based on differences in crop planting management, comprising: The first determining module is used to acquire surface reflectance data of the target area during the historical annual mulching period and determine the surface reflectance image sequence; The second determining module is used to determine the film covering difference index and the average texture feature index based on the surface reflectance image sequence. The film covering difference index and the average texture feature index are both used to characterize the difference in the proportion of film covering area between cotton fields and non-cotton fields. The training module is used to train the classifier based on the mulch difference index, the average texture feature index, and the location information of cotton fields and non-cotton fields in historical years to obtain an early identification model for cotton fields. The identification module is used to obtain the early identification results of cotton fields output by the early identification model of cotton fields based on the surface reflectance data of the target area during the current year's mulching period and the early identification model of cotton fields.

[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the cotton field early identification method based on crop planting management differences as described above.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the cotton field early identification method based on crop planting management differences as described above.

[0015] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the cotton field early identification method based on crop planting and management differences as described above.

[0016] This invention provides a method and apparatus for early identification of cotton fields based on differences in crop planting and management. The method acquires surface reflectance data of the target area during the historical annual mulching period, and determines a surface reflectance image sequence. Based on the surface reflectance image sequence, it determines a mulching difference index and a sum-mean texture feature index, both of which characterize the difference in the proportion of mulched area between cotton fields and non-cotton fields. Based on the mulching difference index, the sum-mean texture feature index, and the location information of cotton fields and non-cotton fields in historical years, a classifier is trained to obtain an early identification model for cotton fields. Based on the surface reflectance data of the target area during the current year's mulching period and the early identification model, the early identification result of the cotton field is obtained from the output of the early identification model. This invention, based on the difference in the proportion of mulched area, trains an early identification model for cotton fields, enabling identification of cotton fields during the mulching period without waiting for the cotton canopy to develop or close. This allows for early mapping and meets the early management decision-making needs of cotton in areas such as irrigation water allocation and agricultural insurance. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 This is one of the flowcharts of the cotton field early identification method based on crop planting and management differences provided by the present invention.

[0019] Figure 2 This is a schematic diagram of simulated scenarios for different planting patterns under different angles between the pixel and the crop row sowing direction.

[0020] Figure 3 It shows a schematic diagram of the spectral curves of mulched farmland and other land features, as well as statistical results of reflectance and mulching difference index (PMI) for each band of DOY122-132.

[0021] Figure 4 It is a schematic diagram of the time series curves of the average texture characteristics of mulched farmland and other land features.

[0022] Figure 5 This is a schematic diagram showing the time-series curves and statistical results of the time-series maximum values ​​of the different characteristics of mulching for various crops.

[0023] Figure 6 This is a schematic diagram of the early identification results of cotton fields provided by the present invention.

[0024] Figure 7 This is a schematic diagram of the spectral response analysis results for different coating area ratios provided by the present invention.

[0025] Figure 8 This is a schematic diagram of the early identification device for cotton fields based on differences in crop planting and management provided by the present invention.

[0026] Figure 9 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0028] The Earliest Identifiable Timing (EIT) is defined as the moment when the classification accuracy reaches a certain reliable level, such as when the accuracy index reaches 0.9. Currently, crop EITs are generally concentrated during the peak growth period. For example, existing studies have used phenological differences to advance the cotton EIT to 2-2.5 months before harvest, which improves timeliness but is insufficient to meet the needs of early management decisions such as irrigation water allocation and agricultural insurance.

[0029] To address the aforementioned shortcomings, this invention provides an early identification scheme for cotton fields based on differences in crop planting and management, namely, an early mapping scheme for cotton fields based on differences in crop planting and management. The following is a detailed description in conjunction with the accompanying drawings.

[0030] Figure 1 This is one of the flowcharts illustrating the early identification method for cotton fields based on differences in crop planting and management provided by this invention, such as... Figure 1 As shown, the present invention provides an early identification method for cotton fields based on differences in crop planting management, where differences in crop planting management refer to the difference in mulch area between cotton fields and non-cotton fields. The method may include steps S110-S140.

[0031] Step S110: Obtain surface reflectance data of the target area during the historical annual mulching period and determine the surface reflectance image sequence.

[0032] In this embodiment of the invention, the target region is Xinjiang. The mulching period refers to the time range from the sowing period to the seedling stage, before the crop canopy closes and the mulch is still visible. The mulching period varies slightly in different regions. Based on the sowing and emergence time of mulched crops in Xinjiang, this embodiment of the invention sets the mulching period to April 1 to May 31.

[0033] Sentinel-2 surface reflectance data for the target area over the years prior to the mulch cover period (N is an integer), after radiometric calibration and atmospheric correction, are obtained. Preprocessing is performed on this data, including cloud and cloud shadow masking, spatial resampling, temporal median composite, and temporal filtering, to obtain 10-day composite time-series data with a spatial resolution of 10 m. Surface reflectance data from different years can be preprocessed separately to obtain subsets of 10-day composite time-series data for each year. Alternatively, a unified preprocessing can be performed on mixed surface reflectance data from different years to obtain a total set of 10-day composite time-series data for all historical years.

[0034] Step S120: Based on the surface reflectance image sequence, determine the film covering difference index and the average texture feature index. The film covering difference index and the average texture feature index are both used to characterize the difference in the proportion of film covering area between cotton fields and non-cotton fields.

[0035] Due to different crop management practices, there are differences in the proportion of mulched area between cotton fields and non-cotton fields. These differences in mulched area proportions are also spectral. Based on this, this invention proposes two types of features to characterize the differences in the proportion of mulched area between cotton and non-cotton fields, specifically including the mulching difference index and the average texture feature index. Considering the differences in the sowing time of mulching in different farmlands, in order to better characterize the differences in mulching signals between cotton fields and non-cotton fields, the temporal maximum values ​​of the mulching difference index and the average texture feature index are calculated. The features characterizing the differences in the proportion of mulched area and their temporal maximum values ​​are determined as early identification features of cotton fields.

[0036] Optionally, the average texture feature index refers to the blue band and the average texture feature.

[0037] Optionally, the film covering difference index and the average texture feature index are both determined based on partial band data of the surface reflectance image sequence. Therefore, in step S110, the surface reflectance data of partial bands of the target area during the film covering period in historical years can be obtained. Based on the surface reflectance data of partial bands, the surface reflectance image sequence corresponding to the partial bands can be determined to reduce the amount of calculation.

[0038] Step S130: Based on the film covering difference index, the average texture feature index, and the location information of cotton fields and non-cotton fields in historical years, the classifier is trained to obtain an early identification model for cotton fields.

[0039] Optionally, non-cotton fields include three categories: corn, wheat, and other crops. A classifier is trained using the mulch difference index, the average texture feature index, and the location information of cotton and non-cotton fields from historical years as input to obtain an early identification model for cotton fields. The classifier in this invention can be a random forest classifier or a support vector machine.

[0040] Optionally, based on the mulch difference index, the average texture feature index, and the location information of cotton fields and non-cotton fields in historical years, a classifier is trained to obtain an early cotton field identification model, including: performing image segmentation on the surface reflectance image, identifying plots with clear physical boundaries, treating each plot as a sample, constructing a feature matrix for each sample based on the mulch difference index and the average texture feature index, labeling each sample based on the location information of cotton fields and non-cotton fields in historical years, and determining the label vector for each sample; training the classifier based on the samples, and using the trained classifier as the early cotton field identification model.

[0041] Step S140: Based on the surface reflectance data of the target area during the current year's mulching period and the cotton field early identification model, obtain the cotton field early identification result output by the cotton field early identification model.

[0042] Optionally, based on the surface reflectance data of the target area during the current year's mulching period, the early identification characteristics of the target area for the current year are determined. These early identification characteristics refer to the mulching difference index, blue band, and average texture features, along with their temporal maximum values, calculated from the surface reflectance data of the target area during the current year's mulching period after preprocessing. Using these early identification characteristics as input, the cotton field early identification model is fed to obtain the cotton field early identification results. The final output of the cotton field early identification results shows the distribution of cotton fields and non-cotton fields.

[0043] It is understood that this invention utilizes differences in crop planting management, namely the difference in the proportion of mulched area between cotton fields and non-cotton fields, to identify cotton fields in the early stages. Specifically, this invention uses the difference in the proportion of mulched area between cotton fields and non-cotton fields, reflected in the difference in surface reflectance, to train a classifier to obtain an early cotton field identification model. This model can identify the distribution of cotton fields in the target area based on surface reflectance data during the mulching period, without waiting for the cotton canopy to develop or the canopy to close. It can achieve early mapping and meet the early management decision-making needs of cotton in areas such as irrigation water allocation and agricultural insurance.

[0044] As an example, the proportion of cotton fields covered with plastic film is greater than the proportion of non-cotton fields covered with plastic film.

[0045] Optionally, the proportion of the cotton field covered with plastic film ranges from 67–73.6% or 62–69.6%; the proportion of the non-cotton field covered with plastic film ranges from 59.97–65.35% or 38.23–41.18%.

[0046] The Xinjiang mulched crops involved in this invention include four types: cotton, corn, tomato, and chili pepper. Cotton fields refer to fields where cotton is grown, while non-cotton fields refer to fields where one or more of corn, tomatoes, and chili peppers are grown. Cotton planting patterns mainly include two types: a one-film, three-pipe, six-row pattern, with a row spacing of (66+10) cm, and ideally, a mulched area ratio of 67–73.6%; and a one-film, three-pipe, three-row planting pattern, with a row spacing of 76 cm, and ideally, a mulched area ratio of 62–69.6%. Other mulched crop planting patterns in Xinjiang mainly fall into two categories: the "one film, two pipes, four rows" planting pattern, with a row spacing of (60+30) cm and an ideal mulched area ratio of 59.97–65.35%, primarily used for corn and tomatoes; and the "one film, one pipe, two rows" planting pattern, with a row spacing of (70+40) cm and an ideal mulched area ratio of 38.23–41.18%, primarily used for corn and chili peppers. Currently, the main mulched crops in Xinjiang are cotton (100% mulched) and corn (80% mulched). Cotton mainly uses a one-film, six-row planting method, while corn mainly uses a one-film, two-row planting method. The mulched area ratios for these two planting methods differ significantly. Therefore, even if the one-film, three-row cotton planting pattern overlaps with the one-film, four-row planting pattern of other crops in terms of mulched area ratio, the difference in mulched area ratio can still help identify most cotton fields early on. Among them, the calculation of the proportion of agricultural film coverage area within 10 m pixels corresponding to various planting modes is obtained by computer simulation. Under ideal conditions, the maximum exposed area of ​​agricultural film accounts for the proportion of the total pixel area. The simulation process takes into account various situations where the angle between the pixel direction and the sowing row direction changes.

[0047] Computer simulations can be used to calculate the percentage of mulched area within 10 m pixels of various types of mulched farmland in Xinjiang under ideal conditions. The analysis shows that the percentage of mulched area in cotton fields is significantly higher than that in other crop fields.

[0048] The calculation of the proportion of agricultural film coverage area within 10m pixels for various planting patterns is obtained through computer simulation. This simulation represents the maximum exposed area of ​​agricultural film to the total pixel area under ideal conditions. The range and average value of the coverage area proportion under various scenarios are statistically analyzed. Specifically, the simulation process considers multiple scenarios where the angle between the pixel direction and the sowing row direction varies from 0° to 90° (the 0°~360° variation can be simplified to the 0°~90° variation), with a simulation step size set at 5°.

[0049] Reference Figure 2The left two columns show a simulated scene with the angle between the pixel and the crop row direction being 0°, and the right two columns show a simulated scene with the angle between the pixel and the crop row direction being 45°. The computer simulation takes into account the cases where the film coverage area is at its maximum and minimum.

[0050] Through simulation, it can be determined that the mulched area of ​​cotton fields using the "one film, three pipes, six rows" planting method is 67–73.6%, the mulched area of ​​cotton fields using the "one film, three pipes, three rows" planting method is 62–69.6%, the mulched area of ​​other crop fields using the "one film, two pipes, four rows" planting method is 59.97–65.35%, and the mulched area of ​​other crop fields using the "one film, one pipe, two rows" planting method is 38.23–41.18%. Currently, the main crops using mulching are cotton (100% mulched) and corn (80% mulched). Among them, cotton mainly uses the "one film, six rows" planting method, and corn mainly uses the "one film, two rows" planting method. There is a significant difference in the mulched area percentage corresponding to these two planting methods. Therefore, it is believed that even if the "one film, three rows" pattern of cotton overlaps with the "one film, four rows" pattern of other crops in terms of mulched area percentage, the difference in mulched area percentage can still be used to identify most large-scale, mechanized cotton fields at an early stage.

[0051] It is understood that this invention, through computer simulation, determines that there is a difference in the proportion of mulched area between cotton fields and non-cotton fields. This difference in mulched area proportion leads to differences in their spectra. Based on this, this invention proposes two types of features to characterize the difference in the proportion of mulched area between cotton fields and other crop fields, including a mulching difference index and blue band and average texture features. Specifically, based on the 10-day synthetic time-series data, the mulching difference index, blue band, and average texture features are calculated for each time phase. Considering the differences in mulching and sowing times among different fields, to better characterize the differences in mulching signals between cotton and non-cotton fields, the temporal maximum values ​​of the mulching difference index, blue band, and average texture features are further extracted. Finally, the mulching difference index, blue band, and average texture features for each time phase, along with the corresponding temporal maximum values, are collectively determined as the early identification features of cotton fields.

[0052] As one embodiment, determining the overlay difference index and the average texture feature index based on the surface reflectance image sequence includes: Based on the first band data of the surface reflectance image sequence, the film covering difference index is determined; Based on the second band data of the surface reflectance image sequence, the texture feature index is determined and averaged. The first band data includes surface reflectance data with center wavelengths of 490nm, 560nm, 842nm and 865nm, and the second band data includes surface reflectance data with a center wavelength of 490nm.

[0053] Optionally, the formula for calculating the coating difference index is as follows: ; PMI represents the coating difference index. The data are surface reflectance data with center wavelengths of 490nm, 560nm, 842nm, and 865nm, respectively.

[0054] Optional, refer to Figure 4 SAVG is obtained by summing the pixel pairs in four directions within a neighborhood and then averaging them by frequency weights. Its value is influenced by the number and clustering of high-value pixels within the GLCM neighborhood; the more high-value pixels in the neighborhood, the larger the calculated SAVG. Mulched farmland exhibits high reflectivity in the blue band, and the SAVG calculated using the blue band can characterize the proportion and clustering degree of mulched farmland within a specified spatial range. The formula for calculating the sum-average texture feature index is as follows: ; Wherein, SAVG represents the average texture feature index, pixel pairs in the gray-level co-occurrence matrix calculated for surface reflectance data with a center wavelength of 490 nm The sum of the corresponding gray values, For pixel pairs in the gray-level co-occurrence matrix The probability of its occurrence, The grayscale level is denoted by .

[0055] This invention proposes two indicators to characterize the difference in the proportion of mulched area between cotton fields and non-cotton fields. One is the Plastic-mulched Index (PMI), and the other is the Sum Average (SAVG) of blue bands. The determination of the two types of indices is involved.

[0056] Based on high-resolution imagery from 2021, three types of land cover were selected through visual interpretation: plastic-mulched farmland (PMF), non-plastic-mulched farmland (nPMF), and crop sampling points, to analyze their spectral characteristics. (Reference) Figure 3 a- Figure 3c. The differences in the temporal changes of the spectra of the three types of land cover are as follows: Spring-sown crops, represented by wheat, have emerged from April to June and have closed the canopy during DOY (Day of Year) 092-152 (April 2, 2021 - June 1, 2021), showing significant differences in spectral curves compared to the other two types. Both PMF and nPMF exhibit a bare soil surface cover in DOY 092 (April 2, 2021 - April 12, 2021), with relatively small spectral differences between the two. In DOY 122 (May 2, 2021 - May 12, 2021), the mulched farmland gradually completes the mulching and sowing operations, and PMF shows higher reflectance in visible light and red edge. As the crops emerge, the differences between the two decrease again in DOY 142 (May 22, 2021 - June 1, 2021). In this context, DOY stands for Day of Year, indicating the day of the year. DOY142 indicates the 142nd day of the year, and so on.

[0057] To quantify the differences among the three types of land features, this embodiment of the invention uses the M-index to assess the separability of features. The M-index is the ratio of the absolute value of the difference between the two means to the sum of the two standard deviations. When M > 1.0, it indicates good separability; when M < 1.0, it indicates poor separability. The calculation method of the M-index is as follows: ; in, and This represents the feature mean of two categories. and This represents the standard deviation of the features of the two categories. In this embodiment of the invention, reference is made to... Figure 3 d- Figure 3 h. To quantify the differences in spectral characteristics between covered farmland and the other two types of land cover, the M index in the B2, B3, B4, B8, and B8A bands of Sentinel-2 was calculated by combining covered farmland with uncovered farmland and covered farmland with crops as categories.

[0058] During the mulching period, PMF showed the greatest difference from the other two categories in the blue and green bands, and the greatest difference from farmland in the near-infrared band. Based on this, the inventors further constructed a PMI index to characterize the difference in the proportion of mulched area for cotton fields. This index can maximize the distinction between PMF and the other two categories. The M index for PMF and nPMF is 1.04, and the M index for PMF and Crop is 2.78.

[0059] Reference Figure 5This presentation displays time-series curves and statistical results of early identification characteristics of cotton, corn, wheat, and other crops in cotton fields. These results are calculated based on ground survey samples obtained in 2020. The statistical curves show the 25th, 50th, and 75th percentile lines for each characteristic across different crops. During the mulching period, cotton fields exhibit higher PMI and SAVG values ​​due to the higher proportion of mulched area compared to other crops. Furthermore, the maximum time-series characteristics PMImax and SAVGmax are also higher in cotton fields than in other crops. Utilizing these early mulching differences can effectively distinguish cotton from other crops.

[0060] Reference Figure 6 By inputting satellite remote sensing images of the mulching period in 2021 into the trained model, the overall classification accuracy (OA) reached 84.70%, and the F1-score for cotton reached 88.16%, demonstrating the accuracy and timeliness advantages of the present invention.

[0061] As an optional embodiment, the mulch area index can also be determined based on the differences in spectral response of different mulch film colors. Specifically, a standard spectral curve library is established for mulch films of different colors, such as transparent, black, and silver-gray. The spectrum of transparent film is greatly affected by the moisture content of the underlying soil, so a mixed spectral model of soil with different moisture content and transparent film is established; black film has extremely low reflectivity in the visible-near-infrared band; silver-gray film has significantly higher reflectivity than soil in the blue-green band. Based on the standard spectral curve library, the mulch area index corresponding to the surface reflectivity image sequence is determined. The mulch area index is used to verify the mulch difference index to improve the accuracy of identification.

[0062] As an optional embodiment, a coating difference index can also be determined based on the differences in spectral response of different coating area proportions. (Refer to...) Figure 7 Transparent mulch films exhibit high reflectivity in the visible-near-infrared band. Their spectral response characteristics can be analyzed through ground-based or UAV spectral measurements. A linear mixing model is established using mulched soil and bare soil as endmembers to establish the relationship between the mulch area ratio and the mulch difference index, thus verifying the mulch difference index's ability to characterize differences in mulch area ratio. In this embodiment, endmembers can be understood as reference spectra representing mulched soil and bare soil, respectively. The linear mixing model... The construction method is as follows: ; ; in, and The reference spectra of the covered soil and the bare soil are shown respectively. and These represent the area percentages of covered soil and bare soil, respectively.

[0063] In summary, this invention provides a remote sensing method for early mapping of cotton fields in Xinjiang based on differences in crop planting and management. By exploring the early cultivation and management characteristics of cotton in Xinjiang and utilizing the difference in the proportion of mulch film coverage area between cotton and other crops within a 10m pixel during the mulching period, early mapping of cotton in Xinjiang can be achieved. The proposed method can advance the EIT (early incubation period) of cotton fields to the mulching period, i.e., 4–4.5 months before harvest, which has a significant advantage in terms of timeliness and can provide data support for early management, monitoring, and decision-making in cotton.

[0064] The early identification device for cotton fields based on differences in crop planting management provided by the present invention will be described below. The early identification device for cotton fields based on differences in crop planting management described below can be referred to in correspondence with the early identification method for cotton fields based on differences in crop planting management described above.

[0065] Figure 8 This is a schematic diagram of the early identification device for cotton fields based on differences in crop planting and management provided by the present invention, as shown in the figure. Figure 8 As shown, the present invention also provides an early identification device for cotton fields based on differences in crop planting and management, comprising: The first determining module 810 is used to acquire surface reflectance data of the target area during the historical annual mulching period and determine the surface reflectance image sequence. The second determining module 820 is used to determine the film covering difference index and the average texture feature index based on the surface reflectance image sequence. The film covering difference index and the average texture feature index are both used to characterize the difference in the proportion of film covering area between cotton fields and non-cotton fields. Training module 830 is used to train a classifier to obtain an early identification model for cotton fields based on the film covering difference index, the average texture feature index, and the location information of cotton fields and non-cotton fields in historical years. The identification module 840 is used to obtain the early identification result of cotton field output by the early identification model of cotton field based on the surface reflectance data of the target area during the current year's mulching period and the early identification model of cotton field.

[0066] As one embodiment, the second determining module 820 is used for: Based on the first band data of the surface reflectance image sequence, the film covering difference index is determined; Based on the second band data of the surface reflectance image sequence, the texture feature index is determined and averaged. The first band data includes surface reflectance data with center wavelengths of 490 nm, 560 nm, 842 nm and 865 nm, and the second band data includes surface reflectance data with a center wavelength of 490 nm.

[0067] As an example, the formula for calculating the coating difference index is as follows: ; PMI represents the coating difference index. The data are surface reflectance data with center wavelengths of 490nm, 560nm, 842nm, and 865nm, respectively.

[0068] As an example, the formula for calculating the sum and average texture feature index is as follows: ; Wherein, SAVG represents the average texture feature index, pixel pairs in the gray-level co-occurrence matrix calculated for surface reflectance data with a center wavelength of 490 nm The sum of the corresponding gray values, For pixel pairs in the gray-level co-occurrence matrix The probability of its occurrence, The grayscale level is denoted by .

[0069] As an example, the proportion of cotton fields covered with plastic film is greater than the proportion of non-cotton fields covered with plastic film.

[0070] As an example, the proportion of the cotton field covered with plastic film ranges from 67–73.6% or 62–69.6%; the proportion of the non-cotton field covered with plastic film ranges from 59.97–65.35% or 38.23–41.18%.

[0071] It should be noted that the cotton field early identification device based on crop planting management differences provided by this invention has the same technical effects as the cotton field early identification method based on crop planting management differences, which will not be elaborated further.

[0072] Figure 9 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 9 As shown, the electronic device may include: a processor 910, a communications interface 920, a memory 930, and a communication bus 940, wherein the processor 910, the communications interface 920, and the memory 930 communicate with each other via the communication bus 940. The processor 910 can call logical instructions in the memory 930 to execute an early identification method for cotton fields based on differences in crop planting management. This method includes: Obtain surface reflectance data of the target area during historical annual mulching periods to determine the surface reflectance image sequence; Based on the surface reflectance image sequence, the film covering difference index and the average texture feature index are determined. The film covering difference index and the average texture feature index are used to characterize the difference in the proportion of film covering area between cotton fields and non-cotton fields. Based on the mulch difference index, the average texture feature index, and the location information of cotton fields and non-cotton fields in historical years, the classifier is trained to obtain an early identification model for cotton fields. Based on the surface reflectance data of the target area during the current year's mulching period and the cotton field early identification model, the cotton field early identification results output by the cotton field early identification model are obtained.

[0073] Furthermore, the logical instructions in the aforementioned memory 930 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0074] On the other hand, the present invention also provides a computer program product, the computer program product comprising a computer program that can be stored on a non-transitory computer-readable storage medium, wherein when the computer program is executed by a processor, the computer is able to execute the cotton field early identification method based on crop planting and management differences provided by the above methods, the method comprising: Obtain surface reflectance data of the target area during historical annual mulching periods to determine the surface reflectance image sequence; Based on the surface reflectance image sequence, the film covering difference index and the average texture feature index are determined. The film covering difference index and the average texture feature index are used to characterize the difference in the proportion of film covering area between cotton fields and non-cotton fields. Based on the mulch difference index, the average texture feature index, and the location information of cotton fields and non-cotton fields in historical years, the classifier is trained to obtain an early identification model for cotton fields. Based on the surface reflectance data of the target area during the current year's mulching period and the cotton field early identification model, the cotton field early identification results output by the cotton field early identification model are obtained.

[0075] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the cotton field early identification method based on crop planting and management differences provided by the above methods, the method comprising: Obtain surface reflectance data of the target area during historical annual mulching periods to determine the surface reflectance image sequence; Based on the surface reflectance image sequence, the film covering difference index and the average texture feature index are determined. The film covering difference index and the average texture feature index are used to characterize the difference in the proportion of film covering area between cotton fields and non-cotton fields. Based on the mulch difference index, the average texture feature index, and the location information of cotton fields and non-cotton fields in historical years, the classifier is trained to obtain an early identification model for cotton fields. Based on the surface reflectance data of the target area during the current year's mulching period and the cotton field early identification model, the cotton field early identification results output by the cotton field early identification model are obtained.

[0076] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0077] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0078] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for early identification of cotton fields based on differences in crop planting and management, characterized in that, include: Obtain surface reflectance data of the target area during historical annual mulching periods to determine the surface reflectance image sequence; Based on the surface reflectance image sequence, the film covering difference index and the average texture feature index are determined. The film covering difference index and the average texture feature index are used to characterize the difference in the proportion of film covering area between cotton fields and non-cotton fields. Based on the mulch difference index, the average texture feature index, and the location information of cotton fields and non-cotton fields in historical years, the classifier is trained to obtain an early identification model for cotton fields. Based on the surface reflectance data of the target area during the current year's mulching period and the cotton field early identification model, the cotton field early identification results output by the cotton field early identification model are obtained.

2. The method for early identification of cotton fields based on differences in crop planting and management as described in claim 1, characterized in that, The determination of the film cover difference index and the average texture feature index based on the surface reflectance image sequence includes: Based on the first band data of the surface reflectance image sequence, the film covering difference index is determined; Based on the second band data of the surface reflectance image sequence, the texture feature index is determined and averaged. The first band data includes surface reflectance data with center wavelengths of 490nm, 560nm, 842nm and 865nm, and the second band data includes surface reflectance data with a center wavelength of 490nm.

3. The method for early identification of cotton fields based on differences in crop planting and management as described in claim 2, characterized in that, The formula for calculating the coating difference index is as follows: ； PMI represents the coating difference index. The data are surface reflectance data with center wavelengths of 490nm, 560nm, 842nm, and 865nm, respectively.

4. The method for early identification of cotton fields based on differences in crop planting and management according to claim 2, characterized in that, The formula for calculating the average texture feature index is as follows: ； Wherein, SAVG represents the average texture feature index, pixel pairs in the gray-level co-occurrence matrix calculated for surface reflectance data with a center wavelength of 490 nm The sum of the corresponding gray values, For pixel pairs in the gray-level co-occurrence matrix The probability of its occurrence, The grayscale level is denoted by .

5. The cotton distribution identification method based on differences in crop planting management according to any one of claims 1 to 4, characterized in that, The proportion of cotton fields covered with plastic film is greater than that of non-cotton fields.

6. The cotton distribution identification method based on differences in crop planting management according to claim 4, characterized in that, The percentage of the cotton field covered with plastic film ranges from 67–73.6% or 62–69.6%; the percentage of the non-cotton field covered with plastic film ranges from 59.97–65.35% or 38.23–41.18%.

7. A cotton distribution identification device based on differences in crop planting management, characterized in that, include: The first determining module is used to acquire surface reflectance data of the target area during the historical annual mulching period and determine the surface reflectance image sequence; The second determining module is used to determine the film covering difference index and the average texture feature index based on the surface reflectance image sequence. The film covering difference index and the average texture feature index are both used to characterize the difference in the proportion of film covering area between cotton fields and non-cotton fields. The training module is used to train the classifier based on the mulch difference index, the average texture feature index, and the location information of cotton fields and non-cotton fields in historical years to obtain an early identification model for cotton fields. The identification module is used to obtain the early identification results of cotton fields output by the early identification model of cotton fields based on the surface reflectance data of the target area during the current year's mulching period and the early identification model of cotton fields.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the cotton field early identification method based on crop planting management differences as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the cotton field early identification method based on crop planting management differences as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the cotton field early identification method based on crop planting management differences as described in any one of claims 1 to 6.

Images