Non-grain monitoring method, device, equipment and medium based on discrete wavelets

Abstract

The embodiment of the invention provides a non-grain monitoring method and device based on discrete wavelets, equipment and a medium. The method comprises the steps that a first EVI index and a second EVI index corresponding to first remote sensing image data and second remote sensing image data respectively are calculated; establishing two data sets respectively corresponding to the two indexes, and respectively carrying out discrete wavelet transform to obtain two corresponding vegetation index sequences; calculating a first signal component and a second signal component of the two vegetation index sequences on each signal mark point of the target monitoring area in quarterly, half-year and above-year scales respectively; determining an abnormal region according to a reference feature value calculated by the first signal component and a target feature value calculated by the second signal component; and determining non-grain generation time and change feature information based on the non-grain mutation point corresponding to the abnormal region. According to the method, remote sensing data is detected through time sequence change, the cultivated land non-grain is monitored in combination with phenological characteristics, time elements are considered, and the monitoring accuracy is improved.

Description

Technical Field

[0001] The present application relates to, but is not limited to, the field of natural resource monitoring technology, and in particular to a discrete wavelet-based non-grain monitoring method, device, equipment, and medium. Background Art

[0002] Monitoring the “non-grainization” of cultivated land is an important task in the protection of cultivated land, and provides a technical tool for safeguarding national food security. The existing monitoring methods for “non-grainization” mainly include: (1) “non-grainization” information statistics methods based on sample sampling, questionnaire surveys and yearbook statistics; (2) land type information interpretation based on remote sensing images, and methods for quickly obtaining large-scale land use changes through remote sensing image analysis. However, the information acquisition efficiency of the “non-grainization” information statistics methods based on sample sampling, questionnaire surveys and yearbook statistics is difficult to meet the timeliness requirements of cultivated land protection; although the land use extraction method based on remote sensing images improves the efficiency of obtaining non-grainization detection results, this method is highly dependent on the accuracy of image classification, and the image noise and the selection of classification algorithms lead to uncertainty in the classification results. This uncertainty is amplified in the trajectory of land cover change, so it is difficult to judge whether the change is real or caused by the misclassification of a certain period of images. In summary, the monitoring accuracy of the existing non-grainization monitoring methods is poor. Summary of the invention

[0003] The embodiments of the present application provide a method, device, equipment, and medium for monitoring non-grain conversion based on discrete wavelets, which combines the time series change detection of remote sensing index with phenological characteristics to determine the non-grain conversion monitoring results, fully considers the time factors of non-grain conversion of cultivated land, and effectively improves the accuracy of monitoring results.

[0004] In a first aspect, an embodiment of the present application provides a non-grain monitoring method based on discrete wavelet, comprising:

[0005] Acquire the first remote sensing image data and the second remote sensing image data according to the historical time series, calculate the first EVI index of the target monitoring area based on the band synthesis method and the first remote sensing impact data, and calculate the second EVI index of the target monitoring area based on the band synthesis method and the second remote sensing impact data, wherein the land use type corresponding to the target monitoring area is cultivated land, the first remote sensing image data is the sample data of the target monitoring area, and the second remote sensing image data is the overall data of the target monitoring area;

[0006] Establishing a first time series data set within a preset monitoring time period based on the first EVI index, and establishing a second time series data set within the preset monitoring time period based on the second EVI index;

[0007] Performing discrete wavelet transform on each first vegetation index in the first time series data set to obtain each first vegetation index sequence, and performing discrete wavelet transform on each second vegetation index in the second time series data set to obtain each second vegetation index sequence, wherein there is a first time period relationship between the first vegetation index sequence and the corresponding discrete wavelet transform number, and there is a second time period relationship between the second vegetation index sequence and the corresponding discrete wavelet transform number;

[0008] Determine each signal marker point in the target monitoring area, and calculate the first signal component of the first vegetation index sequence corresponding to each signal marker point on the quarterly, semi-annual and inter-annual scales according to the first time period relationship, and calculate the second signal component of the second vegetation index sequence corresponding to each signal marker point on the quarterly, semi-annual and inter-annual scales according to the second time period relationship, wherein the signal marker point is used to indicate the signal marker point of the known food crop type in the cultivated land, and different signal marker points correspond to different food crop types;

[0009] A reference eigenvalue is calculated based on all of the first signal components, and a target eigenvalue is calculated based on all of the second signal components, wherein the reference eigenvalue includes the mean of the first inter-annual or larger scale signal component, the number of valid cycles of the first quarter signal component, and the number of valid cycles of the first half-year signal component corresponding to each of the signal landmark points, and the target eigenvalue is used to indicate the mean of the second inter-annual or larger scale signal component, the number of valid cycles of the second quarter signal component, and the number of valid cycles of the second half-year signal component corresponding to each pixel in the second remote sensing image data;

[0010] An abnormal area is determined based on the first inter-annual or higher scale signal component mean, the first quarter signal component effective cycle number, the first semi-annual signal component effective cycle number, the second inter-annual or higher scale signal component mean, the second quarter signal component effective cycle number, and the second semi-annual signal component effective cycle number, wherein the abnormal area is used to indicate a sub-area in the target monitoring area where the phenomenon of cultivated land being converted to non-grain occurs;

[0011] The MK mutation point test algorithm is used to detect the non-grain mutation point corresponding to the abnormal area, and the non-grain occurrence time and the corresponding non-grain change characteristic information are determined based on the non-grain mutation point.

[0012] In some embodiments, the mother wavelet corresponding to the discrete wavelet transform operation is a Meyer wavelet, and the first time period relationship is constructed according to the following method:

[0013] Determining the center frequency of the Meyer wavelet;

[0014] determining a temporal resolution of the first time series data set;

[0015] The first time period relationship is constructed according to the center frequency, the time resolution, and the number of discrete wavelet transforms.

[0016] In some embodiments, the expression of the first time period relationship is as follows:

[0017] P = 2 n-1 △t / V c ;

[0018] Wherein, P is the first time period, n is the number of discrete wavelet transforms, Δt is the time resolution of the first time series data set, V c is the center frequency of the Meyer wavelet.

[0019] In some embodiments, calculating the first signal components of the first vegetation index sequence corresponding to each of the signal marker points at the quarterly, semi-annual and interannual scales or above according to the first time period relationship includes:

[0020] Determine a first transformation number, a second transformation number, and a third transformation number, wherein the first transformation number corresponds to the number of discrete wavelet transformations corresponding to the first signal component on a quarterly scale, the second transformation number corresponds to the number of discrete wavelet transformations corresponding to the first signal component on a semi-annual scale, and the third transformation number corresponds to the number of discrete wavelet transformations corresponding to the first signal component on an inter-annual scale or above, the first transformation number is less than the second transformation number, and the second transformation number is less than the third transformation number;

[0021] Performing discrete wavelet transform on the first vegetation index sequence corresponding to the signal marker point to obtain corresponding low-frequency signal and high-frequency signal, and recording the current wavelet transform times;

[0022] Performing discrete wavelet transform on the low-frequency signal and the high-frequency signal again until the current number of wavelet transforms reaches the third number of transforms;

[0023] Obtain a first high-frequency signal corresponding to the first number of transformations, a second high-frequency signal corresponding to the second number of transformations, and a first low-frequency signal corresponding to the third number of transformations, wherein the first high-frequency signal is the first signal component on a quarterly scale, the second high-frequency signal is the first signal component on a semi-annual scale, and the first low-frequency signal is the first signal component on a scale above an annual scale.

[0024] In some embodiments, the first vegetation index sequence corresponding to one of the signal marker points corresponds to a plurality of the first high-frequency signals, a plurality of the second high-frequency signals, and a plurality of the first low-frequency signals, and the reference eigenvalue is calculated based on all of the first signal components, including:

[0025] Calculate the mean of all the first low-frequency signals to obtain the mean of the first signal component at a scale above interannual;

[0026] Calculate the effective cycle number of the first quarter signal component by using the zero-crossing detection method and all the first high-frequency signals;

[0027] The number of effective cycles of the first semi-annual signal component is calculated using a zero-crossing detection method and all of the second high-frequency signals.

[0028] In some embodiments, determining an abnormal area based on the first inter-annual or higher scale signal component mean, the first quarter signal component valid cycle number, the first semi-annual signal component valid cycle number, the second inter-annual or higher scale signal component mean, the second quarter signal component valid cycle number, and the second semi-annual signal component valid cycle number includes:

[0029] When the mean value of the second inter-annual or larger scale signal component is less than or equal to the mean value of the first inter-annual or larger scale signal component, the area of ​​the current pixel in the second remote sensing image data is determined as the abnormal area;

[0030] or,

[0031] When the mean value of the second inter-annual or higher-scale signal component is greater than the mean value of the first inter-annual or higher-scale signal component, when the number of valid cycles of the second quarterly signal component is less than the number of valid cycles of the first quarterly signal component, or the number of valid cycles of the second semi-annual signal component is less than the number of valid cycles of the first semi-annual signal component, the area of ​​the current pixel in the second remote sensing image data is determined as the abnormal area.

[0032] In some embodiments, the MK mutation point detection algorithm is used to detect the non-grain mutation point corresponding to the abnormal area, including:

[0033] Acquire a first sequence set and a second sequence set corresponding to the abnormal area from the second time series data set, wherein the first sequence set is a sequence set arranged in chronological order, and the sequence order in the second sequence set is opposite to that in the first sequence set;

[0034] Performing a discrete wavelet transform process on each sequence in the first sequence set to obtain a third sequence set;

[0035] Performing a discrete wavelet transform process on each sequence in the second sequence set to obtain a fourth sequence set;

[0036] Acquire a preset number of first sequence samples from the third sequence set, and calculate a first sequence mean and a first sequence variance of the first sequence samples;

[0037] Acquire a preset number of second sequence samples from the fourth sequence set, and calculate a second sequence mean and a second sequence variance of the second sequence samples;

[0038] Calculating a first statistic based on the first sequence samples, the first sequence mean, and the first sequence variance, and drawing a first curve based on the first statistic;

[0039] Calculating a second statistic based on the second sequence samples, the second sequence mean, and the second sequence variance, multiplying the second statistic by -1 to obtain a third statistic, and drawing a second curve based on the third statistic;

[0040] The intersection point between the first curve and the second curve is determined as the non-grain mutation point.

[0041] In a second aspect, an embodiment of the present application provides a control device, comprising at least one control processor and a memory for communicating with the at least one control processor; the memory stores instructions that can be executed by the at least one control processor, and the instructions are executed by the at least one control processor so that the at least one control processor can execute the non-grain monitoring method based on discrete wavelets as described in the first aspect.

[0042] In a third aspect, an embodiment of the present application further provides an electronic device, comprising the control device of the second aspect.

[0043] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions are used to execute the non-grain monitoring method based on discrete wavelets as described in the first aspect.

[0044] The embodiment of the present application provides a non-grain monitoring method, device, equipment, and medium based on discrete wavelet, the method comprising: acquiring first remote sensing image data and second remote sensing image data according to a historical time series, calculating a first EVI index of a target monitoring area based on a band synthesis method and the first remote sensing impact data, and calculating a second EVI index of the target monitoring area based on the band synthesis method and the second remote sensing impact data, wherein the land use type corresponding to the target monitoring area is cultivated land, the first remote sensing image data is sample data of the target monitoring area, and the second remote sensing image data is overall data of the target monitoring area; establishing a preset monitoring time period based on the first EVI index A first time series data set is obtained, and a second time series data set within the preset monitoring time period is established based on the second EVI index; discrete wavelet transform is performed on each first vegetation index in the first time series data set to obtain each first vegetation index sequence, and discrete wavelet transform is performed on each second vegetation index in the second time series data set to obtain each second vegetation index sequence, wherein there is a first time period relationship between the first vegetation index sequence and the corresponding discrete wavelet transform times, and there is a second time period relationship between the second vegetation index sequence and the corresponding discrete wavelet transform times; each signal marker point of the target monitoring area is determined, and each signal marker is calculated according to the first time period relationship. The first signal components of the first vegetation index sequence corresponding to the signal marker point at the quarterly, semi-annual and inter-annual scales are calculated, and the second signal components of the second vegetation index sequence corresponding to each signal marker point are calculated according to the second time period relationship at the quarterly, semi-annual and inter-annual scales, wherein the signal marker point is used to indicate the signal marker point of the known food crop type in the cultivated land, and different signal marker points correspond to different food crop types; the reference eigenvalues ​​are calculated based on all the first signal components, and the target eigenvalues ​​are calculated based on all the second signal components, wherein the reference eigenvalues ​​include the mean of the first inter-annual scale signal components corresponding to each of the signal marker points, the first quarter signal component average, and the target eigenvalue average. The target characteristic value is used to indicate the mean value of the signal component at a scale of second interannual or above, the effective cycle number of the signal component at a scale of second quarter, and the effective cycle number of the signal component at a scale of second half-year corresponding to each pixel in the second remote sensing image data; based on the mean value of the signal component at a scale of first interannual or above, the effective cycle number of the signal component at a scale of first quarter, the effective cycle number of the signal component at a scale of first half-year, the mean value of the signal component at a scale of second interannual or above, the effective cycle number of the signal component at a scale of second quarter, and the effective cycle number of the signal component at a scale of second half-year, the abnormal area is determined, wherein the abnormal area is used to indicate the sub-area in the target monitoring area where the phenomenon of cultivated land being converted to non-grain has occurred;The MK mutation point test algorithm is used to detect the non-grain mutation point corresponding to the abnormal area, and the non-grain occurrence time and the corresponding non-grain change characteristic information are determined based on the non-grain mutation point. According to the scheme provided in the embodiment of the present application, the remote sensing image data is subjected to time series change detection, and the non-grain conversion of cultivated land is monitored in combination with the phenological characteristics of cultivated land, namely, the EVI index, which fully considers the time factor of the non-grain conversion of cultivated land, and effectively improves the accuracy of the monitoring results. ; BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a flowchart of the steps of a non-grain monitoring method based on discrete wavelet provided by an embodiment of the present application;

[0046] Figure 2 It is a structural diagram of a control device provided in another embodiment of the present application. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0048] It is understood that although the functional modules are divided in the device schematic diagram and the logical order is shown in the flow chart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flow chart. The terms "first", "second", etc. in the specification, claims or the above drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0049] Monitoring the “non-grainization” of cultivated land is an important task in the protection of cultivated land, and provides a technical tool for safeguarding national food security. The existing monitoring methods for “non-grainization” mainly include: (1) “non-grainization” information statistics methods based on sample sampling, questionnaire surveys and yearbook statistics; (2) land type information interpretation based on remote sensing images, and methods for quickly obtaining large-scale land use changes through remote sensing image analysis. However, the information acquisition efficiency of the “non-grainization” information statistics methods based on sample sampling, questionnaire surveys and yearbook statistics is difficult to meet the timeliness requirements of cultivated land protection; although the land use extraction method based on remote sensing images improves the efficiency of obtaining non-grainization detection results, this method is highly dependent on the accuracy of image classification, and the image noise and the selection of classification algorithms lead to uncertainty in the classification results. This uncertainty is amplified in the trajectory of land cover change, so it is difficult to judge whether the change is real or caused by the misclassification of a certain period of images. In summary, the monitoring accuracy of the existing non-grainization monitoring methods is poor.

[0050] In order to solve the above-mentioned problems, the embodiments of the present application provide a non-grain monitoring method, device, equipment, and medium based on discrete wavelet, and the method includes: acquiring first remote sensing image data and second remote sensing image data according to a historical time series, calculating a first EVI index of a target monitoring area based on a band synthesis method and the first remote sensing impact data, and calculating a second EVI index of the target monitoring area based on the band synthesis method and the second remote sensing impact data, wherein the land use type corresponding to the target monitoring area is cultivated land, the first remote sensing image data is sample data of the target monitoring area, and the second remote sensing image data is overall data of the target monitoring area; establishing a first EVI index based on the first EVI index A first time series data set within a preset monitoring time period is established, and a second time series data set within the preset monitoring time period is established based on the second EVI index; discrete wavelet transform is performed on each first vegetation index in the first time series data set to obtain each first vegetation index sequence, and discrete wavelet transform is performed on each second vegetation index in the second time series data set to obtain each second vegetation index sequence, wherein there is a first time period relationship between the first vegetation index sequence and the corresponding discrete wavelet transform number, and there is a second time period relationship between the second vegetation index sequence and the corresponding discrete wavelet transform number; each signal landmark point of the target monitoring area is determined, and each signal landmark point is calculated according to the first time period relationship. The first signal components of the first vegetation index sequence corresponding to the signal landmark points at the quarterly, semi-annual and inter-annual scales are calculated, and the second signal components of the second vegetation index sequence corresponding to each of the signal landmark points at the quarterly, semi-annual and inter-annual scales are calculated according to the second time period relationship, wherein the signal landmark points are used to indicate signal landmark points of known food crop types in cultivated land, and different signal landmark points correspond to different food crop types; reference eigenvalues ​​are calculated based on all of the first signal components, and target eigenvalues ​​are calculated based on all of the second signal components, wherein the reference eigenvalues ​​include the first inter-annual scale signal component mean, the first quarterly scale signal component mean, and the second quarterly scale signal component mean. The target characteristic value is used to indicate the second inter-annual or higher scale signal component mean, the second quarter signal component effective cycle number, and the second half-year signal component effective cycle number corresponding to each pixel in the second remote sensing image data; based on the first inter-annual or higher scale signal component mean, the first quarter signal component effective cycle number, the first half-year signal component effective cycle number, the second inter-annual or higher scale signal component mean, the second quarter signal component effective cycle number, and the second half-year signal component effective cycle number, the abnormal area is determined, wherein the abnormal area is used to indicate the sub-area in the target monitoring area where the phenomenon of cultivated land non-grainization occurs;The MK mutation point test algorithm is used to detect the non-grain mutation point corresponding to the abnormal area, and the non-grain occurrence time and the corresponding non-grain change characteristic information are determined based on the non-grain mutation point. According to the scheme provided in the embodiment of the present application, the remote sensing image data is subjected to time series change detection, and the non-grain conversion of cultivated land is monitored in combination with the phenological characteristics of cultivated land, namely, the EVI index, which fully considers the time factor of the non-grain conversion of cultivated land, and effectively improves the accuracy of the monitoring results. ;

[0051] The embodiments of the present application are further described below in conjunction with the accompanying drawings.

[0052] refer to Figure 1 , Figure 1 1 is a flowchart of a method for monitoring non-grain production based on discrete wavelets provided by an embodiment of the present application. The present application embodiment provides a method for monitoring non-grain production based on discrete wavelets, which includes but is not limited to the following steps:

[0053] Step S10, obtaining the first remote sensing image data and the second remote sensing image data according to the historical time series, calculating the first EVI index of the target monitoring area based on the band synthesis method and the first remote sensing impact data, and calculating the second EVI index of the target monitoring area based on the band synthesis method and the second remote sensing impact data, wherein the land use type corresponding to the target monitoring area is cultivated land, the first remote sensing image data is the sample data of the target monitoring area, and the second remote sensing image data is the overall data of the target monitoring area.

[0054] Specifically, the first remote sensing image data and the second remote sensing image data of this embodiment are multiple remote sensing images superimposed according to the historical time series in the target monitoring area, which can provide an effective data basis for the subsequent detection of the change characteristics of cultivated land from the time dimension. The difference between the first remote sensing image data and the second remote sensing image data is that the first remote sensing image data corresponds to the image of a few known sample points selected in the target monitoring area, which is used to calculate the reference characteristic value to determine whether the current target monitoring area has the phenomenon of non-grainization of cultivated land, while the second remote sensing image data corresponds to the full range of data points in the target monitoring area, and the target characteristic value calculated based on the second remote sensing image data is compared with the reference characteristic value.

[0055] Step S20: establishing a first time series data set within a preset monitoring time period based on the first EVI index, and establishing a second time series data set within a preset monitoring time period based on the second EVI index.

[0056] Specifically, this embodiment does not limit the specific preset monitoring time period, which can be three months, six months, etc., and can be determined by those skilled in the art according to actual needs.

[0057] Specifically, the first time series data set established based on the first EVI index and the second time series data set established based on the second EVI index can reflect the growth status of grain and cash crops in the target monitoring area, providing an effective data basis for subsequent guarantee of the accuracy of non-grain monitoring results.

[0058] Step S30, performing discrete wavelet transform on each first vegetation index in the first time series data set to obtain each first vegetation index sequence, and performing discrete wavelet transform on each second vegetation index in the second time series data set to obtain each second vegetation index sequence, wherein there is a first time period relationship between the first vegetation index sequence and the corresponding number of discrete wavelet transforms, and there is a second time period relationship between the second vegetation index sequence and the corresponding number of discrete wavelet transforms.

[0059] Specifically, this embodiment selects a Meyer wavelet with a high similarity to the vegetation growth characteristic waveform as the mother wavelet, and performs discrete wavelet transform on each vegetation index in the first time series data set and the second time series data set, so as to extract the time element information of each vegetation index, and provide an effective data basis for the subsequent calculation of the corresponding reference eigenvalues ​​and target eigenvalues, so that the time element is fully considered when monitoring the non-grainization of cultivated land, and the accuracy of the monitoring results of the non-grainization of cultivated land is guaranteed.

[0060] Specifically, the first time period relationship of this embodiment is constructed according to the following steps:

[0061] Step S31, determining the center frequency of the Meyer wavelet;

[0062] Step S32, determining the time resolution of the first time series data set;

[0063] Step S33, constructing a first time period relationship according to the center frequency, the time resolution, and the number of discrete wavelet transforms.

[0064] Specifically, the expression of the first time period relationship is as follows:

[0065] P = 2 n-1 △t / V c ;

[0066] Where P is the first time period, n is the number of discrete wavelet transforms, △t is the time resolution of the first time series data set, V cis the center frequency of the Meyer wavelet. Similarly, the expression for the second time period relationship is the same as that for the first time period relationship. The difference is that when P is the second time period, △t is the time resolution of the second time series data set, which provides support for calculating the first signal component of the first vegetation index sequence corresponding to each signal landmark point in the target monitoring area at the quarterly, semi-annual and interannual scales, as well as the second signal component of the second vegetation index sequence corresponding to each signal landmark point at the quarterly, semi-annual and interannual scales.

[0067] Step S40, determine each signal marker point in the target monitoring area, and calculate the first signal component of the first vegetation index sequence corresponding to each signal marker point on the quarterly, semi-annual and inter-annual scales according to the first time period relationship, and calculate the second signal component of the second vegetation index sequence corresponding to each signal marker point on the quarterly, semi-annual and inter-annual scales according to the second time period relationship, wherein the signal marker points are used to indicate the signal marker points of known food crop types in the cultivated land, and different signal marker points correspond to different food crop types.

[0068] Specifically, since there are multiple first signal components and second signal components, including signal components at quarterly, semi-annual and inter-annual scales, they are used to indicate the change characteristics of cultivated land vegetation at quarterly, semi-annual and inter-annual scales.

[0069] Specifically, Figure 1 Step S40 of calculating the first signal components of the first vegetation index sequence corresponding to each signal marker point at the quarterly, semi-annual and inter-annual scales or above according to the first time period relationship includes but is not limited to the following steps:

[0070] Step S41, determining the first transformation number, the second transformation number and the third transformation number, wherein the first transformation number corresponds to the number of discrete wavelet transformations corresponding to the first signal component on the quarterly scale, the second transformation number corresponds to the number of discrete wavelet transformations corresponding to the first signal component on the semi-annual scale, and the third transformation number corresponds to the number of discrete wavelet transformations corresponding to the first signal component on the annual scale or above, the first transformation number is less than the second transformation number, and the second transformation number is less than the third transformation number;

[0071] Step S42, performing discrete wavelet transform on the first vegetation index sequence corresponding to the signal marker point to obtain corresponding low-frequency signals and high-frequency signals, and recording the current wavelet transform times;

[0072] Step S43, performing discrete wavelet transform on the low-frequency signal and the high-frequency signal again until the current wavelet transform times reaches the third transform times;

[0073] Step S44, obtaining the first high-frequency signal corresponding to the first number of transformations, the second high-frequency signal corresponding to the second number of transformations, and the first low-frequency signal corresponding to the third number of transformations, wherein the first high-frequency signal is the first signal component on a quarterly scale, the second high-frequency signal is the first signal component on a semi-annual scale, and the first low-frequency signal is the first signal component on a scale above an annual scale.

[0074] Specifically, referring to the description of the above embodiments, the first time period relationship is associated with the time resolution of the first time series data set, the number of discrete wavelet transforms calculated by the center frequency of the mother wavelet, and the time period of monitoring corresponding to the high-frequency signal component sequence after the wavelet transform. In some embodiments, the first vegetation index sequence corresponding to each signal marker point is calculated according to the first time period relationship. The specific method is as follows: when the Meyer wavelet is used as the mother wavelet in the discrete wavelet transform, its center frequency is 0.6634, and the time resolution of the first time series data set used is 16 days, the first transformation number, the second transformation number and the third transformation number are determined, wherein the first transformation number corresponds to obtaining the first signal component corresponding to the quarterly scale The number of discrete wavelet transforms is 3, the second number of transforms corresponds to the number of discrete wavelet transforms corresponding to the first signal component on the semi-annual scale, and the third number of transforms corresponds to the number of discrete wavelet transforms corresponding to the first signal component on the inter-annual scale or above. The first number of transforms is 3, the second number of transforms is 4, and the third number of transforms is 5. The low-frequency signal A1 and the high-frequency signal D1 are generated by the first vegetation index sequence corresponding to the signal marker through a wavelet transform. The low-frequency signal A1 is transformed again by wavelet transform to generate the low-frequency signal A2 and the high-frequency signal D2. And so on. D3, D4, and A5 are calculated in this way and roughly correspond to the change characteristics on the quarterly, semi-annual, and inter-annual scales or above, among which D3 corresponds to 96 days, D4 corresponds to 193 days, and A5 corresponds to more than 386 days. This provides an effective data basis for the subsequent calculation of reference characteristic values ​​for judging whether there is non-grain land conversion.

[0075] It should be noted that the calculation principle of the second signal component of the second vegetation index sequence corresponding to each signal marker point at the quarterly, semi-annual and interannual scales according to the second time period relationship in this embodiment is similar to the calculation principle of the first signal component. The difference lies in the different data sources, which will not be elaborated here.

[0076] Step S50, calculate the reference eigenvalue based on all the first signal components, and calculate the target eigenvalue based on all the second signal components, wherein the reference eigenvalue includes the first inter-annual or higher scale signal component mean, the first quarter signal component effective cycle number, and the first half-year signal component effective cycle number corresponding to each signal landmark point, and the target eigenvalue is used to indicate the second inter-annual or higher scale signal component mean, the second quarter signal component effective cycle number, and the second half-year signal component effective cycle number corresponding to each pixel in the second remote sensing image data.

[0077] Specifically, in some embodiments, a first vegetation index sequence corresponding to a signal marker point corresponds to a plurality of first high-frequency signals, a plurality of second high-frequency signals, and a plurality of first low-frequency signals. Figure 1 The step S50 of calculating the reference characteristic value based on all the first signal components includes but is not limited to the following steps:

[0078] Step S51, calculating the mean of all first low-frequency signals to obtain the mean of the first annual or higher scale signal component;

[0079] Step S52, using the zero-crossing detection method and all the first high-frequency signals to calculate the number of effective cycles of the first quarter signal component;

[0080] Step S53, using the zero-crossing detection method and all the second high-frequency signals to calculate the number of effective cycles of the first semi-annual signal component.

[0081] It can be understood that the step of calculating the reference eigenvalue based on all the first signal components in this embodiment includes: calculating the mean α of all the first low-frequency signals (A5) of the first vegetation index sequence corresponding to the selected signal landmark point, calculating the effective period number of the first quarter signal component by using the zero-crossing detection method and all the first high-frequency signals, denoted as θ1, calculating the effective period number of the first half-year signal component by using the zero-crossing detection method and all the second high-frequency signals, denoted as θ2, and α, θ1 and θ2 are used for judgment with the target eigenvalue.

[0082] Specifically, the target characteristic value of this embodiment is used to indicate the characteristic value corresponding to each pixel in the second remote sensing image data containing the full range of data points in the target monitoring area, and compared with the reference characteristic value of the corresponding position, it can determine whether there is a phenomenon of non-grainization of cultivated land at a certain location in the target monitoring area.

[0083] It can be understood that the calculation principle of calculating the target eigenvalue based on all the second signal components in this embodiment is similar to the calculation principle of calculating the reference eigenvalue. The difference lies in the different data sources (the reference eigenvalue is calculated based on the first signal component, and the target eigenvalue is calculated based on the second signal component). No further details will be given here.

[0084] Step S60, determining the abnormal area based on the first inter-annual or higher scale signal component mean, the first quarter signal component effective cycle number, the first half-year signal component effective cycle number, the second inter-annual or higher scale signal component mean, the second quarter signal component effective cycle number and the second half-year signal component effective cycle number, wherein the abnormal area is used to indicate the sub-area where the phenomenon of non-grainization of cultivated land occurs in the target monitoring area.

[0085] Specifically, in one embodiment, Figure 1 Step S60 includes but is not limited to the following steps:

[0086] Step S61, when the second inter-annual or higher scale signal component mean value is less than or equal to the first inter-annual or higher scale signal component mean value, the current pixel is determined as an abnormal area in the second remote sensing image data;

[0087] or,

[0088] Step S62, when the mean value of the signal component at the second inter-annual scale or above is greater than the mean value of the signal component at the first inter-annual scale or above, when the number of valid cycles of the signal component in the second quarter is less than the number of valid cycles of the signal component in the first quarter, or the number of valid cycles of the signal component in the second half of the year is less than the number of valid cycles of the signal component in the first half of the year, the area of ​​the current pixel in the second remote sensing image data is determined as an abnormal area.

[0089] It can be understood that based on the comparison result of the calculated target eigenvalue and the reference eigenvalue, it can be determined whether there is a phenomenon of non-grainization of cultivated land in the target monitoring area. The specific judgment method is: when the mean value of the signal component of the second inter-annual scale or above is less than or equal to the mean value of the signal component of the first inter-annual scale or above, the current pixel is determined as an abnormal area in the second remote sensing image data; when the mean value of the signal component of the second inter-annual scale or above is greater than the mean value of the signal component of the first inter-annual scale or above, when the effective cycle number of the signal component of the second quarter is less than the effective cycle number of the signal component of the first quarter, or the effective cycle number of the signal component of the second half of the year is less than the effective cycle number of the signal component of the first half of the year, the current pixel is determined as an abnormal area in the second remote sensing image data. The target eigenvalue and the reference eigenvalue of this embodiment are calculated by combining the time series detection result (corresponding to the remote sensing image data obtained according to the time series, and the operation of obtaining the time characteristics by discrete wavelet transform) with the phenological characteristics (corresponding to the growth status of food crops and cash crops in the target monitoring area), that is, to determine whether there is a phenomenon of non-grainization of cultivated land in the target monitoring area, the time factor and the phenological factor of cultivated land are fully considered, and the accuracy of the monitoring of non-grainization of cultivated land can be guaranteed.

[0090] Step S70, using the MK mutation point detection algorithm to detect the non-grain mutation point corresponding to the abnormal area, and determining the non-grain occurrence time and the corresponding non-grain change characteristic information based on the non-grain mutation point.

[0091] Specifically, in one embodiment, Figure 1 Step S70 includes but is not limited to the following steps:

[0092] Step S71, obtaining a first sequence set and a second sequence set corresponding to the abnormal area from the second time series data set, wherein the first sequence set is a sequence set arranged in chronological order, and the sequence order in the second sequence set is opposite to that in the first sequence set;

[0093] Step S72, performing a discrete wavelet transform process on each sequence in the first sequence set to obtain a third sequence set;

[0094] Step S73, performing a discrete wavelet transform process on each sequence in the second sequence set to obtain a fourth sequence set;

[0095] Step S74, obtaining a preset number of first sequence samples from the third sequence set, and calculating the first sequence mean and the first sequence variance of the first sequence samples;

[0096] Step S75, obtaining a preset number of second sequence samples from the fourth sequence set, and calculating the second sequence mean and the second sequence variance of the second sequence samples;

[0097] Step S77, calculating a first statistic based on the first sequence samples, the first sequence mean and the first sequence variance, and drawing a first curve based on the first statistic;

[0098] Step S77, calculating a second statistic based on the second sequence samples, the second sequence mean and the second sequence variance, multiplying the second statistic by -1 to obtain a third statistic, and drawing a second curve based on the third statistic;

[0099] Step S78: determine the intersection point between the first curve and the second curve as the non-grain mutation point.

[0100] Specifically, the method of obtaining the non-grain mutation point in this embodiment is as follows: a first sequence set and a second sequence set corresponding to the abnormal area are obtained from the second time series data set, and each sequence in the first sequence set is subjected to a discrete wavelet transform process to achieve signal denoising to obtain a third sequence set. The third sequence set r i The corresponding number of samples is less than the number of samples in the first sequence set and the second sequence set; a preset number of first sequence samples are obtained from the third sequence set, defined as x i =(x1,x2,…,x m ). Construct the order sequence S k , the expression is as follows:

[0101]

[0102] In the formula, k = 2, 3, ..., m, m is the sample size of the first sequence, r i Represents x i Greater than x j The cumulative number of (j=1,…,i);

[0103] Calculate the first sequence mean E(S k ) and the first series variance Var(S k ), the corresponding expression is as follows:

[0104]

[0105] Calculate the first statistic UF based on the first sequence sample, the first sequence mean and the first sequence variance k , and based on the first statistic UF k Draw the first curve and the first statistic UF k According to the following formula:

[0106]

[0107] Among them, UF1=0;

[0108] A preset number of second sequence samples are obtained from the fourth sequence set, a second sequence mean and a second sequence variance of the second sequence samples are calculated, a second statistic is calculated based on the second sequence samples, the second sequence mean and the second sequence variance, and a calculation principle for drawing a second curve based on the second statistic is similar to the calculation principle for calculating the first statistic and drawing the first curve, the difference being that the data sources are different, the data source for drawing the first curve corresponds to the first sequence set, while the data source for drawing the second curve corresponds to the reversed first sequence set, i.e., the second sequence set, and the second statistic is multiplied by -1 to obtain a third statistic, and the second curve is drawn according to the third statistic.

[0109] It is understandable that after monitoring the abnormal area, further monitoring the non-grain mutation points in the abnormal area and determining the non-grain occurrence time and corresponding non-grain change characteristic information based on the non-grain mutation points can provide data support for the subsequent formulation of response measures for the abnormal area.

[0110] like Figure 2 As shown, Figure 2 2 is a structural diagram of a control device provided in an embodiment of the present application. The present invention also provides a control device 200, including:

[0111] The processor 210 may be implemented by a general-purpose central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0112] The memory 220 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 220 can store an operating system and other application programs. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 220, and the processor 210 calls and executes the non-grain monitoring method based on discrete wavelet in the embodiment of this application;

[0113] Input / output interface 230, used to implement information input and output;

[0114] Communication interface 240, used to realize communication interaction between the apparatus and other devices, which can be realized by wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.);

[0115] bus 250 , which transmits information between the various components of the device (e.g., processor 210 , memory 220 , input / output interface 230 , and communication interface 240 );

[0116] The processor 210 , the memory 220 , the input / output interface 230 , and the communication interface 240 are connected to each other in communication within the device via the bus 250 .

[0117] In addition, an embodiment of the present application further provides an electronic device, including the control device 200 of the above embodiment.

[0118] In addition, an embodiment of the present application further provides a storage medium, which is a computer-readable storage medium and stores a computer program. When the computer program is executed by a processor, the above-mentioned non-grain monitoring method based on discrete wavelets is implemented.

[0119] As a non-transient computer-readable storage medium, the memory can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage devices. In some embodiments, the memory may optionally include a memory remotely arranged relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned networks include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof. The device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and are implemented to be located in one place, or may also be distributed to multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment.

[0120] It will be appreciated by those skilled in the art that all or some of the steps and systems in the disclosed method above may be implemented as software, firmware, hardware and appropriate combinations thereof. Some physical components or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor or a microprocessor, or may be implemented as hardware, or may be implemented as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or a non-transitory medium) and a communication medium (or a temporary medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that may be used to store desired information and may be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically include computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0121] The above is a specific description of the preferred implementation of the present invention, but the present invention is not limited to the above-mentioned implementation mode. Technical personnel familiar with the field can also make various equivalent deformations or substitutions under the shared conditions without violating the spirit of the present invention. These equivalent deformations or substitutions are all included in the scope defined by the claims of the present invention.

Claims

1. A method for monitoring non-grain production based on discrete wavelet, characterized in that: include: Acquire the first remote sensing image data and the second remote sensing image data according to the historical time series, calculate the first EVI index of the target monitoring area based on the band synthesis method and the first remote sensing impact data, and calculate the second EVI index of the target monitoring area based on the band synthesis method and the second remote sensing impact data, wherein the land use type corresponding to the target monitoring area is cultivated land, the first remote sensing image data is the sample data of the target monitoring area, and the second remote sensing image data is the overall data of the target monitoring area; Establishing a first time series data set within a preset monitoring time period based on the first EVI index, and establishing a second time series data set within the preset monitoring time period based on the second EVI index; Performing discrete wavelet transform on each first vegetation index in the first time series data set to obtain each first vegetation index sequence, and performing discrete wavelet transform on each second vegetation index in the second time series data set to obtain each second vegetation index sequence, wherein there is a first time period relationship between the first vegetation index sequence and the corresponding discrete wavelet transform number, and there is a second time period relationship between the second vegetation index sequence and the corresponding discrete wavelet transform number; Determine each signal marker point in the target monitoring area, and calculate the first signal component of the first vegetation index sequence corresponding to each signal marker point on the quarterly, semi-annual and inter-annual scales according to the first time period relationship, and calculate the second signal component of the second vegetation index sequence corresponding to each signal marker point on the quarterly, semi-annual and inter-annual scales according to the second time period relationship, wherein the signal marker point is used to indicate the signal marker point of the known food crop type in the cultivated land, and different signal marker points correspond to different food crop types; A reference eigenvalue is calculated based on all of the first signal components, and a target eigenvalue is calculated based on all of the second signal components, wherein the reference eigenvalue includes the mean of the first inter-annual or larger scale signal component, the number of valid cycles of the first quarter signal component, and the number of valid cycles of the first half-year signal component corresponding to each of the signal landmark points, and the target eigenvalue is used to indicate the mean of the second inter-annual or larger scale signal component, the number of valid cycles of the second quarter signal component, and the number of valid cycles of the second half-year signal component corresponding to each pixel in the second remote sensing image data; An abnormal area is determined based on the first inter-annual or higher scale signal component mean, the first quarter signal component effective cycle number, the first semi-annual signal component effective cycle number, the second inter-annual or higher scale signal component mean, the second quarter signal component effective cycle number, and the second semi-annual signal component effective cycle number, wherein the abnormal area is used to indicate a sub-area in the target monitoring area where the phenomenon of cultivated land being converted to non-grain occurs; The MK mutation point test algorithm is used to detect the non-grain mutation point corresponding to the abnormal area, and the non-grain occurrence time and the corresponding non-grain change characteristic information are determined based on the non-grain mutation point.

2. The non-grain monitoring method based on discrete wavelet according to claim 1 is characterized in that: The mother wavelet corresponding to the discrete wavelet transform operation is the Meyer wavelet, and the first time period relationship is constructed according to the following method: Determining the center frequency of the Meyer wavelet; determining a temporal resolution of the first time series data set; The first time period relationship is constructed according to the center frequency, the time resolution, and the number of discrete wavelet transforms.

3. The non-grain monitoring method based on discrete wavelet according to claim 2 is characterized in that: The expression of the first time period relationship is as follows: P=2 n-1 △t / V c ; Wherein, P is the first time period, n is the number of discrete wavelet transforms, Δt is the time resolution of the first time series data set, and V is the center frequency of the Meyer wavelet.

4. The non-grain monitoring method based on discrete wavelet according to claim 2 is characterized in that: Calculating the first signal components of the first vegetation index sequence corresponding to each of the signal marker points at the quarterly, semi-annual and interannual scales or above according to the first time period relationship includes: Determine a first transformation number, a second transformation number, and a third transformation number, wherein the first transformation number corresponds to the number of discrete wavelet transformations corresponding to the first signal component on a quarterly scale, the second transformation number corresponds to the number of discrete wavelet transformations corresponding to the first signal component on a semi-annual scale, and the third transformation number corresponds to the number of discrete wavelet transformations corresponding to the first signal component on an inter-annual scale or above, the first transformation number is less than the second transformation number, and the second transformation number is less than the third transformation number; Performing discrete wavelet transform on the first vegetation index sequence corresponding to the signal marker point to obtain corresponding low-frequency signal and high-frequency signal, and recording the current wavelet transform times; Performing discrete wavelet transform on the low-frequency signal and the high-frequency signal again until the current number of wavelet transforms reaches the third number of transforms; Obtain a first high-frequency signal corresponding to the first number of transformations, a second high-frequency signal corresponding to the second number of transformations, and a first low-frequency signal corresponding to the third number of transformations, wherein the first high-frequency signal is the first signal component on a quarterly scale, the second high-frequency signal is the first signal component on a semi-annual scale, and the first low-frequency signal is the first signal component on a scale above an annual scale.

5. The method for monitoring non-grain consumption based on discrete wavelet according to claim 4 is characterized in that: The first vegetation index sequence corresponding to one of the signal marker points corresponds to a plurality of the first high-frequency signals, a plurality of the second high-frequency signals, and a plurality of the first low-frequency signals, and a reference eigenvalue is calculated based on all of the first signal components, including: Calculate the mean of all the first low-frequency signals to obtain the mean of the first signal component at a scale above interannual; Calculate the effective cycle number of the first quarter signal component by using the zero-crossing detection method and all the first high-frequency signals; The number of effective cycles of the first semi-annual signal component is calculated using a zero-crossing detection method and all of the second high-frequency signals.

6. The method for monitoring non-grain waste based on discrete wavelet according to claim 1, characterized in that: Determining an abnormal area based on the first inter-annual or higher scale signal component mean, the first quarter signal component effective cycle number, the first semi-annual signal component effective cycle number, the second inter-annual or higher scale signal component mean, the second quarter signal component effective cycle number, and the second semi-annual signal component effective cycle number includes: When the mean value of the second inter-annual or larger scale signal component is less than or equal to the mean value of the first inter-annual or larger scale signal component, the area of ​​the current pixel in the second remote sensing image data is determined as the abnormal area; or, When the mean value of the second inter-annual or higher-scale signal component is greater than the mean value of the first inter-annual or higher-scale signal component, when the number of valid cycles of the second quarterly signal component is less than the number of valid cycles of the first quarterly signal component, or the number of valid cycles of the second semi-annual signal component is less than the number of valid cycles of the first semi-annual signal component, the area of ​​the current pixel in the second remote sensing image data is determined as the abnormal area.

7. The method for monitoring non-grain consumption based on discrete wavelet according to claim 1, characterized in that: The MK mutation point test algorithm is used to detect the non-grain mutation point corresponding to the abnormal area, including: Acquire a first sequence set and a second sequence set corresponding to the abnormal area from the second time series data set, wherein the first sequence set is a sequence set arranged in chronological order, and the sequence order in the second sequence set is opposite to that in the first sequence set; Performing a discrete wavelet transform process on each sequence in the first sequence set to obtain a third sequence set; Performing a discrete wavelet transform process on each sequence in the second sequence set to obtain a fourth sequence set; Acquire a preset number of first sequence samples from the third sequence set, and calculate a first sequence mean and a first sequence variance of the first sequence samples; Acquire a preset number of second sequence samples from the fourth sequence set, and calculate a second sequence mean and a second sequence variance of the second sequence samples; Calculating a first statistic based on the first sequence samples, the first sequence mean, and the first sequence variance, and drawing a first curve based on the first statistic; Calculating a second statistic based on the second sequence samples, the second sequence mean, and the second sequence variance, multiplying the second statistic by -1 to obtain a third statistic, and drawing a second curve based on the third statistic; The intersection point between the first curve and the second curve is determined as the non-grain mutation point.

8. A control device, characterized in that: It includes at least one control processor and a memory for communicating with the at least one control processor; the memory stores instructions that can be executed by the at least one control processor, and the instructions are executed by the at least one control processor so that the at least one control processor can execute the discrete wavelet-based non-grain monitoring method as described in any one of claims 1 to 7.

9. An electronic device, characterized in that: Comprising the control device as claimed in claim 8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the non-grain monitoring method based on discrete wavelets as described in any one of claims 1 to 7.

Images