Soybean flowering period extraction method and device based on shape model

Abstract

The application discloses a soybean flowering period extraction method and device based on a shape model, and the method comprises the following steps: obtaining a corresponding flowering period shape model reference curve of soybeans in a sampling area according to remote sensing observation data; wherein the flowering period shape model reference curve comprises a reference flowering period of the soybeans in the sampling area; obtaining a target feature curve of a target pixel in the sampling area according to the remote sensing observation data; wherein the target feature curve comprises a vegetation index time series curve of the soybeans in the target pixel; performing shape matching on the flowering period shape model reference curve and the target feature curve to obtain a matching relationship between the flowering period shape model reference curve and the target feature curve; and extracting a target flowering period of the soybeans in the target pixel according to the matching relationship. The application simplifies the soybean flowering period extraction steps, saves the cost of the soybean flowering period extraction, and improves the accuracy of the soybean flowering period extraction.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of data processing, in particular to a soybean flowering period extraction method and device based on a shape model. BACKGROUND

[0002] Soybean is an important economic crop, and the extraction of the flowering period of soybean plays a crucial role in the construction of agricultural information systems and effective crop management. The growth period of soybean generally refers to the number of days from emergence to maturity, and the whole process from seed germination to emergence, seedling growth, flower bud differentiation, flowering and pod setting, grain filling, and finally mature new seeds. The flowering period is a key growth period for dividing soybean from the growth and development stage to the reproductive development stage.

[0003] In related technologies, the extraction of the flowering period of soybean is mainly realized by field observation by observation personnel. However, due to the subjective factors of observation personnel, it is difficult to ensure the quality of the obtained data, and the labor cost and time cost are high, the observation area is small, and it is difficult to realize large-scale monitoring. SUMMARY

[0004] The embodiments of the present application provide a soybean flowering period extraction method and device based on a shape model, which can reduce the cost of soybean flowering period extraction.

[0005] Therefore, the first aspect of the present application provides a soybean flowering period extraction method based on a shape model, which comprises:

[0006] According to the remote sensing observation data, a flowering period shape model reference curve corresponding to the soybean in the sampling area is obtained; wherein the flowering period shape model reference curve includes a reference flowering period of the soybean in the sampling area;

[0007] According to the remote sensing observation data, a target feature curve of a target pixel in the sampling area is obtained; wherein the target feature curve includes a vegetation index time series curve of the soybean in the target pixel;

[0008] The flowering period shape model reference curve and the target feature curve are shape matched to obtain a matching relationship between the flowering period shape model reference curve and the target feature curve;

[0009] According to the matching relationship, a target flowering period of the soybean in the target pixel is extracted.

[0010] Preferably, the shape matching of the flowering period shape model reference curve and the target feature curve to obtain the matching relationship between the flowering period shape model reference curve and the target feature curve comprises:

[0011] time-shifting and / or scaling the flowering period shape model reference curve to obtain a first feature curve; wherein the flowering period shape model reference curve comprises a time series curve of a vegetation index of soybeans in all pixels in the sampling area;

[0012] calculating a correlation between each curve segment in the target feature curve and the first feature curve in the target feature curve;

[0013] determining a curve segment in the target feature curve that meets a correlation condition as a second feature curve, and obtaining a matching relationship between the first feature curve and the second feature curve.

[0014] Preferably, the target flowering period of soybeans in the target pixel according to the matching relationship comprises:

[0015] determining a feature point corresponding to the reference flowering period in the second feature curve according to the matching relationship between the first feature curve and the second feature curve, and determining a time corresponding to the feature point as the target flowering period of soybeans in the target pixel.

[0016] Preferably, the obtaining of the flowering period shape model reference curve of soybeans in the sampling area according to remote sensing observation data comprises:

[0017] obtaining time series data of a vegetation index of soybeans in all pixels in the sampling area according to remote sensing observation data;

[0018] obtaining a shape model reference curve and a reference flowering period of soybeans in the sampling area according to the time series data of the vegetation index;

[0019] cutting a curve segment corresponding to a preset time period in the shape model reference curve as a flowering period shape model reference curve according to the reference flowering period; wherein the time corresponding to the reference flowering period is a time in the preset time period.

[0020] Preferably, the cutting of the curve segment corresponding to the preset time period in the shape model reference curve as the flowering period shape model reference curve according to the reference flowering period comprises:

[0021] cutting a curve segment corresponding to a preset time period in the shape model reference curve as a flowering period shape model reference curve with the reference flowering period as a midpoint in the shape model reference curve; wherein the time corresponding to the reference flowering period is a midpoint of the preset time period.

[0022] Preferably, the obtaining of the flowering period shape model reference curve of soybeans in the sampling area according to remote sensing observation data comprises:

[0023] obtain spatial distribution data of soybeans in the sampling area according to remote sensing observation data, wherein the spatial distribution data of soybeans comprises a soybean planting density of each pixel in the sampling area;

[0024] obtain vegetation index time series data of a first pixel according to the spatial distribution data of soybeans, wherein the first pixel is a pixel with a soybean planting density not less than a density threshold;

[0025] obtain the flowering period shape model reference curve according to the vegetation index time series data of the first pixel.

[0026] The second aspect of the application provides a soybean flowering period extraction device based on a shape model, characterized in that the device comprises:

[0027] a data analysis unit configured to obtain a flowering period shape model reference curve of soybeans in a sampling area according to remote sensing observation data, wherein the flowering period shape model reference curve comprises a reference flowering period of soybeans in the sampling area;

[0028] The data analysis unit is further configured to obtain a target feature curve of a target pixel in the sampling area according to the remote sensing observation data, wherein the target feature curve comprises a vegetation index time series curve of soybeans in the target pixel;

[0029] a shape model matching unit configured to perform shape matching between the flowering period shape model reference curve and the target feature curve to obtain a matching relationship between the flowering period shape model reference curve and the target feature curve;

[0030] a flowering period extraction unit configured to extract a target flowering period of soybeans in the target pixel according to the matching relationship.

[0031] Preferably, the shape model matching unit is specifically configured to:

[0032] perform time translation and / or stretching on the flowering period shape model reference curve to obtain a first feature curve, wherein the flowering period shape model reference curve comprises a vegetation index time series curve of soybeans in all pixels in the sampling area;

[0033] calculate a correlation between each curve segment in the target feature curve and the first feature curve in the target feature curve;

[0034] determine a curve segment in the target feature curve satisfying a correlation condition as a second feature curve, and obtain a matching relationship between the first feature curve and the second feature curve.

[0035] Preferably, the flowering period extraction unit is specifically configured to:

[0036] According to the matching relationship between the first feature curve and the second feature curve, a feature point corresponding to the reference flowering period is determined in the second feature curve, and a time corresponding to the feature point is determined as the target flowering period of the soybean in the target pixel.

[0037] Preferably, the data analysis unit is specifically used for:

[0038] According to remote sensing observation data, vegetation index time series data of soybeans in all pixels in the sampling area are obtained;

[0039] According to the vegetation index time series data, a shape model reference curve and a reference flowering period of soybeans in the sampling area are obtained;

[0040] According to the reference flowering period, a curve segment corresponding to a preset time period in the shape model reference curve is intercepted as a flowering period shape model reference curve; wherein the time corresponding to the reference flowering period is a time in the preset time period.

[0041] From the above technical solutions, the present application has the following advantages: according to remote sensing observation data, the present application obtains a flowering period shape model reference curve corresponding to soybeans in a sampling area; wherein the flowering period shape model reference curve includes a reference flowering period of soybeans in the sampling area; according to remote sensing observation data, a target feature curve of a target pixel in the sampling area is obtained; wherein the target feature curve includes a vegetation index time series curve of soybeans in the target pixel; through analysis of a large range of remote sensing observation data, the flowering period of soybeans is extracted pixel by pixel, and the rasterization of soybean remote sensing observation data is realized; the flowering period shape model reference curve and the target feature curve are shape-matched to obtain a matching relationship between the flowering period shape model reference curve and the target feature curve; according to the matching relationship, the target flowering period of soybeans in the target pixel is extracted; through the vegetation index time series curve, the growth phenology of soybeans is embodied, based on the stability of the growth phenology of soybeans, the embodiment of the flowering period of soybeans on the vegetation index time series curve has stability, which simplifies the extraction steps of the flowering period of soybeans, saves the cost of the extraction of the flowering period of soybeans, and improves the accuracy of the extraction of the flowering period of soybeans. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 A flowchart of a soybean flowering period extraction method based on a shape model is provided for an embodiment of the present application;

[0043] Figure 2 A flowchart of a soybean flowering period extraction method based on a shape model is provided for a scene embodiment of the present application;

[0044] Figure 3A shape model reference curve provided by the scenario embodiment of the present application and a vegetation index time series curve of a first pixel are shown in the figure;

[0045] Figure 4 A shape model reference curve provided by the scenario embodiment of the present application and a target feature curve are shown in the figure;

[0046] Figure 5 A spatial distribution map of soybean flowering period provided by the scenario embodiment of the present application is shown in the figure;

[0047] Figure 6 A device for extracting soybean flowering period based on a shape model provided by the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0048] Embodiments of the present application will be described in more detail by referring to the attached drawings. Although certain embodiments of the present application are shown in the drawings, it should be understood that the present application can be implemented in various forms and should not be interpreted as being limited to the embodiments set forth herein, rather these embodiments are provided to make the present application more thorough and complete. It should be understood that the drawings and embodiments of the present application are only for exemplary purposes and are not intended to limit the scope of protection of the present application.

[0049] Referring to Figure 1 The embodiment of the present application provides a method for extracting soybean flowering period based on a shape model, which comprises the following steps:

[0050] Step 101: According to remote sensing observation data, a shape model reference curve of soybean flowering period in a sampling area is obtained.

[0051] The shape model reference curve of flowering period includes a reference flowering period of soybean in the sampling area. The reference flowering period is the time when all the soybeans in the sampling area enter the flowering period, for example, when the observation is performed, the time when 50% of the area proportion of soybeans in the sampling area enter the flowering period can be taken as the time when all the soybeans in the sampling area enter the flowering period, of course, the reference flowering period of soybeans in the sampling area can also be determined by other ways, which is not specifically limited in the present application.

[0052] In the collection of remote sensing observation data, a pixel is the minimum unit of the ground scene scanned and sampled by the sensor, and the remote sensing observation data includes the vegetation index corresponding to each pixel in the sampling area at different times. The vegetation index can be used to detect the growth period of plants. In the embodiments of the present application, the extraction of the flowering period of soybeans is performed by a vegetation index time series curve. The vegetation index time series curve is a curve drawn in an x-o-y coordinate system, the x-axis is the observation time, and the y-axis is the vegetation index obtained at the aforementioned observation time. For example, the x-axis is the day of year (DOY), and the y-axis is the vegetation index corresponding to each day of year.

[0053] It should be noted that the remote sensing observation data can be remote sensing observation data obtained by a visible infrared imaging radiometer (VIIRS) remote sensing satellite. The VIIRS remote sensing satellite has a high revisit frequency and can continuously observe the ground surface. The spatial resolution of 500 m can initially meet the needs of crop management.

[0054] In a possible implementation, step 101 can be implemented in the following manner:

[0055] Step 1011: According to the remote sensing observation data, obtain the vegetation index time series data of soybeans in all pixels in the sampling area.

[0056] In the vegetation index time series data obtained by the remote sensing observation data, the vegetation index corresponding to each pixel in the sampling area at each day of year is included. According to the aforementioned vegetation index, a vegetation index time series image of soybeans in the sampling area can be drawn, which is used as a reference shape model for extracting the flowering period of soybeans in the sampling area.

[0057] It should be noted that the vegetation index time series data described in the embodiments of the present application can be drawn in the form of a curve in an x-o-y coordinate system, i.e., a vegetation index time series curve, which is only one form of a vegetation index time series image and should not be understood as a limitation on the methods described in the present application.

[0058] Further, since the remote sensing observation data is affected by the weather, when the remote sensing observation data is used to obtain the vegetation index time series image, the vegetation index time series image can be preprocessed, such as smoothing the image by linear interpolation and filtering.

[0059] Step 1012: According to the vegetation index time series data, obtain the shape model reference curve and the reference flowering period of soybeans in the sampling area.

[0060] The shape model reference curve of the soybeans in the sampling area and the reference flowering period are obtained by sampling the vegetation index time series data of each pixel in the sampling area; the shape model reference curve is a vegetation index time series curve of the soybeans in the sampling area in units of pixels, used to describe the growth and development of the soybeans in the sampling area and as a reference shape model for extracting the flowering period of the soybeans in the sampling area; and the reference flowering period is the time corresponding to the overall flowering period of the soybeans in the sampling area.

[0061] For example, in the shape model reference curve, the vegetation index corresponding to the ith accumulated day can be obtained by averaging the vegetation indices of all pixels on the ith accumulated day, where i is an integer greater than or equal to 1; of course, a weighted average can also be obtained according to the planting density of the soybeans in each pixel; the shape model reference curve can also be obtained by other calculation methods, which are not specifically limited in the present application.

[0062] Step 1013: According to the reference flowering period, the curve segment corresponding to the preset time period in the shape model reference curve is intercepted as the flowering period shape model reference curve.

[0063] The time corresponding to the reference flowering period is the time in the preset time period. In order to facilitate shape matching, the curve segment corresponding to the preset time period including the reference flowering period in the shape model reference curve is intercepted as the flowering period shape model reference curve for extracting the flowering period of the soybeans in the sampling area.

[0064] The curve segment including the reference flowering period is selected from the shape model reference curve as the flowering period shape model reference curve for shape matching, which improves the efficiency of shape matching.

[0065] Further, when the flowering period shape model reference curve is intercepted from the shape model reference curve, it can be achieved by the following way:

[0066] In the shape model reference curve, the curve segment corresponding to the preset time period is intercepted as the flowering period shape model reference curve with the reference flowering period as the midpoint.

[0067] The time corresponding to the reference flowering period is the midpoint of the preset time period. With the reference flowering period as the midpoint on the time series, the same length of time period is intercepted forward and backward respectively, the curve segment corresponding to the preset time period is obtained and used as the flowering period shape model reference curve.

[0068] In one possible implementation, the reference flowering period can also be used as the midpoint of the curve segment to intercept the flowering period shape model reference curve.

[0069] Since the reference flowering period describes the time when the whole soybeans in the sampling area enter the flowering period, and when the flowering period of each pixel in the sampling area is extracted, there may be some error between the flowering period corresponding to each pixel and the reference flowering period. Therefore, the time corresponding to the reference flowering period is taken as the midpoint of the preset time period, and the flowering period shape model reference curve corresponding to the preset time period is intercepted, so as to retain more information on both sides of the reference flowering period, improve the shape matching efficiency, and improve the accuracy of soybean flowering period extraction.

[0070] In a possible implementation, step 101 can be implemented by the following steps, specifically including:

[0071] Step S1: obtaining soybean spatial distribution data in the sampling area according to remote sensing observation data.

[0072] The soybean spatial distribution data includes the soybean planting density of each pixel in the sampling area.

[0073] Step S2: obtaining the vegetation index time series data of the first pixel according to the soybean spatial distribution data.

[0074] The first pixel is a pixel with a soybean planting density not less than a density threshold. From the sampling area, a pixel with a higher soybean planting density is selected, and the vegetation index time series data of the pixel is obtained according to the remote sensing observation data.

[0075] Step S3: obtaining the flowering period shape model reference curve according to the vegetation index time series data of the first pixel.

[0076] By judging the pixel with a soybean planting density not less than the density threshold as the first pixel, and generating the flowering period shape model reference curve according to the vegetation index time series data of all the first pixels, the pixel with a higher soybean planting density is used as the first pixel for generating the flowering period shape model reference curve, so that the flowering period shape model reference curve is closer to the real situation, and thus the accuracy of soybean flowering period extraction is improved.

[0077] Step 102: obtaining a target feature curve of a target pixel in the sampling area according to remote sensing observation data.

[0078] The target feature curve includes the vegetation index time series curve of the soybeans in the target pixel. The target pixel is a pixel whose flowering period is to be extracted. In the remote sensing observation data, the vegetation index corresponding to the soybeans in the target pixel at different observation times is obtained, and the vegetation index time series curve of the soybeans in the target pixel, i.e. the target feature curve, is drawn according to the correspondence between the observation time and the vegetation index, which is used for shape matching with the flowering period shape model reference curve in the subsequent step.

[0079] Step 103: Shape matching the flowering stage shape model reference curve with the target feature curve to obtain the matching relationship between the flowering stage shape model reference curve and the target feature curve.

[0080] The flowering stage shape model reference curve of the soybean in the sampling area obtained in step 101 is taken as a shape model, and shape matching is performed between the shape model and the target feature curve of the target pixel at the flowering stage to be extracted. In the matching process, the flowering stage shape model reference curve is translated and / or stretched in time, the curve segment most similar to the shape model is searched for in the target feature curve, and the corresponding relationship, i.e., the matching relationship, between the aforementioned curve segment and the flowering stage shape model reference curve is obtained, and each point on the aforementioned curve segment is one-to-one corresponding to each point on the flowering stage shape model reference curve.

[0081] Specifically, step 103 can be implemented in the following manner:

[0082] Step 1031: The flowering stage shape model reference curve is translated and / or stretched in time to obtain a first feature curve.

[0083] The flowering stage shape model reference curve includes the vegetation index time series curve of the soybean in all pixels in the sampling area. Since there are certain differences in the growth environment of the soybean in different pixels, the growth and development of the soybean in different pixels also have time differences. In the pixel with superior environment, the flowering stage of the soybean can be earlier than the reference flowering stage; and in the pixel with poor environment, the flowering stage of the soybean can be later than the reference flowering stage. Therefore, by translating the flowering stage shape reference curve in time, the time difference between the flowering stages of the soybean in different pixels can be reduced. Meanwhile, the growth and development cycle of the soybean can also have certain differences between different pixels. The growth and development cycle of the soybean in some pixels is longer, and the growth and development cycle of the soybean in some pixels is shorter. Therefore, by stretching the flowering stage shape model reference curve in time, the vegetation index time series curve of the soybean between pixels with different growth and development cycles can be matched.

[0084] Step 1032: In the target feature curve, the correlation between each curve segment in the target feature curve and the first feature curve is calculated.

[0085] The first feature curve obtained by translating and / or stretching the flowering stage shape model reference curve is matched with the target feature curve. In the target feature curve, different curve segments are intercepted, and the correlation between each curve segment and the first feature curve is calculated to determine the matching degree between each curve segment and the first feature curve.

[0086] Specifically, in order to improve the matching efficiency, when the curve segment is intercepted in the target feature curve, a curve segment with the same length on the x-axis as the first feature curve can be selected for shape matching with the first feature curve. For example, when the first feature curve corresponds to a time period between t1 and t2 on the x-axis, a certain curve segment intercepted in the target feature curve corresponds to a time period between t3 and t4 on the x-axis, and the length of the time period between t1 and t2 is equal to the length of the time period between t3 and t4.

[0087] Step 1033: determining a curve segment in the target feature curve that meets the correlation condition as a second feature curve, and obtaining a matching relationship between the first feature curve and the second feature curve.

[0088] By the correlation between the first feature curve and each curve segment in the target feature curve, the matching degree between the first feature curve and each curve segment is determined, the curve segment that meets the correlation condition is determined as the second feature curve, the points on the first feature curve and the points on the second feature curve are one-to-one corresponding, and the matching relationship between the first feature curve and the second feature curve is obtained. The curve segment that meets the correlation condition can be the curve segment with the greatest correlation with the first feature curve. Of course, other correlation conditions can also be set to obtain the second feature curve, which is not limited in the present application.

[0089] By translating and / or scaling the flowering period shape model reference curve in time to obtain the first feature curve, shape matching the target feature curve using the first feature curve, and determining the second feature curve through the correlation condition to obtain the matching relationship between the first feature curve and the second feature curve, the error of shape matching is reduced, and the accuracy of soybean flowering period extraction is improved.

[0090] Step 104: extracting the target flowering period of soybean in the target pixel according to the matching relationship.

[0091] In the flowering period shape model reference curve, the reference flowering period is included. After shape matching, according to the matching relationship between the flowering period shape model reference curve and the target feature curve, the corresponding relationship between each point on the flowering period shape model reference curve and each point on the target feature curve can be obtained. According to the aforementioned corresponding relationship, the point on the target feature curve corresponding to the reference flowering period is extracted as the target flowering period of soybean in the target pixel.

[0092] Specifically, on the basis of steps 1031 to 1033, step 104 can be implemented in the following way:

[0093] According to the matching relationship between the first characteristic curve and the second characteristic curve, a characteristic point corresponding to the reference flowering period is determined in the second characteristic curve, and a time corresponding to the characteristic point is determined as the target flowering period of the soybean in the target pixel.

[0094] According to the remote sensing observation data, the flowering period shape model reference curve corresponding to the soybean in the sampling area is obtained; the flowering period shape model reference curve includes the reference flowering period of the soybean in the sampling area; according to the remote sensing observation data, the target characteristic curve of the target pixel in the sampling area is obtained; the target characteristic curve includes the vegetation index time series curve of the soybean in the target pixel; through analysis of a large range of remote sensing observation data, the soybean flowering period is extracted pixel by pixel, and the rasterization of the soybean remote sensing observation data is realized; the flowering period shape model reference curve and the target characteristic curve are shape matched to obtain the matching relationship between the flowering period shape model reference curve and the target characteristic curve; according to the matching relationship, the target flowering period of the soybean in the target pixel is extracted; the vegetation index time series curve reflects the growth phenology of the soybean, and based on the stability of the growth phenology of the soybean, the reflection of the soybean flowering period on the vegetation index time series curve has stability, which simplifies the soybean flowering period extraction step, saves the cost of soybean flowering period extraction, and improves the accuracy of soybean flowering period extraction.

[0095] Referring to Figure 2 The scene embodiment of the present application provides a soybean flowering period extraction method based on a shape model, including the following steps:

[0096] Step 201: Obtain the spatial distribution data of soybean in the sampling area and the flowering period time observation data of soybean.

[0097] The spatial distribution data of soybean in the sampling area is obtained according to the remote sensing observation data obtained by the remote sensing satellite, the flowering period time observation data of soybean is obtained according to the ground observation data, and the time when all the pixels in the sampling area enter the flowering period as a whole can be obtained according to the flowering period time observation data of soybean, and the time when 50% of the soybean in the sampling area enters the flowering period is taken as the reference flowering period.

[0098] The spatial distribution data of soybean is the crop data layer (CDLs) data, and the area proportion of the soybean entering the flowering period in the sampling area is recorded in the form of area proportion in the flowering period time observation data of soybean. The time when the area proportion of the soybean entering the flowering period reaches 50% is obtained by linear interpolation, which is taken as the reference flowering period of the soybean in the sampling area.

[0099] Step 202: Obtain the vegetation index time series curve and data preprocessing.

[0100] According to the soybean spatial distribution data, the vegetation index time series curve of the soybean in each pixel in the sampling area is obtained, and the pixel with the soybean planting density of more than 75% is selected as the target pixel, and the vegetation index time series curve of the target pixel, i.e., the target feature curve, is obtained.

[0101] The calculation formula of the vegetation index is as follows:

[0102]

[0103] EVI2 is a dual-band enhanced vegetation index, which can make up for the lack of blue light band in the remote sensing observation data obtained by the VIIRS remote sensing satellite; ρ NIR is the near-infrared reflectance, ρ red is the red reflectance.

[0104] In the data preprocessing, since the resolution of the CDLs data is 30 meters, and the resolution of the remote sensing observation data obtained by the VIIRS remote sensing satellite is 500 meters, in order to unify the data resolution, the CDLs data is resampled, and the resolution of the CDLs is aggregated from 30 meters to 500 meters by the density average method. Further, the soybean spatial distribution data can also be filtered to filter out inaccurate data obtained due to weather and other reasons.

[0105] Step 203: Obtain the shape model reference curve.

[0106] In the sampling area, the pixel with the soybean planting density close to 100% is selected as the first pixel, the shape model reference curve of the soybean in the sampling area is obtained according to the vegetation index time series curve of all the first pixels, and the feature point corresponding to the reference flowering period is marked on the shape model reference curve according to the soybean flowering period observation data. As shown in Figure 3 , in Figure 3 , the x-axis is the accumulated day, the y-axis is the vegetation index corresponding to each accumulated day, the reference curve is the shape model reference curve of the soybean in the sampling area, the first curve is the vegetation index time series curve of all the first pixels, the feature point corresponding to the reference flowering period is marked on the reference curve, and the time when the whole soybean in the sampling area enters the flowering period is marked on the horizontal axis, and the vegetation index of the soybean in the sampling area at the time is marked on the vertical axis.

[0107] Step 204: Shape model fitting.

[0108] Referring to Figure 4As shown, the reference curve is a shape model reference curve of soybeans in the sampling area, and the target curve is a vegetation index time series curve of soybeans in the target pixel, i.e., a target feature curve. According to the shape model reference curve, a time series of half-window length is respectively intercepted before and after the reference flowering period, and the curve segment corresponding to the time series in the shape model reference curve is taken as the flowering period shape model reference curve. The length of the time series corresponding to the flowering period shape model reference curve is twice the half-window length, i.e., a preset window length, and the time period corresponding to the preset window length is a preset time period. The best estimation of the target flowering period of the soybeans to be extracted in the target pixel is determined according to the correlation between the target feature curve and the flowering period shape model reference curve. After the flowering period shape model reference curve is translated and / or scaled in time with the reference flowering period as the center, a first feature curve is obtained. According to the correlation between the first feature curve and the corresponding curve segment in the target feature curve, the fitting parameter corresponding to the maximum correlation is taken as the target flowering period of the soybeans in the target pixel. The first feature curve and the corresponding curve segment in the target feature curve have the same length of time series.

[0109] In the process of shape model fitting, the translation and / or scaling of the flowering period shape model reference curve in time is calculated by the following formula:

[0110]

[0111] wherein, is the first feature curve obtained after the flowering period shape model reference curve is translated and / or scaled, g(x) is the flowering period shape model reference curve, tshift is the translation parameter of the flowering period shape model reference curve in the x-axis direction, xscale is the scaling parameter of the flowering period shape model reference curve in the x-axis direction, and p0 is the reference flowering period.

[0112] The correlation between the first feature curve and the corresponding curve segment in the target feature curve is calculated, and the calculation formula is as follows:

[0113]

[0114] x∈(p0-tshift-w,p0-tshift+w)

[0115]

[0116] wherein, r is the correlation between the first feature curve and the corresponding curve segment h(x) in the target feature curve, h(x) is the curve segment in the target feature curve corresponding to the first feature curve; x is the daily corresponding time series value, y is the daily corresponding vegetation index, N represents the length of the time series within the preset window length, and w represents the half-window length.

[0117] Step 205: Extract the soybean flowering period time from the target pixel.

[0118] Based on the tshift and xscale values ​​corresponding to the highest correlation, the target flowering period is extracted from the target feature curve. The formula for extracting the target flowering period is as follows:

[0119] xscale×p est +(1-xscale)×p0+xscale×tshift=p0

[0120] p est =p0-tshift

[0121] Where, p est The target flowering period is defined in the target feature curve. Using the method provided in this application embodiment, the soybean flowering period is extracted from three target pixels respectively. Compared with the flowering period obtained from actual observation data, the extraction errors of the target flowering period from the three target pixels are 1.61 days, 1.64 days and 1.93 days respectively. It can be seen that the method provided in this application embodiment improves the extraction accuracy and precision of soybean flowering period.

[0122] The time series curves of soybean vegetation index in each pixel of the sampling area obtained in step 202 are processed pixel by pixel using steps 204 and 205 to extract the flowering period of soybean in each pixel. A spatial distribution map of soybean flowering period is then drawn based on the flowering period, as shown below. Figure 5 As shown, the flowering period of soybeans in different pixels within the sampling area was rasterized to generate a spatial distribution map, which better reflects the spatial distribution of soybean crop growth progress and provides guidance for agricultural water and fertilizer management.

[0123] Although the operations are described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or in a sequential order. In certain environments, multitasking and parallel processing may be advantageous.

[0124] It should be understood that the steps described in the method embodiments of this application may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of this application is not limited in this respect.

[0125] See Figure 6 As shown in the figure, this application provides a soybean flowering period extraction device based on a shape model, including: a data analysis unit 601, a shape model matching unit 602, and a flowering period extraction unit 603.

[0126] The data analysis unit 601 is configured to: obtain a flowering period shape model reference curve of soybeans in the sampling area according to remote sensing observation data; and wherein the flowering period shape model reference curve comprises a reference flowering period of the soybeans in the sampling area.

[0127] The data analysis unit 601 is further configured to: obtain a target feature curve of the target pixel in the sampling area according to the remote sensing observation data; and wherein the target feature curve comprises a vegetation index time series curve of the soybeans in the target pixel.

[0128] The shape model matching unit 602 is configured to: perform shape matching between the flowering period shape model reference curve and the target feature curve to obtain a matching relationship between the flowering period shape model reference curve and the target feature curve.

[0129] The flowering period extraction unit 603 is configured to: extract a target flowering period of the soybeans in the target pixel according to the matching relationship.

[0130] Further, the shape model matching unit 602 is specifically configured to:

[0131] perform time translation and / or stretching on the flowering period shape model reference curve to obtain a first feature curve; and wherein the flowering period shape model reference curve comprises a vegetation index time series curve of the soybeans in all pixels in the sampling area.

[0132] In the target feature curve, calculate a correlation between each curve segment in the target feature curve and the first feature curve.

[0133] determine a curve segment in the target feature curve that satisfies a correlation condition as a second feature curve, and obtain a matching relationship between the first feature curve and the second feature curve.

[0134] Further, the flowering period extraction unit 603 is specifically configured to:

[0135] determine a feature point corresponding to the reference flowering period in the second feature curve according to the matching relationship between the first feature curve and the second feature curve, and determine a time corresponding to the feature point as the target flowering period of the soybeans in the target pixel.

[0136] Further, the data analysis unit 601 is specifically configured to:

[0137] obtain vegetation index time series data of the soybeans in all pixels in the sampling area according to the remote sensing observation data;

[0138] obtain the shape model reference curve and the reference flowering period of the soybeans in the sampling area according to the vegetation index time series data;

[0139] According to the reference flowering period, a curve segment corresponding to the preset time period is intercepted in the shape model reference curve as the flowering period shape model reference curve; wherein the time corresponding to the reference flowering period is the time in the preset time period.

[0140] Further, the data analysis unit 601 is specifically configured to:

[0141] In the shape model reference curve, the reference flowering period is taken as a midpoint, and a curve segment corresponding to the preset time period is intercepted as the flowering period shape model reference curve; wherein the time corresponding to the reference flowering period is the midpoint of the preset time period.

[0142] Further, the data analysis unit 601 is specifically configured to:

[0143] According to the remote sensing observation data, the spatial distribution data of soybeans in the sampling area is obtained; wherein the spatial distribution data of soybeans includes the planting density of soybeans in each pixel in the sampling area;

[0144] According to the spatial distribution data of soybeans, the vegetation index time series data of the first pixel is obtained; wherein the first pixel is a pixel with a soybean planting density not less than a density threshold;

[0145] According to the vegetation index time series data of the first pixel, the flowering period shape model reference curve is obtained.

[0146] The names of the messages or information exchanged between the plurality of devices in the embodiments of the application are only used for illustrative purposes, and are not used to limit the scope of the messages or information.

[0147] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described system, device and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be repeated here.

[0148] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or other forms.

[0149] The units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, i.e. may be located in one place, or may be distributed on multiple network units. Part or all of the units may be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0150] In addition, the functional units in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0151] The integrated unit, if realized in the form of a software functional unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (Read-Only Memory, ROM), a random access memory (Random Access Memory, RAM), a magnetic disk or an optical disk, and various computer program storage media.

[0152] The above-described embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements for part of the technical features; and these modifications or replacements do not make the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for soybean flowering stage extraction based on shape model, characterized in that, The method comprises: According to remote sensing observation data, obtain the corresponding flowering period shape model reference curve of soybean in the sampling area; wherein, the flowering period shape model reference curve includes the reference flowering period of soybean in the sampling area; According to the remote sensing observation data, obtain the target feature curve of the target pixel in the sampling area; wherein, the target feature curve includes the vegetation index time series curve of soybean in the target pixel; The shape of the flowering period shape model reference curve and the target feature curve is matched to obtain the matching relationship between the flowering period shape model reference curve and the target feature curve; Wherein, the shape of the flowering period shape model reference curve and the target feature curve is matched to obtain the matching relationship between the flowering period shape model reference curve and the target feature curve, including: the flowering period shape model reference curve is translated and / or scaled in time to obtain a first feature curve; wherein, the flowering period shape model reference curve includes the vegetation index time series curve of soybean in all pixels in the sampling area; in the target feature curve, the correlation between each curve segment in the target feature curve and the first feature curve is calculated; the curve segment in the target feature curve that meets the correlation condition is determined as a second feature curve, and the matching relationship between the first feature curve and the second feature curve is obtained; According to the matching relationship, the target flowering period of soybean in the target pixel is extracted; Wherein, according to the matching relationship between the first feature curve and the second feature curve, the feature point corresponding to the reference flowering period is determined in the second feature curve, and the time corresponding to the feature point is determined as the target flowering period of soybean in the target pixel.

2. The method of claim 1, wherein, According to remote sensing observation data, obtain the corresponding flowering period shape model reference curve of soybean in the sampling area; wherein, the flowering period shape model reference curve includes the reference flowering period of soybean in the sampling area; According to remote sensing observation data, obtain the vegetation index time series data of soybean in all pixels in the sampling area; According to the vegetation index time series data, obtain the shape model reference curve and the reference flowering period of soybean in the sampling area; According to the reference flowering period, the curve segment corresponding to the preset time period in the shape model reference curve is intercepted as the flowering period shape model reference curve; wherein, the time corresponding to the reference flowering period is the time in the preset time period.

3. The method of claim 2, wherein, According to the reference flowering period, the curve segment corresponding to the preset time period in the shape model reference curve is intercepted as the flowering period shape model reference curve; wherein, the time corresponding to the reference flowering period is the midpoint of the preset time period. According to remote sensing observation data, obtain the corresponding flowering period shape model reference curve of soybean in the sampling area; wherein, the flowering period shape model reference curve includes the reference flowering period of soybean in the sampling area; 4. The method of claim 1, wherein, ​ According to remote sensing observation data, obtain the spatial distribution data of soybeans in the sampling area; wherein the spatial distribution data of soybeans includes the planting density of soybeans in each pixel in the sampling area; According to the spatial distribution data of soybeans, obtain the vegetation index time series data of the first pixel; wherein the first pixel is a pixel with a soybean planting density not less than a density threshold; According to the vegetation index time series data of the first pixel, obtain the flowering period shape model reference curve.

5. A shape model based soybean flowering stage extraction apparatus, characterized by, The device comprises: a data analysis unit configured to: obtain, according to remote sensing observation data, a flowering period shape model reference curve of soybeans in a sampling area; wherein the flowering period shape model reference curve includes a reference flowering period of soybeans in the sampling area; The data analysis unit is further configured to: obtain, according to the remote sensing observation data, a target feature curve of a target pixel in the sampling area; wherein the target feature curve includes a vegetation index time series curve of soybeans in the target pixel; a shape model matching unit configured to: perform shape matching between the flowering period shape model reference curve and the target feature curve to obtain a matching relationship between the flowering period shape model reference curve and the target feature curve; The shape model matching unit is specifically configured to: perform time translation and / or stretching on the flowering period shape model reference curve to obtain a first feature curve; wherein the flowering period shape model reference curve includes vegetation index time series curves of soybeans in all pixels in the sampling area; calculate the correlation between each curve segment in the target feature curve and the first feature curve in the target feature curve; determine the curve segment in the target feature curve that meets the correlation condition as a second feature curve, and obtain the matching relationship between the first feature curve and the second feature curve; a flowering period extraction unit configured to: extract a target flowering period of soybeans in the target pixel according to the matching relationship; The flowering period extraction unit is specifically configured to: determine a feature point corresponding to the reference flowering period in the second feature curve according to the matching relationship between the first feature curve and the second feature curve, and determine the time corresponding to the feature point as the target flowering period of soybeans in the target pixel.

6. The apparatus of claim 5, wherein, The data analysis unit is specifically configured to: obtain, according to remote sensing observation data, vegetation index time series data of soybeans in all pixels in the sampling area; obtain a shape model reference curve and a reference flowering period of soybeans in the sampling area according to the vegetation index time series data; According to the reference flowering period, intercept the curve segment corresponding to a preset time period in the shape model reference curve as the flowering period shape model reference curve; wherein the time corresponding to the reference flowering period is a time in the preset time period.

Images