A remote sensing identification method and system for crop rotation information

Abstract

The present invention relates to the field of plant growth monitoring technology, and specifically discloses a remote sensing identification method and system for crop rotation information, the method comprising acquiring spectral information of a region based on remote sensing technology, determining a crop area based on the spectral information; selecting characteristic crops in the crop area based on the spectral information, and synchronously determining the growth sequence of the characteristic crops; connecting the growth sequences of characteristic crops at the same position to obtain rotation characteristic sequences at different positions; comparing the rotation characteristic sequences at different positions to cluster the positions; determining the analysis value of each position according to the distribution of similar positions, selecting a position with high analysis value, and reading the historical environmental data at the position; and feeding back the historical environmental data and rotation characteristic sequence of the position to staff. The present invention selects the most valuable crops for analysis from large-scale crops based on remote sensing technology, and the screening is extremely accurate, which greatly reduces the data analysis pressure of staff.

Description

Technical Field

[0001] The invention relates to the technical field of plant growth monitoring, and in particular to a remote sensing identification method and system for crop rotation information. Background Art

[0002] Crop rotation is an agricultural planting method that refers to the planting of different crops on the same piece of land in a certain pattern, usually changing the crop type every year or every few seasons. This method helps to improve soil fertility, reduce pests and diseases, improve land structure, and increase agricultural production efficiency.

[0003] In actual scenarios, crops are planted in large quantities and over large areas. The growth data of these crops are different, and the planning process of crop rotation information needs to rely on the growth data of actual crops. It is obviously impossible to analyze all crops. Therefore, how to select crops with higher analysis value from existing crops and provide a reference for the crop rotation planning process is the technical problem that the present invention aims to solve. Summary of the invention

[0004] The purpose of the present invention is to provide a method and system for remote sensing identification of crop rotation information to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A method for remote sensing identification of crop rotation information, the method comprising:

[0007] Acquiring spectral information of a region based on remote sensing technology, and determining a crop area based on the spectral information; the spectral information is spectral reflectance characteristics under different remote sensing bands, and the remote sensing bands include visible light bands, near infrared bands, and short-wave infrared bands;

[0008] Select characteristic crops in the crop area based on spectral information, and simultaneously determine the growth sequence of the characteristic crops; the growth sequence is a time series, the elements in the growth sequence represent growth characteristics, and the element number corresponds to time;

[0009] Connect the growth sequences of characteristic crops at the same location to obtain the rotation characteristic sequences at different locations;

[0010] Compare the rotation feature sequences of different locations and cluster the locations;

[0011] Determine the analysis value of each location based on the distribution of similar locations, select a location whose analysis value reaches a preset threshold, and read the historical environmental data at the location from a preset environmental data record library; the historical environmental data is obtained in real time by sensors installed in the crop area;

[0012] The longest rotation characteristic sequence is selected from all rotation characteristic sequences corresponding to the location, and the historical environmental data is used as the independent variable and the longest rotation characteristic sequence as the dependent variable, which are fed back to the staff.

[0013] As a further solution of the present invention: the step of selecting characteristic crops in the crop area based on spectral information and simultaneously determining the growth sequence of the characteristic crops comprises:

[0014] At a preset first time interval, a short-wave infrared image of the crop area is acquired based on the short-wave infrared band, and the crop area is segmented according to differential information of each position in the short-wave infrared image to obtain sub-areas;

[0015] At every second preset time period, a visible light image of the crop area is acquired based on the visible light band, and each sub-area is identified according to differential information of each position in the visible light image to locate the crop;

[0016] At every third preset time period, a near infrared wave image of the crop area is acquired based on the near infrared band, and a vegetation index of each crop is extracted;

[0017] Compare the vegetation index of each crop, determine the characteristic crop, and read the vegetation index of the characteristic crop as a growth sequence;

[0018] Among them, the first duration is longer than the second duration, and the second duration is longer than the third duration;

[0019] The calculation process of the vegetation index of characteristic crops is:

[0020] ; In the formula, is the vegetation index, is the gain factor, the value is 2.5, is the reflectivity in the near-infrared band, is the red band reflectivity, is the reflectance of the blue band; are preset constants with values ​​of 6 and 7.5 respectively; is the background light constant, the value is 10000;

[0021] The process of determining characteristic crops is as follows:

[0022] For any crop, locate other crops within a preset range;

[0023] Calculate the current crop Value vs. other crops The difference ratio of the values;

[0024] Calculate the mean of the difference ratio. When the mean meets the preset mean condition, mark the current crop as a characteristic crop. The comparison process between the mean and the mean condition is:

[0025] ; In the formula, Indicates the most recent time, is the preset backtracking time, Indicates the total number of other crops within the preset range. express At time , the difference ratio between the current crop and other crops; represents the preset threshold; where, All are multiples of the third duration.

[0026] As a further solution of the present invention: the step of connecting the growth sequences of characteristic crops at the same location to obtain the rotation characteristic sequences at different locations comprises:

[0027] Get the coordinates of each characteristic crop;

[0028] When the distance between the coordinates of the two characteristic crops is less than a preset distance threshold, the growth sequences of the two characteristic crops are read;

[0029] The two growth sequences are connected based on the time order to obtain the rotation characteristic sequence of the coordinate.

[0030] As a further solution of the present invention: the step of comparing the crop rotation feature sequences at different positions and clustering the positions comprises:

[0031] For the rotation feature sequences of two different locations, obtain their maximum common sequence;

[0032] Align the largest common sequence and calculate the sequence similarity;

[0033] The rotation difference distance between the two locations was determined based on the length of the largest common sequence and sequence similarity;

[0034] Clustering all locations based on the rotation difference distance;

[0035] The process of determining the maximum common sequence adopts a dynamic programming algorithm. When the difference ratio of two vegetation indices is less than a preset ratio threshold, the two vegetation indices are regarded as the same vegetation indices in the process of determining the maximum common sequence.

[0036] The sequence similarity adopts cosine similarity;

[0037] The process of determining the rotation difference distance is as follows:

[0038] ; In the formula, is the rotation difference distance, is the length of the maximum common sequence, Indicates the maximum length of two rotation feature sequences; is the similarity of the two largest common sequences; These are preset correction factors.

[0039] As a further solution of the present invention: the step of determining the analysis value of each location according to the distribution of similar locations, selecting a location whose analysis value reaches a preset threshold, and reading the historical environmental data at the location from a preset environmental data record library comprises:

[0040] Read the center of each type of location;

[0041] Calculating the distance between the center and other locations, and determining the analytical value of the other locations based on the distance; the analytical value is proportional to the distance;

[0042] Comparing the analysis value with a preset threshold, and when the analysis value reaches the preset threshold, marking the position as a reference position;

[0043] The historical environmental data at the reference location is read from a preset environmental data record library.

[0044] As a further solution of the present invention: the step of selecting the longest crop rotation feature sequence from all crop rotation feature sequences corresponding to the location, taking the historical environmental data as the independent variable and the longest crop rotation feature sequence as the dependent variable, and feeding back to the staff includes:

[0045] Query similar locations of the reference location;

[0046] Obtain the rotation feature sequence of the same position and select the longest rotation feature sequence;

[0047] Historical environmental data were used as independent variables and the longest rotation characteristic sequence as the dependent variable, which were fed back to the staff.

[0048] The technical solution of the present invention also provides a crop rotation information remote sensing identification system, the system comprising:

[0049] A crop area determination module is used to obtain spectral information of a region based on remote sensing technology, and determine the crop area based on the spectral information; the spectral information is spectral reflectance characteristics under different remote sensing bands, and the remote sensing bands include visible light bands, near infrared bands, and short-wave infrared bands;

[0050] A growth sequence determination module is used to select characteristic crops in the crop area based on spectral information and simultaneously determine the growth sequence of the characteristic crops; the growth sequence is a time series, the elements in the growth sequence represent growth characteristics, and the element sequence number corresponds to time;

[0051] The growth sequence connection module is used to connect the growth sequences of characteristic crops at the same location to obtain the rotation characteristic sequences at different locations;

[0052] The location clustering module is used to compare the rotation feature sequences of different locations and cluster the locations;

[0053] A location selection module is used to determine the analysis value of each location based on the distribution of similar locations, select a location whose analysis value reaches a preset threshold, and read the historical environmental data at the location from a preset environmental data record library; the historical environmental data is obtained in real time by sensors installed in the crop area;

[0054] The data extraction and feedback module is used to select the longest crop rotation feature sequence from all crop rotation feature sequences corresponding to the location, use the historical environmental data as the independent variable and the longest crop rotation feature sequence as the dependent variable, and feed it back to the staff.

[0055] As a further solution of the present invention: the growth sequence determination module includes:

[0056] A first recognition unit is used to obtain a short-wave infrared image of the crop area based on the short-wave infrared band at a preset first time interval, and divide the crop area according to differential information of each position in the short-wave infrared image to obtain sub-areas;

[0057] A second identification unit is used to obtain a visible light image of the crop area based on the visible light band at every second preset time period, identify each sub-area according to differential information of each position in the visible light image, and locate the crop;

[0058] A third identification unit is used to obtain a near-infrared wave image of the crop area based on the near-infrared band at a preset third time interval, and extract a vegetation index of each crop;

[0059] A data processing unit is used to compare the vegetation index of each crop, determine the characteristic crop, and read the vegetation index of the characteristic crop as a growth sequence;

[0060] Among them, the first duration is longer than the second duration, and the second duration is longer than the third duration;

[0061] The calculation process of the vegetation index of characteristic crops is:

[0062] ; In the formula, is the vegetation index, is the gain factor, the value is 2.5, is the reflectivity in the near-infrared band, is the red band reflectivity, is the reflectance of the blue band; are preset constants with values ​​of 6 and 7.5 respectively; is the background light constant, the value is 10000;

[0063] The process of determining characteristic crops is as follows:

[0064] For any crop, locate other crops within a preset range;

[0065] Calculate the current crop Value vs. other crops The difference ratio of the values;

[0066] Calculate the mean of the difference ratio. When the mean meets the preset mean condition, mark the current crop as a characteristic crop. The comparison process between the mean and the mean condition is:

[0067] ; In the formula, Indicates the most recent time, is the preset backtracking time, Indicates the total number of other crops within the preset range. express At time , the difference ratio between the current crop and other crops; represents the preset threshold; where, All are multiples of the third duration.

[0068] As a further solution of the present invention: the growing sequence connection module includes:

[0069] A coordinate acquisition unit, used for acquiring the coordinates of each characteristic crop;

[0070] A growth sequence reading unit, used for reading the growth sequences of the two characteristic crops when the distance between the coordinates of the two characteristic crops is less than a preset distance threshold;

[0071] The connection execution unit is used to connect two growth sequences based on the time sequence to obtain the rotation characteristic sequence of the coordinate.

[0072] As a further solution of the present invention: the location clustering module includes:

[0073] A common sequence extraction unit is used to obtain the maximum common sequence of the rotation feature sequences of two different positions;

[0074] A first calculation unit is used to compare the maximum common sequence and calculate the sequence similarity;

[0075] A second calculation unit is used to determine the rotation difference distance of two positions according to the length of the largest common sequence and the sequence similarity;

[0076] A clustering execution unit, used for clustering all positions based on the crop rotation difference distance;

[0077] The process of determining the maximum common sequence adopts a dynamic programming algorithm. When the difference ratio of two vegetation indices is less than a preset ratio threshold, the two vegetation indices are regarded as the same vegetation indices in the process of determining the maximum common sequence.

[0078] The sequence similarity adopts cosine similarity;

[0079] The process of determining the rotation difference distance is as follows:

[0080] ; In the formula, is the rotation difference distance, is the length of the maximum common sequence, Indicates the maximum length of two rotation feature sequences; is the similarity of the two largest common sequences; These are preset correction factors.

[0081] Compared with the prior art, the beneficial effects of the present invention are: the present invention obtains crop growth data based on remote sensing technology, selects characteristic crops based on the growth data, merges the growth data of characteristic crops at the same location, extends single crop growth data into growth data containing crop rotation information, further screens the growth data containing crop rotation information to obtain the final high-value location, obtains the environmental data of the location, uses it as a reference for the crop rotation planning process, and feeds it back to the staff; the screening of the present invention is extremely accurate, which greatly reduces the data analysis pressure of the staff. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention.

[0083] Figure 1 This is a flowchart of the remote sensing identification method for crop rotation information.

[0084] Figure 2 This is a flowchart of the first sub-process of the remote sensing identification method for crop rotation information.

[0085] Figure 3 This is a flowchart of the second sub-process of the remote sensing identification method for crop rotation information.

[0086] Figure 4 This is the third sub-process flowchart of the remote sensing identification method for crop rotation information.

[0087] Figure 5 This is a fourth sub-process flowchart of the remote sensing identification method for crop rotation information.

[0088] Figure 6This is the fifth sub-process flowchart of the remote sensing identification method for crop rotation information.

[0089] Figure 7 This is a structural block diagram of the crop rotation information remote sensing identification system. DETAILED DESCRIPTION

[0090] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention 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 invention and are not intended to limit the present invention.

[0091] Figure 1 The flowchart of the method for remote sensing identification of crop rotation information is as follows. In an embodiment of the present invention, a method for remote sensing identification of crop rotation information includes:

[0092] Step S100: acquiring spectral information of a region based on remote sensing technology, and determining a crop area based on the spectral information; the spectral information is spectral reflectance characteristics under different remote sensing bands, and the remote sensing bands include a visible light band, a near infrared band, and a short-wave infrared band;

[0093] The process of collecting and identifying crop rotation information based on spectral-spatial and temporal information involves the use of remote sensing technology, data processing and analysis models to collect and analyze data in the agricultural production process from multiple dimensions (spectral, spatial and temporal). This process can help agricultural producers effectively identify important information such as crop types, health status, rotation patterns, etc. at different time points and spatial locations; in the technical solution of the present invention, the spectral information of the region is obtained based on remote sensing technology, and the spectral information is identified to determine the area where crops exist, which is called the crop area; the recognition accuracy requirements of this process are very low. In general, a simple recognition of remote sensing images in the visible light band can obtain the area where crops exist.

[0094] Step S200: Select characteristic crops in the crop area based on spectral information, and simultaneously determine the growth sequence of the characteristic crops; the growth sequence is a time sequence, the elements in the growth sequence represent growth characteristics, and the element sequence number corresponds to time;

[0095] Further processing of the spectral information can locate crops in the crop area, identify crops, and select some crops worthy of analysis, called characteristic crops. By identifying the spectral information and extracting the growth information of characteristic crops, such as vegetation index, the vegetation index is integrated to obtain the growth sequence of characteristic crops. Each element in the growth sequence represents the growth situation at a moment.

[0096] Step S300: connecting the growth sequences of characteristic crops at the same location to obtain the rotation characteristic sequences at different locations;

[0097] In the case of crop rotation, the plants planted in different positions are different in different time periods. By counting all characteristic crops, some of these characteristic crops will have the same position (close enough). The growth sequences of the characteristic crops in the same position are connected. At this time, the plant growth conditions at different times in each position are obtained, which is called the crop rotation characteristic sequence. In this process, the growth conditions of the characteristic crops are converted into the growth conditions at each position. Since the evaluation parameters of the growth conditions of different plants may be different, when constructing the growth sequence, the evaluation parameters need to be dimensionless to make the connection process smoother.

[0098] Step S400: comparing the crop rotation feature sequences at different locations and clustering the locations;

[0099] By comparing the rotation feature sequences of different locations, we can determine the similarities between different locations and classify similar locations as one category.

[0100] Step S500: determining the analysis value of each position according to the distribution of similar positions, selecting a position whose analysis value reaches a preset threshold, and reading the historical environmental data of the position from a preset environmental data record library; the historical environmental data is obtained in real time by sensors installed in the crop area;

[0101] After clustering the positions, the distribution of similar positions is obtained. Similar positions indicate that the growth conditions of their crops are similar. In the crop area, the growth conditions of crops are similar, and their positions are likely to be similar. Based on this, the analysis value of each position is determined according to the distribution of similar positions. Generally, the more special the position, the higher the analysis value. The analysis value is compared with the preset threshold. When the analysis value reaches the preset threshold, it is necessary to perform a specific analysis on the position. In the technical solution of the present invention, the selection process of characteristic crops is the first step of the screening process. On this basis, the positions of characteristic crops are further screened to obtain the final positions. The advantage of doing so is that the most worthy positions for analysis can be selected from a large number of crop positions.

[0102] After the final location is selected, the historical environmental data at the location is read from the preset environmental data record library; wherein the historical environmental data is obtained in real time by sensors installed in the crop area, and the sensors include temperature sensors, humidity sensors, light intensity sensors, etc., which are used to obtain temperature, humidity and light intensity. The temperature includes air temperature and soil temperature, and the humidity includes air humidity and soil humidity. It is worth mentioning that environmental information can also be obtained through spectral information. The spectral data corresponding to the short-wave infrared (SWIR) band can be used to provide accurate soil and crop moisture status information.

[0103] Step S600: Select the longest crop rotation feature sequence from all crop rotation feature sequences corresponding to the location, use the historical environmental data as an independent variable and the longest crop rotation feature sequence as a dependent variable, and feed it back to the staff;

[0104] For the selected position, it is determined which type of position it belongs to. Compared with positions of the same type, the rotation feature sequence is very similar (because position clustering is implemented based on the rotation feature sequence), and the lengths of different rotation feature sequences may be different. In the technical solution of the present invention, the longest rotation feature sequence is selected from all rotation feature sequences of positions of the same type as the predicted rotation feature sequence of the current position, and then its historical environmental data is read, and the historical environmental data is used as the independent variable, and the longest rotation feature sequence is used as the dependent variable to feed back to the staff; the actual significance of the feedback data is what the crop growth conditions at the position are under the current historical environmental data, wherein the crop growth conditions at the position are highly accurate prediction data, and for a position, assuming that it has only planted one season of crops, since the selection process of the rotation feature sequence is to select the longest rotation feature sequence of positions of the same type, it is actually expanded into a long sequence with extremely high accuracy, that is, if it rotates, the crop growth conditions can be obtained.

[0105] Figure 2 This is a first sub-flow chart of the remote sensing identification method for crop rotation information. The steps of selecting characteristic crops in the crop area based on spectral information and synchronously determining the growth sequence of the characteristic crops include:

[0106] Step S201: acquiring a shortwave infrared image of the crop area based on the shortwave infrared band at a preset first time interval, and dividing the crop area according to differential information of each position in the shortwave infrared image to obtain sub-areas;

[0107] Step S202: acquiring a visible light image of the crop area based on the visible light band at every second preset time period, identifying each sub-area according to differential information of each position in the visible light image, and locating the crop;

[0108] Step S203: acquiring near infrared wave images of the crop area based on the near infrared band at every third preset time period, and extracting the vegetation index of each crop;

[0109] Step S204: compare the vegetation index of each crop, determine the characteristic crop, and read the vegetation index of the characteristic crop as a growth sequence.

[0110] In an example of the technical solution of the present invention, a short-wave infrared image of a crop area is acquired based on a short-wave infrared band at a preset first time interval, and the crop area is divided according to differential information of each position in the short-wave infrared image to obtain sub-areas; the short-wave infrared image can represent accurate soil and crop moisture status information, and by analyzing it, areas with different soil conditions can be separated to obtain sub-areas with different soil conditions.

[0111] Every second preset time period, a visible light image of the crop area is acquired based on the visible light band, and each sub-area is identified according to the differential information of each position in the visible light image to locate the crop; the visible light image is very sensitive to monitoring the chlorophyll content of plants and is used to locate crops.

[0112] Every third preset time period, a near-infrared wave image of the crop area is acquired based on the near-infrared band, and the vegetation index of each crop is extracted. The near-infrared wave image reflects the moisture content and growth condition of the plant and is often used to calculate vegetation indices (such as NDVI) to determine the growth condition of each crop.

[0113] It should be noted that the application frequencies of the three images are different. Step S201 may be performed once every month, step S202 may be performed once every week, and step S203 may be performed once every hour.

[0114] Finally, by comparing the vegetation index (growth conditions) of each crop, we can select crops that are more worthy of analysis, called characteristic crops, and read the vegetation index of characteristic crops at different times as the growth sequence.

[0115] Among them, the first duration is longer than the second duration, and the second duration is longer than the third duration;

[0116] The calculation process of the vegetation index of characteristic crops is:

[0117] ; In the formula, is the vegetation index, is the gain factor, the value is 2.5, is the reflectivity in the near-infrared band, is the red band reflectivity, is the reflectance of the blue band; are preset constants with values ​​of 6 and 7.5 respectively; is the background light constant, the value is 10000;

[0118] The calculation process of vegetation index is not complicated and can be extracted from spectral information. EVI is an improved vegetation index that aims to reduce the interference of atmospheric influence, soil background effect and other external factors, and provide more reliable vegetation monitoring results than NDVI. EVI not only considers the reflectivity of near infrared (NIR) and red band (Red), but also introduces the reflectivity of blue band (Blue), as well as multiple constants for correction, which is especially suitable for high-density vegetation areas. It should be noted that the calculation of this application is , it is necessary to locate the crops first and analyze them individually, rather than analyzing the entire area as a whole.

[0119] The process of determining characteristic crops is as follows:

[0120] For any crop, locate other crops within a preset range;

[0121] Calculate the current crop Value vs. other crops The difference ratio of the values;

[0122] Calculate the mean of the difference ratio. When the mean meets the preset mean condition, mark the current crop as a characteristic crop. The comparison process between the mean and the mean condition is:

[0123] ; In the formula, Indicates the most recent time, is the preset backtracking time, Indicates the total number of other crops within the preset range. express At time , the difference ratio between the current crop and other crops; represents the preset threshold; where, All are multiples of the third duration.

[0124] For any crop, locate other crops within a preset range (the preset range is usually a circular range) and calculate the current crop's Value vs. other crops The difference ratio of the values ​​is calculated as follows: the difference is divided by the current crop value.

[0125] After obtaining the difference ratio, the average of the difference ratios between the current crop and all other crops is calculated. Then, each identification can get a mean value. A backtracking time is preset by the staff, and the mean of the means within the backtracking time is calculated to get the final mean value. It is compared with the preset threshold. When the mean reaches the preset threshold, it means that the crop at this location is more common and can be used as a feature crop.

[0126] Figure 3 This is a second sub-flow chart of the method for remote sensing identification of crop rotation information. The step of connecting the growth sequences of characteristic crops at the same location to obtain the rotation characteristic sequences at different locations includes:

[0127] Step S301: obtaining the coordinates of each characteristic crop;

[0128] Step S302: when the distance between the coordinates of the two characteristic crops is less than a preset distance threshold, reading the growth sequences of the two characteristic crops;

[0129] Step S303: Connect the two growth sequences based on the time sequence to obtain the rotation feature sequence of the coordinate.

[0130] After obtaining the characteristic crops, the coordinates of each characteristic crop are obtained, and the distance between any two characteristic crops is calculated. When the distance is small enough, the growth sequences of the two characteristic crops are read; the elements in the growth sequence correspond to time, and the two growth sequences are connected based on the time order, and the obtained sequence is the crop rotation characteristic sequence at the coordinate; it should be noted that the merged growth sequence is used as the new growth sequence at the coordinate, and it is merged with other growth sequences with the same coordinate to obtain a longer crop rotation characteristic sequence.

[0131] Figure 4 This is a third sub-flow chart of the crop rotation information remote sensing identification method, wherein the step of comparing the crop rotation feature sequences at different locations and clustering the locations comprises:

[0132] Step S401: for the crop rotation feature sequences at two different locations, obtain their maximum common sequence;

[0133] Step S402: aligning the largest common sequence and calculating sequence similarity;

[0134] Step S403: determining the rotation difference distance between two locations according to the length of the largest common sequence and the sequence similarity;

[0135] Step S404: clustering all positions based on the crop rotation difference distance.

[0136] The above content explains the comparison process of crop rotation characteristic sequences. For crop rotation characteristic sequences in two different positions, different plants may be planted in the two positions. Therefore, the crop rotation characteristic sequences may be different, and even the lengths may be different. For crop rotation characteristic sequences of different lengths, their maximum common sequence is obtained, the maximum common sequence is compared, and the sequence similarity is calculated; then, the difference between the crop rotation characteristic sequences in the two positions is determined based on the length of the maximum common sequence and the sequence similarity, which is called the difference distance; finally, all positions are clustered based on the crop rotation difference distance; the clustering scheme can adopt the K-means clustering scheme; the K value can be a preset value or a preset incremental value, and the optimal K value is determined according to the clustering result.

[0137] The process of determining the maximum common sequence adopts a dynamic programming algorithm. When the difference ratio of two vegetation indices is less than a preset ratio threshold, the two vegetation indices are regarded as the same vegetation indices in the process of determining the maximum common sequence.

[0138] The sequence similarity adopts cosine similarity;

[0139] The process of determining the rotation difference distance is as follows:

[0140] ; In the formula, is the rotation difference distance, is the length of the maximum common sequence, Indicates the maximum length of two rotation feature sequences; is the similarity of the two largest common sequences; These are preset correction factors.

[0141] The rotation difference distance is inversely proportional to the similarity. The length of the maximum common sequence affects the weight of the similarity. The longer the length of the maximum common sequence, the greater the calculated similarity weight and the greater the distance amplification.

[0142] Figure 5 The fourth sub-process flowchart of the remote sensing identification method for crop rotation information is as follows: the steps of determining the analysis value of each position according to the distribution of similar positions, selecting a position whose analysis value reaches a preset threshold, and reading the historical environmental data at the position in a preset environmental data record library include:

[0143] Step S501: Read the center of each type of position;

[0144] Step S502: Calculate the distance between the center and other positions, and determine the analysis value of other positions according to the distance; the analysis value is proportional to the distance;

[0145] Step S503: comparing the analysis value with a preset threshold, and when the analysis value reaches the preset threshold, marking the position as a reference position;

[0146] Step S504: reading historical environmental data at a reference location from a preset environmental data record library.

[0147] After clustering the locations, locations of different categories are obtained. For each category of locations, a center is selected once (if the K-means clustering scheme is used, the center is known), and the distance between the center and other locations of the same category is calculated, and the analysis value of other locations is determined based on the distance; the analysis value is proportional to the distance, and a linear function with a positive coefficient can be used; the analysis value is compared with a preset threshold, and when the analysis value reaches the preset threshold, the location is marked as a reference location, and the historical environmental data at the reference location is read from the preset environmental data record library.

[0148] Figure 6 This is a fifth sub-flow chart of the crop rotation information remote sensing identification method, wherein the steps of selecting the longest crop rotation feature sequence from all crop rotation feature sequences corresponding to the position, taking the historical environmental data as the independent variable and the longest crop rotation feature sequence as the dependent variable, and feeding back to the staff include:

[0149] Step S601: querying similar locations of the reference location;

[0150] Step S602: obtaining crop rotation feature sequences of the same type of position, and selecting the longest crop rotation feature sequence;

[0151] Step S603: Use the historical environmental data as the independent variable and the longest crop rotation feature sequence as the dependent variable, and feed them back to the staff.

[0152] In one example of the technical solution of the present invention, since the crop rotation characteristic sequence at the reference position may be very short and it is difficult to observe the changes in different crops, the present application queries similar positions of the reference position, obtains the crop rotation characteristic sequences of similar positions, selects the longest crop rotation characteristic sequence, uses the longest crop rotation characteristic sequence as the prediction result, synchronously reads historical environmental data, uses the historical environmental data as the independent variable, and the longest crop rotation characteristic sequence as the dependent variable, and feeds back to the staff.

[0153] Figure 7 1 is a structural block diagram of a crop rotation information remote sensing identification system. In an embodiment of the present invention, a crop rotation information remote sensing identification system is provided. The system 10 includes:

[0154] The crop area determination module 11 is used to obtain spectral information of the area based on remote sensing technology, and determine the crop area based on the spectral information; the spectral information is the spectral reflectance characteristics under different remote sensing bands, and the remote sensing bands include visible light bands, near infrared bands and short-wave infrared bands;

[0155] A growth sequence determination module 12 is used to select characteristic crops in the crop area based on the spectral information and simultaneously determine the growth sequence of the characteristic crops; the growth sequence is a time sequence, the elements in the growth sequence represent the growth characteristics, and the element sequence number corresponds to the time;

[0156] A growth sequence connection module 13 is used to connect the growth sequences of characteristic crops at the same location to obtain rotation characteristic sequences at different locations;

[0157] The location clustering module 14 is used to compare the rotation feature sequences of different locations and cluster the locations;

[0158] The location selection module 15 is used to determine the analysis value of each location according to the distribution of similar locations, select the location whose analysis value reaches a preset threshold, and read the historical environmental data at the location from a preset environmental data record library; the historical environmental data is obtained in real time by sensors installed in the crop area;

[0159] The data extraction and feedback module 16 is used to select the longest crop rotation feature sequence from all crop rotation feature sequences corresponding to the location, use the historical environmental data as an independent variable and the longest crop rotation feature sequence as a dependent variable, and feed it back to the staff.

[0160] Furthermore, the growth sequence determination module 12 includes:

[0161] A first recognition unit is used to obtain a short-wave infrared image of the crop area based on the short-wave infrared band at a preset first time interval, and divide the crop area according to differential information of each position in the short-wave infrared image to obtain sub-areas;

[0162] A second identification unit is used to obtain a visible light image of the crop area based on the visible light band at every second preset time period, identify each sub-area according to differential information of each position in the visible light image, and locate the crop;

[0163] A third identification unit is used to obtain a near-infrared wave image of the crop area based on the near-infrared band at a preset third time interval, and extract a vegetation index of each crop;

[0164] A data processing unit is used to compare the vegetation index of each crop, determine the characteristic crop, and read the vegetation index of the characteristic crop as a growth sequence;

[0165] Among them, the first duration is longer than the second duration, and the second duration is longer than the third duration;

[0166] The calculation process of the vegetation index of characteristic crops is:

[0167] ; In the formula, is the vegetation index, is the gain factor, the value is 2.5, is the reflectivity in the near-infrared band, is the red band reflectivity, is the reflectance of the blue band; are preset constants with values ​​of 6 and 7.5 respectively; is the background light constant, the value is 10000;

[0168] The process of determining characteristic crops is as follows:

[0169] For any crop, locate other crops within a preset range;

[0170] Calculate the current crop Value vs. other crops The difference ratio of the values;

[0171] Calculate the mean of the difference ratio. When the mean meets the preset mean condition, mark the current crop as a characteristic crop. The comparison process between the mean and the mean condition is:

[0172] ; In the formula, Indicates the most recent time, is the preset backtracking time, Indicates the total number of other crops within the preset range. express At time , the difference ratio between the current crop and other crops; represents the preset threshold; where, All are multiples of the third duration.

[0173] Specifically, the growth sequence connection module 13 includes:

[0174] A coordinate acquisition unit, used for acquiring the coordinates of each characteristic crop;

[0175] A growth sequence reading unit, used for reading the growth sequences of the two characteristic crops when the distance between the coordinates of the two characteristic crops is less than a preset distance threshold;

[0176] The connection execution unit is used to connect two growth sequences based on the time sequence to obtain the rotation characteristic sequence of the coordinate.

[0177] Furthermore, the location clustering module 14 includes:

[0178] A common sequence extraction unit is used to obtain the maximum common sequence of the rotation feature sequences of two different positions;

[0179] A first calculation unit is used to compare the maximum common sequence and calculate the sequence similarity;

[0180] A second calculation unit is used to determine the rotation difference distance of two positions according to the length of the largest common sequence and the sequence similarity;

[0181] A clustering execution unit, used for clustering all positions based on the crop rotation difference distance;

[0182] The process of determining the maximum common sequence adopts a dynamic programming algorithm. When the difference ratio of two vegetation indices is less than a preset ratio threshold, the two vegetation indices are regarded as the same vegetation indices in the process of determining the maximum common sequence.

[0183] The sequence similarity adopts cosine similarity;

[0184] The process of determining the rotation difference distance is as follows:

[0185] ; In the formula, is the rotation difference distance, is the length of the maximum common sequence, Indicates the maximum length of two rotation feature sequences; is the similarity of the two largest common sequences; These are preset correction factors.

[0186] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for remote sensing identification of crop rotation information, characterized in that: The method comprises: Acquiring spectral information of a region based on remote sensing technology, and determining a crop area based on the spectral information; the spectral information is spectral reflectance characteristics under different remote sensing bands, and the remote sensing bands include visible light bands, near infrared bands, and short-wave infrared bands; Select characteristic crops in the crop area based on spectral information, and simultaneously determine the growth sequence of the characteristic crops; the growth sequence is a time series, the elements in the growth sequence represent growth characteristics, and the element number corresponds to time; Connect the growth sequences of characteristic crops at the same location to obtain the rotation characteristic sequences at different locations; Compare the rotation feature sequences of different locations and cluster the locations; Determine the analysis value of each location based on the distribution of similar locations, select a location whose analysis value reaches a preset threshold, and read the historical environmental data at the location from a preset environmental data record library; the historical environmental data is obtained in real time by sensors installed in the crop area; The longest rotation characteristic sequence is selected from all rotation characteristic sequences corresponding to the location, and the historical environmental data is used as the independent variable and the longest rotation characteristic sequence is used as the dependent variable, which is fed back to the staff; The step of connecting the growth sequences of characteristic crops at the same location to obtain the rotation characteristic sequences at different locations comprises: Get the coordinates of each characteristic crop; When the distance between the coordinates of the two characteristic crops is less than a preset distance threshold, the growth sequences of the two characteristic crops are read; Connect two growth sequences based on the time sequence to obtain the rotation characteristic sequence of the coordinate; The step of determining the analysis value of each location according to the distribution of similar locations, selecting a location whose analysis value reaches a preset threshold, and reading the historical environmental data at the location from a preset environmental data record library comprises: Read the center of each type of location; Calculating the distance between the center and other locations, and determining the analytical value of the other locations based on the distance; the analytical value is proportional to the distance; Comparing the analysis value with a preset threshold, and when the analysis value reaches the preset threshold, marking the position as a reference position; The historical environmental data at the reference location is read from a preset environmental data record library.

2. The method for remote sensing identification of crop rotation information according to claim 1, characterized in that: The step of selecting characteristic crops in the crop area based on spectral information and simultaneously determining the growth sequence of the characteristic crops comprises: At a preset first time interval, a short-wave infrared image of the crop area is acquired based on the short-wave infrared band, and the crop area is segmented according to differential information of each position in the short-wave infrared image to obtain sub-areas; At every second preset time period, a visible light image of the crop area is acquired based on the visible light band, and each sub-area is identified according to differential information of each position in the visible light image to locate the crop; At every third preset time period, a near infrared wave image of the crop area is acquired based on the near infrared band, and a vegetation index of each crop is extracted; Compare the vegetation index of each crop, determine the characteristic crop, and read the vegetation index of the characteristic crop as a growth sequence; Among them, the first duration is longer than the second duration, and the second duration is longer than the third duration; The calculation process of the vegetation index of characteristic crops is: ; In the formula, is the vegetation index, is the gain factor, the value is 2.5, is the reflectivity in the near-infrared band, is the red band reflectivity, is the reflectance of the blue band; are preset constants with values ​​of 6 and 7.5 respectively; is the background light constant, the value is 10000; The process of determining characteristic crops is as follows: For any crop, locate other crops within a preset range; Calculate the current crop Value vs. other crops The difference ratio of the values; Calculate the mean of the difference ratio. When the mean meets the preset mean condition, mark the current crop as a characteristic crop. The comparison process between the mean and the mean condition is: ; In the formula, Indicates the most recent time, is the preset backtracking time, Indicates the total number of other crops within the preset range. express At time , the difference ratio between the current crop and other crops; represents the preset threshold; where, All are multiples of the third duration.

3. The method for remote sensing identification of crop rotation information according to claim 1, characterized in that: The step of comparing the crop rotation feature sequences at different positions and clustering the positions comprises: For the rotation feature sequences of two different locations, obtain their maximum common sequence; Align the largest common sequence and calculate the sequence similarity; The rotation difference distance between the two locations was determined based on the length of the largest common sequence and the sequence similarity; Clustering all locations based on the crop rotation difference distance; The process of determining the maximum common sequence adopts a dynamic programming algorithm. When the difference ratio of two vegetation indices is less than a preset ratio threshold, the two vegetation indices are regarded as the same vegetation indices in the process of determining the maximum common sequence. The sequence similarity adopts cosine similarity; The process of determining the rotation difference distance is as follows: ; In the formula, is the rotation difference distance, is the length of the maximum common sequence, Indicates the maximum length of two rotation feature sequences; is the similarity of the two largest common sequences; These are preset correction factors.

4. The method for remote sensing identification of crop rotation information according to claim 1, characterized in that: The step of selecting the longest crop rotation feature sequence from all crop rotation feature sequences corresponding to the location, taking the historical environmental data as the independent variable and the longest crop rotation feature sequence as the dependent variable, and feeding back to the staff includes: Query similar locations of the reference location; Obtain the rotation feature sequence of the same position and select the longest rotation feature sequence; Historical environmental data were used as independent variables and the longest rotation characteristic sequence as the dependent variable, which were fed back to the staff.

5. A crop rotation information remote sensing identification system, characterized in that: The system comprises: A crop area determination module is used to obtain spectral information of a region based on remote sensing technology, and determine the crop area based on the spectral information; the spectral information is spectral reflectance characteristics under different remote sensing bands, and the remote sensing bands include visible light bands, near infrared bands, and short-wave infrared bands; A growth sequence determination module is used to select characteristic crops in the crop area based on spectral information and simultaneously determine the growth sequence of the characteristic crops; the growth sequence is a time series, the elements in the growth sequence represent growth characteristics, and the element sequence number corresponds to time; The growth sequence connection module is used to connect the growth sequences of characteristic crops at the same location to obtain the rotation characteristic sequences at different locations; The location clustering module is used to compare the rotation feature sequences of different locations and cluster the locations; A location selection module is used to determine the analysis value of each location based on the distribution of similar locations, select a location whose analysis value reaches a preset threshold, and read the historical environmental data at the location from a preset environmental data record library; the historical environmental data is obtained in real time by sensors installed in the crop area; A data extraction and feedback module is used to select the longest rotation feature sequence from all rotation feature sequences corresponding to the location, use the historical environmental data as the independent variable and the longest rotation feature sequence as the dependent variable, and feed it back to the staff; The growing sequence connection module comprises: A coordinate acquisition unit, used for acquiring the coordinates of each characteristic crop; A growth sequence reading unit, used for reading the growth sequences of the two characteristic crops when the distance between the coordinates of the two characteristic crops is less than a preset distance threshold; A connection execution unit is used to connect two growth sequences based on the time sequence to obtain the rotation characteristic sequence of the coordinate; The step of determining the analysis value of each location according to the distribution of similar locations, selecting a location whose analysis value reaches a preset threshold, and reading the historical environmental data at the location from a preset environmental data record library comprises: Read the center of each type of location; Calculating the distance between the center and other locations, and determining the analytical value of the other locations based on the distance; the analytical value is proportional to the distance; Comparing the analysis value with a preset threshold, and when the analysis value reaches the preset threshold, marking the position as a reference position; The historical environmental data at the reference location is read from a preset environmental data record library.

6. The crop rotation information remote sensing identification system according to claim 5, characterized in that: The growth sequence determination module comprises: A first recognition unit is used to obtain a short-wave infrared image of the crop area based on the short-wave infrared band at a preset first time interval, and divide the crop area according to differential information of each position in the short-wave infrared image to obtain sub-areas; A second identification unit is used to obtain a visible light image of the crop area based on the visible light band at every second preset time period, identify each sub-area according to differential information of each position in the visible light image, and locate the crop; A third identification unit is used to obtain a near-infrared wave image of the crop area based on the near-infrared band at a preset third time interval, and extract a vegetation index of each crop; A data processing unit is used to compare the vegetation index of each crop, determine the characteristic crop, and read the vegetation index of the characteristic crop as a growth sequence; Among them, the first duration is longer than the second duration, and the second duration is longer than the third duration; The calculation process of the vegetation index of characteristic crops is: ; In the formula, is the vegetation index, is the gain factor, the value is 2.5, is the reflectivity in the near-infrared band, is the red band reflectivity, is the reflectance of the blue band; are preset constants with values ​​of 6 and 7.5 respectively; is the background light constant, the value is 10000; The process of determining characteristic crops is as follows: For any crop, locate other crops within a preset range; Calculate the current crop Value vs. other crops The difference ratio of the values; Calculate the mean of the difference ratio. When the mean meets the preset mean condition, mark the current crop as a characteristic crop. The comparison process between the mean and the mean condition is: ; In the formula, Indicates the most recent time, is the preset backtracking time, Indicates the total number of other crops within the preset range. express At time , the difference ratio between the current crop and other crops; represents the preset threshold; where, All are multiples of the third duration.

7. The crop rotation information remote sensing identification system according to claim 5, characterized in that: The location clustering module comprises: A common sequence extraction unit is used to obtain the maximum common sequence of the rotation feature sequences of two different positions; A first calculation unit is used to compare the maximum common sequence and calculate the sequence similarity; A second calculation unit is used to determine the rotation difference distance of the two positions according to the length of the largest common sequence and the sequence similarity; A clustering execution unit, used for clustering all positions based on the crop rotation difference distance; The process of determining the maximum common sequence adopts a dynamic programming algorithm. When the difference ratio of two vegetation indices is less than a preset ratio threshold, the two vegetation indices are regarded as the same vegetation indices in the process of determining the maximum common sequence. The sequence similarity adopts cosine similarity; The process of determining the rotation difference distance is as follows: ; In the formula, is the rotation difference distance, is the length of the maximum common sequence, Indicates the maximum length of two rotation feature sequences; is the similarity of the two largest common sequences; These are preset correction factors.

Images