阅读说明：本技术 一种基于农机gps定位数据的农田作业面积测算方法 (Agricultural machine GPS positioning data-based farmland operation area measuring and calculating method ) 是由 邱大伟 谭雯雯 刘子辰 张玉成 石晶林 于 2019-10-14 设计创作，主要内容包括：本发明公开了一种基于农机GPS定位数据的农田作业面积测算方法,能够对农田作业面积进行高效准确面积测算、且具有去重耕效果。包括如下步骤：在农田作业期间,农机GPS定位终端每隔设定时间获取农机的GPS定位信息传输至后台数据存储中心,农田作业完毕,选取设定时间段内的GPS定位点,对相邻GPS定位点间的矩形耕地区域进行随机插值,随机插值后的所有GPS定位点进行谱聚类得到至少两个第一次聚类的分块区域。对每个第一次聚类的分块区域再进行谱聚类,得到至少两个第二次聚类的分块区域。对于每个第二次聚类的分块区域,去除其中无GPS定位点的空白区域,并进行耕地拐点消除,由此划分得到最终的分块区域。所有最终的分块区域对应面积总和即为最终作业面积。(The invention discloses a farmland operation area measuring and calculating method based on GPS positioning data of agricultural machinery, which can carry out efficient and accurate area measurement and calculation on farmland operation area and has the effect of removing heavy plowing. The method comprises the following steps: during farmland operation, the GPS positioning terminal of the agricultural machine acquires GPS positioning information of the agricultural machine every set time and transmits the GPS positioning information to the background data storage center, after farmland operation is finished, the GPS positioning points in a set time period are selected, random interpolation is carried out on rectangular farmland areas between adjacent GPS positioning points, and all the GPS positioning points after random interpolation are subjected to spectral clustering to obtain at least two first-clustered block areas. And performing spectral clustering on the block areas of each first clustering to obtain at least two block areas of second clustering. And removing blank areas without GPS positioning points in each block area of the second clustering, and eliminating cropland inflection points, thereby obtaining the final block area through division. And the sum of the areas corresponding to all the final block areas is the final operation area.)

1. A farmland operation area measuring and calculating method based on agricultural machinery GPS positioning data is characterized by comprising the following steps:

s1, during farmland operation of the agricultural machine, the agricultural machine GPS positioning terminal acquires GPS positioning information of the agricultural machine every set time, wherein the GPS positioning information comprises GPS positioning points and time of the agricultural machine, the acquired GPS positioning information is input to the background data storage center, and S2 is executed after the farmland operation is finished;

s2, selecting a GPS positioning point in a set time period from the background data storage center;

s3, carrying out random interpolation on the rectangular cultivated land area between adjacent GPS positioning points;

s4, performing spectral clustering on all the GPS positioning points after random interpolation, wherein the spectral clustering result is a block area for obtaining at least two first clusters;

s5, performing spectral clustering on each block area of the first clustering to obtain at least two block areas of the second clustering;

s6, removing blank areas without GPS positioning points in each block area of the second clustering, and performing arable inflection point elimination, so as to obtain final block areas through division;

directly taking the block area of the second clustering without blank area and arable inflection point as the final block area;

and S7, obtaining sequence boundary points of all the final block areas, solving the area enclosed by the sequence boundary points of each final block area, and obtaining the sum of the areas corresponding to all the final block areas as the final operation area.

2. The method of claim 1, after S2 and before S3, further comprising removing outliers using a K-means algorithm for the selected GPS fix.

3. The method of claim 1, wherein the rectangular arable area between adjacent GPS fixes is:

and establishing rectangles for the adjacent GPS positioning points A and B, wherein the rectangles A and B are respectively used as two wide midpoints of the rectangles, the width of the cultivation tool is used as the width of the rectangles, the distance between the A and the B is used as the length of the rectangles, and the established rectangles of the A and the B are rectangular cultivation areas between the adjacent GPS positioning points A, B.

4. The method according to any one of claims 1 to 3, wherein after the random interpolation of the rectangular cultivated land area between adjacent GPS positioning points, the method further comprises:

sequencing all the GPS positioning points in the rectangular ploughing area after random interpolation according to the distance between the GPS positioning points and the initial GPS positioning point from small to large;

and the initial GPS positioning point is the GPS positioning point with the earliest time in the rectangular cultivated land area.

5. The method of claim 4, wherein the blank area without GPS fix is specifically:

and for all GPS positioning points in the block area of the second clustering, if the distance between two adjacent points is greater than a set threshold value, the current two adjacent points are blank areas.

6. The method according to claim 5, characterized in that the arable point is in particular:

and calculating the distance between each point and the initial point for all the GPS positioning points in the partitioned area of the second clustering, performing curve fitting on the distance values, and taking the GPS positioning point corresponding to the extreme value point of the fitting curve as the arable land inflection point.

7. The method according to claim 6, wherein S7, in particular

Solving the sequential boundary points of all the final block areas by adopting a ConvexHull function of a scipy packet, wherein the sequential boundary points are respectively { (x)₀,y₀),(x₁,y₁),(x₂,y₂),…,(x_N,y_N),(x₀,y₀) N +1 sequential boundary points from 0 to N), (x)₀,y₀)～(x_N,y_N) Sequentially setting the coordinates of 0-Nth sequential boundary points;

the final working Area is Area;

Area＝|A÷2|×9101160000.085981；

wherein A is the area intermediate quantity;

a₀is an initial value of area, a₀＝y₀×(x₀-x₁)。

Technical Field

The invention relates to the technical field of agricultural machinery data analysis, in particular to a farmland operation area measuring and calculating method based on agricultural machinery GPS positioning data.

Background

With the more and more centralization of rural land, large farmland provides favorable conditions for mechanized cultivation, mechanized and intelligent modern agriculture is continuously developed, and under the development trend of precision agriculture, the agricultural machinery operation area measurement and calculation can be used for counting crop yield on one hand, and on the other hand, the agricultural machinery operation area measurement and calculation is also a basis for public agricultural machinery operation charging, so that an efficient and accurate operation area measurement and calculation method is urgently needed.

The existing agricultural machine working area measuring and calculating method mainly has the problem that the measuring and calculating of the heavy tillage working area cannot be effectively solved, so that the working area measuring and calculating result is large, and when a large number of heavy tillage areas exist, the measuring and calculating result of the agricultural machine working area is inaccurate, so that economic disputes of agricultural machine drivers and employers are caused. At present, the measurement and calculation of the working area of agricultural machinery are carried out by an artificial measurement method, a working path measurement method, a polygonal area boundary modeling method, a cell processing method, a space grid splitting method and the like.

The manual measurement method is used for measuring by manually holding the working area measuring and calculating equipment, so that the efficiency is low, and unnecessary labor cost is increased; the method for measuring the operating path of the agricultural machine is to calculate the operating area by multiplying the width of the cultivator by the length of the path traveled by the agricultural machine, and the method has the problem of replanning, and the calculated operating area is larger than the actual operating area.

The method for establishing the boundary of the polygonal farmland collects n vertexes of the boundary of the polygonal farmland operation region, an upper boundary function model and a lower boundary function model are established to complete boundary modeling, and then the area between the upper boundary function and the lower boundary function is calculated.

The cell processing method is to divide the working area into many rectangles according to the GPS positioning data, then calculate the area of the rectangles, and at the same time calculate the area of the overlapping portion, but because the direction of each rectangle is different, each overlapping portion is irregular, and because of the problem of GPS positioning offset, it is difficult to determine the position of the overlapping portion.

The spatial grid division method mainly utilizes a machine learning method to divide regions of cultivated land, for example, a clustering algorithm is utilized to cluster and divide the regions, but the prior method does not solve the problem of operation replanning, and the clustered regions have a plurality of blank regions without positioning points due to irregular cultivated land paths, so that the clustering method and the improved method thereof can not effectively divide the regions into blocks by direct application.

Therefore, a method capable of efficiently and accurately measuring the area of the farmland operation area and having the effect of removing the heavy ploughing is needed.

Disclosure of Invention

In view of the above, the invention provides a farmland operation area measuring and calculating method based on agricultural machinery GPS positioning data, which can perform efficient and accurate area measurement and calculation on farmland operation area and has a heavy plowing removing effect.

In order to achieve the purpose, the technical scheme of the invention comprises the following steps:

s1, during the farmland operation of the agricultural machine, the agricultural machine GPS positioning terminal acquires the GPS positioning information of the agricultural machine at set time intervals, wherein the GPS positioning information comprises the GPS positioning point and the time of the agricultural machine, transmits the acquired GPS positioning information to the background data storage center, and executes S2 after the farmland operation is finished.

And S2, selecting a GPS positioning point in a set time period from the background data storage center.

And S3, carrying out random interpolation on the rectangular cultivated land area between the adjacent GPS positioning points.

And S4, performing spectral clustering on all the GPS positioning points after the random interpolation, wherein the result of the spectral clustering is a block area for obtaining at least two first clusters.

And S5, performing spectral clustering on each block region of the first clustering to obtain at least two block regions of the second clustering.

And S6, removing blank areas without GPS positioning points in the block areas of each secondary clustering, and performing arable inflection point elimination, so as to obtain final block areas through division.

And directly taking the block area of the second clustering without the blank area or the arable inflection point as the final block area.

And S7, obtaining sequence boundary points of all the final block areas, solving the area enclosed by the sequence boundary points of each final block area, and obtaining the sum of the areas corresponding to all the final block areas as the final operation area.

Further, after S2 and before S3, the method further comprises removing outliers from the selected GPS fix using the K-means algorithm.

Further, the rectangular cultivated land area between adjacent GPS positioning points is as follows: and establishing rectangles for the adjacent GPS positioning points A and B, wherein the rectangles A and B are respectively used as two wide middle points of the rectangles, the width of the cultivator is used as the width of the rectangle, the distance between the A and the B is used as the length of the rectangle, and the established rectangles of the A and the B are rectangular cultivated land areas between the adjacent GPS positioning points A, B.

Further, after the random interpolation is performed on the rectangular arable land area between the adjacent GPS positioning points, the method further includes: sequencing all the GPS positioning points in the rectangular ploughing area after random interpolation according to the distance between the GPS positioning points and the initial GPS positioning point from small to large; the initial GPS positioning point is the GPS positioning point with the earliest time in the rectangular cultivated land area.

Further, the blank area without the GPS positioning point specifically includes: and for all GPS positioning points in the block area of the second clustering, if the distance between two adjacent points is greater than a set threshold value, the current two adjacent points are blank areas.

Further, the turning point of the farmland is as follows: and calculating the distance between each point and the initial point for all the GPS positioning points in the partitioned area of the second clustering, performing curve fitting on the distance values, and taking the GPS positioning point corresponding to the extreme value point of the fitting curve as the turning point of the farmland.

Further, S7 specifically includes: solving the sequential boundary points of all the final block areas by adopting a ConvexHull function of a scipy packet, wherein the sequential boundary points are respectively { (x)₀,y₀),(x₁,y₁),(x₂,y₂),…,(x_N,y_N),(x₀,y₀) N +1 sequential boundary points from 0 to N), (x)₀,y₀)～(x_N,y_N) Sequentially setting the coordinates of 0-Nth sequential boundary points;

the final working Area is Area;

Area＝|A÷2|×9101160000.085981；

wherein A is the area intermediate quantity;

a₀is an initial value of area, a₀＝y₀×(x₀-x₁)。

Has the advantages that:

the farmland operation area measuring and calculating method based on the GPS positioning data of the agricultural machinery provided by the invention adopts a random interpolation mode to carry out uniform random interpolation on each rectangular farmland area, carries out twice spectral clustering and blocking on the GPS positioning points after interpolation, and can carry out blocking on the GPS positioning points by clustering all the currently obtained GPS positioning points, but the re-ploughing area can not be repeatedly calculated, so as to obtain the purpose of removing the duplication of the operation area measuring and calculating result; the method obtains more accurate farmland block areas through two times of spectral clustering and removal of blank areas and farmland inflection points, obtains farmland operation areas with higher precision through area solution of boundary points, and has the advantages of less time consumption and higher efficiency.

Drawings

FIG. 1 is a flow chart of a farmland operation area measurement method based on GPS positioning data of agricultural machinery provided by an embodiment of the invention;

FIG. 2 is a schematic diagram of random interpolation according to an embodiment of the present invention;

FIG. 3 is a schematic diagram of GPS point random interpolation according to an embodiment of the present invention;

FIG. 4 is a schematic diagram of a region with a cropland inflection point in a block region, a schematic diagram of a distance scatter point, a schematic diagram of a fitting curve and a schematic diagram of a new block after the inflection point is removed in the embodiment of the invention;

FIG. 5 is a schematic diagram of a clockwise boundary vertex of a blocking area according to an embodiment of the present invention.

Detailed Description

The invention is described in detail below by way of example with reference to the accompanying drawings.

The invention provides a farmland operation area measuring and calculating method based on agricultural machinery GPS positioning data, which comprises the following steps as shown in figure 1:

s1, during farmland operation of the agricultural machinery, the agricultural machinery GPS positioning terminal acquires GPS positioning information of the agricultural machinery at set time intervals, wherein the GPS positioning information comprises GPS positioning points and time of the agricultural machinery, and transmits the acquired GPS positioning information to the background data storage center. Specifically, GPS positioning information can be wirelessly transmitted to the background data storage center

After the field work is completed, S2 is executed.

S2, selecting a GPS positioning point in a set time period from the background data storage center; specifically, due to the existence of the GPS positioning offset, the abnormal data in the GPS positioning data is removed first, and the K-means algorithm may be used to remove the abnormal point for the selected GPS positioning point.

The K-means algorithm mainly finds the point farthest from the clustering center point, filters and adjusts the farthest point from the data set, then uses the remaining points to find a new clustering center point, and then filters the point farthest from the clustering center point.

And S3, carrying out random interpolation on the rectangular cultivated land area between the adjacent GPS positioning points.

In the embodiment of the invention, the rectangular cultivated land area between adjacent GPS positioning points is as follows: and establishing rectangles for the adjacent GPS positioning points A and B, wherein the rectangles A and B are respectively used as two wide middle points of the rectangles, the width of the cultivator is used as the width of the rectangle, the distance between the A and the B is used as the length of the rectangle, and the established rectangles of the A and the B are rectangular cultivated land areas between the adjacent GPS positioning points A, B.

After the random interpolation, the method further comprises the following steps: sequencing all the GPS positioning points in the rectangular ploughing area after random interpolation according to the distance between the GPS positioning points and the initial GPS positioning point from small to large; wherein the initial GPS positioning point is the GPS positioning point with the earliest time in the rectangular cultivated land area.

However, due to the re-tillage phenomenon in the tillage process and the offset of the positioning points, an overlapping area exists between the rectangular frames, and the overlapping area does not take a regular shape, as shown in fig. 2, a schematic diagram of random interpolation between two GPS positioning points exists, and an irregular re-tillage area exists in the adjacent rectangular area, which is shown as a rectangle L in the drawing₁And L₂Rectangular L for the area ploughed by the ploughing tools₁Is two GPS points (x)₁,y₁) And (x)₂,y₂) The width of the distance between the two ploughs is the width of the ploughing tool. Wherein (x)₁,y₁) Precision and latitude of GPS positioning, two GPS points (x), respectively₁,y₁) And (x)₂,y₂) The distance between them is calculated as shown in equation (1).

And S4, performing spectral clustering on all the GPS positioning points after the random interpolation, wherein the result of the spectral clustering is a block area for obtaining at least two first clusters.

The purpose of the step is to divide cultivated land blocks of the GPS positioning points; because the arable land path presents bar distribution after the random interpolation, and the clustering and blocking of the bar-shaped regions are considered to be good at spectral clustering, the first spectral clustering is carried out on all the GPS positioning points after the random interpolation in S4, the primary block division of the GPS positioning points is completed, the determination of the number of the blocking regions uses the contour coefficient as the measurement index to carry out self-adaptive selection, and the number of the clustering blocks when the contour coefficient can obtain the maximum value is selected.

Because the result of the random interpolation in the step S3 completely covers the cultivated land area, all the currently obtained GPS positioning points are clustered, the GPS positioning points are partitioned, and the re-cultivated area is not repeatedly calculated, so that the aim of removing the weight of the operation area measuring and calculating result is achieved.

And S5, performing spectral clustering on each block region of the first clustering to obtain at least two block regions of the second clustering.

In order to divide the first clustering block result more finely, the second spectral clustering is carried out, and the clustering block number also uses the contour coefficient as a measurement index to carry out self-adaptive selection.

And S6, removing blank areas without GPS positioning points in the block areas of each secondary clustering, and performing arable inflection point elimination, so as to obtain final block areas through division.

The blank area without the GPS positioning point specifically comprises the following steps: and for all GPS positioning points in the block area of the second clustering, if the distance between two adjacent points is greater than a set threshold value, the current two adjacent points are blank areas.

Because a blank area without the coverage of positioning points exists in the new block area, which has a great influence on the calculation of the cultivated land area, the blank area in each block area needs to be eliminated, and the method comprises the following steps:

(1) calculating the distance l between adjacent GPS positioning points in each block;

(2) when the distance value l is larger than a certain threshold value, judging that a blank area exists between the positioning points, and generating a new block area;

(3) and (3) sequentially and circularly processing the (1) and the (2) until all data in the block are processed, and obtaining new block area division.

Because the GPS points are subjected to uniform interpolation in the step C and the adjacent GPS positioning points are subjected to distance sequencing, the distance between the adjacent GPS positioning points is within 3 meters, and the threshold value is obtained according to actual debugging and can be 4.

In the embodiment of the invention, the turning point of the farmland is specifically as follows: and calculating the distance between each point and the initial point for all the GPS positioning points in the partitioned area of the second clustering, performing curve fitting on the distance values, and taking the GPS positioning point corresponding to the extreme value point of the fitting curve as the turning point of the farmland.

As the arable land inflection point region continuously covered by the GPS positioning points exists in the arable land blocks, as shown in fig. 4(1), the region surrounded by the boundary vertices of the region has a region without positioning points, so that the area surrounded by the boundary vertices cannot represent the arable land area of the region, and therefore, the arable land inflection points existing in each block region are eliminated in step G, the method includes the following steps:

(1) respectively calculating the distance between each GPS positioning point in each block area and the first positioning point in each block to form a distance discrete value combination L;

(2) then carrying out minimum two-dimensional polynomial curve fitting on the L to obtain a fitting curve p;

(3) and solving an extreme point of the fitting curve p to obtain a GPS positioning point of a farmland inflection point, and obtaining new farmland block area division.

Fig. 4(1) is a schematic diagram of a region with an inflection point of arable land in a certain partitioned region, and due to the existence of the inflection point of arable land, a convex hull boundary point of a positioning point in the region cannot be effectively drawn; fig. 4(2) is a schematic diagram from a discrete point and a schematic diagram of a fitting curve, and it can be seen that the schematic diagram from the discrete point can effectively reflect the inflection point position of the farmland, and the inflection point position of the farmland can be obtained by solving each extreme point of the fitting curve, and the inflection point position is obtained by calculating the corresponding argument value of each extreme point of the fitting curve, where the derivative of the fitting curve is 0; FIGS. 4(3) and 4(6) are new farmland block areas obtained after the turning point of the farmland is removed.

Directly taking the block area of the second clustering without blank area and arable inflection point as the final block area;

and S7, obtaining sequence boundary points of all the final block areas, solving the area enclosed by the sequence boundary points of each final block area, and obtaining the sum of the areas corresponding to all the final block areas as the final operation area.

FIG. 5 is a schematic diagram of clockwise boundary points of two block regions in the embodiment of the present invention, and the sequential boundary points of all final block regions are solved by using a ConvexHull function of a scipy packet, which are { (x) respectively₀,y₀),(x₁,y₁),(x₂,y₂),…,(x_N,y_N),(x₀,y₀) N +1 sequential boundary points from 0 to N), (x)₀,y₀)～(x_N,y_N) Sequentially setting the coordinates of 0-Nth sequential boundary points;

the final working Area is Area;

Area＝|A÷2|×9101160000.085981；

wherein A is the area intermediate quantity;

a₀is an initial value of area, a₀＝y₀×(x₀-x₁)。

In summary, the above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.