A harvester operating speed control method

Abstract

The present application relates to the technical field of crop harvesting, and discloses a harvester operation speed control method, comprising the following steps: constructing a three-dimensional grid of a crop harvesting area to be measured, uniformly dividing the three-dimensional grid units into n grids, using a laser radar to perform three-dimensional scanning on crops, processing the collected radar data to generate a radar image; the method constructs a three-dimensional grid of a crop harvesting area to be measured, and further uniformly divides the grid units into multiple sub-grids, thereby achieving fine division of the harvesting area; this division method helps to more accurately assess the weight of crops in each grid, thereby providing a more optimal solution for adjusting the operating speed of the harvester; in subsequent steps, multiple grids are combined into fewer grids through correlation analysis of the weight of crops in the grids, thereby reducing the number of times the harvester needs to adjust the speed during operation, and improving the overall efficiency of the harvesting operation.

Description

Technical Field

[0001] This invention relates to the field of crop harvesting technology, specifically a method for controlling the operating speed of a harvester. Background Technology

[0002] In modern agricultural production, crop harvesting is a crucial step, its efficiency and quality directly impacting agricultural yield and economic benefits. Traditional crop harvesting operations rely heavily on manual experience, with harvester speeds often adjusted based on the operator's subjective judgment. However, this method has several drawbacks. First, human judgment struggles to accurately adapt to the actual growth conditions of crops in different regions, such as variations in crop height, density, and weight, easily leading to harvesting speeds that are too fast or too slow. Excessive speed may result in incomplete harvesting and increased crop losses; excessive speed reduces harvesting efficiency and increases operating costs. Second, with the development of large-scale and intensive agriculture, farmland areas are constantly expanding, and crop varieties are becoming increasingly diverse, making traditional manual control methods insufficient to meet the demands of large-scale, high-efficiency harvesting. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a harvester operating speed control method, which has the advantages of facilitating precise control of the harvester's operating speed and having a high degree of intelligence, thus solving the problems mentioned in the background art.

[0004] To achieve the aforementioned goals of precise control of harvester operating speed and a high degree of intelligence, this invention provides the following technical solution: a harvester operating speed control method, comprising the following steps:

[0005] S1: Construct a three-dimensional grid of the crop harvesting area to be tested, and divide the three-dimensional grid cells into n grids evenly;

[0006] S2: Use LiDAR to perform three-dimensional scanning of crops, process the collected radar data, and generate radar images;

[0007] S3: Perform threshold segmentation on the radar image to separate the crops from the radar image, and extract the height features of the crops from the segmented image. In each grid, interpolate the known height features to obtain the height value of any point in the grid.

[0008] S4: Calculate the volume V of crops within each grid. n Through crop volume V n The weight F of the crop in each grid was calculated. n The weight F of each grid crop n All correspond to the harvester's operating speed γ;

[0009] S5: Analyze the weight F of the crop in each grid. α The correlation between them is determined, and the n grids are merged according to the algorithm to obtain new n' grids, where n' < n;

[0010] S6: Control the harvester's operating speed γ' after merging the grids based on the weight of the crops in the new n' grids;

[0011] S7: The weight uniformity (CWDU) of crops is analyzed based on the n grids before merging and the n' grids after merging.

[0012] Preferably, step S3 includes the following steps:

[0013] S3.1: Denoise or enhance the contrast of radar images to improve segmentation accuracy;

[0014] S3.2: Use threshold segmentation to segment crops from radar images. The expression is:

[0015]

[0016] Where f(x,y) is the pixel value at position (x,y) in the radar image, T is the set threshold, g(x,y) is the segmented binary image, 1 represents the crop region, and 0 represents the background region;

[0017] S3.3: Extract the height features of crops from the segmented binary image g(x,y), using the following expression:

[0018]

[0019] Where H(x,y) is the height value of position (x,y) in the radar image, M is the mask of the crop area, and H is the height feature of the crop.

[0020] S3.4: Use bilinear interpolation to interpolate the crop height feature within a single grid cell. Let (x0, y0) be the point to be interpolated, and (x1, y1), (x1, y2), (x2, y1), and (x2, y2) be four known points surrounding the point to be interpolated. The expression is:

[0021]

[0022] Where a, b, c, and d are interpolation coefficients.

[0023] Preferably, step S4 specifically includes:

[0024] S4.1: Based on the average height H of crops within the grid grid Area A of crops within the grid n *B nThe volume of the crop can be obtained using the following expression:

[0025]

[0026] S4.2: Based on volume V α The weight F of the crop in each grid is calculated from the density ρ of the crop. α The expression is:

[0027] F n =V n *ρ

[0028] Preferably, step S5 specifically includes:

[0029] S5.1: Analyze crop weight F for each grid using Pearson correlation coefficient. n The correlation between them is expressed as:

[0030]

[0031] Where i and j are any two grids, and their crop weights correspond to F respectively. i F j , and These represent the average crop weights at grids i and j, respectively.

[0032] S5.2: and The expression is:

[0033]

[0034]

[0035] S5.3: Calculate the Pearson correlation coefficient r between each pair of grids in the n grids. ij We can obtain a correlation coefficient matrix R, R(i,j) = r ij ;

[0036] S5.4: Merge grids based on the correlation coefficient matrix R and the merging threshold τ. For example, setting τ=0.8 means that when the correlation coefficient of crop weight between two grids is greater than 0.8, the two grids are considered to have a strong correlation and can be considered for merging.

[0037] Preferably, the specific step of merging the meshes in step S5 is as follows:

[0038] Step 1: Initialization, treating each grid as an independent cluster;

[0039] Step 2: Based on the correlation coefficient matrix R, convert the correlation coefficients into a distance matrix D, where the distance matrix D(i,j) = 1 - |r|. ij |;

[0040] Step 3: Find the two clusters C with the smallest distance in the distance matrix D. i and C j If the distance D (C) i C j If the number of clusters is less than the merging threshold τ, then the two clusters will be merged into a new C. new =C i ∪C j ;

[0041] Step 4: After merging clusters, recalculate the distances between the new cluster and other clusters, and update the distance matrix D'(C). new C k );

[0042] Step 5: Calculate the distance between them using the average link method, expressed as:

[0043]

[0044] Among them, |C new | and | C k | Respectively, cluster C new and cluster C k The number of elements in the middle;

[0045] Step 6: Repeat steps 3 to 6 until there are no clusters that can be merged if the distance D is less than the merging threshold τ.

[0046] Preferably, the specific steps of step S7 are as follows:

[0047]

[0048]

[0049] Where θF' is the standard deviation of crop weight within each grid after merging. This represents the weight of the crops in the i-th grid after merging. This represents the average weight of crops within all grids after merging.

[0050] The weight uniformity (CWDU) of crops within the grid before and after merging is calculated using the following expression:

[0051]

[0052] Compared with the prior art, the present invention provides a method for controlling the operating speed of a harvester, which has the following beneficial effects:

[0053] 1. This harvester operating speed control method constructs a three-dimensional grid of the crop harvesting area to be measured, and further divides the grid cells into multiple sub-grids to achieve a fine division of the harvesting area. This division method helps to more accurately assess the crop weight in each grid, thereby providing a better solution for adjusting the harvester's operating speed. In subsequent steps, by analyzing the correlation of crop weight within the grid, multiple grids are merged into fewer grids, reducing the number of times the harvester needs to frequently adjust its speed during operation and improving the overall efficiency of the harvesting operation.

[0054] 2. This harvester speed control method uses a grid merging algorithm to combine grids with similar loads, allowing the harvester to maintain a relatively stable load within the merged grids. This uniform distribution helps reduce malfunctions and damage caused by uneven loads, improving harvesting quality. Precise harvesting control can reduce crop losses during the harvesting process. For example, by adjusting the harvester's operating speed, excessive crushing or omission of crops in the field can be avoided, thereby increasing the harvest rate. Attached Figure Description

[0055] Figure 1 This is a schematic diagram of a harvester operation speed control method proposed in this invention. Detailed Implementation

[0056] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] Please see Figure 1 A method for controlling the operating speed of a harvester, comprising the following steps:

[0058] S1: Construct a three-dimensional grid of the crop harvesting area to be tested, and divide the three-dimensional grid cells into n grids evenly;

[0059] S2: Use LiDAR to perform three-dimensional scanning of crops, process the collected radar data, and generate radar images;

[0060] S3: Perform threshold segmentation on the radar image to separate the crops from the radar image, and extract the height features of the crops from the segmented image. In each grid, interpolate the known height features to obtain the height value of any point in the grid.

[0061] S4: Calculate the volume V of crops within each grid.n Through crop volume V n The weight F of the crop in each grid was calculated. n The weight F of each grid crop n All correspond to the harvester's operating speed γ;

[0062] S5: Analyze the weight F of the crop in each grid. α The correlation between them is determined, and the n grids are merged according to the algorithm to obtain new n' grids, where n' < n;

[0063] S6: Control the harvester's operating speed γ' after merging the grids based on the weight of the crops in the new n' grids;

[0064] S7: The weight uniformity (CWDU) of crops is analyzed based on the n grids before merging and the n' grids after merging.

[0065] Accurately Calculating Crop Weight: A 3D scan of the crop is performed using LiDAR to generate a LiDAR image. Thresholding segmentation is then used to segment the crop from the image. The height features of the crop are further extracted, and interpolation calculations are performed within each grid to obtain the height value of any point within the grid. Based on these height values, the volume and weight of the crop within each grid can be accurately calculated.

[0066] Dynamic speed adjustment for harvesters: The harvester's operating speed is calculated based on the weight of the crops in each grid. During harvesting, the harvester can automatically adjust its speed according to the weight of the crops in its current grid, achieving precise harvesting. This dynamic adjustment method helps prevent harvesters from missing crops due to excessive speed or reducing operational efficiency due to insufficient speed.

[0067] Evaluating the effectiveness of grid merging: By calculating indicators such as the reduction rate of grid number, uniformity of crop weight distribution, and stability of harvester operating speed, the effectiveness of the grid merging algorithm can be comprehensively evaluated. These evaluation results can provide decision support for subsequent algorithm optimization and parameter adjustment.

[0068] Harvesting strategies can be guided by the distribution of crop weight and the operating speed of the harvester. For example, in areas with heavier crops, the harvester's operating time can be increased or harvesting efficiency can be improved; in areas with lighter crops, the harvester's operating time can be reduced or harvesting efficiency can be decreased.

[0069] Step S3 includes the following steps:

[0070] S3.1: Denoise or enhance the contrast of radar images to improve segmentation accuracy;

[0071] S3.2: Use threshold segmentation to segment crops from radar images. The expression is:

[0072]

[0073] Where f(x,y) is the pixel value at position (x,y) in the radar image, T is the set threshold, g(x,y) is the segmented binary image, 1 represents the crop region, and 0 represents the background region;

[0074] S3.3: Extract the height features of crops from the segmented binary image g(x,y), using the following expression:

[0075]

[0076] Where H(x,y) is the height value of position (x,y) in the radar image, M is the mask of the crop area, and H is the height feature of the crop.

[0077] S3.4: Use bilinear interpolation to interpolate the crop height feature within a single grid cell. Let (x0, y0) be the point to be interpolated, and (x1, y1), (x1, y2), (x2, y1), and (x2, y2) be four known points surrounding the point to be interpolated. The expression is:

[0078]

[0079] Where a, b, c, and d are interpolation coefficients.

[0080] Step S4 specifically includes:

[0081] S4.1: Based on the average height H of crops within the grid grid Area A of crops within the grid n *B n The volume of the crop can be obtained using the following expression:

[0082]

[0083] S4.2: Based on volume V α The weight F of the crop in each grid is calculated from the density ρ of the crop. α The expression is:

[0084] F n =V n *ρ

[0085] Step S5 specifically includes:

[0086] S5.1: Analyze crop weight F for each grid using Pearson correlation coefficient. n The correlation between them is expressed as:

[0087]

[0088] Where i and j are any two grids, and their crop weights correspond to F respectively. i F j , and These represent the average crop weights at grids i and j, respectively.

[0089] S5.2: and The expression is:

[0090]

[0091]

[0092] S5.3: Calculate the Pearson correlation coefficient r between each pair of grids in the n grids. ij We can obtain a correlation coefficient matrix R, R(i,j) = r ij ;

[0093] S5.4: Merge grids based on the correlation coefficient matrix R and the merging threshold τ.

[0094] The choice of the merging threshold τ directly affects the quality of clustering or segmentation. If τ is set too large, different categories that should be separated may be incorrectly merged together, resulting in over-merging, making the clustering results too coarse and losing detailed information in the data. Conversely, if τ is set too small, data that should belong to the same category may be split into multiple smaller categories, resulting in under-merging, making the clustering results too fragmented and unable to accurately reflect the inherent structure of the data.

[0095] The possible range of values ​​for τ is divided into a grid. Experiments are then conducted for each τ value within the grid, and the corresponding performance metrics (such as internal or external metrics) are calculated. By comparing the performance metrics under different τ values, the optimal range of τ values ​​can be found. For example, assuming the range of τ values ​​is [0, 1], this range can be divided into 100 equally spaced points, i.e., τ = 0.01, 0.02, ..., 1.00. Then, clustering or segmentation experiments are performed for each τ value, and the corresponding silhouette coefficient is calculated. Finally, the intervals containing the τ values ​​with the highest silhouette coefficients are selected as candidate ranges.

[0096] If the quality of the clustering results changes only slightly when the value of τ changes within a certain range, it indicates that the range is a relatively stable candidate range. However, if a small change in the value of τ causes a significant change in the clustering results, it is necessary to carefully consider the selection of the value of τ within that range, or to adjust the candidate range.

[0097] This technical solution can be adjusted and optimized according to different harvesting areas and crop types. For example, in areas with high crop density, the number of grid divisions can be increased to improve the accuracy of the assessment; in areas with low crop density, the number of grid divisions can be reduced to simplify the calculation process.

[0098] The specific steps for merging the meshes in step S5 are as follows:

[0099] Step 1: Initialization, treating each grid as an independent cluster;

[0100] Step 2: Based on the correlation coefficient matrix R, convert the correlation coefficients into a distance matrix D, where the distance matrix D(i,j) = 1 - |r|. ij |;

[0101] Step 3: Find the two clusters C with the smallest distance in the distance matrix D. i and C j If the distance is D (C i C j If the number of clusters is less than the merging threshold τ, then the two clusters will be merged into a new C. new =C i ∪C j ;

[0102] Step 4: After merging clusters, recalculate the distances between the new cluster and other clusters, and update the distance matrix D'(C). new C k );

[0103] Step 5: Calculate the distance between them using the average link method, expressed as:

[0104]

[0105] Among them, |C new | and | C k | Respectively, cluster C new and cluster C k The number of elements in the middle;

[0106] Step 6: Repeat steps 3 to 6 until there are no clusters that can be merged if the distance D is less than the merging threshold τ.

[0107] The specific steps of step S7 are as follows:

[0108]

[0109]

[0110] Where θF' is the standard deviation of crop weight within each grid after merging. This represents the weight of the crops in the i-th grid after merging. This represents the average weight of crops within all grids after merging.

[0111] The weight uniformity (CWDU) of crops within the grid before and after merging is calculated using the following expression:

[0112]

[0113] The reciprocal of the ratio of standard deviation to mean is used here to make the CWDU value larger when the uniformity is higher, so as to more intuitively reflect the quality of uniformity.

[0114] In summary, this harvester speed control method constructs a three-dimensional grid of the crop harvesting area and further divides the grid cells into multiple sub-grids, achieving a refined division of the harvesting area. This division method helps to more accurately assess the crop weight within each grid, thus providing a better solution for adjusting the harvester's operating speed. In subsequent steps, by analyzing the correlation of crop weight within the grid, multiple grids are merged into fewer grids, reducing the number of times the harvester needs to frequently adjust its speed during operation, improving the overall efficiency of the harvesting operation. At the same time, the stable load distribution also reduces the wear and tear on key components of the harvester, lowering maintenance costs and downtime.

[0115] Furthermore, by using a grid merging algorithm, grids with similar loads are combined, allowing the harvester to maintain a relatively stable load while operating within the merged grid. This uniform distribution helps reduce malfunctions and damage to the harvester caused by uneven loads, improving harvesting quality. Precise harvesting control can also reduce crop losses during the harvesting process. For example, by adjusting the harvester's operating speed, excessive crushing or omission of crops in the field can be avoided, thereby increasing the crop yield.

[0116] It should be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0117] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method of controlling the operating speed of a harvester, characterised by, The method comprises the following steps: S1: constructing a three-dimensional grid of a crop harvesting area to be measured, and uniformly dividing the three-dimensional grid cells into n grids; S2: using a laser radar to perform three-dimensional scanning on the crops in the n grids, processing the collected radar data to generate a radar image; S3: performing threshold segmentation on the radar image to segment the crops from the radar image, and extracting the height features of the crops from the segmented image, and in each grid, interpolating the known height features to obtain the height values of any point in the grid; S4: Calculate the volume V of crops within each grid. n Through crop volume V n The weight F of the crop in each grid was calculated. n The weight F of each grid crop n All of these correspond to the harvester's operating speed γ. For example, in areas where crops are heavy, the harvester's operating time can be increased or the harvesting efficiency can be improved; in areas where crops are light, the harvester's operating time can be reduced or the harvesting efficiency can be decreased. S5: analyze the weight F of each grid crop n correlation between the n grids, and according to the algorithm, the n grids are merged to obtain new n' grids, n' < n; S6: controlling the running speed γ' of the harvester after the grids are merged according to the weight of the crops in the new n' grids; S7: analyzing the weight uniformity CWDU of the crops based on the n grids before merging and the n' grids after merging; In step S5, the following steps are specifically performed: S5.1: Analyze crop weight F for each grid using Pearson correlation coefficient. n The correlation between them is expressed as: Wherein, i, j are any two grids, the crop weight of which respectively corresponds to F i , F j , and represent the average value of the crop weight of grid i and grid j respectively; S5.2: and The expression for is: S5.3: Calculate the Pearson correlation coefficient r between each pair of the n grids ij A correlation coefficient matrix R can be obtained, R(i, j) = r ij ; S5.4: merging the grids according to the correlation coefficient matrix R and the merging threshold τ; The step of merging the grids in step S5 is specifically as follows: Step one: initialization, regarding each grid as an independent cluster; Step two: convert the correlation coefficients to a distance matrix D according to the correlation coefficient matrix R, distance matrix D(i,j) = 1 - |r ij |; Step 3: Find the two clusters C with the smallest distance in the distance matrix D. i and C j If the distance D (C) i C j If the number of clusters is less than the merging threshold τ, then the two clusters will be merged into a new C. new =C i ∪C j ; Step four: After merging clusters, re-compute the distances between the new cluster and other clusters, and update the distance matrix D'(C new , C k ); Step five: calculating the distance between clusters using the average linkage method, and the expression is: wherein |C new | and |C k | are the number of elements in the cluster C new and the cluster C k respectively. Step six: repeating steps three to six until there are no clusters with a distance D < merging threshold τ that can be merged; The specific steps of step S7 are as follows: wherein θF' is the standard deviation of the weight of the crop in each grid after merging, is the weight of the crop in the i-th grid after merging, is the average of the weight of the crop in all grids after merging; Calculate the weight uniformity CWDU of the crops in the grids before and after merging, and the expression is: 。 2. A method of controlling the speed of a harvester according to claim 1, characterised in that, The step S3 includes the following steps: S3.1: denoising or enhancing the contrast of the radar image to improve the accuracy of segmentation; S3.2: using threshold segmentation method to segment the crops from the radar image; S3.3: extracting the height features of the crops in the segmented binary image g(x, y), and the expression is: Where H(x, y) is the height value of position (x, y) in the radar image, M is the mask of the crop area, and H is the height feature of the crop; S3.4: using bilinear interpolation to interpolate the height features of the crops in a single grid cell, and setting (x0, y0) as the interpolation point, (x1, y1), (x1, y2), (x2, y1), (x2, y2) as the four known points around the interpolation point, and the expression is: Where a, b, c, and d are interpolation coefficients.

3. A method of controlling the speed of a harvester according to claim 1, characterised in that, In the step S4, the following steps are specifically performed: S4.1: The average height H of the crops within the grid grid The area A of the crops within the grid n * x B n The volume of the crops is obtained, expressed by the formula: S4.2: According to the volume V n The weight F of the crop in each grid is calculated with the density p of the crop n , the expression is: F n = F n x p.

Images