In-field path planning method of corn planter for soil mechanical compaction and reduction

Abstract

The invention discloses a corn planter in-field path planning method for soil mechanical compaction and reduction, and relates to the technical field of agricultural machinery path planning. The invention aims to solve the problems that the existing corn planter path planning method is low in seeding efficiency and the growth of corn crops is influenced by multiple times of soil compaction. The method comprises the following steps: acquiring in-field furrow information and boundary point information of irregular field parcels, clustering the irregular field parcels, and dividing the irregular field parcels into a plurality of regular field parcels; setting a seed and fertilizer supplementing point position in each regular field; and planning the global seeding path based on the position of the seed fertilizer supplement point to obtain a final seeding path. The method is used for planning the corn seeding path.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural machinery path planning, and particularly to a method for in-field path planning of a corn seeder for reducing soil mechanical compaction. Background Art

[0002] With the continuous advancement of agricultural modernization, the application of intelligent agricultural technologies has gradually become a key means to improve agricultural production efficiency and sustainable development. Traditional mechanized operations, although significantly improving agricultural production efficiency, have also brought problems such as soil compaction and deterioration of the crop growth environment. As an important device in agricultural production, the optimization of the operation path of a corn seeder is of great significance for reducing soil compaction and improving operation efficiency.

[0003] The operation path of traditional corn seeders is generally designed based on fixed rules, and this method often has uneven operation. Some fields may have multiple sowing or unsown situations, resulting in the need for the seeder to operate multiple times. Therefore, the driving path length and sowing time of the corn seeder are increased, leading to low sowing efficiency. Multiple operations will also cause some soil to be compacted multiple times, making the soil less breathable, thus affecting the growth of corn crops. Summary of the Invention

[0004] The purpose of the present invention is to solve the problems that the existing corn seeder path planning method still has low sowing efficiency and multiple soil compactions affecting the growth of corn crops, and a method for in-field path planning of a corn seeder for reducing soil mechanical compaction is proposed.

[0005] The method for in-field path planning of a corn seeder for reducing soil mechanical compaction is specifically as follows:

[0006] Step 1: Obtain the in-field ridge ditch information and boundary point information of the irregular field, and cluster the irregular field to divide it into multiple regular fields, specifically as follows:

[0007] Step 1.1: Obtain the boundary point information of the irregular field;

[0008] The boundary point information is the set Ρ of boundary points of the irregular field = {(x1,y1),(x2,y2),...,(x s ,y s ),...,(x n ,y n )};

[0009] Among them, n is the total number of boundary points of the irregular field, and (x s ,y s ) is the coordinate of the s-th boundary point of the irregular field, and (x n ,y n) is the coordinate of the nth boundary point of the irregular field;

[0010] Steps 1 and 2: Use the boundary point information of the irregular field for clustering, and divide the irregular field into multiple regular fields;

[0011] Step 2: Set the positions of the seed fertilizer replenishment points within each regular field;

[0012] Step 3: Plan the global sowing path based on the positions of the seed fertilizer replenishment points to obtain the final sowing path.

[0013] Furthermore, the use of the boundary point information of the irregular field for clustering in Steps 1 and 2 to divide the irregular field into multiple regular fields is specifically as follows:

[0014] Step 1.1: Obtain the number of points N(p s (x s ,y s ) within the ε-neighborhood of each point p in the set of boundary points of the irregular field, specifically: s ) is as follows:

[0015] N(p s ) = {q ∈ Ρ|||p s - q|| ≤ ε}

[0016] where s is the label of the point in the set of boundary points of the irregular field, q is the point within the ε-neighborhood of point p s (x s ,y s ), s ranges from 1 to n, and n is the total number of boundary points of the irregular field;

[0017] Step 1.2: Use N(p s ) to determine whether p s is a core point. If p s is a core point, add p s to the core point set C, and then execute Step 1.3; otherwise, mark p s as a noise point;

[0018] Step 1.3: Obtain the total number of points in N(p s ). If the total number of points in N(p s ) is greater than the preset threshold, it means that point p s is covered by density, and mark the points in N(p s ) as the same cluster C k ; otherwise, it means that point p s is not covered by density, and mark point p s as a noise point;

[0019] Step 1.4: For each cluster Ck Divided into regular rectangular fields with an area of A s , specifically:

[0020]

[0021] Among them, A s is the maximum area sown by the seeder in a single pass, S k is the set of all small regular fields in cluster C k . is the j k -th small regular field corresponding to the k clusters C k , is the area of the j k -th small regular field, m k is the number of small regular fields corresponding to cluster C k .

[0022] Furthermore, the use of N(p s ) to determine whether p s is a core point in step 1-3 is specifically:

[0023] If N(p s ) ≥ MinPts, then p s is a core point; otherwise, p s is not a core point;

[0024] Among them, MinPts is the threshold of the number of points in the minimum neighborhood;

[0025] The set of core points C is specifically: C = {p s |N(p s ) ≥ MinPts}.

[0026] Furthermore, setting the positions of seed and fertilizer replenishment points in each regular field in step 2 is specifically:

[0027] Step 2-1, initialize parameters and set constraints;

[0028] The specific initialization parameters are:

[0029] Set the initial solution The initial temperature T = T0, the initial objective function value F(x) = D0, and initialize the iteration number N = 1;

[0030] Among them, is the set of randomly set initial positions of seed and fertilizer replenishment points, D0 is the total length of the initial sowing path, T0 > 1000, T0 is the initial temperature, t is the total number of replenishment points, and p is the replenishment point label;

[0031] The constraints include: obstacle area constraint, reciprocating path constraint for ox plowing, and field switching path constraint;

[0032] The obstacle area constraint is:

[0033] where O s is the obstacle area, and X is the set of labels of irregular field boundary points;

[0034] The reciprocating path constraint for ox plowing is:

[0035] where D 切换 is the field switching distance, distance(x s ,y s ,P 起点 ) is the distance from the point (x s ,y s ) to the starting point of the next field, and P 起点 is the starting point of the field;

[0036] Step 22. Randomly perturb the positions of the supplementary points based on the currently existing positions of the supplementary points to obtain a new solution

[0037] where is the set of positions of the supplementary points after perturbation, δ~U(-Δ,Δ), δ is a random perturbation, R p ′=R p +δ, R p ′ is the supplementary point after random perturbation, R p is the position of the p-th supplementary point, -Δ is the lower limit of random perturbation, Δ is the upper limit of random perturbation, and U(-Δ,Δ) is a uniform distribution within the range of (-Δ,Δ);

[0038] Step 23. Obtain the objective function value F(x) before random perturbation and the objective function value F(x′) after random perturbation, and obtain the objective function difference ΔF = F(x′) - F(x);

[0039] Step 24. Determine whether to accept the new solution based on the objective function difference. If the new solution is accepted, let x = x', and at the same time update the temperature T = αT according to the cooling schedule; then execute Step 25; if the new solution is not accepted, directly return to Step 22;

[0040] The determination of whether to accept the new solution based on the objective function difference is specifically:

[0041] First, obtain the probability P using the objective function difference:

[0042]

[0043] Among them, α ∈ (0, 1), and α is the temperature decrease coefficient;

[0044] Then, judge whether to accept the new solution according to the probability P:

[0045] If P = 1, accept the new solution; otherwise, generate a random number, compare the magnitudes of P and the random number. If P is less than or equal to the random number, accept the new solution; if P is greater than the random number, do not accept the new solution;

[0046] Step 2.5: Judge whether the current temperature has reached the termination temperature T min or whether the current number of iterations has reached the maximum number of iterations N max ; if T ≤ T min or N ≥ N max then output the set of positions of the optimal supplementary points and the corresponding minimum path length D'; otherwise, let N = N + 1 and return to Step 2.2.

[0047] Furthermore, the constraints of the supplementary points are as follows:

[0048]

[0049] Among them, is the set of boundary points of the small regular field , (x b , y b ) are the coordinates of the boundary points in the small regular field , b is the label of the boundary point in the small regular field , is the total number of boundary points in the small regular field .

[0050] Furthermore, obtaining the objective function value F(x) before the random perturbation and the objective function value F(x') after the random perturbation in Step 2.3 is specifically as follows:

[0051]

[0052] Among them, D is the sowing path length before the random perturbation, k is the cluster label, K is the total number of clusters, j k is the label of the small regular field in the set of regular fields formed by the k-th cluster, m k is the total number of small regular fields in the set of regular fields formed by the k-th cluster, is the sowing path length of the bull tillage reciprocating method in the field before the random perturbation, is the path length from the boundary point where sowing ends in the field before the random perturbation to the nearest supplementary point; D' is the sowing path length after the random perturbation, is the sowing path length of the bull tillage reciprocating method in the field after the random perturbation The path length of sowing in is the path length from the boundary of the field after random perturbation to the nearest supplementary point.

[0053] Furthermore, the global sowing path is planned based on the positions of the seed and fertilizer supplementary points in step three to obtain the final sowing path, specifically as follows:

[0054] Step three one: According to the search step size α f in the search area Ω f randomly initialize the positions of N' f furrows to generate N f sowing paths, and regard each sowing path as a firefly Ρ i :

[0055] Ρ i ={(x i,1 , y i,1 ), (x i,2 , y i,2 ),..., (x i,kf , y i,kf ),..., (x i,KF , y i,KF}}, i = 1, 2,..., N f

[0056] Initialize the light intensity L0 of the firefly; initialize the iteration number t' = 1;

[0057] where i is the sowing path label, (x i,kf , y i,kf ) is the position in the kf-th furrow of the i-th path, KF is the total number of positions in the furrows of the i-th sowing path, Ρ i is the i-th firefly; the positions of N f furrows include all the positions of the optimal supplementary points;

[0058] Step three two: Obtain the fitness value F f (Ρ i );

[0059] Step three three: Obtain the light intensity L of each firefly according to the fitness value of each firefly i , specifically as follows:

[0060]

[0061] where γ f is the light intensity absorption coefficient, and L0 is the initial light intensity;

[0062] Step three four: Obtain the attractiveness between fireflies, specifically as follows:

[0063]

[0064] Among them, β ij,f is the attractiveness of firefly Ρ i to firefly Ρ j , β 0f is the preset maximum attractiveness value between fireflies, d ij,f is the relative light intensity between firefly Ρ i and firefly Ρ j , (x i,kf , y i,kf ) is the kf-th point in the seeding path corresponding to firefly Ρ i , (x j,kf , y j,kf ) is the kf-th point in the seeding path corresponding to firefly Ρ j ;

[0065] Step 35: Update the positions of fireflies using the attractiveness between fireflies;

[0066] Step 36: Select elite fireflies according to the light intensity of fireflies and the updated positions of fireflies, perform chaotic perturbation on the positions of elite fireflies to update the positions of elite fireflies, and then execute Step 37; re-initialize the fireflies except the elite fireflies randomly, and return to Step 32;

[0067] Step 37: Update the step factor, specifically:

[0068]

[0069] Among them, κ f ∈(0,1), κ f is the step dynamic attenuation coefficient;

[0070] Step 38: If the number of iterations reaches the maximum value T max,firefly or the change in the fitness function value is less than the preset threshold, output as the optimal path Ρ best ; otherwise, let t' = t' + 1, and return to Step 32;

[0071] The change in the fitness function value being less than the preset threshold is specifically:

[0072]

[0073] Among them, ε f is the search precision threshold, is the position of the elite firefly in the (t' + 1)-th iteration, is the position of the elite firefly in the t'-th iteration, is The fitness value of is the fitness value of

[0074] Furthermore, the obtaining of the fitness value F of each firefly in Step 3-2 f (Ρ i ) is specifically as follows:

[0075]

[0076] where C comp,f (Ρ i ) is the number of compacted ridges of path Ρ i , ω a is the weight of the number of compacted ridges, D total,f (Ρ i ) is the total driving distance of path Ρ i , ω b is the weight of the driving distance, Area f (Ρ i ) is the sowing area covered by path Ρ i , ω c is the weight of the sowing area, is the internal path length of the i-th sowing path in the small regular plot j k , l 压实 is the degree of compaction of the i-th sowing path in the small regular plot j k , is the sowing area of the i-th sowing path in the small regular plot j k , is the boundary path length of the i-th sowing path in the small regular plot j k .

[0077] Furthermore, the updating of the firefly positions by using the attractiveness between fireflies in Step 3-5 is specifically as follows:

[0078]

[0079] where is the position of firefly Ρ i at time t', is the position of firefly Ρ i at time t'+1, is the position of firefly Ρ j at time t', α f is the step size factor, rand f is a random number obeying [0,1], ΔΡ chaos,i is the chaotic search perturbation, r c is the chaotic mapping control parameter, is the chaotic perturbation variable at time t' is the chaotic disturbance variable at time t'+1.

[0080] Furthermore, in step 3-6, select elite fireflies according to the light intensity of the fireflies and the updated positions of the fireflies, perform chaotic disturbance on the positions of the elite fireflies to update the positions of the elite fireflies, and then execute step 3-7; re-randomly initialize the fireflies except the elite fireflies, and return to step 3-2, specifically:

[0081] Step 3-6-1: Sort the positions of the fireflies in ascending order according to the light intensity of the fireflies, and select the first p elite ·N f fireflies as elite fireflies;

[0082] where p elite is the proportion of the elite group;

[0083] Step 3-6-2: Perform chaotic disturbance on the positions of the elite fireflies to update the positions of the elite fireflies, and then execute step 3-7:

[0084]

[0085] where Ρ elite is the proportion of the elite group, ΔΡ chaos,i is the chaotic search disturbance, is the position of the elite firefly after update;

[0086] Step 3-6-3: Re-randomly initialize (1 - p elite )·N f fireflies that are not elite fireflies, and return to step 3-2.

[0087] The beneficial effects of the present invention are:

[0088] The present invention proposes a method for path planning of a corn planter in the field, clusters irregular fields, and optimizes the path based on a multi-agent optimization algorithm, which can not only achieve compaction control during the operation of the planter, but also ensure the high efficiency and coverage rate of the operation to the greatest extent. The method of the present invention can not only effectively reduce the impact of compaction on the soil, but also improve the efficiency of agricultural machinery path planning. The present invention conducts in-field path planning for maximizing the operation coverage area of the corn planter on the premise of minimizing soil mechanical compaction, optimizes the in-field path planning of the corn planter, reduces the impact of compaction on the soil, achieves the operation goal of high efficiency and low loss, and avoids the negative impact on the growth of corn crops. Description of the Drawings

[0089] Figure 1 is the flowchart of the present invention;

[0090] Figure 2It is the flow chart of the DBSCAN algorithm;

[0091] Figure 3 It is the flow chart of the simulated annealing algorithm;

[0092] Figure 4 It is the flow chart of the chaotic firefly algorithm. Specific implementation manners

[0093] Specific implementation manner 1: As Figure 1 shown, the specific process of the in-field path planning method for a corn seeder aiming at soil mechanical compaction reduction is as follows:

[0094] Step 1: As Figure 2 shown, use GNSS technology to obtain the in-field ridge and furrow information and boundary point information of the irregular field block, and cluster the irregular field block to divide it into multiple regular field blocks. Specifically:

[0095] Step 1-1: Use GNSS technology to obtain the boundary point information of the irregular field block;

[0096] The boundary point information is the set Ρ of boundary points of the irregular field block = {(x1, y1), (x2, y2),..., (x s , y s ),..., (x n , y n )};

[0097] Among them, n is the total number of boundary points of the irregular field block, (x n , y n ) is the coordinate of the nth boundary point of the irregular field block, and (x s , y s ) is the coordinate of the sth boundary point of the irregular field block;

[0098] Step 1-2: Use the boundary point information of the irregular field block for clustering to divide the irregular field block into multiple regular field blocks. Specifically:

[0099] Step 1-2-1: Obtain the number of points N(p s (x s , y s )) in the ε-neighborhood of each point p s in the set of boundary points of the irregular field block according to the preset neighborhood radius ε = {q ∈ Ρ |||p s - q|| ≤ ε};

[0100] Among them, s is the label of the points in the set of boundary points of the irregular field block, q is the point in the ε-neighborhood of point p s (x s , y s ), s takes values from 1 to n, and n is the total number of boundary points of the irregular field block;

[0101] Step 122: Use N(p s ) to determine whether p s is a core point. If p s is a core point, add p s to the core point set C = {p s |N(p s ) ≥ MinPts}, and then execute Step 123; otherwise, mark p s as a noise point;

[0102] If N(p s ) ≥ MinPts, then p s is a core point; otherwise, p s is not a core point;

[0103] where MinPts is the threshold of the number of points in the minimum neighborhood;

[0104] Step 123: Obtain the total number of points in N(p s ). If the total number of points in N(p s ) is greater than the preset threshold 5, it means that the point p s is density-covered. Mark the points in N(p s ) as the same cluster C k ; otherwise, it means that the point p s is not density-covered, and then mark the point p s as a noise point;

[0105] Step 124: Divide each cluster C k into regular rectangular fields with an area of A s :

[0106] where A s is the maximum area sown by the seeder in one pass, S k is the set of all small regular fields in cluster C k , is the j k -th small regular field corresponding to cluster C k , is the area of the j k -th small regular field, and m k is the number of small regular fields corresponding to cluster C k .

[0107] When dividing each cluster C k into regular rectangular fields with an area of A s , the area A s should contain as many points in cluster C k as possible; A sThe maximum area that can be sown with a full box of seeds and fertilizers.

[0108] Step 2. As Figure 3 shown, set the positions of the seed and fertilizer replenishment points within each regular field block, specifically:

[0109] Step 2-1. Initialize the parameters and set the constraints:

[0110] Set the initial solution The initial temperature T = T0, the initial objective function value F(x) = D0, initialize the iteration count N = 1; set the maximum iteration count N max = 100;

[0111] Among them, is the set of randomly set initial positions of the seed and fertilizer replenishment points, D0 is the total length of the initial sowing path, T0 > 1000 is a relatively large initial temperature, t is the total number of replenishment points, t can be set to a relatively large value, and then unnecessary replenishment points are reduced through optimization, p is the label of the replenishment point;

[0112] The constraints include:

[0113] Obstacle area constraint: The path of the seeding machine cannot pass through the obstacle area O s :

[0114] Among them, X is the set of labels of the boundary points of the irregular field block;

[0115] Ox-plowing reciprocating path constraint: The path of the seeding machine should maintain the rules of the ox-plowing reciprocating method, and the continuity of the path points (x s , y s ) satisfies x s+1 = x s + Δx, y s+1 = y s or x s+1 = x s , y s+1 = y s + Δy;

[0116] Among them, Δx is the distance the seeding machine moves in the x direction, and Δy is the distance the seeding machine moves in the y direction;

[0117] Field block switching path constraint: After the seeding machine completes the current field block, it should go from the current end point to the rectangular vertex P of the next field block 起点 Specifically:

[0118] Among them, D 切换 is the field block switching distance;

[0119] Step 2-2. Randomly perturb the positions of the replenishment points based on the currently existing positions of the replenishment points to obtain a new solution

[0120] Among them, is the set of positions of the supplementary points after perturbation, δ~U(-Δ,Δ), δ is a random perturbation, R p ′ = R p +δ, R p ′ is the supplementary point after random perturbation, R p ′ satisfies the boundary constraint conditions, R p is the position of the p-th supplementary point, -Δ is the lower limit of the random perturbation, Δ is the upper limit of the random perturbation, and U(-Δ,Δ) is the uniform distribution within the range of (-Δ,Δ);

[0121] Among them, the supplementary points must be located on the boundary of the regular field block, specifically:

[0122]

[0123] Among them, is the set of boundary points of the small regular field block of, (x b , y b ) is the coordinate of the boundary point in the small regular field block , b is the label of the boundary point in the small regular field block , is the total number of boundary points in the small regular field block ;

[0124] During the seeding process, the consumption of seeds and fertilizers cannot exceed the capacity of the seeder, specifically:

[0125]

[0126] Among them, is the area of, Q s is the maximum capacity of the seed box, C s is the consumption of seeds per unit area, C f is the consumption of fertilizers per unit area, Q f is the maximum capacity of the fertilizer box;

[0127] Step 2. Obtain the objective function value F(x) before random perturbation and the objective function value F(x′) after random perturbation, and obtain the objective function difference ΔF = F(x′) - F(x);

[0128] The objective function value F(x) before random perturbation is specifically:

[0129]

[0130] Among them, D is the seeding path length before random perturbation, k is the cluster label, K is the total number of clusters, and j k is the label of the small regular field block in the set of regular field blocks formed by the k-th cluster, and m k is the total number of small regular field blocks in the set of regular field blocks formed by the k-th cluster. is the path length of the ox-plowing reciprocating method in the field before random perturbation during seeding, is the path length from the boundary point where seeding ends in the field before random perturbation to the nearest replenishment point;

[0131] The objective function value F(x′) after random perturbation is specifically:

[0132]

[0133] Among them, D' is the seeding path length after random perturbation, is the path length of the ox-plowing reciprocating method in the field after random perturbation during seeding, is the path length from the boundary of the field after random perturbation to the nearest replenishment point;

[0134] Step 24: Determine whether to accept the new solution according to the difference of the objective function. If the new solution is accepted, let x = x', and at the same time update the temperature T = αT according to the cooling schedule; then execute Step 25; if the new solution is not accepted, directly return to Step 22;

[0135] The determination of whether to accept the new solution according to the difference of the objective function is specifically:

[0136] First, obtain the probability P using the difference of the objective function:

[0137]

[0138] Among them, α ∈ (0, 1), and α is the temperature reduction coefficient;

[0139] Then, determine whether to accept the new solution according to the probability P:

[0140] If P = 1, accept the new solution; otherwise, generate a random number, compare the magnitudes of P and the random number. If P is less than or equal to the random number, accept the new solution; if P is greater than the random number, do not accept the new solution;

[0141] Step 25: Determine whether the current temperature has reached the termination temperature T min or whether the current number of iterations has reached the maximum number of iterations N max ; if T ≤ T min or N ≥ N max then output the set of optimal replenishment point positions and the corresponding minimum path length D'; otherwise, let N = N + 1 and return to Step Two.

[0142] Step Three: As Figure 4 shown, use the firefly algorithm to plan the global seeding path to obtain the final seeding path, specifically:

[0143] Step 3-1: According to the search step α f within the search area Ω f randomly initialize the positions of N' f ridge furrows, generate N f seeding paths, and regard each seeding path as a firefly Ρ i ={(x i,1 ,y i,1 ),(x i,2 ,y i,2 ),...,(x i,kf ,y i,kf ),...,(x i,KF ,y i,KF ),i = 1,2,...,N f ;

[0144] Initialize the firefly light intensity L0; initialize the number of iterations t' = 1; set the maximum number of searches T max,firefly = 200;

[0145] where i is the seeding path label, (x i,kf ,y i,kf ) is the position in the kf-th ridge furrow of the i-th path, KF is the total number of positions in the ridge furrows of the i-th seeding path, Ρ i is the i-th firefly; the positions of N f ridge furrows include all the positions of the optimal supplementary points;

[0146] Step 3-2: Obtain the fitness value of each firefly:

[0147] F f (Ρ i ) = ω a ·C comp,f (Ρ i ) + ω b ·D total,f (Ρ i ) - ω c ·Area f (Ρ i )

[0148]

[0149] where C comp,f (Ρ i) is the path Ρ i is the number of compacted ridges, ω a is the weight of the number of compacted ridges, D total,f (Ρ i ) is the path Ρ i total driving distance, ω b is the weight of the driving distance, Area f (Ρ i ) is the path Ρ i sown area covered, ω c is the weight of the sown area, is the small regular plot j k internal path length of the i-th sowing path within, l 压实 is the small regular plot j k compaction degree of the i-th sowing path within, is the small regular plot j k sown area of the i-th sowing path within, is the small regular plot j k boundary path length of the i-th sowing path within;

[0150] Step Three: Obtain the light intensity L of each firefly according to the fitness value of each firefly i :

[0151]

[0152] where γ f is the light intensity absorption coefficient, L0 is the initial light intensity;

[0153] Step Four: Obtain the attractiveness between fireflies:

[0154]

[0155] where β ij,f is the attractiveness of firefly Ρ i to firefly Ρ j , β 0f is the maximum attractiveness between preset fireflies, set to 0.8 in the present invention, d ij,f is the Euclidean distance between firefly Ρ i and firefly Ρ j i.e., the relative light intensity, (x i,kf ,y i,kf ) is the kf-th point in the sowing path corresponding to firefly Ρ i , (x j,kf ,y j,kf ) is the kf-th point in the sowing path corresponding to firefly Ρ j ;

[0156] Step 35. Update the positions of fireflies based on the attractiveness between fireflies:

[0157]

[0158] Among them, is the position of firefly Ρ i at time t', is the position of firefly Ρ i at time t'+1, is the position of firefly Ρ j at time t', α f is the step size factor, rand f is a random number obeying [0,1], ΔΡ chaos,i is the chaotic search perturbation, generated by the Logistic map, r c is the chaotic map control parameter (for the Logistic map), is the chaotic perturbation variable at time t', is the chaotic perturbation variable at time t'+1;

[0159] In this step, the initial value of the chaotic perturbation variable

[0160] Step 36. Select elite fireflies according to the light intensity of fireflies and the updated positions of fireflies, perform chaotic perturbation on the positions of elite fireflies to update the positions of elite fireflies, re-initialize the fireflies except elite fireflies randomly, and return to Step 32;

[0161] Step 361. Sort the positions of fireflies in ascending order according to the light intensity of fireflies, and select the first p elite ·N f firefly individuals as elite fireflies;

[0162] Among them, p elite is the elite group ratio;

[0163] Step 362. Perform chaotic perturbation on the positions of elite fireflies to update the positions of elite fireflies, and then execute Step 37:

[0164]

[0165] Among them, Ρ elite is the elite group ratio, ΔΡ chaos is the chaotic search perturbation, is the position of the elite firefly after update;

[0166] Step 363. Re-initialize the worst (1 - p elite )·N f individuals randomly and return to Step 32;

[0167] Step 37. Dynamically adjust the step size factor α f : κ f ∈(0, 1);

[0168] where κ f is the step size dynamic attenuation coefficient;

[0169] Step 38. If the number of iterations reaches the maximum search times T max,firefly or the change in the fitness function value is less than the preset threshold: then end, and take as the optimal path Ρ best and output; otherwise, let t' = t' + 1, and return to Step 32;

[0170] where ε f is the search precision threshold, T max,firefly is the maximum search times, is the position of the elite firefly in the (t' + 1)-th iteration, is the position of the elite firefly in the t'-th iteration.

Claims

1. A corn planter field path planning method for reducing soil mechanical compaction, characterized in that The specific process of the method is: Step 1: Obtain the ridge and boundary point information of the irregular field, cluster the irregular field, and divide the irregular field into multiple regular fields. Specifically: Step 11: Obtain boundary point information of irregular fields; The boundary point information is an irregular field boundary point set P = {(x1, y1), (x2, y2), ..., (x s ,y s ),...,(x n ,y n )}; Where n is the total number of irregular field boundary points, (x s ,y s ) is the coordinate of the sth boundary point of the irregular field, (x n ,y n ) is the coordinate of the nth boundary point of the irregular field; Step 1 and 2: cluster the boundary point information of the irregular field and divide the irregular field into multiple regular fields; Step 2: Set the location of seed and fertilizer replenishment points in each regular field; Step 3: Plan the global sowing path based on the location of the seed and fertilizer replenishment points to obtain the final sowing path.

2. The method for corn planter field path planning for reducing soil mechanical compaction according to claim 1, characterized in that: In the steps 1 and 2, the boundary point information of the irregular field is used for clustering to divide the irregular field into a plurality of regular fields, specifically: Step 121: Obtain each point p in the irregular field boundary point set according to the preset neighborhood radius ε s (x s ,y s )’s ε-neighborhood N(p s ), specifically: N(p s )={q∈Ρ|||p s -q||≤ε} Among them, s is the number of the collection point of the irregular field boundary points, q is the point p s (x s ,y s )ε-points in the neighborhood, s ranges from 1 to n, n is the total number of irregular field boundary points; Step 122: Using N(p s ) judge p s Is it a core point? If p s As the core point, p s Add the core point set C, and then execute steps 1, 2, and 3; otherwise, set p s Mark as noise point; Step 1, 2, 3, obtain N(p s ) the total number of midpoints, if N(p s ) is greater than the preset threshold, which means that point p s Covered by density, N(p s ) are marked as the same cluster C k ; Otherwise, it means point p s If the point p is not covered by density, s Mark as noise point; Step 124: Each cluster C k Divide into area A s Regular rectangular fields, specifically: Among them, A s is the maximum area of ​​a single sowing by the seeder, S k It is cluster C k The set of all small regular plots in is a k-cluster C k The corresponding j k A small regular plot of land, is the jth k The area of ​​a small regular field, m k It is cluster C k The corresponding number of small regular plots.

3. The method for corn planter field path planning for reducing soil mechanical compaction according to claim 2, characterized in that: The use of N(p s ) judge p s Whether it is a core point, specifically: If N(p s )≥MinPts then p s is the core point; otherwise, p s It’s not the core point; Among them, MinPts is the threshold of the number of points in the minimum neighborhood; The core point set C is specifically: C = {p s |N(p s )≥MinPts}.

4. The method for corn planter field path planning for reducing soil mechanical compaction according to claim 3, characterized in that: The step 2 of setting the seed and fertilizer supplementation point position in each regular field is specifically as follows: Step 21: Initialize parameters and set constraints; The initialization parameters are specifically: Setting the initial solution Initial temperature T = T0, initial objective function value F(x) = D0, initialization iteration number N = 1; in, is the randomly set initial seed and fertilizer supplementation point location set, D0 is the total length of the initial sowing path, T0>1000, T0 is the initial temperature, t is the total number of supplementation points, and p is the supplementation point number; The constraints include: obstacle area constraints, cattle plowing reciprocating path constraints and field switching path constraints; The obstacle zone constraints are: Among them, O s is the obstacle area, X is the set of label numbers of the irregular field boundary points; The reciprocating path constraint of the ox plowing is: Among them, D 切换 is the field switching distance, distance(x s ,y s ,P 起点 ) is the point (x s ,y s ) to the starting point of the next field, P 起点 It is the starting point of the field; Step 2: Randomly perturb the position of the supplementary point based on the current existing supplementary point position to obtain a new solution in, is the set of supplementary point positions after disturbance, δ~U(-Δ,Δ), δ is a random disturbance, R p ′=R p +δ, R p ′ is the supplementary point after random disturbance, R p is the position of the pth supplementary point, -Δ is the lower limit of random perturbation, Δ is the upper limit of random perturbation, U(-Δ,Δ) is the uniform distribution within the range of (-Δ,Δ), (x' p ,y' p ) are the coordinates of the supplementary points after random perturbation; Step 23: Obtain the objective function value F(x) before random disturbance and the objective function value F(x′) after random disturbance, and obtain the objective function difference ΔF=F(x′)-F(x); Step 24: determine whether to accept the new solution based on the difference in the objective function. If the new solution is accepted, set x=x' and update the temperature T=αT according to the cooling schedule; then execute step 25; if the new solution is not accepted, return directly to step 22; The method of judging whether to accept a new solution based on the objective function difference is as follows: First, the probability P is obtained by using the difference of the objective function: Where, α∈(0,1), α is the temperature drop coefficient; Then, determine whether to accept the new solution based on the probability P: If P = 1, accept the new solution. Otherwise, generate a random number and compare P with the random number. If P is less than or equal to the random number, accept the new solution. If P is greater than the random number, do not accept the new solution. Step 25: Determine whether the current temperature reaches the end temperature T min Or whether the current number of iterations has reached the maximum number of iterations N max ; If T≤T min Or N ≥ N max Then output the optimal supplementary point location set and the corresponding minimum path length D'; otherwise, set N=N+1 and return to step 22.

5. The method for corn planter field path planning for reducing soil mechanical compaction according to claim 4, characterized in that: The constraints of the supplementary points are as follows: in, Small regular plots The boundary point set, (x b ,y b ) is a small regular plot The coordinates of the boundary points in, b is a small regular field The boundary point number, Small regular plots The total number of boundary points in .

6. The method for corn planter field path planning for reducing soil mechanical compaction according to claim 5, characterized in that: The objective function value F(x) before random disturbance and the objective function value F(x′) after random disturbance in the second and third steps are obtained as follows: Where D is the seeding path length before random perturbation, k is the cluster number, K is the total number of clusters, and j k is the number of the small regular plot in the regular plot set formed by the kth cluster, m k is the total number of small regular plots in the regular plot set formed by the kth cluster, is the reciprocating ox-ploughing method on the field before random disturbance The path length for seeding, is the random perturbation of the field block The path length from the boundary point where sowing ends to the nearest supplementary point; D' is the length of the sowing path after random disturbance, is the reciprocating ox-ploughing method on the field after random disturbance The path length for seeding, is the field after random disturbance The path length from the boundary to the nearest supplementary point.

7. The method for corn planter field path planning for reducing soil mechanical compaction according to claim 6, characterized in that: The global sowing path is planned based on the seed and fertilizer supplementation point positions in step 3 to obtain the final sowing path, specifically: Step 31: According to the search step α f In the search area Ω f Inside, randomly initialize N' f The positions of the ridges generate N f There are sowing paths, each sowing path is a firefly i : P i ={(x i,1 ,and i,1 ),(x i,2 ,and i,2 ),...,(x i,kf ,and i,kf ),...,(x i,KF ,and i,KF )},i=1,2,...,N f Initialize the firefly light intensity L0; initialize the number of iterations t'=1; Among them, i is the seeding path number, (x i,kf ,y i,kf ) is the kfth position in the ith path, KF is the total number of positions in the ith seeding path, P i is the i-th firefly; N f The positions include all the optimal supplementary point positions; Step 32: Get the fitness value F of each firefly f (P i ); Step 3. Obtain the light intensity L of each firefly according to its fitness value i , specifically: Among them, γ f is the light intensity absorption coefficient, L0 is the initial light intensity; Step 34: Get the attraction between fireflies, specifically: Among them, β ij,f It's Firefly i Firefly j The attractiveness, β 0f is the preset value of the maximum attraction between fireflies, d ij,f It's Firefly i With Firefly j Relative light intensity between i,kf ,y i,kf ) is a firefly i The kfth point in the corresponding sowing path, (x j,kf ,y j,kf ) is a firefly j The kfth point in the corresponding sowing path; Step 35: Update the firefly position using the attraction between fireflies; Step 36: Select elite fireflies according to the light intensity of the fireflies and the updated positions of the fireflies, and perform chaotic perturbation on the positions of the elite fireflies to update the positions of the elite fireflies, and then execute step 37; re-randomly initialize the fireflies except the elite fireflies, and return to step 32; Step 37: Update the step size factor, specifically: Among them, κ f ∈(0,1),κ f is the step-size dynamic attenuation coefficient; Step 38: If the number of iterations reaches the maximum search number T max,firefly Or the fitness function value changes less than the preset threshold, As the optimal path best Output; otherwise, set t'=t'+1 and return to step 32; The change in fitness function value is less than the preset threshold: Among them, ε f is the search precision threshold, is the position of the elite firefly in the t'+1th iteration, is the position of the elite firefly in the t'th iteration, yes The fitness value of yes The fitness value of .

8. The method for corn planter field path planning for reducing soil mechanical compaction according to claim 7, characterized in that: The fitness value F of each firefly is obtained in step 32. f (P i ), specifically: Among them, C comp,f (P i ) is the path P i The number of compacted ridges, ω a is the weight of the number of compacted ridges, D total,f (P i ) is the path P i Total distance traveled, ω b is the driving distance weight, Area f (P i ) is the path P i The sown area covered, ω c is the sown area weight, It is a small regular field j k The internal path length of the i-th sowing path, l 压实 It is a small regular field j k The compaction degree of the i-th sowing path, It is a small regular field j k The sowing area of ​​the i-th sowing path, It is a small regular field j k The boundary path length of the i-th seeding path.

9. The method for corn planter field path planning for reducing soil mechanical compaction according to claim 8, characterized in that: The method of updating the firefly positions by using the attraction between fireflies in step 35 is as follows: in, It's time for fireflies i location, is the firefly P at time t'+1 i location, It's time for fireflies j The position of f is the step size factor, rand f is a random number in [0,1], ΔΡ chaos,i is the chaotic search perturbation, r c is the chaotic map control parameter, is the chaotic disturbance variable at time t', is the chaotic disturbance variable at time t'+1.

10. The method for corn planter field path planning for reducing soil mechanical compaction according to claim 9, characterized in that: In step 36, elite fireflies are selected according to the light intensity of the fireflies and the updated positions of the fireflies, and the positions of the elite fireflies are chaotically disturbed to update the positions of the elite fireflies, and then step 37 is executed; the fireflies other than the elite fireflies are randomly initialized again, and the process returns to step 32, specifically: Step 361: Sort the positions of the fireflies in ascending order according to their light intensity, and select the first p elite ·N f Firefly as an elite firefly; Among them, p elite is the proportion of elite groups; Step 362: Perform chaotic perturbation on the position of the elite firefly to update the position of the elite firefly, and then execute step 37: Among them, elite is the proportion of the elite group, ΔΡ chaos,i is the chaotic search perturbation, It is the location of the elite fireflies after the update; Step 363: (1-p elite )·N f The fireflies that are not elite fireflies are randomly initialized again and return to step 32.

Images