Efficient and accurate pleurotus eryngii automatic root cutting and shaping method

Abstract

The invention discloses an efficient and accurate pleurotus eryngii automatic root cutting and shaping method. Belongs to the field of agricultural automation equipment. Contour extraction and 3D representation; defining a parameter space and searching a strategy; root cutting simulation and evaluation; carrying out parameter optimization solution; performing multi-cycle search; and verifying a parameter optimization result. The method has the characteristics that the root cutting precision is improved: the shape of the root of the pleurotus eryngii is accurately adapted through multi-view image acquisition and a complex parameter optimization algorithm, so that the root subjected to root cutting reaches relatively high precision; (2) the production efficiency is improved, the time and workload of manual operation are reduced through the automatic root cutting process, and meanwhile the optimal root cutting parameters can be rapidly determined; (3) the flexibility is enhanced: root cutting parameters can be automatically adjusted according to the actual forms of different pleurotus eryngii samples to adapt to the processing of pleurotus eryngii of various specifications; and (4) the product quality is stabilized, the quality consistency of each batch of products is guaranteed through accurate root cutting and shape modification, and standardized production is facilitated.

Description

Technical Field

[0001] The present invention belongs to the field of agricultural automation equipment, and is particularly related to the automation technology of root trimming and shape modification during the processing of Pleurotus eryngii. Specifically, it relates to an efficient and precise method for automatic root trimming and shape modification of Pleurotus eryngii. Background Art

[0002] During the production and processing of Pleurotus eryngii, root trimming and shape modification is an important step. At present, most factories still use manual methods for root trimming, which requires a large amount of labor input, and it is difficult to ensure the accuracy and consistency of root trimming by manual operation. Although some enterprises have tried to use automated equipment, there are many problems with the existing automated root trimming technologies. For example, the control and adjustment of the cutting tools in the existing automated root trimming equipment are not flexible enough. The angles and depths of the cutting tools are often fixed or can only be adjusted within a limited range, and they cannot adapt to the roots of Pleurotus eryngii with different shapes and sizes. Moreover, the inclination angle of the cutting tool spindle of some equipment cannot be changed within a large range according to actual needs, resulting in an inability to achieve an ideal root trimming effect when processing Pleurotus eryngii with various root shapes.

[0003] Specifically: (1) Insufficient accuracy: Due to the inability to precisely adapt to the morphological changes of the roots of Pleurotus eryngii, it is difficult to meet the high-precision requirements for indicators such as the flatness of the roots, cross-sectional angles, and roundness after root trimming. For example, there may be a large height difference in the flatness of the roots, affecting the appearance and quality of the product; (2) Low efficiency: Whether it is manual operation or existing automated equipment, it is difficult to achieve efficient processing while ensuring quality. Manual operation is slow, and the overall processing efficiency of existing equipment is not high because of frequent adjustments or the inability to effectively process Pleurotus eryngii with different morphologies; (3) Poor flexibility: It cannot quickly adjust the root trimming parameters according to different Pleurotus eryngii samples, and requires manual intervention or replacement of components such as cutting tools, increasing the operation complexity and cost; (4) Unstable product quality: Due to poor root trimming accuracy and consistency, the quality of each batch of products fluctuates greatly, which is not conducive to standardized production and market promotion. Summary of the Invention

[0004] In view of the above problems, the object of the present invention is to provide a method that can automatically optimize the root trimming parameters according to the actual morphology of Pleurotus eryngii, realizing efficient and precise automatic root trimming and shape modification, and improving product quality and production efficiency.

[0005] The technical solution of the present invention is as follows: A method for efficient and precise automatic root trimming and shape modification of Pleurotus eryngii according to the present invention, and its operation steps are as follows:

[0006] Step (1): Image acquisition and preprocessing;

[0007] Step (2): Contour extraction and 3D representation;

[0008] Step (3): Parameter space definition and search strategy;

[0009] Step (4): Root trimming simulation and evaluation;

[0010] Step (5): Parameter optimization solution;

[0011] Step (6): Multiple loop search;

[0012] Step (7): Verification of parameter optimization results.

[0013] Furthermore, in step (1), the specific steps of the image acquisition and preprocessing are as follows:

[0014] (1.1): Obtain a set of Pleurotus eryngii side pictures I = {I 1 , I 2 ,..., I n} at n different angles, where n ≥ 6 and is an integer;

[0015] (1.2): Preprocess each picture I i ∈ I to obtain a set of preprocessed pictures P = {P 1 , P 2 ,..., P n}.

[0016] 3. An efficient and accurate automatic root trimming and shaping method for Pleurotus eryngii according to claim 1, characterized in that, in step (2), the specific steps of the contour extraction and 3D representation are as follows:

[0017] (1.1): Perform contour detection on each preprocessed picture P i ∈ P to obtain a set of 2n 2D contour lines L = {L 1 , L 2 ,..., L 2n};

[0018] (1.2): Construct a three-dimensional representation M of Pleurotus eryngii based on L, where M consists of a vertex set V and a face set F.

[0019] Furthermore, in step (3), the specific parameter space definition and search strategy are as follows:

[0020] Define the root trimming parameter space Ω = C × A × D × R, including the following parameters:

[0021] I. The number of root trimming times C of the tool:

[0022] Initial value: C 0 = 3;

[0023] Search range: [Cmin, Cmax];

[0024] Progressive step: ΔC;

[0025] II. Tilt angle A of the tool spindle:

[0026] Initial value: A 0 = 90°;

[0027] Search range: [Amin, Amax];

[0028] Progressive step: ΔA;

[0029] III. Depth D of the tool spindle:

[0030] Initial value: D 0 = 5 mm;

[0031] Search range: [Dmin, Dmax];

[0032] Progressive step: ΔD;

[0033] IV. Rotation angle R of the Pleurotus eryngii:

[0034] Initial value: R 0 = 0°;

[0035] Search range: (Rmin, Rmax);

[0036] Progressive step: ΔR.

[0037] Furthermore, in step (4), the root trimming simulation and evaluation are specifically as follows:

[0038] For the parameter combination ω = (c, a, d, r), perform the following operations:

[0039] I. Apply the transformation T(ω) to the 3D model M to simulate the root trimming process;

[0040] II. Calculate the evaluation function E(M'(ω)), where M'(ω) is the model after root trimming;

[0041] The evaluation function is defined as: E(M') = w 1 S(M') + w 2 V(M') - w 3 D(M, M')

[0042] Among them: The calculation of the surface smoothness S(M'):

[0043] In the formula, F' is the face set of the model after root trimming, n i is the normal vector of face i, n m is the normal vector of the ideal root trimming surface;

[0044] Calculation of volume retention rate V(M'): V(M') = Vol(M') / Vol(M), where Vol() represents the volume calculation function of the 3D model and is defined as:

[0045] Vol(M) = |(1 / 6)∑(xi1(yi2zi3 - yi3zi2) + xi2(yi3zi1 - yi1zi3) + xi3(yi1zi2 - yi2zi1))|

[0046] Calculation of model difference degree D(M, M'):

[0047] In the formula, V and V' are the vertex sets before and after root trimming respectively; where |V| is the total number of vertices of the model before root trimming, v i is the i-th vertex in the vertex set V of the model before root trimming, V' is the vertex set of the model after root trimming, ||v i - v'|| represents the Euclidean distance, that is, √[(x i - x') 2 + (y i - y') 2 + (z i - z') 2 , min{||v i - v'||: v' ∈ V'} represents the minimum distance from vertex v i to all vertices in the vertex set V' of the model after root trimming;

[0048] w 1 、w 2 、w 3 are weight coefficients, satisfying w 1 + w 2 + w 3 = 1 and w i > 0 (i = 1, 2, 3).

[0049] Furthermore, in step (5), the parameter optimization solution is specifically:

[0050] Adopt an iterative optimization strategy to solve the parameters:

[0051] I. Initialize the parameter combination ω 0 = (c 0 , a 0 , d 0 , r 0 );

[0052] II. Calculate the gradient of the current point:

[0053] III. Adaptive step size update strategy: η k = η β (1 - βk ) / (1 + γ k );

[0054] In the formula:

[0055] η β is the basic step size, with an initial value of 0.1;

[0056] β k = |E(ω k ) - E(ω k-1 )| / |E(ω k-1 )| is the relative improvement rate;

[0057] is the gradient change rate;

[0058] IV. Update parameters:

[0059] V. Calculate the parameter update amount: Δω k = ||ω k+1 - ω k ||;

[0060] VI. Termination condition judgment:

[0061] If Δω k < ε 1 ;

[0062] or |E(ω k+1 ) - E(ω k )| < ε 2 ;

[0063] or the maximum number of iterations K is reached max then terminate the optimization; otherwise, return to step (2).

[0064] Furthermore, in step (6), the multi-loop search means that: during the entire optimization process, perform the root trimming simulation and evaluation steps for each parameter combination ω = (c, a, d, r);

[0065] Specifically, apply the transformation T(ω) to the 3D model M to simulate the root trimming process, and evaluate the trimmed model M'(ω) through the evaluation function E(M'(ω));

[0066] In the evaluation function, comprehensively consider the surface smoothness S(M'), the volume retention rate V(M'), and the model difference degree D(M, M'), and evaluate the root trimming effect from multiple dimensions through different calculation methods (such as S(M') = 1 - (1 / |F'|), V(M') = Vol(M') / Vol(M), D(M, M') = (1 / |V|)).

[0067] Further, in step (7), the verification of the parameter optimization result is specifically as follows:

[0068] First, apply the optimal parameter combination ω to the actual Pleurotus eryngii samples;

[0069] Second, collect the images of the Pleurotus eryngii after root trimming and compare them with the ideal model M;

[0070] Among them, the ideal model M is defined as:

[0071] Root flatness: The height difference between any two points ≤ 0.5 mm;

[0072] Root cross-section angle: The angle with the vertical plane ≤ 5°;

[0073] Remaining mushroom body length: Error ≤ 1 mm;

[0074] Root roundness: The difference between the maximum radius and the minimum radius ≤ 1 mm;

[0075] Root trimming area ratio: Actual root trimming area / Theoretical root trimming area ∈ [0.95, 1.05];

[0076] Third, calculate the deviation δ between the actual root trimming effect and the expected effect:

[0077] δ = ∑w i |M′ - M| / σ i

[0078] In the formula: M′ is the measured value of the i-th evaluation index;

[0079] M is the ideal value of the i-th evaluation index;

[0080] σ i is the allowable deviation of the i-th evaluation index;

[0081] w i is the weight coefficient, Σw i = 1;

[0082] Fourth, if δ exceeds the preset threshold ε, then return to step (5) for parameter fine-tuning,

[0083] The fine-tuning strategy is: If the deviation of a certain index exceeds 2σ i , then focus on optimizing the corresponding parameter;

[0084] Adopt the bisection method to conduct a fine search in the neighborhood of this parameter;

[0085] Re-evaluate all indexes after each fine-tuning.

[0086] The beneficial effects of the present invention are as follows: 1. Improve the root trimming accuracy: Through multi-perspective image acquisition and complex parameter optimization algorithms, it can accurately adapt to the morphology of the Pleurotus eryngii roots, enabling indicators such as the flatness of the trimmed roots, the cross-section angle, and the roundness to reach a high level of accuracy. For example, the root flatness can be controlled within 0.5 mm, the angle between the root cross-section and the vertical plane is within 5°, the difference between the maximum and minimum radii of the root roundness is within 1 mm, and the root trimming area ratio is between [0.95, 1.05]; 2. Improve production efficiency: The automated root trimming process reduces the time and workload of manual operations. At the same time, the efficient algorithm can quickly determine the optimal root trimming parameters, shortening the processing time of a single Pleurotus eryngii and overall improving production efficiency; 3. Enhance flexibility: It can automatically adjust the root trimming parameters according to the actual morphology of different Pleurotus eryngii samples, without the need to frequently change tools or perform complex manual interventions, and is suitable for processing Pleurotus eryngii of various specifications; 4. Stabilize product quality: Precise root trimming and shaping ensure the quality consistency of each batch of products, which is conducive to standardized production and market promotion, and improves the market competitiveness of products. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] Figure 1 is the overall operation block diagram of the present invention;

[0088] Figure 2 is the algorithm flowchart in the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0089] The following further elaborates on the specific technical solutions of the present invention with specific examples.

[0090] As shown in the figure, an efficient and precise automated root trimming and shaping method for Pleurotus eryngii according to the present invention has the following operating steps:

[0091] S1. Image acquisition and preprocessing:

[0092] (1). Obtain a set of side images of Pleurotus eryngii at n different angles I = {I 1 , I 2 ,..., I n}, where n ≥ 6 and is an integer;

[0093] (2). Preprocess each image I i ∈ I to obtain the preprocessed image set P = {P 1 , P 2 ,..., P n};

[0094] S2. Contour extraction and 3D representation:

[0095] (1). For each preprocessed image P i∈P performs contour detection to obtain a set L of 2n 2D (two-dimensional) contour lines, where L = {L 1 , L 2 ,..., L 2n};

[0096] (2) Based on L, construct a three-dimensional representation M of Pleurotus eryngii, where M consists of a vertex set V and a face set F;

[0097] S3. Parameter Space Definition and Search Strategy:

[0098] Define the root trimming parameter space Ω = C × A × D × R, which includes the following parameters:

[0099] (1) The number of times C of the tool for root trimming:

[0100] Initial value: C 0 = 3 (determined based on the ratio of the thickness removed by a single root trimming to the total root thickness);

[0101] Search range: [Cmin, Cmax];

[0102] Step size for progression: ΔC;

[0103] (2) The inclination angle A of the tool spindle:

[0104] Initial value: A 0 = 90° (based on the statistical average of the angle between the root and the cap of Pleurotus eryngii);

[0105] Search range: [Amin, Amax];

[0106] Step size for progression: ΔA;

[0107] (3) The depth D of the tool spindle:

[0108] Initial value: D 0 = 5 mm (based on 80% of the average root thickness of Pleurotus eryngii);

[0109] Search range: [Dmin, Dmax];

[0110] Step size for progression: ΔD;

[0111] (4) The rotation angle R of Pleurotus eryngii:

[0112] Initial value: R 0 = 0° (with the main axis of Pleurotus eryngii as the reference);

[0113] Search range: (Rmin, Rmax);

[0114] Step size for progression: ΔR;

[0115] S4. Root Trimming Simulation and Evaluation:

[0116] For the parameter combination ω = (c, a, d, r), perform the following operations:

[0117] (1) Apply the transformation T(ω) to the 3D model M to simulate the root trimming process;

[0118] (2) Calculate the evaluation function E(M'(ω)), where M'(ω) is the model after root trimming;

[0119] The evaluation function is defined as: E(M') = w 1 S(M') + w 2 V(M') - w 3 D(M, M');

[0120] Where:

[0121] Calculation of the surface smoothness S(M'):

[0122] In the formula, F' is the face set of the model after root trimming, n i is the normal vector of face i, n m is the normal vector of the ideal root trimming surface;

[0123] Calculation of the volume retention rate V(M'): V(M') = Vol(M') / Vol(M), where Vol() (volume) represents the volume calculation function of the 3D (three-dimensional) model and is defined as:

[0124] Vol(M) = |(1 / 6)∑(xi1(yi2zi3 - yi3zi2) + xi2(yi3zi1 - yi1zi3) + xi3(yi1zi2 - yi2zi1))|

[0125] Calculation of the model difference degree D(M, M'):

[0126] In the formula, V and V' are the vertex sets before and after root trimming respectively; where |V| is the total number of vertices of the model before root trimming, v i is the i-th vertex in the vertex set V of the model before root trimming, V' is the vertex set of the model after root trimming, ||v i - v'|| represents the Euclidean distance, that is, √[(x i - x') 2 + (y i - y') 2 + (z i - z') 2 , min{||v i - v'||: v' ∈ V'} represents the minimum distance from vertex v i to all vertices in the vertex set V' of the model after root trimming;

[0127] w 1 , w 2 , w 3 are weight coefficients, satisfying w 1 + w 2 + w 3 = 1 and w i > 0 (i = 1, 2, 3);

[0128] S5. Parameter optimization and solution:

[0129] Use the iterative optimization strategy to solve the parameters:

[0130] (1). Initialize the parameter combination ω 0 = (c 0 , a 0 , d 0 , r 0 );

[0131] (2). Calculate the gradient at the current point:

[0132] (3). Adaptive step size update strategy: η k = η β (1 - β k ) / (1 + γ k );

[0133] In the formula: η β is the base step size, with an initial value of 0.1;

[0134] β k = |E(ω k ) - E(ω k-1 )| / |E(ω k-1 )| is the relative improvement rate;

[0135] is the gradient change rate; Gradient; ω: Parameter combination;

[0136] (4). Update the parameters:

[0137] (5). Calculate the parameter update amount: Δω k = ||ω k+1 - ω k ||;

[0138] (6). Judgment of termination conditions:

[0139] If Δω k < ε 1 (Parameter change threshold);

[0140] Or |E(ωk+1 ) - E(ω k ) | < ε 2 (Threshold for improving the objective function);

[0141] or reach the maximum number of iterations K max Then terminate the optimization; otherwise return to step (2);

[0142] S6. Multi - loop search:

[0143] During the entire optimization process, perform the root - cutting simulation and evaluation steps for each parameter combination ω = (c, a, d, r);

[0144] Specifically, apply the transformation T(ω) to the 3D model M to simulate the root - cutting process, and evaluate the root - cut model M'(ω) through the evaluation function E(M'(ω)); in the evaluation function, multiple aspects such as surface smoothness S(M'), volume retention rate V(M'), and model difference degree D(M, M') are comprehensively considered. Through different calculation methods (such as S(M') = 1 - (1 / |F'|), V(M') = Vol(M') / Vol(M), D(M, M') = (1 / |V|)), S(M'): surface smoothness; V(M'): volume retention rate; D(M, M'): model difference degree;

[0145] Evaluate the root - cutting effect from multiple dimensions;

[0146] S7. Verification of parameter optimization results:

[0147] (1). Apply the optimal parameter combination ω to the actual Pleurotus eryngii samples;

[0148] (2). Collect the images of the root - cut Pleurotus eryngii and compare them with the ideal model M;

[0149] Among them, the ideal model M is defined as:

[0150] Flatness of the root: The height difference between any two points ≤ 0.5 mm;

[0151] Angle of the root cross - section: The angle with the vertical plane ≤ 5°;

[0152] Retained mushroom body length: Error ≤ 1 mm;

[0153] Roundness of the root: The difference between the maximum radius and the minimum radius ≤ 1 mm;

[0154] Ratio of root - cutting area: Actual root - cutting area / Theoretical root - cutting area ∈ [0.95, 1.05];

[0155] (3). Calculate the deviation δ between the actual root - cutting effect and the expected effect:

[0156] δ = ∑wi |M′ - M| / σ i

[0157] Where: M′ is the measured value of the i-th evaluation index;

[0158] M is the ideal value of the i-th evaluation index;

[0159] σ i is the allowable deviation of the i-th evaluation index;

[0160] w i is the weight coefficient, Σw i = 1;

[0161] (4), If δ exceeds the preset threshold ε, return to step S5 for parameter fine-tuning. The fine-tuning strategy is:

[0162] If the deviation of a certain index exceeds 2σ i , then focus on optimizing the corresponding parameter;

[0163] Adopt the bisection method to perform a fine search in the neighborhood of this parameter;

[0164] Re-evaluate all indexes after each fine-tuning.

[0165] Embodiment

[0166] 1. In actual production, first place the Pleurotus eryngii on the instantaneous rotation platform, start the image acquisition system, control the rotation platform to rotate at a predetermined angular interval, and the industrial camera synchronously acquires side pictures of the Pleurotus eryngii at different angles. For example, one picture is acquired every 60°, and more than 6 pictures are acquired in total;

[0167] 2. Preprocess the acquired pictures, including operations such as removing image noise, converting the color image to a grayscale image, and enhancing the edge contour, to obtain clear preprocessed pictures;

[0168] 3. Use the image processing algorithm to perform contour detection on the preprocessed pictures, extract the 2D contour lines, and construct a three-dimensional model of the Pleurotus eryngii based on these contour lines;

[0169] 4. According to the defined root-trimming parameter space and initial parameter values, start parameter optimization. By simulating the root-trimming process, calculate the evaluation function, and continuously adjust the parameter combination until the termination condition is met to obtain the optimal root-trimming parameters;

[0170] 5. Apply the optimal parameters to the actual root-trimming operation. The adjustable tool system adjusts the tilt angle and depth of the tool spindle according to the parameters. At the same time, the rotation platform rotates the Pleurotus eryngii at a specified angle to complete the root-trimming and shaping process.

[0171] After root trimming is completed, collect the images of Pleurotus eryngii after root trimming, compare them with the ideal model, and calculate the deviation. If the deviation exceeds the preset threshold, re-optimize the parameters according to the fine-tuning strategy to ensure that the root trimming effect meets the requirements; repeat the above process to achieve efficient, precise, and automated root trimming and shaping of a batch of Pleurotus eryngii.

Claims

1. An efficient and accurate method for automated root pruning and shaping of Pleurotus eryngii, characterized in that: The operation steps are as follows: Step (1): image acquisition and preprocessing; Step (2): contour extraction and 3D representation; Step (3): parameter space definition and search strategy; Step (4): root cutting simulation and evaluation; Step (5): parameter optimization solution; Step (6): multiple loop search; Step (7): Verification of parameter optimization results.

2. The method for efficient and accurate automatic root pruning and shaping of Pleurotus eryngii according to claim 1, characterized in that: In step (1), the specific steps of image acquisition and preprocessing are: (1.1): Get a set of n side images of Pleurotus eryngii from different angles I = {I1, I2, ..., I n }, where n ≥ 6 and is an integer; (1.2): For each image I i ∈I for preprocessing, and obtain the preprocessed picture set P = {P1,P2,...,P n }.

3. The efficient and accurate automatic root pruning and shaping method of Pleurotus eryngii according to claim 1, characterized in that: In step (2), the specific steps of contour extraction and 3D representation are: (1.1): For each preprocessed image P i ∈P to perform contour detection and obtain a set of 2n 2D contour lines L = {L1,L2,...,L 2n }; (1.2): Based on L, construct the three-dimensional representation M of Pleurotus eryngii, where M consists of a vertex set V and a face set F.

4. The efficient and accurate automatic root pruning and shaping method of Pleurotus eryngii according to claim 1, characterized in that: In step (3), the parameter space definition and search strategy are specifically: Define the root cutting parameter space Ω = C × A × D × R, including the following parameters: the number of root cutting times of the tool C, the inclination angle of the tool spindle A, the depth of the tool spindle D, and the rotation angle R of the Pleurotus eryngii.

5. The efficient and accurate automatic root pruning and shaping method of Pleurotus eryngii according to claim 4, characterized in that: The parameters are specifically:

1. Tool root cutting times C: Initial value: C0=3; Search range: [Cmin,Cmax]; Progressive step length: ΔC; 2. Tool spindle tilt angle A: Initial value: A0 = 90°; Search range: [Amin,Amax]; Progressive step length: ΔA; 3. Tool spindle depth D: Initial value: D0 = 5 mm; Search range: [Dmin,Dmax]; Progressive step length: ΔD; 4. Pleurotus eryngii rotation angle R: Initial value: R0 = 0°; Search range: (Rmin, Rmax); Progressive step length: ΔR.

6. The efficient and accurate automatic root pruning and shaping method of Pleurotus eryngii according to claim 1, characterized in that: In step (4), the root pruning simulation and evaluation specifically includes: For the parameter combination ω=(c,a,d,r), do the following: First, apply the transformation T(ω) to the 3D model M to simulate the root shaving process; 2. Calculate the evaluation function E(M'(ω)), where M'(ω) is the model after root pruning; The evaluation function is defined as: E(M') = w1S(M') + w2V(M') - w3D(M, M') Where: Calculation of surface smoothness S(M'): Where F' is the face set of the model after root pruning, n i is the normal vector of face i, n m is the normal vector of the ideal root cutting surface; Calculation of volume retention rate V(M'): V(M') = Vol(M') / Vol(M), where Vol() represents the volume calculation function of the 3D model, defined as: Vol(M)=|(1 / 6)∑(xi1(yi2zi3-yi3zi2)+xi2(yi3zi1-yi1zi3)+xi3(yi1zi2-yi2zi1))| Calculation of model difference D(M,M'): Where V and V' are the vertex sets before and after root pruning, respectively; |V| is the total number of vertices of the model before root pruning, and v i is the i-th vertex in the vertex set V of the model before pruning, V' is the vertex set of the model after pruning, ||v i -v'|| represents the Euclidean distance, that is, √[(x i -x')2+(y i -y')2+(z i -z') 2 ],min{||v i -v'||:v'∈V'} represents vertex v i The minimum distance to all vertices in the vertex set V' of the model after root pruning; w1, w2, w3 are weight coefficients, satisfying w1+w2+w3=1 and w i >0(i=1,2,3).

7. The efficient and accurate automatic root pruning and shaping method of Pleurotus eryngii according to claim 1, characterized in that: In step (5), the parameter optimization solution is specifically: Adopt iterative optimization strategy to solve the parameters:

1. Initialize the parameter combination ω0 = (c0, a0, d0, r0); 2. Calculate the gradient of the current point:

3. Adaptive step size update strategy: η k =η β (1-β k ) / (1+γ k ); Where: η β is the basic step size, with an initial value of 0.1; β k =|E(ω k )-E(ω k-1 )| / |E(ω k-1 )| is the relative improvement rate; is the gradient change rate; 4. Update parameters:

5. Calculate the parameter update amount: Δω k =||ω k+1 -ω k ||; 6. Termination condition judgment: I see. k <e1; or |E(ω k+1 ) - E(ω k )| < ε2; Or reach the maximum number of iterations K max Then the optimization is terminated; otherwise, return to step (2).

8. The efficient and accurate automatic root pruning and shaping method of Pleurotus eryngii according to claim 1, characterized in that: In step (6), the multiple loop search means that during the entire optimization process, a root pruning simulation and evaluation step is performed for each parameter combination ω=(c, a, d, r).

9. The efficient and accurate automatic root pruning and shaping method of Pleurotus eryngii according to claim 8, characterized in that: Specifically, the transformation T(ω) is applied to the 3D model M to simulate the root shaving process, and the root shaving model M'(ω) is evaluated by the evaluation function E(M'(ω)); In the evaluation function, the surface smoothness S(M'), volume retention rate V(M') and model difference D(M,M') are comprehensively considered, and the root pruning effect is evaluated from multiple dimensions through different calculation methods (such as S(M') = 1-(1 / |F'|), V(M') = Vol(M') / Vol(M), D(M,M') = (1 / |V|)).

10. The efficient and accurate automatic root pruning and shaping method of Pleurotus eryngii according to claim 1, characterized in that: In step (7), the parameter optimization result verification is specifically:

1. Apply the optimal parameter combination ω to the actual Pleurotus eryngii samples; 2. Collect the image of Pleurotus eryngii after root pruning and compare it with the ideal model M; Among them, the ideal model M is defined as: Root flatness: height difference between any two points ≤ 0.5mm; Root section angle: Angle with vertical plane ≤5°; Retain mushroom body length: error ≤ 1mm; Root roundness: the difference between the maximum radius and the minimum radius is ≤1mm; Root cutting area ratio: actual root cutting area / theoretical root cutting area∈[0.95,1.05]; 3. Calculate the deviation δ between the actual root pruning effect and the expected effect: δ=∑w i |M′-M| / σ i Where: M′ is the measured value of the i-th evaluation index; M is the ideal value of the i-th evaluation index; σ i is the allowable deviation of the i-th evaluation index; w i is the weight coefficient, Σw i =1; 4. If δ exceeds the preset threshold ε, return to step (5) to fine-tune the parameters. The fine-tuning strategy is: if the deviation of an indicator exceeds 2σ i , then focus on optimizing the corresponding parameters; A binary search is performed in the parameter neighborhood; Re-evaluate all metrics after each fine-tuning.

Images