A wheelbase optimization control method for seedling pruning machine based on machine learning

Abstract

The present invention discloses a wheelbase optimization control method for a seedling pruning machine based on machine learning, which relates to the field of artificial intelligence technology. The method collects basic machine parameters and seedling environmental parameters; performs formal extraction on the seedling environmental parameters to obtain a captured seedling signal sequence, and converts the sequence into a seedling characteristic coefficient diagram; screens and assimilates the basic machine parameters to obtain a wheelbase regulation coefficient; constructs a seedling twin space for virtual generation to obtain twin pruned seedlings and a virtual pruning machine; optimizes seedling pruning for the twin pruned seedlings according to the virtual pruning machine to obtain an optimal pruning wheelbase; performs virtual driving on the virtual pruning machine; and optimizes the distance transportation during the virtual driving process to obtain an optimal driving wheelbase; enhances the stability of the seedling pruning machine during driving, and improves the pruning efficiency and pruning quality of the pruning machine.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a seedling pruning machine wheelbase optimization control method based on machine learning. Background Art

[0002] Sapling trimmers play an important role in forestry, horticulture, and agriculture. They help workers efficiently complete sapling pruning tasks, improve efficiency, ensure pruning consistency, and reduce labor intensity and injury risks. However, existing sapling trimmers have numerous problems. Some have fixed wheelbases, making them incapable of adapting to varying terrain and sapling growth conditions. Others, while equipped with wheelbase adjustment, lack intelligent and efficient methods. Traditional adjustment methods often rely on manual experience, making it difficult to accurately and quickly adjust the wheelbase based on real-time changes in terrain, sapling density, height, and other factors.

[0003] By utilizing the powerful data processing and analysis capabilities of machine learning and real-time perception of the environment and seedling information, not only can the wheelbase be automatically and accurately adjusted to match the pruning object, but the driving wheelbase can also be intelligently controlled during driving, thereby improving the adaptability, flexibility and work efficiency of the pruning machine and better meeting the needs of modern seedling pruning operations. To this end, a wheelbase optimization control method for a seedling pruning machine based on machine learning is provided. Summary of the Invention

[0004] The purpose of the present invention can be achieved through the following technical solutions:

[0005] A method for optimizing wheelbase control of a seedling pruning machine based on machine learning includes the following steps:

[0006] Step S1: collecting machine basic parameters and seedling environmental parameters;

[0007] Step S2: changing the form of the seedling environmental parameters to obtain the seedling environmental signal, setting the resource stability coefficient to perform coefficient transformation to obtain the stage resource coefficient, performing stage extraction on the seedling environmental signal according to the stage resource coefficient, and obtaining the captured seedling signal sequence;

[0008] Step S3: Perform feature conversion on the captured seedling signal sequence to obtain a seedling characteristic coefficient map, and screen and assimilate the machine basic parameters through the stage seedling characteristic coefficients to obtain the wheelbase control coefficient;

[0009] Step S4: Construct a seedling twin space for virtual generation, obtain twin pruned seedlings and virtual pruners, optimize the pruned seedlings according to the virtual pruners, obtain the optimal pruned wheelbase, virtually drive the virtual pruners, and optimize the distance transportation of the virtual driving process to obtain the optimal driving wheelbase.

[0010] Preferably, the process of collecting machine basic parameters and seedling environmental parameters includes:

[0011] Collect data on the seedling pruning machine to obtain basic machine parameters;

[0012] A seedling collection terminal is set up to conduct comprehensive collection of the pruning area through the seedling collection terminal to obtain seedling environmental parameters.

[0013] Preferably, the process of setting the resource stability coefficient and performing coefficient transformation includes:

[0014] Perform local transformation on the resource stability coefficient to obtain local variables;

[0015] Perform statistical decomposition of resource stability coefficients according to local variables to obtain local decomposition parameters;

[0016] The resource stability coefficient is decomposed hierarchically according to the local decomposition parameters to obtain the stage resource coefficient.

[0017] Preferably, the process of extracting the seedling environment signal in stages according to the stage resource coefficient includes:

[0018] Decompose and match the seedling environmental signals according to the stage resource coefficients to obtain the seedling signal segments;

[0019] Performing odd-even division on the obtained seedling signal segments to obtain odd-even signal segments;

[0020] Obtain resource stability coefficients, perform filtering conversion on the resource stability coefficients, and obtain high-order filtering ends;

[0021] Perform target filtering on the odd and even signal segments according to the high-order filter end to obtain the captured signal segment;

[0022] Based on the seedling environmental signal, the seedling signal segment is replaced within the segment according to the obtained capture signal segment to obtain a captured seedling signal sequence.

[0023] Preferably, the process of characterizing the captured seedling signal sequence comprises:

[0024] The characteristic coefficients of the stage seedlings were obtained by reconstructing the characteristics of the captured seedling signal sequence according to the stage resource coefficients;

[0025] Construct a two-dimensional rectangular coordinate system based on the characteristic coefficients of the seedlings at each stage;

[0026] A seedling characteristic coefficient curve is generated according to the obtained stage seedling characteristic coefficients, and the obtained seedling characteristic coefficient curve is uploaded to a two-dimensional rectangular coordinate system to obtain a seedling characteristic coefficient diagram.

[0027] Preferably, the process of screening and assimilating the basic parameters of the machine through the stage seedling characteristic coefficients includes:

[0028] Perform format changes on the basic parameters of the machine to obtain the basic signals of the machine;

[0029] Decompose and match the machine basic signal according to the stage resource coefficient to obtain the machine signal segment;

[0030] Performing odd-even division on the machine signal segment to obtain the machine odd-even signal segment, and performing target filtering on the machine odd-even signal segment according to the high-order filter end to obtain the captured machine signal segment;

[0031] performing intra-segment replacement on the machine signal segment according to the captured machine signal segment based on the machine basic signal to obtain a captured machine signal sequence;

[0032] The captured machine signal sequence is reconstructed according to the stage resource coefficient to obtain the wheelbase control coefficient.

[0033] Preferably, the process of constructing the seedling twin space for virtual generation includes:

[0034] Perform virtual twinning on the pruning area to obtain the seedling twin space, mark the seedlings in the seedling twin space according to the pruning area to obtain twin pruned seedlings, and map the seedling pruners to the seedling twin space to obtain the virtual pruners;

[0035] The wheelbase control coefficient is uploaded to the corresponding virtual pruning machine to obtain the seedling characteristic coefficient map, the seedling characteristic coefficient map is matched with the twin pruned seedlings, and the successfully matched seedling characteristic coefficient map is uploaded to the twin pruned seedlings.

[0036] Preferably, the process of optimizing the pruning of twin pruning seedlings according to the virtual pruning machine includes:

[0037] Graphically constructing the wheelbase control coefficient to obtain a wheelbase control coefficient graph, and performing curve statistics on the wheelbase control coefficient graph to obtain a wheelbase control curve;

[0038] issuing a wheelbase adjustment command to the virtual trimmer, performing simulated adjustment on the virtual trimmer according to the wheelbase adjustment command, recording a wheelbase control coefficient diagram after the simulated adjustment, and recording the wheelbase control curve after the simulated adjustment as a virtual control curve;

[0039] Upload the virtual control curve to the seedling characteristic coefficient map, and make an approximate analogy with the virtual control curve through the seedling characteristic coefficient map to obtain the approximate degree of the adjustment wheelbase;

[0040] The wheelbase adjustment instruction is changed and modified, and the adjusted wheelbase approximations obtained from each change and modification are sorted and selected to obtain the optimal trimmed wheelbase.

[0041] Preferably, the process of performing virtual driving of the virtual mower and optimizing the distance transportation of the virtual driving process includes:

[0042] issuing a driving wheelbase adjustment command to the virtual trimmer, performing simulated adjustment on the virtual trimmer according to the driving wheelbase adjustment command, recording a wheelbase adjustment coefficient diagram after the simulated adjustment, and recording the wheelbase adjustment curve after the simulated adjustment as a virtual driving adjustment curve;

[0043] The obtained virtual driving control curve is uploaded to the seedling characteristic coefficient map, and the virtual driving control curve is approximated by the seedling characteristic coefficient map to obtain the driving wheelbase approximation;

[0044] The wheelbase approximations are sorted and selected according to the obtained wheelbase adjustment instructions to obtain the optimal wheelbase.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] 1. Collect the parameter information of the seedlings in the pruning area and the pruning machine parameters and process them into the same data format, which is conducive to accurate wheelbase matching and improves the wheelbase adjustment speed;

[0047] 2. Construct a seedling twin space. By mapping the seedlings and pruning machines corresponding to the pruning area in the seedling twin space, the optimal pruning wheel width is obtained by adjusting the wheel width adjustment instructions in the virtual seedling twin space and comparing the approximation between the machine characteristic curve and the seedling characteristic curve. By finding the optimal pruning wheel width in the virtual space and applying it to the actual pruning process, work efficiency is greatly improved and it can adapt to seedlings of different sizes and shapes, as well as different pruning requirements.

[0048] 3. In the constructed seedling twin space, the wheelbase is also adjusted for the route between the pruning machine and the next pruning object, which can better adapt to different working environments, improve its adaptability under different conditions, and obtain the most stable driving state, which not only increases the driving speed but also enhances the safety of the pruning process. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0050] Figure 1 This is a schematic diagram of the present invention. DETAILED DESCRIPTION

[0051] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0052] like Figure 1 As shown, a wheelbase optimization control method for a seedling pruning machine based on machine learning includes the following steps:

[0053] Step S1: collecting machine basic parameters and seedling environmental parameters;

[0054] Step S2: changing the form of the seedling environmental parameters to obtain the seedling environmental signal, setting the resource stability coefficient to perform coefficient transformation to obtain the stage resource coefficient, performing stage extraction on the seedling environmental signal according to the stage resource coefficient, and obtaining the captured seedling signal sequence;

[0055] Step S3: Perform feature conversion on the captured seedling signal sequence to obtain a seedling characteristic coefficient map, and screen and assimilate the machine basic parameters through the stage seedling characteristic coefficients to obtain the wheelbase control coefficient;

[0056] Step S4: Construct a seedling twin space for virtual generation, obtain twin pruned seedlings and virtual pruners, optimize the pruned seedlings according to the virtual pruners, obtain the optimal pruned wheelbase, virtually drive the virtual pruners, and optimize the distance transportation of the virtual driving process to obtain the optimal driving wheelbase.

[0057] It should be further explained that, in the specific implementation process, due to the different terrains and environments in which the seedlings are located, the seedling pruning machine requires different wheelbases to improve the stability of the pruning machine on different ground surfaces and to adapt to different working conditions to help the seedling pruning machine work more efficiently. However, the wheelbase adjustment steps in the design of some seedling pruning machines are cumbersome and require special tools or professional knowledge, which is not conducive to the operation of ordinary users. In addition, the operation needs to be stopped during the wheelbase adjustment process, resulting in reduced seedling pruning efficiency. Through machine learning, the working environment of the seedling pruning machine can be monitored in real time, and the wheelbase can be automatically adjusted to adapt to terrain changes, reducing manual intervention, thereby improving pruning quality and the healthy growth of trees.

[0058] The process of collecting machine basic parameters and seedling environmental parameters includes:

[0059] Collecting data on the seedling pruning machine to obtain basic machine parameters, wherein the basic machine parameters represent state parameters of the normal operation of the pruning machine, including but not limited to the center of gravity position, tilt angle, wheel size, and wheelbase range of the pruning machine;

[0060] A seedling collection terminal is set up, and the pruning area is comprehensively collected through the seedling collection terminal to obtain seedling environmental parameters, which include topographic data, seedling size data and environmental obstacle data, wherein the topographic data includes but is not limited to the slope, undulation and hardness of the ground, and the seedling size data includes but is not limited to the height, thickness and branch position of the seedlings. The environmental obstacle data represents the position and size of the surrounding obstacles, and the obtained seedling environmental parameters are associated with the corresponding target pruning seedlings, wherein the target pruning seedlings represent the seedlings that need to be pruned in the pruning area. After associating the seedling environmental parameters, the growth size of each target pruning seedling, the surrounding environment and whether there are obstacles in the surrounding area can be directly obtained, which facilitates the seedling pruning machine to prune the target pruning seedlings without obstacles, quickly and accurately.

[0061] Performing form change on the obtained seedling environmental parameters to obtain seedling environmental signals;

[0062] The form change means processing the obtained seedling environmental parameters into a signal form, and according to the topographic data, seedling size data and environmental obstacle data included in the seedling environmental parameters, the seedling environmental signal includes a topographic signal, a seedling size signal and an environmental obstacle signal;

[0063] Setting a resource stability coefficient, wherein the resource stability coefficient is expressed in a function form and is determined according to the signal characteristics of the seedling environmental signal, and selecting an appropriate wavelet function as the resource stability coefficient based on a wavelet function, wherein the wavelet function includes but is not limited to Symlets wavelet, Morlet wavelet, and Gaussian wavelet;

[0064] Performing local transformation on the obtained resource stability coefficient to obtain a local variable;

[0065] The local transformation represents scaling and translation transformation of the control resource stability coefficient in the time dimension and the frequency dimension, and the distance of the scaling and translation transformation is recorded to obtain a local variable;

[0066] Perform statistical decomposition on the resource stability coefficient according to the obtained local variables to obtain local decomposition parameters;

[0067] The statistical decomposition means performing statistics on the local variables obtained in the resource stability coefficient to obtain the number of variables, averaging the local variables according to the obtained number of variables, and rounding the average value result downward to obtain a local decomposition parameter, that is, the local decomposition parameter is an integer;

[0068] Decompose the resource stability coefficient hierarchically according to the obtained local decomposition parameters to obtain the stage resource coefficient;

[0069] The hierarchical decomposition means dividing the resource stability coefficient into equal parts according to the local decomposition parameter to obtain stage resource coefficients of equal length, and the number of stage resource coefficients is equal to the local decomposition parameter;

[0070] Decompose and match the seedling environmental signal according to the obtained stage resource coefficient to obtain the seedling signal segment;

[0071] The matching decomposition means dividing the seedling environment signal into equal length and quantity according to the number of stage resource coefficients to obtain seedling signal segments of equal length, and the number of seedling signal segments is equal to the number of stage resource coefficients;

[0072] Performing odd-even division on the obtained seedling signal segments to obtain odd-even signal segments, wherein the odd-even signal segments include odd signal segments and even signal segments, wherein the odd-even division means numbering the seedling signal segments obtained by decomposition and matching based on the seedling environment signal, and recording the seedling signal segments at odd positions as odd signal segments, and recording the seedling signal segments at even positions as even signal segments;

[0073] Obtaining a resource stability coefficient, performing filtering conversion on the obtained resource stability coefficient, and obtaining a high-order filtering end;

[0074] The filtering conversion means that according to the resource stability coefficient being determined by the wavelet function, a filter is constructed for the resource stability coefficient, that is, the corresponding wavelet function is converted into a high-pass filter, that is, a high-order filtering end;

[0075] Target filtering is performed on the odd and even signal segments according to the obtained high-order filter end to obtain a captured signal segment;

[0076] The target filtering means applying a high-order filter end to the seedling environment signal and filtering the seedling signal segments whose odd-numbered signal segments are odd-numbered signal segments, capturing the high-frequency features of the odd-numbered signal segments through the high-order filter end to obtain captured signal segments;

[0077] Based on the seedling environmental signal, the seedling signal segment is replaced within the segment according to the obtained capture signal segment to obtain a capture seedling signal sequence;

[0078] The intra-segment replacement means replacing the odd-numbered signal segments at the corresponding positions with the captured signal segments according to the order of the seedling environment signals to obtain a captured seedling signal sequence, wherein the captured seedling signal sequence represents a signal segment sequence composed of the even-numbered signal segments and the captured signal segments, and the arrangement order is the same as the order of the seedling signal segments obtained by decomposition matching;

[0079] In particular, extracting odd signal segments from seedling environmental signals can facilitate the analysis of local changes in signals and be used for multi-faceted analysis. The high-order filtering end of the wavelet function transformation is used for feature capture, which expands the scope of application and effectively improves computational efficiency.

[0080] The process of performing feature conversion on the captured seedling signal sequence to obtain the seedling characteristic coefficient map includes:

[0081] The captured seedling signal sequence is reconstructed according to the stage resource coefficient to obtain the stage seedling characteristic coefficient, and the stage seedling characteristic coefficient is associated with the corresponding seedling environment signal;

[0082] The feature reconstruction means uploading the obtained stage resource coefficients to the captured seedling signal sequence in the order of the stage resource coefficients, and making one-to-one correspondence between the stage resource coefficients and the signal segments in the captured seedling signal sequence, adaptively filtering the stage resource coefficients and the signal segments at the corresponding positions to obtain the stage adaptation coefficients, wherein adaptive filtering means convolving the stage resource coefficients with the signal segments at the corresponding positions to obtain the stage adaptation coefficients, and performing combined statistics on the obtained stage adaptation coefficients based on the order of the captured seedling signal sequence to obtain the stage seedling characteristic coefficients, wherein combined statistics means summing the stage adaptation coefficients in the order of the captured seedling signal sequence to obtain the stage seedling characteristic coefficients; in particular, the signal segments in "adaptively filtering the stage resource coefficients and the signal segments at the corresponding positions" have corresponding different signal segments according to odd and even positions, the signal segments at odd positions are captured signal segments, and the signal segments at even positions are even signal segments;

[0083] Furthermore, according to the topographic and geomorphic signals, the seedling size signals and the environmental obstacle signals included in the seedling environmental signals, the stage seedling characteristic coefficients include the topographic and geomorphic characteristic coefficients, the seedling size characteristic coefficients and the environmental obstacle characteristic coefficients;

[0084] Graphically constructing the obtained stage seedling characteristic coefficients to obtain a seedling characteristic coefficient graph;

[0085] It should be further explained that, in a specific implementation process, the graph construction process includes:

[0086] A two-dimensional rectangular coordinate system is constructed based on the obtained stage seedling characteristic coefficients;

[0087] Generate a seedling characteristic coefficient curve according to the obtained stage seedling characteristic coefficient, upload the obtained seedling characteristic coefficient curve to a two-dimensional rectangular coordinate system, and obtain a seedling characteristic coefficient map, wherein, according to the stage seedling characteristic coefficient including the topography and geomorphology characteristic coefficient, the seedling size characteristic coefficient and the environmental obstacle characteristic coefficient, the seedling characteristic coefficient curve includes the topography and geomorphology coefficient curve, the seedling size coefficient curve and the environmental obstacle coefficient curve, and then, the seedling characteristic coefficient map includes three coefficient curves, namely, the topography and geomorphology coefficient curve, the seedling size coefficient curve and the environmental obstacle coefficient curve;

[0088] In particular, the characteristic curve of the seedling size can be intuitively observed according to the seedling size coefficient curve in the seedling characteristic coefficient diagram, which is convenient for matching the most suitable pruning wheelbase of the pruning machine, and the terrain coefficient curve and the environmental obstacle coefficient curve can provide the most suitable driving wheelbase for the pruning machine. For the pruning machine, different terrains and obstacles will affect the stability and speed of the driving process to different degrees. In order to improve the stability and speed, it is necessary to adjust the wheelbase during driving according to different terrains and obstacles, which can better adapt to different working environments and improve its adaptability under different conditions until the target pruning seedlings are reached, and then a second adjustment is made to adjust to the pruning wheelbase suitable for the target pruning seedlings, so as to perform precise pruning and thus improve the pruning quality;

[0089] Obtain machine basic parameters, screen and assimilate the obtained machine basic parameters through stage seedling characteristic coefficients, and obtain wheelbase control coefficients;

[0090] It should be further explained that, in the specific implementation process, the screening assimilation means that the seedling characteristic coefficient at the stage is obtained by converting and extracting the seedling environmental parameters, and the parameters related to the pruning machine wheelbase in the machine basic parameters are processed in the same way to obtain the same data form as the seedling characteristic coefficient at the stage, that is, the wheelbase control coefficient. The specific process includes:

[0091] Perform format changes on the basic parameters of the machine to obtain the basic signals of the machine;

[0092] According to the center of gravity position, tilt angle, wheel size and wheelbase range of the trimmer included in the basic machine parameters, the basic machine signal is to convert all the data included in the basic machine parameters into a corresponding signal;

[0093] Decomposing and matching the machine basic signal according to the obtained stage resource coefficient to obtain the machine signal segment;

[0094] Performing parity division on the obtained machine signal segments to obtain machine parity signal segments, wherein the machine parity signal segments include machine odd signal segments and machine even signal segments;

[0095] Target filtering is performed on the machine odd and even signal segments according to the obtained high-order filter end to obtain a captured machine signal segment;

[0096] performing intra-segment replacement on the machine signal segments according to the obtained captured machine signal segments based on the machine basic signal to obtain a captured machine signal sequence;

[0097] A phase resource coefficient is obtained, and a feature reconstruction of a captured machine signal sequence is performed according to the obtained phase resource coefficient to obtain a wheelbase control coefficient, and the obtained wheelbase control coefficient is associated with a corresponding machine basic signal.

[0098] The process of constructing a seedling twin space for virtual generation includes:

[0099] Performing virtual twinning on the pruned area to obtain a seedling twin space, and associating the obtained seedling twin space with the corresponding pruned area;

[0100] Furthermore, the virtual twin represents a virtual twin space with exactly the same functional structure as the actual pruning area. In this virtual twin space, the growth conditions and geographical environment of the seedlings are exactly the same as those of the pruning area, and is recorded as the seedling twin space. Then, there is a corresponding seedling twin space for each pruning area.

[0101] Marking the seedlings in the seedling twin space according to the pruning area to obtain twin pruned seedlings, wherein the seedling marking means marking the target pruned seedlings in the seedling twin space and recording them as twin pruned seedlings;

[0102] Map the seedling pruning machine to the seedling twin space to obtain a virtual pruning machine;

[0103] The virtual pruning machine means mapping a real seedling pruning machine to the seedling twin space to obtain a twin virtual pruning machine, and the function and structure of the virtual pruning machine are exactly the same as those of the real seedling pruning machine;

[0104] Obtaining a wheelbase control coefficient, uploading the obtained wheelbase control coefficient to a corresponding virtual trimmer, indicating obtaining the wheelbase control coefficient corresponding to the virtual trimmer, and then controlling the wheelbase adjustment of the virtual trimmer according to the wheelbase control coefficient to obtain the most suitable driving wheelbase and trimming wheelbase;

[0105] Obtaining a seedling characteristic coefficient map, matching the obtained seedling characteristic coefficient map with the twin pruned seedlings, uploading the successfully matched seedling characteristic coefficient map to the twin pruned seedlings, and associating it with the twin pruned seedlings;

[0106] The twin pruning space of the seedlings is pruned according to the obtained twin pruning seedlings to obtain an initial planning path. The pruning plan represents a pruning route set according to the target pruning seedlings that need to be pruned in the pruning area. The target pruning seedlings are pruned sequentially according to the pruning route by the seedling pruning machine. After a target pruning seedling is pruned, there will be a distance to travel before reaching the next target pruning seedling. The wheelbase of the current seedling pruning machine may cause instability in the driving process, resulting in low speed and low pruning efficiency. Therefore, the wheelbase of the seedling pruning machine during the driving process needs to be controlled to adjust to the most suitable driving wheelbase.

[0107] Through the seedling twin space, the twin pruning seedlings are optimized according to the obtained virtual pruning machine to obtain the optimal pruning wheel distance;

[0108] It should be further explained that, in the specific implementation process, the seedling pruning optimization means simulating the wheel spacing adjustment of the target pruning seedlings to be pruned in the seedling twin space to obtain the most appropriate pruning wheel spacing, which can not only reduce damage to the seedlings but also improve pruning quality and efficiency. The specific process includes:

[0109] Graphically constructing the obtained wheelbase control coefficient to obtain a wheelbase control coefficient graph, and performing curve statistics on the wheelbase control coefficient graph to obtain a wheelbase control curve;

[0110] The graphical construction represents generating a wheelbase control curve based on the wheelbase control coefficient, and then uploading the wheelbase control curve to a two-dimensional rectangular coordinate system to obtain a wheelbase control coefficient map, wherein, based on basic machine parameters including but not limited to the center of gravity position, tilt angle, wheel size, and wheelbase range of the trimmer, the corresponding wheelbase control coefficient also includes the corresponding control coefficient, and the generated wheelbase control curve includes curves corresponding to the center of gravity position, tilt angle, wheel size, and wheelbase range of the trimmer, all of which are included in the same wheelbase control coefficient map;

[0111] issuing a wheelbase adjustment instruction to the obtained virtual trimmer, performing simulated adjustment on the virtual trimmer according to the obtained wheelbase adjustment instruction, recording a wheelbase control coefficient diagram after the simulated adjustment, and recording the wheelbase control curve after the simulated adjustment as a virtual control curve, wherein the wheelbase adjustment instruction indicates adjusting the wheelbase of the virtual trimmer according to the center of gravity position, tilt angle, and wheel size in the basic parameters of the machine, and satisfying the wheelbase range, recording the adjusted wheelbase distance, and recording the wheelbase control curve corresponding to the adjusted wheelbase distance in the wheelbase control coefficient diagram as the virtual control curve;

[0112] The obtained virtual control curve is uploaded to the seedling characteristic coefficient map, and the virtual control curve is approximated by the seedling characteristic coefficient map to obtain the approximate degree of the adjustment wheelbase;

[0113] The approximate analogy represents calculating the similarity between the virtual control curve and the seedling size coefficient curve in the seedling characteristic coefficient diagram to obtain the wheelbase adjustment approximation; in this embodiment, the similarity is obtained by using the Euclidean distance between the virtual control curve and the seedling size coefficient curve, that is, the virtual control curve and the seedling size coefficient curve are converted into the same number of points, and the Euclidean distance between the point of the virtual control curve and the corresponding point of the seedling size coefficient curve is calculated. Finally, the Euclidean distance of all points is summed and averaged to obtain the wheelbase adjustment approximation, then each wheelbase adjustment instruction has a corresponding virtual control curve, and each wheelbase adjustment instruction has a corresponding wheelbase adjustment approximation;

[0114] The obtained wheelbase adjustment instructions are changed and modified, and the adjustment wheelbase approximations obtained from each change and modification are sorted and selected to obtain the optimal pruning wheelbase, indicating that each wheelbase adjustment instruction has a corresponding adjustment wheelbase approximation, and the wheelbase adjustment instructions can be simulated and modified, and the obtained adjustment wheelbase approximations are sorted in order from small to large, and the wheelbase adjustment instruction corresponding to the adjustment wheelbase approximation that ranks first is recorded as the optimal pruning wheelbase, indicating the simulated pruning wheelbase that is most suitable for the target pruning seedlings, and the simulated pruning wheelbase is applied to the seedling pruning machine corresponding to the virtual pruning machine, so that the pruning of the seedlings can be optimized;

[0115] In particular, based on the pruning route corresponding to the initial planned path, there is a corresponding driving route when reaching the next target pruning seedling. In order to increase the driving speed of the pruning machine and thus shorten the pruning time, the wheelbase of the virtual pruning machine needs to be adjusted during this driving process. The virtual pruning machine is adjusted for distance and transportation through the seedling twin space to obtain the optimal driving wheelbase. The specific process includes:

[0116] issuing a driving wheelbase adjustment instruction to the obtained virtual trimmer, performing simulated adjustment on the virtual trimmer according to the obtained driving wheelbase adjustment instruction, recording a wheelbase adjustment coefficient diagram after the simulated adjustment, and recording the wheelbase adjustment curve after the simulated adjustment as a virtual driving adjustment curve;

[0117] The obtained virtual driving control curve is uploaded to the seedling characteristic coefficient map, and the virtual driving control curve is approximated by the seedling characteristic coefficient map to obtain the driving wheelbase approximation;

[0118] The approximate analogy means calculating the similarity between the virtual driving control curve and the topography coefficient curve and the environmental obstacle coefficient curve in the seedling characteristic coefficient diagram to obtain the driving wheelbase approximation; in this embodiment, the similarity is obtained by comprehensively utilizing the Euclidean distance between the virtual driving control curve and the topography coefficient curve and the Euclidean distance between the virtual driving control curve and the environmental obstacle coefficient curve, that is, first converting the virtual driving control curve and the topography coefficient curve into the same number of points, calculating the Euclidean distance between the points of the virtual driving control curve and the corresponding points of the topography coefficient curve, and finally calculating the Euclidean distance of all points. The distances are summed and then averaged to obtain the control terrain approximation. Similarly, the virtual driving control curve and the environmental obstacle coefficient curve are converted into the same number of points, and the Euclidean distances between the points of the virtual driving control curve and the corresponding points of the environmental obstacle coefficient curve are calculated. Finally, the Euclidean distances of all points are summed and then averaged to obtain the control obstacle approximation. The obtained control terrain approximation and the control obstacle approximation are added to obtain the driving wheelbase approximation. Therefore, for each driving wheelbase adjustment instruction, there is a corresponding virtual driving control curve, and for each driving wheelbase adjustment instruction, there is a corresponding wheelbase adjustment approximation.

[0119] The wheelbase approximations are sorted and selected according to the obtained wheelbase adjustment instructions to obtain the optimal wheelbase, indicating that each wheelbase adjustment instruction has a corresponding wheelbase approximation. The wheelbase adjustment instructions can be simulated and modified, and the obtained wheelbase approximations are sorted in ascending order. The wheelbase adjustment instruction corresponding to the first-ranked wheelbase approximation is recorded as the optimal wheelbase, indicating the most suitable wheelbase for this section of road. The wheelbase of the seedling trimmer is adjusted to the optimal wheelbase according to the virtual trimmer, so that it can run safely and stably to the next target of trimming seedlings.

[0120] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A seedling pruning machine wheelbase optimization control method based on machine learning, characterized in that: The following steps are involved: Step S1: collecting machine basic parameters and seedling environmental parameters; Step S2: changing the form of the seedling environmental parameters to obtain the seedling environmental signal, setting the resource stability coefficient to perform coefficient transformation to obtain the stage resource coefficient, performing stage extraction on the seedling environmental signal according to the stage resource coefficient, and obtaining the captured seedling signal sequence; Step S3: Perform feature conversion on the captured seedling signal sequence to obtain a seedling characteristic coefficient map, and screen and assimilate the machine basic parameters through the stage seedling characteristic coefficients to obtain the wheelbase control coefficient; Step S4: Construct a seedling twin space for virtual generation, obtain twin pruned seedlings and virtual pruners, optimize the pruned seedlings according to the virtual pruners, obtain the optimal pruned wheelbase, virtually drive the virtual pruners, and optimize the distance transportation of the virtual driving process to obtain the optimal driving wheelbase.

2. The method for optimizing wheelbase control of a seedling pruning machine based on machine learning according to claim 1, wherein: The process of collecting machine basic parameters and seedling environmental parameters includes: Collect data on the seedling pruning machine to obtain basic machine parameters; A seedling collection terminal is set up to conduct comprehensive collection of the pruning area through the seedling collection terminal to obtain seedling environmental parameters.

3. The method for optimizing wheelbase control of a seedling pruning machine based on machine learning according to claim 1, wherein: The process of setting the resource stability coefficient and performing coefficient transformation includes: Perform local transformation on the resource stability coefficient to obtain local variables; Perform statistical decomposition of resource stability coefficients according to local variables to obtain local decomposition parameters; The resource stability coefficient is decomposed hierarchically according to the local decomposition parameters to obtain the stage resource coefficient.

4. The method for optimizing wheelbase control of a seedling pruning machine based on machine learning according to claim 1, wherein: The process of extracting seedling environmental signals according to the stage resource coefficient includes: Decompose and match the seedling environmental signals according to the stage resource coefficients to obtain the seedling signal segments; Performing odd-even division on the obtained seedling signal segments to obtain odd-even signal segments; Obtain resource stability coefficients, perform filtering conversion on the resource stability coefficients, and obtain high-order filtering ends; Perform target filtering on the odd and even signal segments according to the high-order filter end to obtain the captured signal segment; Based on the seedling environmental signal, the seedling signal segment is replaced within the segment according to the obtained capture signal segment to obtain a captured seedling signal sequence.

5. The method for optimizing wheelbase control of a seedling pruning machine based on machine learning according to claim 1, wherein: The process of characterizing the captured seedling signal sequence includes: The characteristic coefficients of the stage seedlings were obtained by reconstructing the characteristics of the captured seedling signal sequence according to the stage resource coefficients; Construct a two-dimensional rectangular coordinate system based on the characteristic coefficients of the seedlings at each stage; Generate a seedling characteristic coefficient curve according to the obtained stage seedling characteristic coefficients, wherein the seedling characteristic coefficient curve includes a topography coefficient curve, a seedling size coefficient curve, and an environmental obstacle coefficient curve; The obtained seedling characteristic coefficient curve is uploaded to a two-dimensional rectangular coordinate system to obtain a seedling characteristic coefficient graph.

6. The method for optimizing wheelbase control of a seedling pruning machine based on machine learning according to claim 4, wherein: The process of screening and assimilation of machine basic parameters through stage seedling characteristic coefficients includes: Perform format changes on the basic parameters of the machine to obtain the basic signals of the machine; Decompose and match the machine basic signal according to the stage resource coefficient to obtain the machine signal segment; Performing odd-even division on the machine signal segment to obtain the machine odd-even signal segment, and performing target filtering on the machine odd-even signal segment according to the high-order filter end to obtain the captured machine signal segment; performing intra-segment replacement on the machine signal segment according to the captured machine signal segment based on the machine basic signal to obtain a captured machine signal sequence; The captured machine signal sequence is reconstructed according to the stage resource coefficient to obtain the wheelbase control coefficient.

7. The method for optimizing wheelbase control of a seedling pruning machine based on machine learning according to claim 2, wherein: The process of constructing the seedling twin space for virtual generation includes: Perform virtual twinning on the pruning area to obtain the seedling twin space, mark the seedlings in the seedling twin space according to the pruning area to obtain twin pruned seedlings, and map the seedling pruners to the seedling twin space to obtain the virtual pruners; The wheelbase control coefficient is uploaded to the corresponding virtual pruning machine to obtain the seedling characteristic coefficient map, the seedling characteristic coefficient map is matched with the twin pruned seedlings, and the successfully matched seedling characteristic coefficient map is uploaded to the twin pruned seedlings.

8. The method for optimizing wheelbase control of a seedling pruning machine based on machine learning according to claim 5, wherein: The process of optimizing seedling pruning for twin-pruned seedlings based on the virtual pruning machine includes: Graphically constructing the wheelbase control coefficient to obtain a wheelbase control coefficient graph, and performing curve statistics on the wheelbase control coefficient graph to obtain a wheelbase control curve; issuing a wheelbase adjustment command to the virtual trimmer, performing simulated adjustment on the virtual trimmer according to the wheelbase adjustment command, recording a wheelbase control coefficient diagram after the simulated adjustment, and recording the wheelbase control curve after the simulated adjustment as a virtual control curve; Upload the virtual control curve to the seedling characteristic coefficient map, and obtain the wheelbase approximation by making an approximate comparison between the virtual control curve and the seedling size coefficient curve in the seedling characteristic coefficient map; The wheelbase adjustment instruction is changed and modified, and the adjusted wheelbase approximations obtained from each change and modification are sorted and selected to obtain the optimal trimmed wheelbase.

9. The method for optimizing wheelbase control of a seedling pruning machine based on machine learning according to claim 8, wherein: The process of virtually driving the virtual trimmer and optimizing the distance transportation of the virtual driving process includes: issuing a driving wheelbase adjustment command to the virtual trimmer, performing simulated adjustment on the virtual trimmer according to the driving wheelbase adjustment command, recording a wheelbase adjustment coefficient diagram after the simulated adjustment, and recording the wheelbase adjustment curve after the simulated adjustment as a virtual driving adjustment curve; The obtained virtual driving control curve is uploaded to the seedling characteristic coefficient map, and the driving wheelbase approximation is obtained by approximating the virtual driving control curve with the topography coefficient curve and the environmental obstacle coefficient curve in the seedling characteristic coefficient map; The wheelbase approximations are sorted and selected according to the obtained wheelbase adjustment instructions to obtain the optimal wheelbase.

Images