Intelligent operation quality evaluation method and system for agricultural equipment

Abstract

The invention discloses an agricultural equipment intelligent operation quality evaluation method and system, and relates to the technical field of quality evaluation. Comprising the following steps: dividing a buffer area into edge grid units; performing segmented coding processing on the original operation trajectory data to form a fragmented path unit set, and extracting a spatial interaction relationship between each path unit and a corresponding edge grid unit; establishing a quality linkage matrix based on a multi-parameter calculation result; calling a path unit set of a historical high-score area of the edge grid unit to reconstruct a trajectory path according to a preset operation specification template trajectory, and generating an offset coding vector group according to a spatial offset vector between an original operation path and the reconstructed trajectory; and performing multi-dimensional quality index clustering analysis according to the consistency score and the offset coding vector group, generating a fragment quality grade label, and forming a linkage quality evaluation layer. According to the method, historical high-quality path reconstruction and a multi-dimensional clustering mechanism are fused, and the accuracy and adaptability of quality evaluation are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of quality assessment, and in particular to a method and system for assessing the quality of intelligent operation of agricultural equipment. Background Art

[0002] With the continuous advancement of agricultural mechanization and intelligence, agricultural equipment is playing an increasingly critical role in multiple processes, including tillage, sowing, fertilization, spraying, and harvesting. However, existing agricultural operation quality assessment methods often use global trajectories as a holistic analysis, which fails to accurately reflect local differences in operation quality and the spatial linkages between regions. This is particularly true near the boundaries of the operation area, where frequent adjustments and diversions in operation paths significantly increase assessment errors. Furthermore, traditional methods often overlook the interaction between operation paths and the spatial structure of the operation area and lack a systematic analysis of trajectory reconstruction, trajectory offsets, and historical normative operation patterns, resulting in poor interpretability and pertinence of assessment results. Furthermore, a unified assessment framework that simultaneously integrates trajectory encoding, spatial interaction analysis, historical high-scoring trajectory mining, and multidimensional clustering algorithms is currently lacking. This makes it difficult to support dynamic quality label generation and layer visualization of operation trajectories, hindering subsequent intelligent decision-making and improvement.

[0003] CN109099925B discloses a method for navigation path planning and operation quality assessment for unmanned agricultural machinery. This method uses an unmanned remote-controlled aircraft to collect vertex information on field boundaries, establishes a field coordinate model using an RTK positioning device, implements path planning on the platform's main controller, and transmits the path to the agricultural machinery for autonomous operation. Simultaneously, visual sensors collect operation images for remote monitoring and operation quality assessment. However, this method primarily focuses on path planning and image feedback for a single agricultural machinery, lacking quantitative assessment methods for track coverage, overlapping operation areas, and task completion when multiple agricultural machinery are working together. This makes it difficult to meet the needs of efficient large-scale farming and transparent management of operation progress.

[0004] CN119515134A discloses a real-time assessment method for the operating area and quality of multiple agricultural machines on the same farmland plot. This method uses drones to acquire plot boundary points and non-arable land information, constructs a farmland graphical model, and uses GPS to collect machine trajectory information to plot the operating paths. By calculating the overlap between the trajectories and the cultivated land, this method enables real-time assessment of each machine's effective operating area and task completion. This method effectively improves the visual management of collaborative operations involving multiple agricultural machines. However, it still relies on trajectory data itself and lacks the ability to identify detailed ground-level operational quality through image analysis. This makes it impossible to immediately identify and provide feedback on abnormal behaviors such as path deviations and missed tillage.

[0005] Therefore, a quality assessment scheme for intelligent operation of agricultural equipment is needed. Summary of the Invention

[0006] In view of the problems existing in the existing agricultural equipment operation quality assessment technology, the present invention is proposed.

[0007] Therefore, the problem to be solved by the present invention is how to improve the accuracy and adaptability of quality assessment.

[0008] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0009] In the first aspect, the present invention provides a method for evaluating the quality of intelligent operation of agricultural equipment, which includes: constructing an edge buffer area based on an outer contour point set of an operation area, and dividing the buffer area into edge grid units; performing segmented coding processing on the original operation trajectory data to form a slice path unit set, and extracting the spatial interaction relationship between each path unit and the corresponding edge grid unit respectively; establishing a quality linkage matrix based on the multi-parameter calculation results of the path unit and the edge grid unit, and each element in the quality linkage matrix reflects the multi-parameter consistency score of the corresponding edge grid unit and the slice path unit under the linkage relationship; based on the preset operation specification template trajectory, calling the path unit set of the edge grid unit's historical high-scoring area, reconstructing the trajectory path through interpolation-correction, and generating an offset coding vector group based on the spatial offset vector between the original operation path and the reconstructed trajectory; performing multi-dimensional quality indicator clustering analysis on the slice path unit set according to the consistency score and the offset coding vector group, generating a slice quality grade label, and forming a linkage quality assessment layer.

[0010] As a preferred solution of the method for evaluating the quality of intelligent operation of agricultural equipment described in the present invention, the division of the edge grid unit includes: performing contour thinning processing on the outer trajectory nodes in the operation path data, extracting the boundary point set and then constructing the minimum outer polygon boundary; calculating the local curvature index of each boundary segment in combination with the directional change rate of the boundary, setting the variant buffer width according to the curvature value, and generating an asymmetric outward expansion buffer area; dividing the outward expansion buffer area into several local area sub-blocks according to the boundary normal direction, and performing grid rotation segmentation based on the local direction in each sub-block to generate multi-directional aligned edge grid units.

[0011] As a preferred solution of the method for evaluating the quality of intelligent operation of agricultural equipment described in the present invention, the segmented coding process is based on the operation timestamp and spatial segment length.

[0012] As a preferred solution of the method for evaluating the quality of intelligent operation of agricultural equipment described in the present invention, the multiple parameters include: spatial coverage, trajectory deviation and trajectory stability.

[0013] As a preferred solution of the method for evaluating the quality of intelligent operation of agricultural equipment described in the present invention, the following is a solution: according to the path unit set corresponding to the edge grid unit in the historical operation round, the score value in the quality linkage matrix is ​​extracted to be greater than the set threshold Q t A high-quality path reference set is constructed based on the path unit subset of the edge grid unit. For each edge grid unit, based on the trajectory point group at the same spatial position in its high-quality path reference set, a regional local interpolation method is used to construct a fitting trajectory track, which is used as the operation specification template trajectory segment, that is, the reconstructed trajectory path.

[0014] As a preferred solution of the method for evaluating the quality of intelligent operation of agricultural equipment described in the present invention, the generation of the offset coding vector group includes: selecting the edge grid unit of each original path unit and its spatial interaction, and calculating the Euclidean distance offset vector ΔV between each trajectory point in the trajectory path and the reconstructed trajectory in a two-dimensional coordinate system. ij (k), where k represents the trajectory point index; the path offset vector sequence ΔV ij (k) Perform segmented sliding window analysis and perform mean-variance normalization according to a fixed number of points to generate a standardized offset vector group ΔV ij (k) e ; Normalize the offset vector group ΔV ij (k) e Divide the direction into multiple quadrant segments, generate the direction code D, and convert the modulus |ΔV ij (k) e | Divide into several amplitude interval segments to generate amplitude code A; splice the direction code D and the amplitude code A to form a composite coding unit.

[0015] As a preferred solution of the intelligent operation quality assessment method of agricultural equipment described in the present invention, the multidimensional quality indicator cluster analysis includes: extracting the consistency score vector of each path unit in the quality linkage matrix; converting the offset coding vector group of the path unit into a frequency linear vector; combining the consistency score vector and the frequency linear vector for each path unit to form a multidimensional indicator feature vector to form a quality feature vector group for cluster analysis; and performing unsupervised clustering on the quality feature vector group.

[0016] In a second aspect, the present invention provides an intelligent operation quality assessment system for agricultural equipment, which includes: an edge construction module for constructing an edge buffer area based on an outer contour point set of an operation area, and dividing the buffer area into edge grid units; a trajectory encoding module for performing segmented encoding processing on the original operation trajectory data to form a slice path unit set, and extracting the spatial interaction relationship between each path unit and the corresponding edge grid unit respectively; a linkage matrix module for establishing a quality linkage matrix based on the multi-parameter calculation results of the path unit and the edge grid unit, wherein each element in the quality linkage matrix reflects the multi-parameter consistency score of the corresponding edge grid unit and the slice path unit under the linkage relationship; a trajectory reconstruction module for calling the path unit set of the historical high-scoring area of ​​the edge grid unit based on a preset operation specification template trajectory, reconstructing the trajectory path through an interpolation-correction method, and generating an offset coding vector group based on the spatial offset vector between the original operation path and the reconstructed trajectory; a quality clustering module for performing multi-dimensional quality indicator clustering analysis on the slice path unit set according to the consistency score and the offset coding vector group, generating a slice quality grade label, and forming a linkage quality assessment layer.

[0017] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of the method for evaluating the quality of intelligent operation of agricultural equipment as described in the first aspect of the present invention are implemented.

[0018] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of the method for evaluating the quality of intelligent operation of agricultural equipment as described in the first aspect of the present invention are implemented.

[0019] The beneficial effects of the present invention are as follows: by constructing an edge buffer zone and dividing it into multi-directionally aligned edge grid cells, the present invention achieves refined modeling of the boundary areas of agricultural equipment operations; by combining trajectory segmentation coding with spatial interaction analysis, a quality linkage matrix is ​​established to quantify the consistency between the trajectory and the operating specifications; and by reconstructing the trajectory and analyzing the offset coding, the error modeling and standardization of the original path are achieved. Compared with existing assessment methods that rely solely on trajectory overlap rate or trajectory offset distance, the present invention integrates historical high-quality path reconstruction with a multidimensional clustering mechanism, effectively improving the accuracy and adaptability of quality assessment, ultimately forming a linkage quality assessment layer with spatial continuity and difference sensitivity, providing highly reliable decision support for the operational quality assessment of agricultural equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. 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 creative work.

[0021] Figure 1 A flow chart of the quality assessment method for intelligent operation of agricultural equipment;

[0022] Figure 2 Flowchart for dividing edge grid units for the quality assessment method of intelligent operation of agricultural equipment;

[0023] Figure 3 This is the structural diagram of the intelligent operation quality assessment system for agricultural equipment. DETAILED DESCRIPTION

[0024] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0025] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0026] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0027] As described in the background technology above, existing agricultural operation quality assessment methods mostly use global trajectories as units for holistic analysis, which makes it difficult to accurately reflect the differences in local operation quality and the spatial linkage relationship between regions. Especially near the boundary of the operation area, the assessment error increases significantly due to frequent adjustments or turns of the operation path. In addition, traditional methods often ignore the interactive characteristics between the operation path and the spatial structure of the operation area, and lack a systematic analysis of trajectory reconstruction, trajectory offset and historical standard operation mode, resulting in poor interpretability and pertinence of the assessment results. On the other hand, there is currently a lack of a unified assessment framework that can simultaneously integrate trajectory encoding, spatial interaction analysis, historical high-scoring trajectory mining and multidimensional clustering algorithms. It is difficult to support the dynamic quality label generation and layer visualization expression of the operation trajectory, which is not conducive to subsequent intelligent decision-making and improvement. Therefore, there is a need for an intelligent operation quality assessment solution for agricultural equipment.

[0028] Figure 1 Flowchart of a method for evaluating the quality of intelligent operation of agricultural equipment according to an embodiment of the present invention.

[0029] like Figure 1 As shown in the figure, the quality assessment method of intelligent operation of agricultural equipment includes:

[0030] S1: Constructing an edge buffer area based on the outer contour point set of the working area, and dividing the buffer area into edge grid units.

[0031] Better, such as Figure 2 As shown, the division of edge grid units includes the following steps: by performing contour thinning processing on the outer trajectory nodes in the operation path data, extracting the boundary point set and constructing the minimum outer polygon boundary, the boundary is used to define the range of the edge operation area to be evaluated; combining the direction change rate of the boundary to calculate the local curvature index of each boundary segment, setting the variant buffer width according to the curvature value, generating an asymmetric outward expansion buffer area, and increasing the buffer width in the boundary corner area to improve the edge resolution; dividing the buffer area into several local area sub-blocks according to the boundary normal direction, and performing grid rotation segmentation based on the local direction in each sub-block to generate multi-directional aligned edge grid units G j , to avoid the non-equidistant error caused by orthogonal segmentation.

[0032] It should be noted that the edge buffer area is constructed based on the outer contour point set of the operation area, which aims to achieve accurate monitoring of the boundary quality of agricultural equipment during operation by structural modeling of the spatial boundary of the operation data. Specifically, first, based on the historical path trajectory data set collected by agricultural equipment during the operation process, the path nodes located at the operation boundary are extracted to form the original contour point set. The path node refers to the sampling point located at the outermost edge of the trajectory in the operation trajectory. Its positioning information is usually provided by a positioning system such as GNSS and RTK, and this embodiment does not make a sole limitation. In actual operation, since the boundary path nodes often have problems such as uneven distribution density and local mutation, directly using these points for boundary construction may cause the contour to be redundant and complex, affecting the accuracy of subsequent calculations. Therefore, in this step, contour thinning processing is adopted, that is, the original contour point set is resampled by setting the angle change threshold and the point distance threshold to remove redundant points in the trajectory and retain the turning points with obvious boundary structural characteristics, so that the boundary data is more regular and has concise expression.

[0033] Then, based on the extracted boundary point set, a minimum enclosing polygon boundary is constructed to limit the geometric boundary range of the edge buffer. Among them, the minimum enclosing polygon boundary refers to the polygon with the smallest area and the least number of vertices that contains all the contour point sets on a two-dimensional plane. This structure has high stability and low computational complexity in computational geometry. In the present invention, the Andrew algorithm is used to construct the minimum enclosing boundary, and the geometric connectivity and closure of the generated boundary are controlled by the arrangement order of the boundary points; for example, assuming that the contour point set is: A(0,0), B(2,1), C(4,0), D(3,3), E(1,4), first arrange the points in ascending order of the horizontal coordinates: A→B→E→D→C; then execute Andrew convex hull construction, automatically eliminate the concave point B, and retain the four points A, C, D, and E to form the minimum enclosing convex polygon ACDE; finally, the boundary closure is ensured by the arrangement order (such as clockwise A→C→D→E→A) to form the basic boundary for buffer calculation. This boundary serves as the basic framework for the subsequent buffer zone construction. It has clear spatial boundary information, which is conducive to defining the edge determination area and providing a spatial reference area for edge trajectory quality analysis.

[0034] After the boundary construction is completed, the local curvature index of each segment of the boundary is calculated by combining the directional change rate between the boundary points as the buffer width variation parameter. Among them, the directional change rate refers to the change in the angle amplitude between the directional vectors of adjacent boundary segments, which is used to measure the degree of turning of the boundary at this position. Conventional buffer zone construction methods usually use a fixed width to equidistantly expand the boundary. However, since agricultural operations are prone to trajectory deviation and insufficient operation coverage in corner areas, it is difficult to cover high-risk areas using a fixed width expansion, resulting in incomplete analysis. Therefore, the present invention designs a curvature-driven buffer width control method, which sets different buffer expansion scales according to the local curvature of each segment of the boundary, expands a larger width in areas where the boundary changes drastically (i.e., corner areas), and maintains a smaller expansion width in linear areas, thereby generating a variant asymmetric buffer zone. For example, assuming that the boundary of a certain operation path is an "L"-shaped corner structure, one segment is a horizontal line segment (curvature ≈ 0) and the other segment is a 90° corner area (high curvature). When constructing buffer zones, a 2-meter buffer width is set for horizontal segments; for corners, a 5-meter buffer width is automatically added based on their high curvature. This ensures that more potential offset trajectories are covered at corners, enhancing the accuracy of boundary detection. This strategy enhances spatial perception of boundary details and improves the spatial resolution of edge risk assessment.

[0035] Furthermore, the constructed buffer area is then divided into local areas according to the normal direction of the boundary. This division method is different from the conventional method of using orthogonal grid segmentation processing strategy. The present invention partitions the buffer into fan-shaped sub-blocks according to the normal direction of each segment of the boundary, so that the division direction of each sub-block area is consistent with the local direction of the boundary. This local direction consistency segmentation strategy can avoid the problem of non-aligned cutting in the corner area of ​​the conventional Cartesian grid, and reduce the problems such as uneven edge grid area and reduced resolution caused by direction deviation. Then, the grid unit generation operation is performed in each local partition. Specifically, in each sub-block, a rotating coordinate system is constructed based on the local normal direction of the boundary, and a two-dimensional equidistant grid is segmented based on this coordinate system. Among them, the size of the grid unit is set according to the point density in the buffer zone and the operating width of the operating equipment, usually maintained at 1 / 3 to 1 / 5 of the equipment coverage width to ensure that there is sufficient track point coverage density in each grid unit to support quality assessment. During the segmentation process, the geometric continuity and overlap ratio control between grid units must also be considered to prevent track overlap statistical errors caused by grid interlacing.

[0036] It should be noted that in order to avoid the problem that the normal direction may change sharply within a short distance in scenarios with high boundary complexity (such as high curvature, multiple broken line segments, and irregular edges), which will lead to a rapid increase in the number of fan-shaped sub-blocks if not controlled, the present invention sets a maximum number of fan-shaped sub-blocks; and if the normal change angle between a batch of consecutive sub-blocks is less than the set threshold, they are merged into a unified sub-block to maintain the stability and anti-disturbance capability of the offset pattern analysis.

[0037] In summary, the present invention not only establishes a high-precision edge space determination framework at the geometric modeling level, but also achieves enhanced spatial perception capabilities and improved edge recognition accuracy through local curvature-driven buffer zone variation construction and rotation direction aligned grid division strategies. This design effectively avoids problems such as directional error amplification, inconsistent resolution, and cutting distortion in traditional boundary processing, and provides high-quality, structured edge data support for agricultural equipment operation quality assessment. Especially in application scenarios where operations in boundary areas are highly discontinuous and equipment travel paths are unstable, the edge grid unit division method provided in this step can effectively support subsequent key processes such as path offset analysis, trajectory quality classification, and operation consistency scoring, and is one of the basic links for realizing intelligent analysis of the overall system.

[0038] S2: Perform segmented encoding processing on the original operation trajectory data to form a set of segmented path units, and extract the spatial interaction relationship between each path unit and the corresponding edge grid unit.

[0039] Preferably, the process of forming the path unit set includes: collecting the timestamps and corresponding coordinate points in the operation trajectory data to generate an ordered trajectory point set, which is used as the basic data source for path segmentation and spatial matching; performing time-space synchronous segmentation on the ordered trajectory point set, setting the spatial segment length L s Control path length and time interval T s Control the node sparsity to form multiple continuous path segments, which are recorded as path unit sets P1~P n , where n is the number of path units.

[0040] Specifically, in the operational process of the present invention, it is first necessary to collect the timestamp sequence and the corresponding two-dimensional geographic coordinate point pairs based on the original operation trajectory data to form a time-series trajectory point set. The trajectory point set usually comes from the operation trajectory data recorded in real time by equipment with positioning modules such as agricultural machinery and garden machinery. The data has the dual characteristics of time + space, which not only includes the spatial distribution of the operation path, but also records the dynamic process of the operation behavior. Therefore, the organization of the trajectory point set is particularly critical in the initial stage. The present invention uses the timestamp as the main index to ensure the temporal consistency of the trajectory point sequence, so as to maintain the continuity of the path direction and the stability of the node connection during the segmentation process. In addition, the trajectory point set must also undergo pre-processing steps such as denoising and interpolation to eliminate abnormal jump points and signal drift, so as to enhance the spatial logic and matching reliability of the path.

[0041] Furthermore, the present invention introduces a time-space synchronization segmentation mechanism in the process of forming a path unit set. The core of this mechanism is to achieve orderly segmentation of the trajectory by setting two control parameters, namely the spatial segment length and the time interval. Among them, the spatial segment length is used to control the maximum length of the path unit to prevent a single path unit from covering too many grid units, resulting in fuzzy matching results; and the time interval is used to control the density and continuity of the trajectory nodes to avoid node concentration or sparseness problems caused by speed changes or residence time during the operation. For example, when the equipment is in the stage of slow U-turn at the boundary, the time interval may be shortened to improve the accuracy of the path points, while in the stage of uniform straight driving, the time interval can be appropriately increased to improve processing efficiency.

[0042] The present invention performs a traversal sliding window operation on the above trajectory point set. In each time-space segmentation, the cumulative path length is calculated from the current starting point in sequence until the set spatial segment length is reached or the trajectory time span reaches the set upper limit, and a new path unit P is generated. i This process is repeated until all trajectory points are covered, and finally a path unit set consisting of several path units is formed.

[0043] Furthermore, each path unit P i The spatial extent and edge grid unit Gj The intersection of the set is determined to identify the interaction grid subset G(P i ), and record its grid code as the interaction key.

[0044] Specifically, in actual operation, it is necessary to first obtain the edge buffer grid division result generated in step S1, that is, the grid unit set, and then for each path unit P i Perform spatial range calculation, construct its minimum envelope rectangle or convex hull, and then compare it with each G in the grid unit set. j There are two ways to determine the intersection of the two surfaces: one is to calculate the intersection of the two surfaces based on geometric Boolean operations, and the other is to determine the intersection of the two surfaces. i Is the path geometry range consistent with G j The boundary contours have non-empty intersections; the other is based on the trajectory point projection judgment, statistical P i Whether each trajectory point falls into G j When the interaction condition is met, it can be determined that P i With G j There is a spatial coupling relationship, denoted as (P i ,G j ).

[0045] S3: Based on the multi-parameter calculation results of the path unit and the edge grid unit, a quality linkage matrix is ​​established, where each element in the quality linkage matrix reflects the multi-parameter consistency score of the corresponding edge grid unit and the slice path unit under the linkage relationship.

[0046] Preferably, for each pair of path units P i and the corresponding interactive grid subset G(P i ), extract the overlapping area in the plane space, and the corresponding interactive grid subset G(P i ) area ratio to calculate the spatial coverage R ij ; Based on the standard trajectory reference line, for each path unit at the edge grid unit G j The average offset distance between the actual trajectory and the standard trajectory reference line is calculated to obtain the offset index O ij , which is used to characterize the degree of deviation of the operation path.

[0047] For example, within the spatial interaction range between the path unit and the grid unit, for each pair of interaction terms (P i ,G j ), it is necessary to extract the overlapping area in the plane space, which is obtained by Boolean geometry operation, that is, for the path unit P i The spatial contour and grid unit G jThe boundary of the grid cell G is intersected and the area of ​​the non-empty intersection is calculated. j Normalize the ratio of its own area to obtain the spatial coverage index R ij , represents the coverage degree of the path unit to the grid unit. ij The larger the value, the more complete the operation coverage of the path unit within the current grid range, and it can be used as a geometric measurement indicator of operation adequacy.

[0048] Secondly, for the accuracy analysis of the operation path, the present invention introduces the trajectory deviation index, for which a standard trajectory reference line needs to be preset. The reference line can be generated by historical high-quality operation trajectories, planned paths or manually defined trajectories. The present invention does not limit the acquisition method. i In grid cell G j The actual set of trajectory points within the range is matched against the preset standard trajectory reference line using the minimum point-to-line distance. The vertical offset distance of each trajectory point is obtained, and the average of all distance values ​​is taken as the average offset of the unit within the grid. A smaller value indicates that the trajectory adheres closely to the reference line and the operation path is highly accurate. Conversely, it indicates that the trajectory has drifted or yawed significantly, affecting operation accuracy.

[0049] Better, use the direction angle variance σ of the path point set ij With the velocity change rate v ij Jointly construct stability scoring function: S ij =exp(-(σ ij +v ij )) reflects whether the operation trajectory has abnormal sudden turns, speed changes and other unstable behaviors.

[0050] It should be noted that the stability score is based on the path point set in the edge grid unit G j The degree of angular fluctuation and the amplitude of speed change are jointly constructed and quantified into the directional volatility index and the speed disturbance index respectively. The smoothness of the trajectory in the grid is judged based on their joint score. The smoother the trajectory and the more uniform the speed, the higher the score.

[0051] Furthermore, based on the above three indicators, a consistency scoring function is constructed, which is expressed as follows:

[0052]

[0053] Among them, O max is the maximum allowable deviation threshold. After calculating the consistency score, a two-dimensional arrangement is performed according to the spatial interaction pairing method to form a quality linkage matrix Q of size n×b, where n is the total number of path units and b is the total number of edge grid units. In the quality linkage matrix Q, the element Q in the i-th row and j-th column is ijRepresents the path unit P i With edge grid cell G j The job consistency score for the current job round. The quality linkage matrix reflects the job matching quality of each edge grid unit and the corresponding path unit.

[0054] It's important to note that the exponential term in the consistency scoring function reflects the sensitivity of the score to deviation. Even with high coverage and stability, if the deviation exceeds the threshold, the score will be rapidly weakened, ensuring that the scoring function has greater robustness and recognition ability when reflecting path quality. This mechanism effectively overcomes the insensitivity of traditional linear weighting functions to deviations, making it particularly suitable for identifying small differences in trajectories in the context of refined management of marginal areas.

[0055] It can be seen that the present invention realizes the quantification of operation trajectory quality and intelligent identification of regional differences by constructing a quality linkage matrix and dynamically evaluating path scores, which helps to quickly screen high-reliability path areas and enhance the accuracy and pertinence of the path generation and optimization process.

[0056] S4: Based on the preset operation specification template trajectory, the path unit set of the historical high-scoring area of ​​the edge grid unit is called, the trajectory path is reconstructed through the interpolation-correction method, and the offset coding vector group is generated according to the spatial offset vector between the original operation path and the reconstructed trajectory.

[0057] It is known that before performing the operation trajectory offset modeling and correction, it is necessary to first determine a reliable historical trajectory reference area to provide a high-quality data source for the reconstruction template. j In the corresponding path unit set of the historical operation round, extract the score value in the quality linkage matrix greater than the set threshold Q t A high-quality path reference set is constructed by using a subset of path units.

[0058] Furthermore, for each edge grid cell G j ,Based on the trajectory point group at the same spatial position in its ,high-quality path reference set, the regional local interpolation method ,is used to construct the fitting trajectory track, which is used as the ,job specification template trajectory fragment, i.e., the reconstructed trajectory path.

[0059] Exemplarily, after obtaining a high-quality path reference set within an edge grid unit, it is necessary to perform standardized aggregation processing on multiple historical paths within the same grid unit. Specifically, for each edge grid unit, based on the spatial distribution of path points in its reference set in a two-dimensional plane coordinate system, pairing and integration are performed according to the consistency of the trajectory point position index to generate a set of trajectory point groups corresponding to the spatial position. In order to ensure the smoothness and accuracy of trajectory reconstruction, local area interpolation methods such as spline interpolation or weighted least squares interpolation technology can be used. The present invention does not limit its acquisition method, and fits the historical trajectory points in each trajectory point position group. The constructed fitting curve can be regarded as a template trajectory segment of the edge grid area, which has the characteristics of directional continuity, morphological stability and compliance with the operation paradigm.

[0060] After the trajectory template is constructed, the trajectory offset calculation operation needs to be performed on the original path unit of the current operation round. Specifically, the edge grid unit of each original path unit and its spatial interaction is selected, and the Euclidean distance offset vector ΔV between each trajectory point in the trajectory path and the reconstructed trajectory is calculated in the two-dimensional coordinate system. ij (k), where k represents the trajectory point index.

[0061] To ensure stability and comparability in the subsequent encoding and modeling processes, the resulting offset vector sequence must be normalized. In this embodiment, a sliding window segmented mean-variance normalization method is used, segmenting the trajectory offset vector sequence into fixed-point windows (e.g., a sliding window length of 10 trajectory points). Within each window, the mean and standard deviation of the offset vector group are calculated and normalized to generate a standardized offset vector group.

[0062] Specifically, the path offset vector sequence ΔV ij (k) Perform segmented sliding window analysis and perform mean-variance normalization according to a fixed number of points to generate a standardized offset vector group ΔV ij (k) e , used as input for the encoding process; based on the normalized offset vector group ΔV ij (k) e , based on the location interval, the path unit P i At each edge grid cell G j The job offset state within the range generates the corresponding offset encoding vector group C ij , the coding structure has directionality and amplitude classification capabilities.

[0063] Furthermore, after generating the standardized offset vector group, this step enters the trajectory offset encoding stage. Its core goal is to discretize the two-dimensional offset information into an interpretable symbol vector to facilitate the use of downstream trajectory diagnosis and behavior classification models. Among them, the corresponding offset encoding vector group C is generated. ij The process is as follows: the normalized offset vector group ΔV ij (k) e The direction is divided into multiple quadrant segments, that is, the direction code D∈{0,1,2,…}. The direction can be divided into 4, 8 or 16 as needed, which is not limited in this embodiment. For example, it corresponds to: east, northeast, north,…, southeast, etc.; and the modulus length |ΔV ij (k) e | is divided into several amplitude intervals, for example, set as: small offset: [0, 0.3); medium offset: [0.3, 0.6); large offset: [0.6, 1]; the amplitude code is: A∈{0, 1, 2}, etc.; the direction code D and the amplitude code A are spliced ​​to form a composite coding unit.

[0064] Arranging these offset coding units in the order of trajectory points forms a set of offset coding vectors for the original path units within the current edge grid area. This coding preserves the directional trends and behavioral characteristics of the path in different spatial segments, and its coding structure has excellent directional differentiation and amplitude level resolution capabilities.

[0065] S5: Perform multi-dimensional quality index cluster analysis on the fragment path unit set according to the consistency score and the offset coding vector group, generate fragment quality grade labels, and form a linkage quality assessment layer.

[0066] Preferably, each path unit P is extracted from the quality linkage matrix i The consistency score vector of the path unit P i The offset coding vector group is converted into a frequency vector, and the cumulative occurrence intensity of different offset modes in each interactive grid is counted; for each path unit P i The multidimensional index feature vector composed of the consistency score vector and the frequency vector is combined to form a quality feature vector group for cluster analysis; the quality feature vector group is subjected to unsupervised clustering (such as DBSCAN or density-based partitioning, which is not limited to the present invention), and the fragment quality level label is generated according to the mean level of the consistency score vector of each cluster center; and the fragment quality level label is mapped to the path unit P i In the edge grid area, a linkage quality assessment layer is constructed to visualize the differences in regional operation quality.

[0067] It should be noted that to facilitate clustering, this coding group needs to be converted into a frequency-based statistical feature. That is, the frequency of occurrence of various offset coding modes (such as a direction code of 2 and an amplitude code of 1, indicating a moderate offset in the north direction) is counted for each path unit and normalized across the entire trajectory interval to form a one-dimensional linear vector. This frequency vector not only retains the distribution information of the offset direction and amplitude, but also has statistical versatility and clusterability. It can effectively characterize the directional preference and amplitude characteristics of path units that deviate from the standard trajectory during actual operation, providing basic support for subsequent quality difference identification.

[0068] Furthermore, for the constructed quality feature vector group, an unsupervised clustering algorithm is used to perform quality classification of path units. Since path units in different spatial regions may have characteristics such as uneven quantity, different density, irregular structure, etc., the present invention gives priority to using density clustering algorithms, such as DBSCAN or its variants, in clustering model selection. The embodiment of the present invention does not limit the algorithm exclusively. The algorithm does not rely on the preset number of categories, can automatically discover high-density areas and identify sparse or abnormal path units, and has strong adaptability and robustness. At the same time, during the clustering process, the fineness of quality grading can be controlled by adjusting parameters such as the minimum number of samples within the cluster and the search radius. Each cluster output by the clustering algorithm corresponds to a path quality grade cluster, and its internal members have high similarity, which can stably portray the comprehensive performance of consistency and offset behavior in trajectory operations.

[0069] After clustering is complete, the present invention further generates shard quality grade labels based on the mean level of the average consistency score vector for each cluster path unit. Specifically, the mean consistency score within each type of path unit is calculated and divided into several grade intervals, with each grade corresponding to a shard quality label, such as Q1, Q2, Q3, Q4, etc. This label division method based on cluster center statistics can effectively avoid the risk of misjudgment caused by manually set thresholds, while ensuring the internal consistency and classification stability of the labels.

[0070] Finally, the generated fragment quality grade labels are mapped to the edge grid unit area where the corresponding path unit is located, and a linkage quality assessment layer is constructed based on the spatial topological relationship. The so-called linkage quality assessment layer refers to a two-dimensional visual graphical structure that displays the spatial distribution of the quality grades of each path unit in different edge grid unit areas. This layer not only has the spatial linkage characteristics of the operation trajectory segments, but also integrates the trajectory quality clustering results, realizing the full process mapping from path characteristics-quality score-spatial labeling, which facilitates users to intuitively understand and analyze the regional operation quality differences at the map level. At the same time, the assessment layer can be further linked with the job scheduling system or feedback mechanism to form a closed-loop quality monitoring system.

[0071] In summary, this step constructs a high-dimensional feature space using the bidirectional indicators of consistency score and offset coding frequency, combines it with a density-based clustering algorithm to achieve unsupervised quality grading, and uses a spatial mapping mechanism to complete layered display. This not only improves the accuracy and interpretability of path quality assessment, but also enhances the expression and operability of spatial trajectory data.

[0072] like Figure 3 As shown, this embodiment also provides an agricultural equipment intelligent operation quality assessment system, including:

[0073] An edge construction module 100 is configured to construct an edge buffer area based on an outer contour point set of an operation area, and divide the buffer area into edge grid units;

[0074] The trajectory encoding module 200 is used to perform segmented encoding processing on the original operation trajectory data to form a set of segmented path units, and respectively extract the spatial interaction relationship between each path unit and the corresponding edge grid unit;

[0075] A linkage matrix module 300 is used to establish a quality linkage matrix based on the multi-parameter calculation results of the path unit and the edge grid unit, wherein each element in the quality linkage matrix reflects the multi-parameter consistency score of the corresponding edge grid unit and the slice path unit under the linkage relationship;

[0076] The trajectory reconstruction module 400 is used to call the path unit set of the historical high-scoring area of ​​the edge grid unit based on the preset operation specification template trajectory, reconstruct the trajectory path through interpolation and correction, and generate an offset coding vector group based on the spatial offset vector between the original operation path and the reconstructed trajectory;

[0077] The quality clustering module 500 is used to perform multi-dimensional quality index clustering analysis on the fragment path unit set according to the consistency score and the offset coding vector group, generate fragment quality level labels, and form a linkage quality assessment layer.

[0078] This embodiment also provides a computer device suitable for the intelligent operation quality assessment method of agricultural equipment, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the intelligent operation quality assessment method of agricultural equipment proposed in the above embodiment.

[0079] The computer device may be a terminal, comprising a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner may be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse.

[0080] This embodiment also provides a storage medium having a computer program stored thereon. When the program is executed by a processor, the method for evaluating the quality of intelligent operation of agricultural equipment proposed in the above embodiment is implemented.

[0081] In summary, the present invention achieves refined modeling of the boundary areas of agricultural equipment operations by constructing an edge buffer zone and dividing it into multi-directionally aligned edge grid units. Combining trajectory segmentation coding with spatial interaction analysis, a quality linkage matrix is ​​established to quantify the consistency between the trajectory and the operating specifications. Furthermore, through trajectory reconstruction and offset coding analysis, error modeling and standardization of the original path are achieved. Compared to existing assessment methods that rely solely on trajectory overlap or trajectory offset distance, the present invention integrates historical high-quality path reconstruction with a multidimensional clustering mechanism, effectively improving the accuracy and adaptability of quality assessment. Ultimately, a linkage quality assessment layer with spatial continuity and difference sensitivity is formed, providing highly reliable decision support for the operational quality assessment of agricultural equipment.

[0082] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for evaluating the quality of intelligent operation of agricultural equipment, characterized by: include: Constructing an edge buffer area based on the outer contour point set of the operation area, and dividing the buffer area into edge grid units; The original operation trajectory data is segmented and encoded to form a set of segmented path units, and the spatial interaction relationship between each path unit and the corresponding edge grid unit is extracted respectively; Based on the multi-parameter calculation results of the path unit and the edge grid unit, a quality linkage matrix is ​​established, wherein each element in the quality linkage matrix reflects the multi-parameter consistency score of the corresponding edge grid unit and the fragment path unit under the linkage relationship; Based on the preset operation specification template trajectory, the path unit set of the historical high-scoring area of ​​the edge grid unit is called, the trajectory path is reconstructed through the interpolation-correction method, and the offset coding vector group is generated according to the spatial offset vector between the original operation path and the reconstructed trajectory; A multi-dimensional quality index cluster analysis is performed on the fragment path unit set according to the consistency score and the offset coding vector group to generate fragment quality grade labels and form a linkage quality assessment layer.

2. The method for evaluating the quality of intelligent operation of agricultural equipment according to claim 1, wherein: The division of the edge grid unit includes: performing contour thinning processing on the outer trajectory nodes in the operation path data, extracting the boundary point set and then constructing the minimum outer polygon boundary; Calculating the local curvature index of each boundary segment based on the directional change rate of the boundary, setting the variant buffer width according to the curvature value, and generating an asymmetric outward expansion buffer area; The outward expansion buffer is divided into several local area sub-blocks according to the boundary normal direction, and grid rotation segmentation is performed within each sub-block based on the local direction to generate edge grid units aligned in multiple directions.

3. The method for evaluating the quality of intelligent operation of agricultural equipment according to claim 2, wherein: The segment encoding process is based on job timestamps and spatial segment lengths.

4. The method for evaluating the quality of intelligent operation of agricultural equipment according to claim 3, wherein: The multiple parameters include: spatial coverage, track deviation and track stability.

5. The method for evaluating the quality of intelligent operation of agricultural equipment according to claim 1, wherein: The reconstruction of the trajectory path includes: extracting the path unit set corresponding to the edge grid unit in the historical operation round, and extracting the score value greater than the set threshold Q in the quality linkage matrix. t A subset of path units is used to build a high-quality path reference set; For each edge grid cell, based on the trajectory point group at the same spatial position in its high-quality path reference set, the regional local interpolation method is used to construct the fitting trajectory track, which is used as the operation specification template trajectory segment, that is, the reconstructed trajectory path.

6. The method for evaluating the quality of intelligent operation of agricultural equipment according to claim 5, wherein: The generation of the offset coding vector group includes: selecting the edge grid unit of each original path unit and its spatial interaction, calculating the Euclidean distance offset vector ΔV between each trajectory point in the trajectory path and the reconstructed trajectory in a two-dimensional coordinate system ij (k), where k represents the trajectory point index; the path offset vector sequence ΔV ij (k) Perform segmented sliding window analysis and perform mean-variance normalization according to a fixed number of points to generate a standardized offset vector group ΔV ij (k) e ; Normalize the offset vector group ΔV ij (k) e Divide the direction into multiple quadrant segments, generate the direction code D, and convert the modulus |ΔV ij (k) e | Divide into several amplitude interval segments to generate amplitude code A; splice the direction code D and the amplitude code A to form a composite coding unit.

7. The method for evaluating the quality of intelligent operation of agricultural equipment according to claim 1, wherein: The multidimensional quality index cluster analysis includes: Extract the consistency score vector of each path unit in the quality linkage matrix; Convert the offset encoding vector group of the path unit into a frequency vector; For each path unit, the consistency score vector and the frequency vector are combined to form a multidimensional index feature vector, forming a quality feature vector group for cluster analysis; Perform unsupervised clustering on groups of quality feature vectors.

8. A system for evaluating the quality of intelligent operation of agricultural equipment, based on the method for evaluating the quality of intelligent operation of agricultural equipment according to any one of claims 1 to 7, characterized in that: Also includes: An edge construction module is used to construct an edge buffer area based on the outer contour point set of the operation area, and divide the buffer area into edge grid units; The trajectory encoding module is used to perform segmented encoding processing on the original operation trajectory data to form a set of segmented path units and extract the spatial interaction relationship between each path unit and the corresponding edge grid unit; A linkage matrix module is used to establish a quality linkage matrix based on the multi-parameter calculation results of the path unit and the edge grid unit. Each element in the quality linkage matrix reflects the multi-parameter consistency score of the corresponding edge grid unit and the fragment path unit under the linkage relationship; The trajectory reconstruction module is used to call the path unit set of the historical high-scoring area of ​​the edge grid unit according to the preset operation specification template trajectory, reconstruct the trajectory path through interpolation-correction, and generate an offset coding vector group based on the spatial offset vector between the original operation path and the reconstructed trajectory; The quality clustering module is used to perform multi-dimensional quality indicator clustering analysis on the fragment path unit set according to the consistency score and offset coding vector group, generate fragment quality grade labels, and form a linkage quality assessment layer.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for evaluating the quality of intelligent operation of agricultural equipment according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for evaluating the quality of intelligent operation of agricultural equipment according to any one of claims 1 to 7 are implemented.

Images