An agricultural equipment intelligent operation quality evaluation method and system

Abstract

The application discloses an agricultural equipment intelligent operation quality evaluation method and system, and relates to the technical field of quality evaluation. The method comprises the following steps: dividing a buffer area into edge grid units; performing segmented coding processing on original operation track data to form a set of fragmented path units, and extracting the spatial interaction relationship between each path unit and the corresponding edge grid unit; establishing a quality linkage matrix based on the calculation results of multiple parameters; calling the path unit set of the historical high-scoring area of the edge grid unit to reconstruct the trajectory path according to a preset operation specification template track, and generating a set of offset coding vectors according to the spatial offset vectors between the original operation path and the reconstructed trajectory; performing multidimensional quality index clustering analysis on the consistency score and the set of offset coding vectors to generate a fragmented quality level label and form a linkage quality evaluation layer. The application combines historical high-quality path reconstruction and multidimensional clustering mechanism, and effectively improves the accuracy and adaptability of quality evaluation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of quality evaluation, in particular to an agricultural equipment intelligent operation quality evaluation method and system. BACKGROUND

[0002] With the continuous improvement of agricultural mechanization and intelligence level, agricultural equipment plays an increasingly key role in multiple links such as plowing, seeding, fertilizing, spraying and harvesting. However, the existing agricultural operation quality evaluation methods are mostly based on global trajectory as a unit for overall analysis, which is difficult to reflect the differences of local operation quality and the spatial linkage relationship between regions in detail, especially near the boundary of the operation area. Due to frequent adjustment or turning of the operation path, the evaluation error is significantly increased. In addition, the traditional method often ignores the interaction between the operation path and the spatial structure of the operation area, lacks systematic analysis of trajectory reconstruction, trajectory deviation and historical standard operation mode, resulting in poor explainability and pertinence of the evaluation results. On the other hand, there is currently a lack of a unified evaluation framework that can integrate trajectory coding, spatial interaction analysis, historical high-score trajectory mining and multi-dimensional clustering algorithm, which is difficult to support dynamic quality label generation and layer visualization expression of operation trajectory, and is not conducive to subsequent intelligent decision-making and improvement.

[0003] CN109099925B discloses a method for unmanned agricultural machine navigation path planning and operation quality evaluation, which uses an unmanned remote control airplane to collect field boundary vertex information, establishes a field coordinate model through an RTK positioning device, realizes path planning on a platform main controller, and sends the path to the agricultural machine for autonomous operation. Meanwhile, it collects operation images through a visual sensor to realize remote monitoring and operation quality evaluation. However, this method mainly faces path planning and image feedback of a single agricultural machine, lacks quantitative evaluation means for trajectory coverage, overlapping operation area and task completion degree in the case of multiple agricultural machines working together, and is difficult to meet the needs of large-area efficient plowing and transparent management of operation progress.

[0004] CN119515134A discloses a method for real-time evaluation of operation area and quality of multiple agricultural machines in the same field, which obtains field boundary points and non-cultivated area information through a drone, constructs a field graphic model, collects agricultural machine trajectory information through a GPS, draws an operation path, and realizes real-time evaluation of effective operation area and task completion degree of each agricultural machine by calculating the overlapping area of the trajectory and the cultivated land. This method effectively improves the visual management level of multiple agricultural machines working together, but it still relies on trajectory data itself and lacks the ability to identify details of ground operation quality based on image analysis, which cannot realize real-time identification and feedback of path deviation, missed plowing and other abnormal behaviors.

[0005] Therefore, there is a need for an agricultural equipment intelligent operation quality evaluation scheme. SUMMARY

[0006] In view of the problems existing in the quality evaluation technology of the existing agricultural equipment, the present application is proposed.

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

[0008] To solve the above technical problems, the present application provides the following technical solutions:

[0009] In a first aspect, the present application provides an intelligent operation quality evaluation method for agricultural equipment, which comprises: constructing an edge buffer zone area according to an outer contour point set of an operation area, dividing the buffer zone into edge grid units; performing segmented encoding processing on original operation trajectory data to form a set of fragmented path units, and extracting the spatial interaction relationship between each path unit and the corresponding edge grid unit; based on the calculation results of multiple parameters of the path unit and the edge grid unit, establishing a quality linkage matrix, each element in the quality linkage matrix reflecting the multiple parameter consistency score of the corresponding edge grid unit and the fragmented path unit under the linkage relationship; according to a preset operation specification template trajectory, calling the path unit set of the historical high-score area of the edge grid unit, reconstructing the trajectory path through interpolation-correction, and generating a set of offset encoding vectors according to the spatial offset vector between the original operation path and the reconstructed trajectory; performing multi-dimensional quality index clustering analysis on the set of fragmented path units according to the consistency score and the set of offset encoding vectors, generating a fragmented quality level label, and forming a linkage quality evaluation layer.

[0010] As a preferred scheme of the intelligent operation quality evaluation method for agricultural equipment, the division of the edge grid unit comprises: constructing a minimum outer polygon boundary by extracting a boundary point set after performing contour thinning processing on the outer trajectory nodes in the operation path data; calculating the local curvature index of each boundary in combination with the direction change rate of the boundary, setting a variable-width buffer width according to the curvature value, and generating an asymmetric outward expansion buffer area; dividing the outward expansion buffer area into a plurality of local area sub-blocks according to the normal direction of the boundary, and performing grid rotation segmentation in each sub-block based on the local direction to generate edge grid units aligned in multiple directions.

[0011] As a preferred scheme of the intelligent operation quality evaluation method for agricultural equipment, the segmented encoding processing is based on operation time stamp and spatial segment length.

[0012] As a preferred scheme of the intelligent operation quality evaluation method for agricultural equipment, the multiple parameters include: spatial coverage, trajectory offset degree, and trajectory stability.

[0013] As a preferred scheme of the agricultural equipment intelligent operation quality evaluation method, wherein: according to the corresponding path unit set of the edge grid unit in the historical operation round, a path unit subset with a score value greater than a set threshold Q in the quality linkage matrix is extracted, and a high-quality path reference set is constructed; for each edge grid unit, based on the trajectory point group at the same spatial position in the high-quality path reference set, a regional local interpolation method is used to construct a fitted trajectory track, which is used as an operation specification template trajectory segment, that is, the trajectory path is reconstructed. t

[0014] As a preferred scheme of the agricultural equipment intelligent operation quality evaluation method, wherein: the generation of the offset encoding vector group includes: selecting the edge grid unit interacting with each original path unit, calculating the Euclidean distance offset vector ΔV ij (k) between each trajectory point in the trajectory path and the reconstructed trajectory in the two-dimensional coordinate system, wherein k represents the trajectory point index; segmenting and sliding window analyzing the path offset vector sequence ΔV ij (k), performing mean-variance normalization processing according to a fixed point number window, generating a standardized offset vector group ΔV ij (k) e ; dividing the standardized offset vector group ΔV ij (k) e into multiple quadrant segments according to the direction to generate direction encoding D, and dividing the module length |ΔV ij (k) e into several amplitude interval segments to generate amplitude encoding A; and splicing the direction encoding D and the amplitude encoding A to form a composite encoding unit.

[0015] As a preferred scheme of the agricultural equipment intelligent operation quality evaluation method, wherein: the multi-dimensional quality index clustering analysis includes: extracting a consistency score vector of each path unit in the quality linkage matrix; converting the offset encoding vector group of the path unit into a frequency straight direction vector; for each path unit, the consistency score vector and the frequency straight direction vector are combined to form a multi-dimensional index feature vector, and a quality feature vector group for clustering analysis is formed; and the quality feature vector group is subjected to non-supervised clustering.

[0016] ​In a second aspect, the present application provides an agricultural equipment intelligent operation quality evaluation system, which comprises: an edge construction module, configured to construct an edge buffer region according to an outer contour point set of an operation area, and divide the buffer region into edge grid units; a trajectory coding module, configured to perform segmented coding processing on original operation trajectory data, form a set of fragmented path units, and extract the spatial interaction relationship between each path unit and a corresponding edge grid unit; a linkage matrix module, configured to establish a quality linkage matrix based on the multi-parameter calculation results of the path units and the edge grid units, wherein each element in the quality linkage matrix reflects the multi-parameter consistency score of the corresponding edge grid unit and the fragmented path unit under the linkage relationship; a trajectory reconstruction module, configured to call the path unit set of the historical high-score region of the edge grid unit according to a preset operation specification template trajectory, reconstruct the trajectory path through an interpolation-correction method, and generate a set of offset coding vectors according to the spatial offset vectors between the original operation path and the reconstructed trajectory; and a quality clustering module, configured to perform multi-dimensional quality index clustering analysis on the set of fragmented path units according to the consistency score and the set of offset coding vectors, generate a fragmented quality level label, and form a linkage quality evaluation layer.

[0017] In a third aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program instructions are executed by the processor, the steps of the agricultural equipment intelligent operation quality evaluation method according to the first aspect of the present application are implemented.

[0018] In a fourth aspect, the present application 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 agricultural equipment intelligent operation quality evaluation method according to the first aspect of the present application are implemented.

[0019] The present application has the following beneficial effects: the present application realizes fine modeling of the boundary region of the agricultural equipment operation by constructing an edge buffer region and dividing it into edge grid units in multiple directions; the quality linkage matrix is established by combining trajectory segmented coding and spatial interaction analysis, and the consistency between the trajectory and the operation specification is quantified; then, the error modeling and standardized processing of the original path are realized through trajectory reconstruction and offset coding analysis. Compared with the existing evaluation methods which only rely on trajectory overlap rate or trajectory offset distance, the present application integrates historical high-quality path reconstruction and multi-dimensional clustering mechanism, effectively improves the accuracy and adaptability of quality evaluation, and finally forms a linkage quality evaluation layer with spatial continuity and difference sensitivity, thereby providing high-reliability decision support for the operation quality evaluation of agricultural equipment. BRIEF DESCRIPTION OF DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments description. Obviously, the drawings in the following description only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0021] Figure 1 Flow chart of the agricultural equipment intelligent operation quality evaluation method;

[0022] Figure 2 Flow chart of the edge grid unit division of the agricultural equipment intelligent operation quality evaluation method;

[0023] Figure 3 Structure diagram of the agricultural equipment intelligent operation quality evaluation system. DETAILED DESCRIPTION

[0024] In order to make the above-mentioned objects, features and advantages of the present application more apparent and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the drawings of the specification.

[0025] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application, but the present application can also be implemented in other ways different from the description, and those skilled in the art can make similar generalizations without departing from the connotation of the present application, therefore the present application is not limited by the specific embodiments disclosed below.

[0026] Secondly, the "one embodiment" or "embodiment" referred to herein means that the specific features, structures or characteristics can be included in at least one implementation of the present application. "In one embodiment" appearing in different places in the specification does not mean the same embodiment, nor is it an independent or selective embodiment that excludes other embodiments.

[0027] As described in the above background, the existing agricultural operation quality evaluation method is mostly analyzed as a whole unit of global trajectory, which is difficult to reflect the difference of local operation quality and the spatial linkage relationship between regions in detail, especially near the boundary position of the operation area, the evaluation error increases significantly due to frequent adjustment or turning of the operation path. In addition, the traditional method often ignores the interaction between the operation path and the spatial structure of the operation area, lacks systematic analysis of trajectory reconstruction, trajectory deviation and historical standard operation mode, resulting in poor explainability and pertinence of the evaluation results. On the other hand, there is still a lack of a unified evaluation framework that can integrate trajectory coding, spatial interaction analysis, historical high-score trajectory mining and multi-dimensional clustering algorithm, which is difficult to support dynamic quality label generation and layer visualization expression of operation trajectory, and is not conducive to subsequent intelligent decision and improvement. Therefore, an agricultural equipment intelligent operation quality evaluation scheme is needed.

[0028] Figure 1 This is a flowchart of an intelligent operation quality assessment method for agricultural equipment according to an embodiment of the present invention.

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

[0030] S1: Construct an edge buffer region based on the set of outer contour points of the work area, and divide the buffer into edge grid units.

[0031] Better, such as Figure 2 As shown, the edge mesh cell division includes the following steps: Contour thinning is performed on the outer trajectory nodes in the work path data to extract the boundary point set and construct a minimum outer polygon boundary, which is used to define the range of the edge work area to be evaluated; the local curvature index of each boundary segment is calculated based on the direction change rate of the boundary, and a variable buffer width is set according to the curvature value to generate an asymmetric outward expansion buffer region. The buffer width is increased in the boundary corner region to improve edge resolution; the buffer is divided into several local region sub-blocks according to the boundary normal direction, and mesh rotation is performed within each sub-block based on the local direction to generate multi-directional aligned edge mesh cells G. j This avoids non-equidistant errors caused by orthogonal segmentation.

[0032] It should be noted that constructing an edge buffer region based on the outer contour point set of the work area aims to achieve accurate monitoring of boundary quality during agricultural equipment operation through structured modeling of the spatial boundary of the work data. Specifically, firstly, based on the historical path trajectory dataset collected by the agricultural equipment during operation, path nodes located at the work boundary are extracted to form the original contour point set. These path nodes refer to sampling points located at the outermost edge of the trajectory, and their positioning information is usually provided by positioning systems such as GNSS and RTK; this embodiment does not impose a unique limitation. In actual operation, due to the uneven distribution density and local abrupt changes of boundary path nodes, directly using these points for boundary construction may result in redundant and complex contours, affecting the accuracy of subsequent calculations. Therefore, contour thinning is adopted in this step, that is, the original contour point set is resampled by setting angle change thresholds and point distance thresholds to remove redundant points in the trajectory and retain turning points with obvious structural features of the boundary, making the boundary data more regular and concise in expression.

[0033] Next, based on the extracted boundary point set, a minimum bounding polygon boundary is constructed to define the geometric boundary range of the edge buffer. The minimum bounding polygon boundary refers to the polygon with the smallest area and the fewest vertices that contains all contour points on the two-dimensional plane. This structure exhibits high stability and low computational complexity in computational geometry. In this invention, the Andrew algorithm is used to construct the minimum bounding boundary, and the geometric connectivity and closure of the generated boundary are controlled by the arrangement order of the boundary points. For example, assuming the contour point set is: A(0,0), B(2,1), C(4,0), D(3,3), E(1,4), the points are first arranged in ascending order of their x-coordinates: A→B→E→D→C; then, the Andrew convex hull construction is performed, automatically removing the concave point B and retaining points A, C, D, and E, forming the minimum bounding convex polygon ACDE; finally, the boundary closure is ensured by the arrangement order (e.g., clockwise A→C→D→E→A), forming the basic boundary for buffer calculation. This boundary serves as the basic framework for the subsequent construction of buffer zones. 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 is constructed, the local curvature index of each boundary segment is calculated by combining the directional change rate between boundary points, serving as a buffer width variation parameter. The directional change rate refers to the change in the angle between the directional vectors of adjacent boundary segments, used to measure the degree of boundary deflection at that location. Conventional buffer construction methods typically use a fixed width to equidistantly expand the boundary. However, due to the tendency for trajectory deviation and insufficient coverage in agricultural operations at corners, a fixed width expansion is insufficient to cover high-risk areas, leading to incomplete analysis. Therefore, this invention designs a curvature-driven buffer width control method, setting different buffer expansion scales based on the local curvature of each boundary segment. A larger width is expanded in areas with drastic boundary changes (i.e., corner areas), while a smaller expansion width is maintained in linear areas, thus generating a variable asymmetric buffer zone. For example, suppose a certain operation path boundary has an "L"-shaped corner structure, with one segment being a horizontal line segment (curvature ≈ 0) and the other a 90° corner area (high curvature). When constructing the buffer, a buffer width of 2 meters is set for horizontal line segments; for corner areas, based on their high curvature characteristics, the buffer width is automatically increased to 5 meters. This ensures that more potential offset trajectories are covered at the corners, enhancing the accuracy of boundary detection. This strategy enhances the spatial perception capability of boundary detail areas and improves the spatial resolution of edge risk assessment.

[0035] Furthermore, the constructed buffer zone is further divided into local regions according to the normal direction of the boundary. This division method differs from the conventional approach of using orthogonal mesh partitioning. This invention partitions the buffer zone into fan-shaped sub-blocks based on the normal direction of each boundary segment, ensuring that the partitioning direction of each sub-block is consistent with the local boundary direction. This locally consistent partitioning strategy avoids the misalignment problem in corner areas of conventional Cartesian meshes, reducing issues such as uneven edge mesh area and resolution degradation caused by directional deviations. Next, mesh cell generation is performed within each local partition. Specifically, a rotating coordinate system is constructed in each sub-block based on the local boundary normal direction, and a two-dimensional equidistant mesh is partitioned based on this coordinate system. The size of the mesh cell is set according to the point density within the buffer zone and the working width of the equipment, typically maintained at 1 / 3 to 1 / 5 of the equipment coverage width, to ensure sufficient trajectory point coverage density within each mesh cell to support quality assessment. During the partitioning process, the geometric continuity and overlap ratio between mesh cells must also be considered to prevent statistical errors in trajectory overlap caused by mesh interlacing.

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

[0037] In summary, this invention not only establishes a high-precision edge space determination framework at the geometric modeling level, but also enhances spatial perception and improves edge recognition accuracy through local curvature-driven buffer variation construction and rotation-aligned mesh generation strategies. This design effectively avoids problems such as amplified directional errors, inconsistent resolution, and cutting distortion in traditional boundary processing, providing high-quality, structured edge data support for agricultural equipment operation quality assessment. Especially in application scenarios with strong discontinuities in boundary area operations and unstable equipment travel paths, the edge mesh unit generation method provided in this step effectively supports subsequent key processes such as path offset analysis, trajectory quality classification, and operation consistency scoring, making it one of the fundamental steps in achieving intelligent analysis of the overall system.

[0038] S2: 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.

[0039] Preferably, the process of forming the path unit set includes: collecting timestamps and corresponding coordinate points from the work trajectory data to generate an ordered trajectory point set, which serves as the basic data source for path segmentation and spatial matching; performing time-space synchronous segmentation on the ordered trajectory point set, and setting the spatial segment length L. s Control path length, time interval T s By controlling the sparsity of nodes, multiple continuous path segments are formed, denoted as path unit sets P1 to P2. n , where n is the number of path units.

[0040] Specifically, in the operational process of this invention, the first step is to collect the timestamp sequence and corresponding two-dimensional geographic coordinate point pairs from the original work trajectory data to form a temporal trajectory point set. This trajectory point set typically originates from the real-time work trajectory data recorded by equipment with positioning modules, such as agricultural machinery and garden machinery. This data has both temporal and spatial characteristics, including both the spatial distribution of the work path and the dynamic process of the work behavior. Therefore, the organization of the trajectory point set in the initial stage is particularly crucial. This invention uses timestamps as the primary index to ensure the temporal consistency of the trajectory point order, 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 also needs to undergo preprocessing steps such as noise reduction and interpolation to eliminate abnormal jump points and signal drift, thereby enhancing the spatial logic and matching reliability of the path.

[0041] Furthermore, this invention introduces a time-space synchronous segmentation mechanism during the formation of the path unit set. The core of this mechanism lies in achieving ordered segmentation of the trajectory by setting two control parameters: spatial segment length and time interval. The spatial segment length controls the maximum length of the path unit, preventing a single path unit from covering too many grid units and causing fuzzy matching results. The time interval controls the density and continuity of trajectory nodes, avoiding node concentration or sparsity issues caused by speed changes or dwell time during operation. For example, when the equipment is in a slow turning phase at the boundary, the time interval may be shortened to improve path point accuracy, while during a constant-speed straight-line phase, the time interval can be appropriately increased to improve processing efficiency.

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

[0043] Furthermore, each path unit P i Spatial range and edge grid unit Gj The set is used to determine the intersection of the meshes and identify the interactive mesh subset G(P) corresponding to the path unit. i ), and record its grid code as an interactive key value.

[0044] Specifically, in actual operation, the first step is to obtain the edge buffer mesh division result generated in step S1, i.e., the mesh cell set, and then process each path cell P. i Perform spatial extent calculations, construct its minimum envelope rectangle or circumscribed convex hull, and then combine it with each G in the mesh cell set. j Perform intersection determination. Spatial intersection determination can be performed in two ways: one is based on geometric Boolean operations for surface intersection calculation, determining P... i Does the path geometry range match G? j The boundary contours have non-empty intersections; another method is based on the projection of trajectory points, statistically analyzing P. i Do the points on each trajectory fall into G? j Inside the boundary. When the interaction condition is met, P can be considered... i With G j There exists a spatial coupling relationship, denoted as (P i G j ).

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

[0046] Better, for each pair of path units P i With the corresponding interactive grid subset G(P) i Extract the area of ​​the overlapping region in the planar space and compare it with the corresponding interactive mesh subset G(P). i The spatial coverage R is calculated from the area ratio of ) ij Based on the standard trajectory reference line, for each path cell, at the edge mesh cell G j The deviation index O is obtained by calculating the average offset distance between the actual trajectory and the standard trajectory reference line. ij It is used to characterize the degree of deviation from the work path.

[0047] For example, within the spatial interaction range of path units and grid units, for each pair of interaction items (P) i G j To extract the area of ​​the overlapping region in planar space, we need to perform Boolean geometric operations, specifically for path unit P. i Spatial profile and grid unit G jPerform an intersection operation on the boundaries and calculate the area of ​​their non-empty intersection. Further, use this intersection area and the mesh cell G... j The ratio of its own area is normalized to obtain the spatial coverage index R. ij , indicating the degree of job coverage of the path cell to that grid cell. Where R ij The larger the value, the more complete the job coverage of the path cell within the current grid range, which can be used as a geometric metric for job sufficiency.

[0048] Secondly, regarding the accuracy analysis of the work path, this invention introduces a trajectory deviation index. For this, a standard trajectory reference line needs to be preset. This reference line can be generated from historical high-quality work trajectories, planned paths, or manually defined trajectories; this invention does not impose a single limitation on its acquisition method. The path unit P... i In grid cell G j The actual trajectory points within the range are matched with 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 that unit within the grid. The smaller this value, the closer the trajectory is to the reference line, and the higher the accuracy of the working path. Conversely, a larger value indicates that the trajectory has drifted or veered significantly, affecting the working accuracy.

[0049] A better approach is to utilize the variance of the direction angle σ of the path point set. ij With the rate of change of velocity v ij Jointly construct the stability scoring function: S ij =exp(-(σ ij +v ij This reflects whether there are unstable behaviors such as abnormal sharp turns or sudden speed changes in the operation trajectory.

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

[0051] Furthermore, based on the above three indicators, a consistency scoring function is constructed, with the following expression:

[0052]

[0053] Among them, O max This represents the maximum allowable offset threshold. After calculating the consistency score, the elements are arranged in two dimensions according to the spatial interaction pairing method, forming a mass linkage matrix Q of size n×b, where n is the total number of path elements and b is the total number of edge grid elements. The element Q in the i-th row and j-th column of the mass linkage matrix Q... ijRepresents path unit P i With edge mesh unit G j The job consistency score under the current job round. The quality linkage matrix reflects the job matching quality between each edge grid cell and its corresponding path cell.

[0054] It should be noted that the exponential term in the consistency scoring function reflects the sensitive impact of deviation on the score. Even with high coverage and stability, if the deviation exceeds the threshold, the score will be rapidly weakened, ensuring that the scoring function has stronger robustness and recognition ability when reflecting path quality. This mechanism effectively overcomes the problem of traditional linear weighting functions being insensitive to deviation behavior recognition, and is particularly suitable for trajectory micro-difference discrimination in scenarios of fine-grained management of edge areas.

[0055] As can be seen, by constructing a quality linkage matrix and dynamically evaluating path scores, this invention achieves the quantification of work trajectory quality and intelligent identification of regional differences, which helps to quickly screen high-reliability path areas and enhances the accuracy and pertinence of the path generation and optimization process.

[0056] S4: Based on the preset operation specification template trajectory, call the path cell set of historical high-scoring areas of edge grid cells, reconstruct the trajectory path through interpolation-correction, and generate an offset encoding vector group based on the spatial offset vector between the original operation path and the reconstructed trajectory.

[0057] As is known, before performing operation trajectory offset modeling and correction, it is first necessary to determine a reliable historical trajectory reference area to provide a high-quality data source upon which the reconstruction template depends. Therefore, based on the edge mesh element G... j From the path unit set corresponding to the historical operation rounds, extract the quality linkage matrix from those with a score value greater than a set threshold Q. t A subset of path units is used to construct a high-quality path reference set.

[0058] Furthermore, for each edge mesh cell G j Based on the trajectory point group at the same spatial location in its high-quality path reference set, the fitting trajectory track is constructed using the regional local interpolation method, which is used as the trajectory segment of the operation specification template, that is, the trajectory path is reconstructed.

[0059] For example, after obtaining a high-quality path reference set within the edge grid cell, multiple historical paths within the same grid cell need to be standardized and aggregated. Specifically, for each edge grid cell, based on the spatial distribution of path points in its reference set in the two-dimensional plane coordinate system, they are paired and integrated according to the consistency of trajectory point position indices to generate a set of trajectory point groups corresponding to spatial positions. To ensure the smoothness and accuracy of trajectory reconstruction, local region interpolation methods can be used, such as spline curve interpolation or weighted least squares interpolation techniques. This invention does not limit the acquisition method, but fits the historical trajectory points in each trajectory point position group. The constructed fitted curve can be regarded as a template trajectory segment of the edge grid region, possessing the characteristics of directional continuity, morphological stability, and conformity to the operational paradigm.

[0060] After the trajectory template is constructed, the trajectory offset of the original path unit in the current job round needs to be calculated. Specifically, the edge grid unit that interacts with each original path unit in space is selected, and the Euclidean distance offset vector ΔV between each trajectory point in the trajectory path and the reconstructed trajectory is calculated in a two-dimensional coordinate system. ij (k), where k represents the trajectory point index.

[0061] To ensure the stability and comparability of subsequent encoding and modeling processes, the obtained offset vector sequence needs to be normalized. In this embodiment, a sliding window segmented mean-variance normalization method is used to segment the trajectory offset vector sequence into windows with a fixed number of points (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 standardized to generate a standardized offset vector group.

[0062] Specifically, the path offset vector sequence ΔV ij (k) Perform piecewise sliding window analysis, and perform mean-variance normalization on a fixed-point window to generate a standardized offset vector set ΔV. ij (k) e , used as input for the encoding process; based on the standardized offset vector group ΔV ij (k) e Based on the location interval, for path unit P i In each edge grid cell G j The job offset status within the range generates the corresponding offset encoding vector group C. ij The encoding structure has directional and amplitude classification capabilities.

[0063] Furthermore, after generating the standardized offset vector set, this step enters the trajectory offset encoding stage. Its core objective is to discretize the two-dimensional offset information into interpretable symbolic vectors, facilitating use by downstream trajectory diagnosis and behavior classification models. This involves generating the corresponding offset encoding vector set C. ij The process is as follows: standardize the offset vector group ΔV ij (k) e The direction is divided into multiple quadrant segments, i.e., the direction code D∈{0,1,2,…}. The direction can be divided into 4, 8, or 16, etc., as needed. This embodiment does not limit this, for example, corresponding to: due east, northeast, due north, …, southeast, etc.; and the modulus |ΔV ij (k) e | Divide into several amplitude intervals, for example, set as: small offset: [0,0.3); medium offset: [0.3,0.6); large offset: [0.6,1]; amplitude encoding: A∈{0,1,2} etc.; concatenate the direction encoding D and the amplitude encoding A to form a composite encoding unit.

[0064] Arranging the aforementioned offset coding units in order of trajectory points forms the offset coding vector group of the original path unit in the current edge grid region. This coding preserves the directional trend and behavioral characteristics of the path in different spatial segments, and its coding structure has good directional discrimination and amplitude level analysis capabilities.

[0065] S5: Perform multi-dimensional quality index clustering analysis on the fragmented path unit set according to the consistency score and offset encoding vector group to generate fragmented quality level labels and form a linked quality assessment layer.

[0066] A better approach is to extract each path unit P from the quality linkage matrix. i Consistency score vector; path unit P i The offset encoding vector group is converted into a frequency vector, and the cumulative occurrence intensity of different offset modes in each interactive grid is statistically analyzed; for each path cell P i A multi-dimensional index feature vector, composed of the joint consistency score vector and the frequency directive, is formed to create a quality feature vector set for cluster analysis. Unsupervised clustering (such as DBSCAN or density-based partitioning, which is not specifically limited to this method) is performed on the quality feature vector set, and piecewise quality level labels are generated based on the mean level of the consistency score vector of each cluster center. These piecewise quality level labels are then mapped to path units P. i Within the edge grid area, a linked quality assessment layer is constructed to visualize the differences in work quality across the area.

[0067] It should be noted that, for ease of clustering, this coding group needs to be converted into a frequency-based statistical feature. This involves statistically analyzing the frequency of various offset coding patterns (e.g., direction coding of 2, amplitude coding of 1, i.e., offset in the north direction) for each path unit, and then normalizing this data across the entire trajectory range to form a one-dimensional linear vector. This frequency vector not only preserves the distribution information of offset direction and amplitude but also possesses statistical universality and clustering ability. It can effectively characterize the directional preferences and amplitude features of path units deviating from the standard trajectory during actual operation, providing fundamental support for subsequent quality difference identification.

[0068] Furthermore, for the constructed quality feature vector group, an unsupervised clustering algorithm is used to classify the quality of path units. Since path units may exhibit uneven quantity, different densities, and irregular structures in different spatial regions, this invention prioritizes density-based clustering algorithms, such as DBSCAN or its variants, in selecting the clustering model. This embodiment does not impose a single algorithm limitation. This algorithm does not rely on a preset number of categories, can automatically discover high-density regions and identify sparse or abnormal path units, and has strong adaptability and robustness. Simultaneously, during the clustering process, the fineness of the quality classification can be controlled by adjusting parameters such as the minimum number of samples within a cluster and the search radius. Each cluster output by the clustering algorithm corresponds to a path quality level cluster, whose members have high similarity, and can stably characterize the comprehensive performance of consistency and deviation behavior in trajectory operations.

[0069] After clustering, this invention further generates segment quality level labels based on the mean level of the average consistency score vector of each cluster path unit. Specifically, the mean consistency score within each type of path unit is calculated and divided into several level intervals, with each level corresponding to a segment 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 setting thresholds, while ensuring the inherent consistency and classification stability of the labels.

[0070] Finally, the generated segmented quality level labels are mapped to the corresponding edge grid cell regions of the path units, and a linked quality assessment layer is constructed based on spatial topological relationships. This linked quality assessment layer is a two-dimensional, visual graphical structure that displays the spatial distribution of quality levels for each path unit across different edge grid cell regions. This layer not only possesses the spatial linkage characteristics of job trajectory segments but also integrates trajectory quality clustering results, achieving a complete mapping from path features to quality scores to spatial labels. This facilitates intuitive understanding and layer-level analysis of regional job quality differences for users. Furthermore, this assessment layer can be further linked with job scheduling systems or feedback mechanisms to form a closed-loop quality monitoring system.

[0071] In summary, this step constructs a high-dimensional feature space using a two-way index of consistency score and offset coding frequency, achieves unsupervised quality classification by combining density clustering algorithm, and completes layered display by using spatial mapping mechanism. This not only improves the accuracy and interpretability of path quality assessment, but also enhances the expressiveness and operability of spatial trajectory data.

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

[0073] The edge construction module 100 is used to construct an edge buffer region based on the outer contour point set of the work area, and divide the buffer 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 to extract the spatial interaction relationship between each path unit and the corresponding edge grid unit.

[0075] The 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. Each element in the quality linkage matrix reflects the multi-parameter consistency score of the corresponding edge grid unit and the piecewise path unit under the linkage relationship.

[0076] The trajectory reconstruction module 400 is used to reconstruct the trajectory path by calling the path cell set of historical high-scoring areas of the edge grid cell according to the preset operation specification template trajectory, and generating an offset encoding 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 fragmented path unit set according to the consistency score and offset encoding vector group, generate fragmented quality level labels, and form a linked quality assessment layer.

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

[0079] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0080] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, it implements the method for evaluating the quality of intelligent operation of agricultural equipment as proposed in the above embodiments.

[0081] In summary, this invention achieves refined modeling of the operational boundary region of agricultural equipment by constructing an edge buffer area and dividing it into multi-directional aligned edge grid cells. It establishes a quality linkage matrix by combining trajectory segmentation coding and spatial interaction analysis to quantify the consistency between the trajectory and operational specifications. Furthermore, through trajectory reconstruction and offset coding analysis, it achieves error modeling and standardization of the original path. Compared to existing evaluation methods that rely solely on trajectory overlap rate or trajectory offset distance, this invention integrates historical high-quality path reconstruction and multi-dimensional clustering mechanisms, effectively improving the accuracy and adaptability of quality assessment. Ultimately, it forms a linked quality assessment layer with spatial continuity and difference sensitivity, 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 it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for evaluating the quality of intelligent operation of agricultural equipment, characterized in that: include: An edge buffer region is constructed based on the set of points on the outer contour of the work area, and the buffer is divided 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. Based on the multi-parameter calculation results of path units and edge mesh units, a quality linkage matrix is ​​established. Each element in the quality linkage matrix reflects the multi-parameter consistency score of the corresponding edge mesh unit and the piecewise path unit under the linkage relationship. Based on the preset operation specification template trajectory, the path cell set of historical high-scoring areas of the edge grid cell is called, and the trajectory path is reconstructed through interpolation-correction. An offset encoding vector group is generated based on the spatial offset vector between the original operation path and the reconstructed trajectory. A multi-dimensional quality index clustering analysis is performed on the fragmented path unit set based on the consistency score and offset encoding vector group to generate fragmented quality level labels and form a linked quality assessment layer. The establishment of the quality linkage matrix based on the multi-parameter calculation results of path cells and edge grid cells includes: the spatial coverage of path cells and edge grid cells. trajectory offset and trajectory stability The consistency score is calculated using the consistency scoring function; The consistency scoring function is as follows: in, The maximum allowable offset threshold is used; after calculating the consistency score, the consistency scores are arranged in a two-dimensional pattern according to the spatial interaction pairing method, forming a shape with a size of [missing information]. The quality linkage matrix, where, This represents the total number of path units. This represents the total number of edge grid cells.

2. The method for evaluating the quality of intelligent operation of agricultural equipment as described in claim 1, characterized in that: The division of the edge mesh unit includes: extracting the boundary point set by performing contour thinning processing on the outer trajectory nodes in the operation path data and constructing the minimum outer polygon boundary; The local curvature index of each boundary segment is calculated based on the direction change rate of the boundary, and the variable buffer width is set according to the curvature value to generate an asymmetric outward expansion buffer region. The outer buffer zone is divided into several local sub-blocks according to the boundary normal direction, and mesh rotation is performed within each sub-block based on the local direction to generate multi-directional aligned edge mesh cells.

3. The method for evaluating the quality of intelligent operation of agricultural equipment as described in claim 2, characterized in that: The segmented encoding process is based on the job timestamp and the spatial segment length.

4. The method for evaluating the quality of intelligent operation of agricultural equipment as described in claim 1, characterized in that: The reconstructed trajectory path includes: extracting the score value greater than a set threshold from the quality linkage matrix based on the path unit set corresponding to the edge grid unit in the historical operation round. A subset of path units is used to construct a high-quality path reference set; For each edge grid cell, based on the trajectory point group at the same spatial location in its high-quality path reference set, a fitted trajectory track is constructed using the regional local interpolation method, which is used as the trajectory segment of the operation specification template, i.e., the trajectory path is reconstructed.

5. The method for evaluating the quality of intelligent operation of agricultural equipment as described in claim 4, characterized in that: The generation of the offset encoding vector group includes: selecting the edge grid cells that interact with each original path cell in space, and calculating the Euclidean distance offset vector between each trajectory point in the trajectory path and the reconstructed trajectory in a two-dimensional coordinate system. ,in, Represents the trajectory point index; the path offset vector sequence Perform piecewise sliding window analysis, and normalize the mean and variance using a fixed number of points to generate a standardized offset vector set. Standardize the offset vector group The system is divided into multiple quadrant segments according to direction, generating a direction code D, and the modulus is... Divide the data into several amplitude intervals to generate amplitude code A; then concatenate the direction code D with the amplitude code A to form a composite coding unit.

6. The method for evaluating the quality of intelligent operation of agricultural equipment as described in claim 1, characterized in that: The multidimensional quality index cluster analysis includes: Extract the consistency score vector of each path unit from the quality linkage matrix; Convert the offset encoding vector group of the path unit into a frequency directive; For each path unit, a multidimensional index feature vector is constructed by combining the consistency score vector and the frequency vector, forming a quality feature vector set for cluster analysis. Perform unsupervised clustering on the quality feature vector group.

7. An intelligent agricultural equipment operation quality assessment system, based on the intelligent agricultural equipment operation quality assessment method according to any one of claims 1 to 6, characterized in that: Also includes: An edge construction module is used to construct an edge buffer region based on the set of outer contour points of the work area, and to divide the buffer into edge grid units; The trajectory encoding module is used to segment and encode 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. The linkage matrix module is used to establish a quality linkage matrix based on the multi-parameter calculation results of path units and edge grid units. Each element in the quality linkage matrix reflects the multi-parameter consistency score of the corresponding edge grid unit and the piecewise path unit under the linkage relationship. The trajectory reconstruction module is used to reconstruct the trajectory path by calling the path cell set of historical high-scoring areas of the edge grid cell according to the preset operation specification template trajectory, and generating an offset encoding 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 index clustering analysis on the fragmented path unit set according to the consistency score and offset encoding vector group, generate fragmented quality level labels, and form a linked quality assessment layer.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the intelligent operation quality assessment method for agricultural equipment as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the intelligent operation quality assessment method for agricultural equipment as described in any one of claims 1 to 6.

Images