Agricultural work vehicle

Abstract

To provide an agricultural work vehicle which is simple, has high reliability, and has a posture detecting function.SOLUTION: An agricultural work vehicle comprises: a point group data acquiring unit 51 which acquires measurement point group data as a distance measurement result to a farm structure 9 which is extended in a farm by a prescribed cross section; a shape template storage part 53 which stores a shape template which is formed of the template point group data; a collation part 54 which collates a spatial distribution between the reference shape template and the measurement point group data which are arranged so that shapes of them correspond each other in the same coordinate space; and a posture data calculation part 55 which calculates posture data to the farm structure 9 of a vehicle body on the basis of a collation result of the collation part 54, and calculates a travel deviation between the vehicle body and the farm structure 9 on the basis of the posture data.SELECTED DRAWING: Figure 3

Description

Technical Field

[0004] , , , , , , , ,

[0001] The present invention relates to an agricultural work vehicle equipped with a posture calculation system for calculating the posture with respect to farm structures such as ridges and furrows in a field.

Background Art

[0002] Patent Document 1 discloses an autonomous driving system for autonomously driving an agricultural work vehicle that travels on the traveling surfaces on both sides of a furrow across the furrow in a field along the furrow. This autonomous driving system includes a LiDAR mounted on the agricultural work vehicle and a control device for driving the agricultural work vehicle along the furrow. The detection area of the LiDAR includes the furrow and the traveling surfaces on both sides of the furrow. The control device estimates the center line of the furrow based on the information from the LiDAR regarding the two traveling surfaces, and drives the agricultural work vehicle along the furrow based on the error between the estimated center line of the furrow and the center line of the agricultural work vehicle.

[0003] Patent Document 2 discloses an in-vehicle processing device for estimating the position of a vehicle using point cloud data. This in-vehicle processing device includes a storage unit that stores point cloud data (parking lot point cloud) including coordinates of a plurality of points representing a part of an object in a first coordinate system (parking lot coordinate system), a sensor input unit that acquires the output of a camera for acquiring information around the vehicle, a movement information acquisition unit that acquires information regarding the movement of the vehicle, a local surrounding information creation unit that generates local surrounding information including the position of the vehicle in a second coordinate system (local coordinate system) and coordinates of a plurality of points representing a part of an object in the second coordinate system, and a position estimation unit that estimates the relationship between the parking lot coordinate system and the local coordinate system based on the parking lot point cloud and the local surrounding information, and estimates the position of the vehicle in the first coordinate system.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Patent Document 2

[0005] In the autonomous driving system described in Patent Document 1, it is necessary to detect the shape of the furrow and the driving surfaces located on both sides of the furrow, which results in a large detection area including the furrow and the two driving surfaces, and thus has the problem of placing a burden on the detection system and calculation system. In the in-vehicle processing device described in Patent Document 2, the preprocessing required before matching the point cloud data of the first coordinate system with the point cloud data of the second coordinate system is complex, so there is a need for a simpler attitude calculation technology.

[0006] In view of the above circumstances, the object of the present invention is to provide an agricultural vehicle equipped with a simpler and more reliable posture detection function. [Means for solving the problem]

[0007] The agricultural vehicle according to the present invention is a vehicle that automatically travels along a farm structure extended in a predetermined cross-section on a farm, and comprises: a point cloud data acquisition unit provided on the vehicle body that acquires measured point cloud data as a result of distance measurement to the farm structure located in front of the vehicle body in the direction of travel; a shape template storage unit that stores a shape template composed of template point cloud data; a reference shape template reading unit that reads the shape template corresponding to the shape of the farm structure located in front of the vehicle body in the direction of travel from the shape template storage unit as a reference shape template; a matching unit that matches the spatial distribution of the reference shape template and the measured point cloud data, which are arranged to correspond in shape in the same coordinate space; and a posture data calculation unit that calculates posture data of the vehicle body relative to the farm structure based on the matching result of the matching unit, and calculates the travel deviation between the vehicle body and the farm structure based on the posture data.

[0008] In this configuration, when the orientation of an agricultural vehicle relative to a farm structure (such as field ridges or furrows) fixedly placed on the farm is required, a shape template corresponding to the shape of the farm structure is read from the shape template storage unit as a reference shape template. The spatial distribution of this reference shape template (template point cloud data) and the measured point cloud data, which is the distance measurement result to the farm structure, in the same coordinate space is then compared. Based on this comparison result, that is, based on the degree of overlap between the template point cloud data and the measured point cloud data, the degree of agreement (i.e., the degree of deviation) of the agricultural vehicle's orientation relative to the farm structure is calculated. In this configuration, a shape template that matches the shape of the farm structure, which serves as a guide or target for the agricultural vehicle, is selected, and this selected shape template becomes the reference shape template to be compared with the measured point cloud data, thus enabling simple and reliable orientation detection.

[0009] The driving deviation calculated by the attitude data calculation unit is to be communicated to the driver or farm work supervisor. In this invention, it is proposed that a notification device for notifying the driving deviation is provided. This notification allows the driver or farm work supervisor to know that a farm vehicle traveling along a farm structure has deviated from the farm structure, and they can take maneuvers to correct the driving deviation.

[0010] By automatically controlling the steering of the farm vehicle to eliminate the driving deviation calculated by the attitude data calculation unit, reliable driving along the farm structure can be achieved. For this reason, the present invention proposes that the vehicle be equipped with a steering control unit that automatically steers the vehicle body to eliminate the aforementioned driving deviation.

[0011] Whether using manual or automatic steering, if a deviation occurs in the vehicle's position relative to the target farm structure, it is important to consider both the directional and lateral deviations of the vehicle body when steering in order to quickly return to the correct position. Therefore, in this invention, it is proposed that the deviation includes both the lateral and directional deviations relative to the extension line of the farm structure.

[0012] In this invention, the measurement point cloud data used in the matching unit is data corresponding to the cross-sectional contours at multiple positions spaced apart in the vehicle's direction of travel of the farm structure, and the shape of the reference shape template is proposed to be the shape corresponding to the cross-sectional contours. The reliability of attitude detection is improved by calculating attitude data from multiple matching results between the measurement point cloud data and template point cloud data at multiple positions of the farm vehicle in the near and far directions relative to the farm structure. Furthermore, in this configuration, even if the shape of the farm structure differs depending on the location, more reliable attitude detection becomes possible by using the measurement point cloud data at multiple locations of the farm structure and the corresponding template point cloud data.

[0013] By using measurement point cloud data acquired at different points in time while moving along the farm structure, it is possible to obtain cross-sectional contours of the farm structure at multiple locations. Therefore, in another embodiment of the present invention, it is also proposed that the attitude data calculation unit calculates the attitude data based on the matching results using the measurement point cloud data acquired at predetermined time intervals or at predetermined travel distances.

[0014] When acquiring cross-sectional contours of farm structures at multiple locations, it is advantageous to acquire cross-sectional contours at locations at different distances from the vehicle body in the vehicle travel direction, which is substantially the same direction as the irradiation direction of the beam used for distance measurement, as this makes it easy to increase the spacing between each cross-sectional contour. For this reason, the present invention proposes that the multiple locations include short-range locations that are close to the vehicle body in the vehicle travel direction and long-range locations that are far from the vehicle body, and that the travel deviation is calculated based on the short-range comparison points determined by the comparison result between the short-range measurement point cloud data, which is the measurement point cloud data in the reference shape template at the short-range locations, and the reference shape template, and the long-range comparison points determined by the comparison result between the long-range measurement point cloud data, which is the measurement point cloud data in the reference shape template at the long-range locations, and the reference shape template. Note that the long-range and short-range locations here are not limited to two locations, but may be three or more locations.

[0015] In a comparison using multiple cross-sectional contours at different distances from the vehicle body in the direction of vehicle travel, obtaining the final comparison result using multiple comparison points calculated is advantageous because it suppresses measurement noise. For this reason, the present invention proposes that the lateral displacement and azimuth displacement between the straight line connecting the near-range comparison point and the far-range comparison point and the extension direction line of the farm structure are calculated as the travel displacement.

[0016] In this invention, it is proposed that the matching unit be configured to calculate an alignment vector for aligning the measured point cloud data with the reference shape template (template point cloud data). Matching point cloud data can be easily performed using a vector calculation algorithm, and the resulting alignment vector is advantageous because it indicates the positional misalignment between the measured point cloud data and the reference shape template, that is, the positional misalignment of the non-attitude calculation body relative to the farm structure. Suitable vector calculation algorithms for this purpose include the ICP (Iterative Closest Point) algorithm or an improved version of ICP. [Brief explanation of the drawing]

[0017] [Figure 1] It is a functional block diagram showing the basic configuration of the attitude detection function adopted in the agricultural work vehicle. [Figure 2] It is a functional block diagram of a tractor equipped with an attitude detection function. [Figure 3] It is a side view of a tractor equipped with a furrow forming machine. [Figure 4] It is a schematic diagram showing the measured point cloud data as the distance measurement result with respect to the furrow. [Figure 5] It is a schematic diagram showing the measured point cloud data of the furrow and the reference shape template. [Figure 6] It is a schematic diagram showing the relationship between the measured point cloud data and the reference shape template at the start of the collation. [Figure 7] It is a schematic diagram showing the relationship between the measured point cloud data and the reference shape template at the end of the collation. [Figure 8] It is a rear view of a tractor equipped with a furrow forming machine for forming a trapezoidal furrow. [Figure 9] It is a schematic diagram showing the relationship between the measured point cloud data and the reference shape template in the attitude calculation process for the trapezoidal furrow. [Figure 10] It is a side view of a vegetable transplanter equipped with an attitude calculation system. [Figure 11] It is a front view of the vegetable transplanter showing the relationship between the vegetable transplanter and the furrow.

Embodiments for Carrying Out the Invention

[0018] FIG. 1 is a functional block diagram showing the basic configuration of the attitude calculation technology adopted in the present application for calculating the attitude of the agricultural work vehicle with respect to the farm structure 9. Here, the farm structure 9 is a structure extending in a predetermined cross section on the farm, for example, a ridge or a furrow. The agricultural work vehicle travels along the farm structure 9 in many farm operations. The control system of the agricultural work vehicle is equipped with an attitude calculation system having an attitude calculation function described below, and is used for calculating the control amount of automatic steering or for steering guidance to assist the operator.

[0019] The basic components of the attitude calculation system are a point cloud data acquisition unit 51, a shape template storage unit 53, a reference shape template reading unit 52, a matching unit 54, and an attitude data calculation unit 55.

[0020] The point cloud data acquisition unit 51 acquires measured point cloud data for the farm structure 9 output from the distance sensor 3, which consists of a ToF camera, stereo camera, LiDAR, etc. The beam-based distance sensor 3 is a sensor that irradiates the farm structure 9 with a beam, receives the reflected beam reflected from the object, and calculates the distance to the reflecting object. Ultrasound, laser light, or electromagnetic waves can be used as the beam. In any case, the distance sensor 3 generates point cloud data corresponding to the shape of the surface of the farm structure 9 as a reflecting object in the target space and provides it to the point cloud data acquisition unit 51. The acquired point cloud data is sent to the matching unit 54 as measured point cloud data.

[0021] The shape template storage unit 53 stores a shape template composed of template point cloud data corresponding to the shape of the farm structure 9. The shape template may be created by another device and stored in the shape template storage unit 53, or a shape template generation unit may be provided that generates a desired shape template based on input data.

[0022] The reference shape template reading unit 52 selects and reads a reference shape template from the shape template storage unit 53 that will be used as the matching target for the measured point cloud data in the matching unit 54. The reference shape template reading unit 52 selects a shape template having a shape corresponding to the shape of the farm structure 9 captured by the distance sensor 3 (e.g., the cross-sectional contour shape) and reads it out as the reference shape template. This selection may be made by a manually input command, but it may also be made automatically based on fixed target information (farm structure information) including the shape and identification code of the farm structure 9 sent via the input information processing unit 41. Alternatively, the shape of the farm structure 9 may be estimated based on the measured point cloud data or image recognition of the captured image of the farm structure 9, and a shape template may be selected based on that estimation.

[0023] The matching unit 54 places the measured point cloud data and a reference shape template (template point cloud data) in a coordinate space representing the locations of the farm structure 9 and the farm vehicle (for example, a coordinate space with the irradiation point of the distance sensor 3 as the origin). The matching unit 54 compares the spatial distribution of the reference shape template and the measured point cloud data, which are arranged to correspond in the same coordinate space. At this time, the reference shape template is arranged to correspond to the actual position of the farm structure 9. It is preferable that the position of the farm structure 9 at the start of the matching process (starting position) be set in advance manually or automatically using captured images. As a process for comparing the spatial distribution of the measured point cloud data and the reference shape template, the matching unit 54 performs a process of superimposing the measured point cloud data onto the reference shape template. The ICP (Iterative Closest Point) algorithm is preferably used in this superposition processing program. As a result of this superposition processing (matching result), a positional alignment vector for superposition is calculated. This positional alignment vector is sent to the attitude data calculation unit 55.

[0024] The alignment vector indicates the attitude deviation of the farm vehicle equipped with the distance sensor 3 relative to the farm structure 9. As the alignment vector, a translation vector indicating the parallel movement necessary to correct lateral deviation and a rotation vector indicating the rotation necessary to correct azimuth deviation can be used. The attitude data calculation unit 55 calculates attitude data indicating the lateral and azimuth deviations of the farm vehicle relative to the farm structure 9 based on the alignment vector generated by the matching unit 54. If the farm structure 9 is a structure that extends along a road, such as a guardrail or fence, the azimuth deviation indicates the angle of the farm structure 9 with respect to the extension direction line, and the lateral deviation indicates the distance from the extension direction line of the farm structure 9 in the direction perpendicular to the said extension direction line.

[0025] The notification unit 42 notifies the user of lateral displacement and azimuth displacement based on the attitude data (lateral displacement and azimuth displacement) sent from the attitude data calculation unit 55. The notification unit 42 generates graphic notification data to visually notify the user of lateral displacement and azimuth displacement and sends it to the notification device, the display 42a. The notification unit 42 also generates audio notification data to audibly notify the user of lateral displacement and azimuth displacement and sends it to the notification device, the speaker 42b.

[0026] Next, the basic configuration of a tractor, which is an example of an agricultural vehicle according to the present invention, and the specific configuration of the attitude calculation system mounted on this tractor will be described. Figure 2 is a side view of the tractor. In the following description, unless otherwise specified, the direction of arrow F in Figure 2 will be referred to as "front," and the direction of arrow B will be referred to as "rear." Also, the direction perpendicular to the plane defined by arrows F and U will be referred to as "side." This tractor has a body 10 equipped with front wheels 1a and rear wheels 1b. A prime mover 11 is mounted in the front region of the body 10, a driver's unit 13 is located in the rear region of the body 10, and a link mechanism 12 for connecting work equipment is provided in the rear end region of the body 10. The driver's unit 13 is equipped with a driver's seat 14, a steering wheel 15, and a front panel 16. A driver's panel 17 is provided on the upper surface of the front panel 16, and a controller box 18 is located inside the front panel 16. A furrowing machine 2 is connected to the link mechanism 12. This tractor can perform tasks such as forming and managing furrows 90, which are an example of farm structures 9, while driving across the field.

[0027] The rib forming machine 2 is equipped with a LiDAR 30 that functions as a distance sensor 3. This LiDAR 30 targets the ribs 90. The controller box 18 houses a tractor control device, which includes the main functional units of the attitude calculation system and the main control units of the tractor.

[0028] Figure 3 is a functional block diagram of the tractor control device. The tractor control device includes the components of the attitude calculation system shown in Figure 1. In the explanation of the functional block diagram in Figure 3, the explanation of the functions described earlier will be omitted as it is repetitive.

[0029] The driver panel 17 includes an operation input section for inputting driving commands and driving information for the tractor, a meter panel section for displaying the tractor's driving status to the driver, and the like. The driving information includes furrow information about the furrows 90 and field information such as a field map. The input information processing unit 41 provides the furrow information input through the driver panel 17 to the reference shape template reading unit 52. Based on the furrow information, the reference shape template reading unit 52 selects a shape template that matches the furrow shape to be formed or the furrow shape to be managed, and provides it to the matching unit 54 as a reference shape template. If appropriate furrow information is not provided, the reference shape template reading unit 52 can also estimate the furrow shape based on the measured point cloud data output from the point cloud data acquisition unit 51 and select a shape template that matches the estimated furrow shape.

[0030] Figure 3 shows the steering control unit 43 and the mileage calculation unit 44 as the main control functions of the tractor that work in conjunction with the attitude calculation system. The steering control unit 43 automatically steers the vehicle body 10 to correct the lateral and azimuth deviations output from the attitude data calculation unit 55. The mileage calculation unit 44, although not shown, uses the satellite navigation function and inertial navigation function equipped on the tractor to detect the position of the vehicle body 10 and to calculate the distance traveled and the vehicle speed of the vehicle body 10. The steering control unit 43 is also capable of automatic steering using the satellite navigation function and inertial navigation function.

[0031] The flow of a simplified attitude calculation process (control steps) for the vehicle body 10 relative to the furrow 90 when performing furrow formation and furrow management with the tractor configured as described above will be explained using Figures 4 to 7. Here, the cross-sectional shape of the furrow 90 is approximately triangular.

[0032] (1) Extraction of three furrow cross-sections from the depth image. Figure 4 shows a distance image based on measurement point cloud data output from a LiDAR 30 that irradiates the furrow 90 with laser light from diagonally above, and a cross-sectional graph showing the outline contours of three furrow cross-sections (near-range furrow cross-section, mid-range furrow cross-section, and far-range furrow cross-section) at specific locations (near-range, mid-range, and far-range) in the distance image. In the cross-sectional graph, the measurement point cloud data (near-range, mid-range, and far-range measurement point cloud data acquired at predetermined time intervals or predetermined travel distances, respectively) is schematically depicted using large points (black circles). In the distance image of Figure 4, the y[pix] axis indicates the direction of extension of the furrow 90, the x[pix] axis indicates the direction of crossing the furrow 90, and the grayscale of the distance image indicates the distance from the LiDAR 30 to the reflection point of the furrow 90. From this point cloud data, it can be estimated that the cross-section of ridge 90 is triangular and extends for at least a predetermined distance. In the cross-sectional graph of Figure 4, the horizontal axis, the x[pix] axis, indicates the direction of crossing the ridge 90, the vertical axis indicates the height of the ridge 90, and the dotted line indicated by the symbol CL represents the pre-set reference plane CL of the ridge 90 (a vertical cross-section passing through the center line in the extension direction of the ridge 90).

[0033] Furthermore, the cross-sectional areas of ridges at multiple specific locations in the depth image may be determined from a single point cloud data set, or they may be determined using point cloud data acquired at predetermined time intervals or at predetermined travel distances.

[0034] (2) Select a triangular shape template as the base shape template. In Figure 5, the reference shape template selected by the reference shape template reading unit 52 is placed in the coordinate system of the schematic measurement point cloud data. The reference shape template is positioned so that its base is parallel to the x[pix] axis (i.e., parallel to the horizontal plane of the field) and its upper vertex lies on the line indicating the reference plane CL. The reference shape template is actually template point data, but here it is represented by a lightly colored triangle.

[0035] (3) Alignment of reference shape template with measurement point cloud data Figure 6 shows the matching process in which a reference shape template is aligned with triangles created by the measured point cloud data. Here, the reference shape template, represented by lightly colored triangles, moves towards the darkly colored triangles during the matching process. In this matching process, the upper vertex of the triangle created by the reference shape template is aligned with the upper vertex of the triangle created by the measured point cloud data. This matching process is performed for three furrow cross-sections.

[0036] (4) Generation of attitude evaluation line AL based on the positions of the three vertices of the reference shape template Figure 7 shows the cross-sectional graphs at the end of the matching process, where the upper vertex of the triangle created by the reference shape template (indicated by a white circle) substantially coincides with the upper vertex of the triangle created by the measurement point cloud data. In the distance image as well, the upper vertex of the triangle created by the reference shape template is indicated by a large white circle. A straight line approximately connecting these three upper vertices (near-range matching point, medium-range matching point, and far-range matching point) is generated as the attitude evaluation line AL. This attitude evaluation line AL represents the tractor's vehicle travel direction line (vehicle longitudinal direction line).

[0037] (5) Calculation of lateral and azimuth deviation of the tractor using the attitude evaluation line The deviation of the attitude evaluation line AL relative to the line indicating the reference plane CL represents the tractor's driving deviation. In an overhead view, the lateral deviation of the attitude evaluation line AL relative to the line indicating the reference plane CL (deviation in the direction of the transverse line in the direction of extension of the furrow 90) represents the lateral deviation of the tractor relative to the furrow 90, and the angular deviation of the attitude evaluation line AL relative to the reference plane CL (the angle of intersection between the line indicating the reference plane CL and the attitude evaluation line AL) represents the azimuth deviation of the tractor relative to the furrow 90.

[0038] In the posture calculation process described above, the movement (matching) was performed only to align the upper vertex of the triangle created by the reference shape template (indicated by a white circle) with the upper vertex of the triangle created by the measured point cloud data. Therefore, matching was performed on three (or two, or four or more) furrow cross-sections, and the directional deviation of the tractor relative to furrow 90 was calculated from the intersection angle between the posture evaluation line AL, which is a straight line that approximately connects the obtained upper vertices, and the line indicating the reference plane CL. Alternatively, a three-dimensional movement (matching) to align the reference shape template with the triangle created by the measured point cloud data in three-dimensional space could be performed on only one furrow cross-section. The lateral deviation of the tractor relative to furrow 90 could then be determined from the translation vector of the alignment vector calculated for this matching, and the directional deviation of the tractor relative to furrow 90 could be determined from the rotation vector.

[0039] Next, using Figures 8 and 9, the process for calculating the tractor's attitude relative to the ridge 90 in a specific ridge-forming operation using a tractor will be explained. Figure 8 is a rear view of the tractor, in which the ridge-forming machine 2 is connected to the link mechanism 12. This ridge-forming machine 2 is equipped with three pairs of ridge-forming plates 2b and four guide wheels 2a, and simultaneously forms three rows of isosceles trapezoidal ridges 90 as the tractor moves. The LiDAR 30 targets the central ridge 90 and the running surfaces of the guide wheels 2a located on both sides of the ridge 90, and irradiates a laser beam from diagonally above. The cross-section of the ridge 90 formed by this ridge-forming machine 2 is an isosceles trapezoid, and a reference shape template that matches this shape and dimensions is selected.

[0040] The left side of Figure 9 shows a distance image based on measured point cloud data of a single row of ridges 90 and the running surface of the guide wheel 2a obtained during ridge formation. This distance image is an overhead view of the measured point cloud data of the ridges 90 as seen from the beam irradiation point of the LiDAR 30, and the distance from the beam irradiation point to the reflection point of the ridges 90 is indicated by the intensity of the grayscale. This distance image shows cross-sectional lines drawn with dashed dots, indicating the positions of the ridge cross-sections at three locations on the ridges 90 set at predetermined intervals.

[0041] In the cross-sectional graph on the right side of Figure 9, the ridge cross-sectional contours (measured point cloud data) at three locations are shown as faint solid lines. Furthermore, two isosceles trapezoids are shown to schematically represent the movement of the reference shape template during the matching process. In this matching process, the reference shape template is aligned with the ridge cross-sectional contour (trapezoid) defined by the measured point cloud data. In other words, the spatial distribution of the measured point cloud data and the template point cloud data defining the reference shape template are matched. At the end of this matching process, the midpoint of the top edge of the reference shape template is shown as a white dot in the distance image on the right side of Figure 9 and in the three ridge cross-sections on the left side of Figure 9. The straight line that approximately connects these three white dots in the distance image becomes the attitude evaluation line AL. Therefore, the lateral displacement of the attitude evaluation line AL with respect to the line representing the reference plane CL is calculated as the lateral displacement of the tractor with respect to the ridge 90, and the angular displacement of the attitude evaluation line AL with respect to the reference plane CL is calculated as the azimuth displacement of the tractor with respect to the ridge 90. The calculated lateral displacement and azimuth displacement are provided to the steering control unit 43 and used for steering control.

[0042] In the above-described embodiment, the posture calculation system was installed on a tractor, which is an agricultural work vehicle. Next, we will describe an embodiment in which the posture calculation system is installed on a transplanter, which is also an agricultural work vehicle. Figure 10 is a side view of a vegetable transplanter as an example of a transplanter, and Figure 11 is a rear view of the vegetable transplanter showing the relationship between the vegetable transplanter and the ridge 90.

[0043] The body elements of this vegetable transplanter, although different in form, are substantially the same as those of the tractor described above, and such body elements are given the same or similar designations. Furthermore, the direction of arrow F in Figure 10 is defined as "forward," the direction of arrow B as "rear," the direction of arrow U in Figure 10 as "up," the direction of arrow D as "down," the direction of arrow L in Figure 11 as "left," and the direction of arrow R as "right."

[0044] As shown in Figure 10, the vegetable transplanter has a vehicle body 10 with front wheels 1c and rear wheels 1d, a prime mover 11 mounted in the front region of the vehicle body 10, an operator's unit 13 located in the middle region of the vehicle body 10, and a link mechanism 12 for connecting work equipment provided in the rear end region of the vehicle body 10. The operator's unit 13 has a driver's seat 14, a steering wheel 15, and a front panel 16. An operator's panel 17 is provided on the upper surface of the front panel 16, and a controller box 18 is located inside the front panel 16.

[0045] A seedling planting device 20 is connected to the linkage mechanism 12. The seedling planting device 20 includes, for example, a planting frame unit 20a, a seedling holder 20b, a planting body 20c, a soil covering wheel 20d, etc., as described in Japanese Patent Publication No. 2020-202802. As the vegetable transplanter travels along the ridges 90 of the field, the seedling planting device 20 plants seedlings (vegetable seedlings) in the ridges 90. Here, the vegetable transplanter is the object whose position is calculated, and the farm structure 9 is the ridges 90.

[0046] The seedling tray 20b, planting body 20c, and soil covering wheel 20d are supported by the planting frame unit 20a, which is the main structure of the seedling planting device 20. Seedlings discharged from the seedling tray 20b are planted on the upper surface of the ridge 90 by the planting body 20c. The areas to the left and right of the planted seedlings are mounded with soil and compacted by the rolling soil covering wheel 20d.

[0047] A LiDAR 30, which functions as a distance sensor 3, is mounted on the front frame of the vehicle body 10 of the vegetable transplanter. This LiDAR 30 targets the furrows 90 in front of the vehicle's direction of travel. The LiDAR 30 is not limited to being mounted on the front frame of the vehicle body 10; for example, it may be mounted on the ceiling member of a gate-shaped spare seedling stand 20e erected at the front of the vehicle body 10. In Figure 11, the LiDAR 30 is mounted in both of the above locations, but of course, either one is sufficient.

[0048] The controller box 18 houses a control device that includes the main functional units of the attitude calculation system and the main control functional units of the vegetable transplanter. The main functional units of the attitude calculation system and the control functional units related to the movement of the vegetable transplanter are substantially the same as the functional units of the tractor control device described with reference to Figure 3, so their description is omitted to avoid repetition. In this vegetable transplanter, measurement point cloud data of the furrows 90 in front of the vehicle's direction of travel is acquired, and attitude data indicating the attitude of the vehicle body 10 relative to the furrows 90 is calculated. When the vegetable transplanter is automatically steered, the vehicle body 10 is automatically steered based on the calculated attitude data. When the vegetable transplanter is manually steered, driving guidance is created and communicated based on the calculated attitude data to assist the driver's operation.

[0049] [Another embodiment] (1) In the two embodiments described above, the farm structure 9 was a furrow 90, but the farm structure 9 may also be a guide belt such as a vinyl tunnel installed on a ridge or field.

[0050] (2) In the above-described embodiment, the matching was performed using a reference polygon in multiple cross-sections cut out from the three-dimensional space created by the measured point cloud data. Alternatively, the matching may be performed using multiple cross-sections based on the measured point cloud data acquired for each predetermined distance traveled by the farm vehicle (or over time). In this case, the measured point cloud data handled corresponds to the three-dimensional shape of the farm structure 9, and a reference shape template corresponding to the three-dimensional contour shape of the farm structure 9 is used.

[0051] (3) In the above-described embodiment, the translation vector at the time of alignment was used as the matching result in the matching unit 54, but a translation vector and a rotation vector may also be used. In that case, the translation vector indicates the lateral displacement, and the rotation vector indicates the azimuth displacement.

[0052] (4) In the embodiments described above, the reference shape template handled was a two-dimensional shape such as a polygon, but it may also be a three-dimensional shape such as a polyhedron.

[0053] (5) The functional block diagrams shown in Figures 1 and 3 are for illustrative purposes only, and each functional unit may be integrated with other functional units or divided into multiple functional units.

[0054] Furthermore, the configurations disclosed in the above embodiments (including other embodiments, the same applies hereinafter) can be applied in combination with configurations disclosed in other embodiments, as long as no inconsistencies arise. Moreover, the embodiments disclosed herein are illustrative, and the embodiments of the present invention are not limited thereto, and can be modified as appropriate without departing from the object of the present invention. [Industrial applicability]

[0055] This invention is applicable to agricultural vehicles equipped with a posture calculation function that calculates the posture relative to farm structures. [Explanation of symbols]

[0056] 2:Ribule forming machine 2a: Guide ring 2b: Ridged board 20: Seedling planting device 20a: Planting frame unit 20b: Seedling stand 20c: Planted body 20d: Covering ring 3: Distance sensor 30: LiDAR, ToF camera 9: Farm structure 90: ridge 10: Vehicle body 17: Operation Panel 18: Controller Box 41: Input Information Processing Unit 42: Hochi Department 42a: Display 42b: Speaker 43: Steering Control Unit 44: Distance calculation unit 51: Point cloud data acquisition unit 52: Reference shape template reading unit 53: Shape template storage unit 54: Verification Unit 55: Posture data calculation unit AL: Posture Evaluation Line CL: Reference plane

Claims

1. An agricultural vehicle that automatically travels along a farm structure extended in a predetermined cross-section on a farm, A point cloud data acquisition unit is provided on the vehicle body and acquires point cloud data as a result of distance measurement to the farm structure located in front of the vehicle's direction of travel. A shape template storage unit that stores a shape template composed of template point cloud data, A reference shape template reading unit reads a shape template from the shape template storage unit as a reference shape template, which corresponds to the shape of the farm structure located in front of the vehicle's direction of travel. A comparison unit that compares the spatial distribution of the reference shape template and the measured point cloud data, which are arranged to correspond in shape within the same coordinate space, A posture data calculation unit calculates posture data of the vehicle body relative to the farm structure based on the matching results of the matching unit, and calculates the driving deviation between the vehicle body and the farm structure based on the posture data. A farm vehicle equipped with [a specific feature].

2. The agricultural vehicle according to claim 1, further comprising a notification device for notifying the aforementioned deviation in driving.

3. The agricultural vehicle according to claim 1, further comprising a steering control unit that automatically steers the vehicle body to eliminate the aforementioned deviation in driving.

4. The agricultural vehicle according to claim 1, wherein the aforementioned deviation in travel includes a lateral deviation and an azimuth deviation with respect to the extension direction line of the farm structure.

5. The agricultural vehicle according to claim 1, wherein the measurement point cloud data used in the matching unit is data corresponding to the cross-sectional contours at multiple positions spaced apart in the vehicle's direction of travel of the farm structure, and the shape of the reference shape template is the shape corresponding to the cross-sectional contour.

6. The agricultural vehicle according to claim 3, wherein the posture data calculation unit calculates the posture data based on the comparison result using the measured point cloud data acquired at predetermined time intervals or for each predetermined travel distance.

7. The aforementioned multiple positions include short-distance positions that are close to the vehicle body in the direction of vehicle travel and long-distance positions that are far from the vehicle body. The agricultural vehicle according to claim 5, wherein the driving deviation is calculated based on a short-range comparison point determined by the comparison result between the short-range measurement point cloud data, which is the measurement point cloud data in the reference shape template at the short-range position, and the reference shape template, and a long-range comparison point determined by the comparison result between the long-range measurement point cloud data, which is the measurement point cloud data in the reference shape template at the long-range position, and the reference shape template.

8. The agricultural vehicle according to claim 7, wherein the lateral displacement and azimuth displacement between the straight line connecting the nearby reference point and the far reference point and the extension direction line of the farm structure are calculated as the travel displacement.

9. The agricultural vehicle according to any one of claims 1 to 8, wherein the matching unit calculates an alignment vector for aligning the measured point cloud data with the reference shape template.

Images