Potato breeding experiment harvesting robot and harvesting method

Abstract

The present invention relates to the technical field of potato breeding experiment harvesting, and in particular to a potato breeding experiment harvesting robot and a harvesting method. The robot comprises: a robot frame, a walking device, a navigation system, a visual system, and an excavating device. A control device receives a signal from the navigation system to control the walking device to drive the harvesting robot to walk along the potato ridges. The visual system collects field information of the plot and provides it to the control device for analysis. The control device analyzes the acquired data to determine the optimal shovel entry position of the excavating device, and the control device drives the excavating device to perform excavation and harvesting operations. The potato breeding experiment robot designed by the invention can realize integrated operations such as automatic navigation, potato plant identification, potato ridge contour extraction, single potato excavation, harvesting and laying in potato breeding experiments. It has a high degree of automation, can effectively reduce the damage to potatoes during the excavation process, and improve the quality of breeding experiments.

Description

Technical Field

[0001] The invention relates to the technical field of potato breeding experiment harvesting, and in particular to a potato breeding experiment harvesting robot and a harvesting method. Background Art

[0002] Potato is one of the world's important crops and is widely grown around the world. The harvesting requirements for potato breeding experiments are different from those in the field. In breeding experiments, plant phenotypic information such as tubers, pests and diseases needs to be extracted after the potatoes are harvested. Therefore, breeding experiments require the whole plant to be harvested, and the harvested potato plants should be placed in the field in the order of planting to facilitate the recording of the tuber enlargement, pest and disease resistance and other results corresponding to different potato varieties.

[0003] At present, the digging and harvesting work of potato breeding experiments is arduous, and manual digging and harvesting is time-consuming and laborious. Traditional potato harvesters are difficult to carry out targeted harvesting of each individual plant, and the potatoes will be damaged during the harvesting process, which affects the accuracy of subsequent test data collection. Therefore, existing harvesters are difficult to be used in potato plot breeding experiments. Summary of the invention

[0004] In view of the technical problems of difficulty in harvesting a single plant and large harvesting damage in potato harvesters in the prior art, the present invention provides a potato breeding test harvesting robot, which can overcome certain or some defects of the prior art.

[0005] A potato breeding experiment harvesting robot according to the present invention comprises:

[0006] Robot rack,

[0007] A walking device, arranged at the bottom of the robot frame, used to drive the robot frame to move in the plot to be harvested;

[0008] A navigation system, arranged at the front of the robot frame, for obtaining spatial information of the plot to be harvested, wherein the spatial information of the plot is used for navigation of the walking device;

[0009] A visual system, disposed at the front of the robot frame, for collecting images of plots on a path of travel when the robot frame moves;

[0010] A digging device, arranged in the middle of the robot frame, for digging individual potatoes in the plot to be harvested;

[0011] A control device, which is in communication with the walking device, the navigation system, the visual system and the excavating device;

[0012] The control device is used to perform global path planning for the entire plot to be harvested according to the pre-acquired plot spatial information of the plot to be harvested, and control the walking device to drive the robot frame to automatically navigate and walk in the plot to be harvested according to the planned global path;

[0013] and for receiving the image data to be detected and the point cloud data collected by the visual system when the robot frame moves, performing potato target detection on the image data to be detected, and calculating the potato ridge point cloud data according to the point cloud data;

[0014] It is also used to control the robot frame to move to the potato target according to the potato ridge point cloud data when a potato target is detected in the image data to be detected, and then control the excavation device to perform single-plant excavation.

[0015] As a preference, it also includes:

[0016] A power device, used to provide power for the excavating device and the traveling device; it includes a diesel oil pipe, a fixing small plate, a diesel oil tank, a tank fixing plate, a hydraulic valve, a hydraulic oil tank outlet nozzle, a diesel engine hydraulic pump assembly, a hydraulic oil pipe and a silencer tail pipe;

[0017] The diesel engine hydraulic pump assembly is installed on the rear end of the robot frame through a fixed small plate, a hydraulic oil tank oil outlet nozzle is installed on the upper end of the diesel engine hydraulic pump assembly, and a silencer tail pipe is installed on the lower end; the hydraulic valve is installed on one side of the diesel engine hydraulic pump assembly, the hydraulic valve is connected to the hydraulic oil tank oil outlet nozzle through a hydraulic oil pipe, the diesel tank is connected to the diesel engine hydraulic pump assembly through a diesel oil pipe, a tank fixing plate is installed on the upper end of the diesel tank, and the diesel tank is installed on one side of the robot frame through the tank fixing plate.

[0018] Preferably, the excavation device comprises excavation components arranged opposite to each other, and the excavation components comprise:

[0019] The excavation upper arm, the top of which is hinged to the upper frame of the robot frame through an upper fixing plate; the middle of which is connected to a first hydraulic cylinder; the first hydraulic cylinder is used to adjust the posture of the excavation upper arm relative to the robot frame, specifically, the first hydraulic cylinder, the fixed end is hinged to the lower frame of the robot frame through a lower fixing plate, and the telescopic end is hinged to the middle of the excavation upper arm;

[0020] The excavation lower arm is arranged at the lower side of the excavation upper arm, and the top of the excavation lower arm is connected to the excavation upper arm through a second hydraulic cylinder; the second hydraulic cylinder is used to adjust the posture of the excavation lower arm relative to the excavation upper arm, and the fixed end of the second hydraulic cylinder is hinged to the inner wall of the excavation upper arm, and the telescopic end of the second hydraulic cylinder is hinged to the top of the excavation lower arm;

[0021] An excavating shovel is disposed at the circular tail end of the excavating lower arm, and a third hydraulic cylinder and an excavating connection mechanism are connected between the excavating shovel and the excavating lower arm; the third hydraulic cylinder is used to adjust the posture of the excavating shovel relative to the excavating lower arm;

[0022] The third hydraulic cylinder has a fixed end hinged to the upper part of the excavation lower arm, and a telescopic end hinged to the upper end of the excavation connection mechanism; the lower end of the excavation connection mechanism is hinged to the upper connecting plate of the excavating shovel.

[0023] Preferably, the upper fixing plate and the lower fixing plate are both mounted on the robot frame via grooves.

[0024] As a preference, it also includes:

[0025] The power supply device is used to supply power to the control device, the navigation system and the visual system; it comprises a battery and a battery mounting plate, the lower end of the battery is fixed on the battery mounting plate, and the battery mounting plate is fixed on the frame.

[0026] The present invention also provides a potato breeding experiment harvesting robot harvesting method, which is performed using the potato plot breeding experiment harvesting robot, and comprises the following steps:

[0027] Step S1, performing global path planning for the entire plot to be harvested according to the pre-acquired plot spatial information of the plot to be harvested, and controlling the walking device to drive the robot frame to perform automatic navigation walking in the plot to be harvested according to the planned global path;

[0028] Step S2, when the robot frame moves, the image data to be detected and the point cloud data collected by the visual system are obtained, potato target detection is performed on the image data to be detected, and potato ridge point cloud data is calculated according to the point cloud data;

[0029] Step S3: if a potato target is detected in the image data to be detected, the robot frame is controlled to move to the potato target according to the potato ridge point cloud data, and then the excavation device is controlled to perform single plant excavation;

[0030] Step S4, repeat steps S2 to S3.

[0031] Preferably, step S3 comprises: if a potato target is detected in the image data to be detected, controlling the robot frame to move to the potato target according to the potato ridge point cloud data, and then controlling the digging device to perform single plant digging;

[0032] Step S31, controlling the harvesting robot to move forward according to the potato plants and potato ridge point cloud data identified by the visual system, so that the digging device moves above a potato plant;

[0033] Step S32, processing the acquired potato ridge point cloud data to calculate the potato ridge height and fit the potato ridge contour line; then determining the best burrowing position of the digging shovel according to the potato ridge height and the fitted potato ridge contour line, and converting the coordinates of the burrowing position into coordinates in the earth world coordinate system;

[0034] Step S33, controlling the digging device to reach the best burial position to dig and obtain potatoes.

[0035] Preferably, in step S31, the acquired potato ridge point cloud data is processed to calculate the height of the potato ridge and fit the potato ridge contour line;

[0036] Wherein, the method for obtaining the potato ridge point cloud data is:

[0037] Performing posture correction on the original point cloud data by point cloud transformation; the original point cloud data includes potato point cloud data, potato ridge point cloud data and ground point cloud;

[0038] The point cloud data obtained after point cloud transformation is denoised using bilateral filtering;

[0039] The ground point cloud is separated by using a CSF algorithm to obtain the potato point cloud data and potato ridge point cloud data.

[0040] The step of processing the acquired potato ridge point cloud data to calculate the height of the potato ridge and fit the potato ridge contour line comprises:

[0041] For the potato ridge point cloud data, 20 point cloud sections were randomly selected in the direction perpendicular to the potato ridge;

[0042] Extracting point cloud data on each section and calculating the mean value to obtain the height of the potato ridge;

[0043] Then, the Gaussian model and least squares method are used to fit the potato ridge contour line to the extracted potato ridge point cloud data.

[0044] Preferably, the excavation device comprises excavation components arranged relatively to each other, and the excavation components comprise: an excavation upper arm, a first hydraulic cylinder, an excavation lower arm, a second hydraulic cylinder, an excavation shovel and a third hydraulic cylinder;

[0045] The step S33 comprises:

[0046] Step S331, controlling the first hydraulic cylinder and the second hydraulic cylinder to make the excavating shovel reach the optimal burrowing position, and then driving the third hydraulic cylinder to make the excavating shovel perform excavation operation;

[0047] Step S332: When the two digging shovels installed opposite to each other have penetrated into the soil to reach half of the width of the potato ridge, the third hydraulic cylinder is controlled to stop working, and the second hydraulic cylinder is controlled to contract to lift the lower digging arm to above the ridge top;

[0048] Step S333, controlling the third hydraulic cylinder to perform three telescopic operations, so as to shake off soil and other impurities in the harvested potatoes by shaking the digging shovel;

[0049] Step S334, controlling the third hydraulic cylinder to contract, driving the digging shovel to open to both sides, and the excavated potatoes fall onto the ridges, so as to complete the harvest;

[0050] Step S335: Control the first hydraulic cylinder and the second hydraulic cylinder to contract, driving the digging lower arm and the digging shovel away from the potato ridge.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] 1. The present invention can realize targeted single-plant harvesting of potatoes through the cooperation of the visual system, the walking device and the digging device to reduce the damage of potatoes during the harvesting process, thereby meeting the harvesting quality requirements of breeding experiments. The robot can realize automatic navigation, potato plant identification, potato ridge contour extraction, single-plant potato digging, harvesting and laying and other integrated operations in potato breeding experiments, with a high degree of automation.

[0053] 2. The visual system of the present invention uses a binocular depth camera Realsense D435i to identify potato plants and potato ridge contours, which not only makes it easier for the control device to control the walking device to accurately move to the position of the potatoes to be harvested, so as to facilitate targeted single-plant harvesting, but also can determine the best entry angle of the digging shovel to ensure that excessive damage to the potatoes is avoided during harvesting.

[0054] 3. In the present invention, a lightweight target recognition model Yolo v8-DSD suitable for potato breeding test robots is designed for potato target image recognition. The method is based on the Yolo v8 recognition model, and the C2f module is improved by using the DualConv convolutional network structure, which is integrated into a new lightweight module C2f-DualConv. The SimAM attention mechanism module is introduced into the neck network neck, and the DBB module is integrated into the Detect module, which is integrated into a new detection head module Detect-DBB. The recognition model has a small number of parameters and calculations, and is suitable for deployment in the control device of the potato breeding test robot.

[0055] 4. The excavation device in the present invention adopts relatively arranged excavation components, and cooperates with multiple hydraulic cylinders to adjust the posture of the excavation shovel. It can not only ensure that the excavation shovel can stably maintain the best posture for excavation to ensure the excavation quality, but also enable the robot to be suitable for potato ridges of different shapes to meet the requirements of single potato plant harvesting in breeding experiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 is a schematic diagram of the structure of a harvesting robot in an embodiment of the present invention;

[0057] Figure 2 yes Figure 1 The schematic diagram of the structure of the power unit;

[0058] Figure 3 yes Figure 1 A schematic diagram of the structure of the power supply device;

[0059] Figure 4 yes Figure 1 A schematic diagram of the structure of the excavation device;

[0060] Figure 5 yes Figure 4 A schematic diagram of the structure from another perspective;

[0061] Figure 6 It is a schematic diagram of the structure of the lightweight model Yolo v8-DSD in an embodiment of the present invention. DETAILED DESCRIPTION Example 1

[0062] Combination Figure 1-Figure 2 The present embodiment provides a potato breeding test harvesting robot, which includes: a robot frame 1; a walking device 8, which is arranged at the bottom of the robot frame 1 and is used to drive the robot frame 1 to move in the plot to be harvested; a navigation system 5, which is arranged at the front of the robot frame 1 and is used to obtain the plot space information of the plot to be harvested, and the plot space information is used for navigation of the walking device 8; a visual system 7, which is arranged at the front of the robot frame 1 and is used to collect images of the plots on the travel path when the robot frame 1 moves; an excavating device 6, which is arranged at the middle of the robot frame 1 and is used to perform single-plant excavation of potatoes in the plot to be harvested; and a control device 3, which is communicatively connected with the walking device 8, the navigation system 5, the visual system 7 and the excavating device 6;

[0063] The control device 3 is first used to perform global path planning for the entire plot to be harvested according to the pre-acquired spatial information of the plot to be harvested, and control the walking device 8 to drive the robot frame 1 to perform automatic navigation walking in the plot to be harvested according to the planned global path;

[0064] and for receiving the image data and point cloud data to be detected collected by the visual system 7 when the robot frame 1 moves, performing potato target detection on the image data to be detected, and calculating potato ridge point cloud data according to the point cloud data;

[0065] It is also used to control the robot frame 1 to move to the potato target according to the potato ridge point cloud data when a potato target is detected in the image data to be detected, and then control the digging device 6 to perform single-plant digging.

[0066] Specifically, when in use, the visual system 7 in this embodiment uses a binocular depth camera Realsense D435i to identify potato plants and potato ridge contours, which not only makes it easier for the control device 3 to control the walking device 8 to accurately move to the position of the potatoes to be harvested, so as to facilitate targeted single-plant harvesting, but also can determine the optimal entry angle of the digging shovel 67 to ensure that excessive damage to the potatoes is avoided during harvesting.

[0067] In this embodiment, the harvesting robot further includes:

[0068] A power device, used to provide power for the excavating device 6 and the walking device 8; it comprises a diesel oil pipe (21), a fixing plate (22), a diesel oil tank (23), an oil tank fixing plate (24), a hydraulic valve (25), a hydraulic oil tank oil outlet nozzle (26), a diesel engine hydraulic pump assembly (27), a hydraulic oil pipe (28) and a silencer tail pipe (29);

[0069] The diesel engine hydraulic pump assembly (27) is installed at the rear end of the robot frame 1 (1) through a fixing small plate (22); a hydraulic oil tank outlet nozzle (26) is installed at the upper end of the diesel engine hydraulic pump assembly (27), and a silencer tail pipe (29) is installed at the lower end; a hydraulic valve (25) is installed at one side of the diesel engine hydraulic pump assembly (27); the hydraulic valve (25) is connected to the hydraulic oil tank outlet nozzle (26) through a hydraulic oil pipe (28); the diesel oil tank (23) is connected to the diesel engine hydraulic pump assembly (27) through a diesel oil pipe (21); a tank fixing plate (24) is installed at the upper end of the diesel oil tank (23); and the diesel oil tank (23) is installed at one side of the robot frame (1) through the tank fixing plate (24).

[0070] In this embodiment, the excavation device 6 includes excavation components arranged relatively to each other, and the excavation components include:

[0071] The top of the excavation upper arm 64 is hinged to the upper frame of the robot frame 1 through an upper fixing plate 61; the middle is connected to a first hydraulic cylinder 62; the first hydraulic cylinder 62 is used to adjust the posture of the excavation upper arm 64 relative to the robot frame 1. Specifically, the first hydraulic cylinder 62 has a fixed end hinged to the lower frame of the robot frame 1 through a lower fixing plate 63, and a telescopic end is hinged to the middle of the excavation upper arm 64;

[0072] The excavation lower arm 65 is arranged at the lower side of the excavation upper arm 64, and the top of the excavation lower arm 65 is connected to the excavation upper arm 64 through the second hydraulic cylinder 66; the second hydraulic cylinder 66 is used to adjust the posture of the excavation lower arm 65 relative to the excavation upper arm 64, and the fixed end of the second hydraulic cylinder 66 is hinged to the inner wall of the excavation upper arm 64, and the telescopic end of the second hydraulic cylinder 66 is hinged to the top of the excavation lower arm 65;

[0073] The digging shovel 67 is provided at the circular tail end of the digging lower arm 65, and a third hydraulic cylinder 68 and an excavation connection mechanism 69 are also connected between the digging shovel 67 and the digging lower arm 65; the third hydraulic cylinder 68 is used to adjust the posture of the digging shovel 67 relative to the digging lower arm 65;

[0074] The third hydraulic cylinder 68 has a fixed end hinged to the upper part of the excavation lower arm 65 and a telescopic end hinged to the upper end of the excavation connection mechanism 69 ; the lower end of the excavation connection mechanism 69 is hinged to the upper connecting plate of the excavation shovel 67 .

[0075] At the same time, the upper fixing plate 61 and the lower fixing plate 63 are both mounted on the robot frame 1 through grooves.

[0076] In this embodiment, the robot further includes: a power supply device 4, including a battery 41 and a battery mounting plate 42, wherein the lower end of the battery 41 is fixed to the battery mounting plate 42, and the battery mounting plate 42 is fixed to the frame. The power supply device 4 supplies power to the control device 3, the navigation system 5, and the visual system 7. Example 2

[0077] This embodiment provides a harvesting method that is beneficial to a potato breeding experiment harvesting robot in Embodiment 1, which comprises the following steps:

[0078] Step S1, based on the pre-acquired spatial information of the plot to be harvested, a global path planning is performed for the entire plot to be harvested, and the walking device 8 is controlled to drive the robot frame 1 to perform automatic navigation walking in the plot to be harvested according to the planned global path;

[0079] Step S2, when the robot frame 1 moves, the image data to be detected and the point cloud data collected by the visual system 7 are obtained, and potato target detection is performed on the image data to be detected, and potato ridge point cloud data is calculated according to the point cloud data;

[0080] Step S3: if a potato target is detected in the image data to be detected, the robot frame 1 is controlled to move to the potato target according to the potato ridge point cloud data, and then the digging device 6 is controlled to perform single-plant digging;

[0081] Step S4, repeat steps S2 to S3.

[0082] Specifically, the robot can realize integrated operations such as automatic navigation, potato plant identification, potato ridge contour extraction, single potato digging, harvesting and laying in potato breeding experiments through the harvesting method in this embodiment, with a high degree of automation.

[0083] In this embodiment, the potato target detection on the image data to be detected in step S2 is performed using a pre-trained lightweight model Yolo v8-DSD based on improved Yolo v8;

[0084] The training process of the lightweight model Yolo v8-DSD based on the improved Yolo v8 includes:

[0085] The data set was captured using a binocular depth camera, the camera model of which and the camera model of the visual system 7 used in the potato plot breeding experiment harvesting robot are both Realsense D435i.

[0086] Screen and expand the acquired data set;

[0087] Preprocess the images in the dataset and divide them into training set, test set and validation set;

[0088] The training set is used to train the lightweight Yolo v8-DSD model, and the validation set is used to validate the trained Yolo v8-DSD model, and finally the trained lightweight model Yolo v8-DSD is obtained;

[0089] The method for improving Yolo v8 to a lightweight Yolo v8-DSD model is:

[0090] The DualConv convolutional network structure is used to improve the C2f module and merge it into a new lightweight module C2f-DualConv;

[0091] And replace the C2f modules in the backbone network backbone and the neck network neck in Yolo v8 with the new lightweight module C2f-DualConv;

[0092] The C2f-DualConv module can process the same input feature map channels by combining 3×3 and 1×1 convolution kernels to optimize information processing and feature extraction;

[0093] The identification model has a small number of parameters and a small amount of calculation, and is suitable for being deployed in the control device 3 of the potato breeding test robot.

[0094] In this embodiment, step S3 includes: if a potato target is detected in the image data to be detected, the robot frame 1 is controlled to move to the potato target according to the potato ridge point cloud data, and then the digging device 6 is controlled to perform single plant digging;

[0095] Step S31, controlling the harvesting robot to move forward according to the potato plants and potato ridge point cloud data identified by the visual system 7, so that the digging device 6 moves above a potato plant;

[0096] Step S32, processing the acquired potato ridge point cloud data to calculate the potato ridge height and fit the potato ridge contour line; then determining the best burrowing position of the digging shovel 67 according to the potato ridge height and the fitted potato ridge contour line, and converting the coordinates of the burrowing position into coordinates in the earth world coordinate system;

[0097] Step S33, control the digging device 6 to reach the best burial position to dig and obtain potatoes.

[0098] Furthermore, in step S31, the obtained potato ridge point cloud data is processed to calculate the height of the potato ridge and fit the potato ridge contour line;

[0099] Wherein, the method for obtaining the potato ridge point cloud data is:

[0100] Performing posture correction on the original point cloud data by point cloud transformation; the original point cloud data includes potato point cloud data, potato ridge point cloud data and ground point cloud;

[0101] The point cloud data obtained after point cloud transformation is denoised using bilateral filtering;

[0102] The ground point cloud is separated by using a CSF algorithm to obtain the potato point cloud data and potato ridge point cloud data.

[0103] The step of processing the acquired potato ridge point cloud data to calculate the height of the potato ridge and fit the potato ridge contour line comprises:

[0104] For the potato ridge point cloud data, 20 point cloud sections were randomly selected in the direction perpendicular to the potato ridge;

[0105] Extracting point cloud data on each section and calculating the mean value to obtain the height of the potato ridge;

[0106] Then, the Gaussian model and least squares method are used to fit the potato ridge contour line to the extracted potato ridge point cloud data.

[0107] Through the above method, the potato ridge height and the fitted potato ridge contour line can be ensured to be accurate, so as to ensure that the walking device 8 walks to the correct position, and the subsequent digging device 6 can adjust its posture to adapt to the digging needs, so as to be suitable for potato ridges of different shapes and meet the potato single plant harvesting requirements in breeding experiments.

[0108] In this embodiment, the excavation device 6 includes excavation components arranged relatively to each other, and the excavation components include: an excavation upper arm 64, a first hydraulic cylinder 62, an excavation lower arm 65, a second hydraulic cylinder 66, an excavation shovel 67 and a third hydraulic cylinder 68;

[0109] The step S33 comprises:

[0110] Step S331, control the first hydraulic cylinder 62 and the second hydraulic cylinder 66 to make the digging shovel 67 reach the best burrowing position, and then drive the third hydraulic cylinder 68 to make the digging shovel 67 perform digging operation;

[0111] Step S332: When the two digging shovels 67 installed opposite to each other have sunk into the soil to half the width of the potato ridge, the third hydraulic cylinder 68 is controlled to stop working, and the second hydraulic cylinder 66 is controlled to contract, so that the digging lower arm 65 is lifted above the ridge top;

[0112] Step S333, controlling the third hydraulic cylinder 68 to perform three telescopic operations, so as to shake off soil and other impurities in the harvested potatoes by shaking the digging shovel 67;

[0113] Step S334, controlling the third hydraulic cylinder 68 to contract, driving the digging shovel 67 to open to both sides, and the excavated potatoes fall onto the ridges, so as to complete the harvest;

[0114] Step S335: Control the first hydraulic cylinder 62 and the second hydraulic cylinder 66 to contract, driving the digging lower arm 65 and the digging shovel 67 away from the potato ridge.

[0115] It can be understood that through the above steps, the control device 3 only needs to control the first hydraulic cylinder 62, the second hydraulic cylinder 66 and the third hydraulic cylinder 68 to adjust the posture of the digging shovel 67, so as to realize the functions of digging out potatoes, shaking off impurities and dropping potatoes at different stages, with a high degree of automation and high potato harvest quality.

[0116] It is easy to understand that those skilled in the art can combine, split, reorganize, etc. the embodiments of the present application to obtain other embodiments based on one or several embodiments provided in the present application, and these embodiments do not exceed the protection scope of the present application.

[0117] The present invention and its implementation methods are described schematically above, and the description is not restrictive. The embodiments shown in the embodiments are only part of the implementation methods of the present invention, and the actual structure is not limited thereto. Therefore, if a person skilled in the art is inspired by the embodiments and designs a structure and an implementation method similar to the technical solution without creativity without departing from the purpose of the invention, they should all fall within the protection scope of the present invention.

Claims

1. A potato breeding experiment harvesting robot, characterized in that: include: Robot frame (1), A walking device (8) is arranged at the bottom of the robot frame (1) and is used to drive the robot frame (1) to move within the plot to be harvested; A navigation system (5) is arranged at the front of the robot frame (1) and is used to obtain spatial information of the plot of land to be harvested, wherein the spatial information of the plot of land is used for navigation of the walking device (8); A visual system (7) is arranged at the front of the robot frame (1) and is used to collect images of land plots on a path of travel when the robot frame (1) moves; A digging device (6) is arranged in the middle of the robot frame (1) and is used to dig individual potatoes in the plot to be harvested; The excavating device (6) comprises excavating components arranged opposite to each other, the excavating components comprising: An excavation upper arm (64), the top of which is hinged to the upper frame of the robot frame (1) via an upper fixing plate (61); a first hydraulic cylinder (62) is connected to the middle thereof; the first hydraulic cylinder (62) is used to adjust the posture of the excavation upper arm (64) relative to the robot frame (1); The excavation lower arm (65) is arranged at the lower side of the excavation upper arm (64), and the top of the excavation lower arm (65) is connected to the excavation upper arm (64) via a second hydraulic cylinder (66); the second hydraulic cylinder (66) is used to adjust the posture of the excavation lower arm (65) relative to the excavation upper arm (64); An excavating shovel (67) is arranged at the circular tail end of the excavating lower arm (65); a third hydraulic cylinder (68) and an excavating connection mechanism (69) are connected between the excavating shovel (67) and the excavating lower arm (65); the third hydraulic cylinder (68) is used to adjust the posture of the excavating shovel (67) relative to the excavating lower arm (65); A control device (3) is communicatively connected to the walking device (8), the navigation system (5), the visual system (7) and the excavating device (6); The control device (3) is used to perform global path planning for the entire plot to be harvested based on pre-acquired plot spatial information of the plot to be harvested, and to control the walking device (8) to drive the robot frame (1) to perform automatic navigation walking within the plot to be harvested according to the planned global path; and for receiving the image data to be detected and the point cloud data collected by the visual system (7) when the robot frame (1) moves, performing potato target detection on the image data to be detected, and calculating potato ridge point cloud data based on the point cloud data; It is also used to control the robot frame (1) to move to the potato target according to the potato ridge point cloud data when a potato target is detected in the image data to be detected, and then control the digging device (6) to perform single-plant digging.

2. A potato breeding experiment harvesting robot according to claim 1, characterized in that: Also includes: A power device (2) for providing power to the excavating device (6) and the walking device (8); It comprises a diesel oil pipe (21), a fixing small plate (22), a diesel oil tank (23), an oil tank fixing plate (24), a hydraulic valve (25), a hydraulic oil tank oil outlet nozzle (26), a diesel engine hydraulic pump assembly (27), a hydraulic oil pipe (28) and a silencer tail pipe (29); The diesel engine hydraulic pump assembly (27) is mounted on the rear end of the robot frame (1) via a fixing plate (22); a hydraulic oil tank outlet nozzle (26) is mounted on the upper end of the diesel engine hydraulic pump assembly (27); and a silencer tail pipe (29) is mounted on the lower end; a hydraulic valve (25) is mounted on one side of the diesel engine hydraulic pump assembly (27); the hydraulic valve (25) is connected to the hydraulic oil tank outlet nozzle (26) via a hydraulic oil pipe (28); a diesel oil tank (23) is connected to the diesel engine hydraulic pump assembly (27) via a diesel oil pipe (21); an oil tank fixing plate (24) is mounted on the upper end of the diesel oil tank (23); and the diesel oil tank (23) is mounted on one side of the robot frame (1) via the oil tank fixing plate (24).

3. A potato breeding experiment harvesting robot according to claim 1, characterized in that: The upper fixing plate (61) and the lower fixing plate (63) are both mounted on the robot frame (1) via grooves.

4. A potato breeding experiment harvesting robot according to claim 1, characterized in that: Also includes: A power supply device (4) is used to supply power to the control device (3), the navigation system (5), and the visual system (7), and comprises a battery (41) and a battery mounting plate (42), wherein the lower end of the battery (41) is fixed to the battery mounting plate (42), and the battery mounting plate (42) is fixed to the frame.

5. A potato breeding experiment harvesting robot harvesting method, using a potato breeding experiment harvesting robot according to any one of claims 1 to 4, characterized in that: The following steps are involved: Step S1, based on the pre-acquired spatial information of the plot to be harvested, a global path planning is performed for the entire plot to be harvested, and the walking device (8) is controlled to drive the robot frame (1) to automatically navigate and walk in the plot to be harvested according to the planned global path; Step S2, when the robot frame (1) moves, acquiring the image data to be detected and the point cloud data collected by the visual system (7), performing potato target detection on the image data to be detected, and calculating potato ridge point cloud data based on the point cloud data; Step S3, if a potato target is detected in the image data to be detected, the robot frame (1) is controlled to move to the potato target according to the potato ridge point cloud data, and then the digging device (6) is controlled to perform single-plant digging; Step S4, repeat steps S2 to S3.

6. A potato breeding experiment harvesting robot harvesting method according to claim 5, characterized in that: Step S3 comprises: if a potato target is detected in the image data to be detected, controlling the robot frame (1) to move to the potato target according to the potato ridge point cloud data, and then controlling the digging device (6) to perform single-plant digging; Step S31, controlling the harvesting robot to move forward according to the potato plants and potato ridge point cloud data identified by the visual system (7), so that the digging device (6) moves above a potato plant; Step S32, processing the acquired potato ridge point cloud data to calculate the potato ridge height and fit the potato ridge contour line; then determining the best burrowing position of the digging shovel (67) according to the potato ridge height and the fitted potato ridge contour line, and converting the coordinates of the burrowing position into coordinates in the earth world coordinate system; Step S33, controlling the digging device (6) to reach the best burial position to dig and obtain potatoes.

7. A potato breeding experiment harvesting robot harvesting method according to claim 6, characterized in that: In step S31, the obtained potato ridge point cloud data is processed to calculate the height of the potato ridge and fit the potato ridge contour line; Wherein, the method for obtaining the potato ridge point cloud data is: Performing posture correction on the original point cloud data by point cloud transformation; the original point cloud data includes potato point cloud data, potato ridge point cloud data and ground point cloud; The point cloud data obtained after point cloud transformation is denoised using bilateral filtering; Using CSF algorithm to separate the ground point cloud to obtain potato point cloud data and potato ridge point cloud data; The step of processing the acquired potato ridge point cloud data to calculate the height of the potato ridge and fit the potato ridge contour line comprises: For the potato ridge point cloud data, 20 point cloud sections were randomly selected in the direction perpendicular to the potato ridge; Extracting point cloud data on each section and calculating the mean value to obtain the height of the potato ridge; Then, the Gaussian model and least squares method are used to fit the potato ridge contour line to the extracted potato ridge point cloud data.

8. The potato breeding experiment harvesting robot harvesting method according to claim 6, characterized in that: The excavating device (6) comprises excavating components arranged relatively to each other, the excavating components comprising: an excavating upper arm (64), a first hydraulic cylinder (62), an excavating lower arm (65), a second hydraulic cylinder (66), an excavating shovel (67) and a third hydraulic cylinder (68); The step S33 comprises: Step S331, controlling the first hydraulic cylinder (62) and the second hydraulic cylinder (66) so that the digging shovel (67) reaches the optimal position for entering the soil, and then driving the third hydraulic cylinder (68) so that the digging shovel (67) performs digging operations; Step S332: When the two digging shovels (67) installed opposite to each other have each penetrated the soil to half the width of the potato ridge, the third hydraulic cylinder (68) is controlled to stop operating, and the second hydraulic cylinder (66) is controlled to contract, so that the digging lower arm (65) is lifted above the ridge top; Step S333, controlling the third hydraulic cylinder (68) to perform three telescopic operations, so as to shake off soil and other impurities in the harvested potatoes by shaking the digging shovel (67); Step S334, controlling the third hydraulic cylinder (68) to contract, driving the digging shovel (67) to open to both sides, so that the dug potatoes fall onto the ridge, thereby completing the harvest; Step S335: Control the first hydraulic cylinder (62) and the second hydraulic cylinder (66) to retract, thereby driving the digging lower arm (65) and the digging shovel (67) away from the potato ridge.

Images