Automatic saffron picking system

Abstract

The invention discloses an automatic crocus sativus picking system. The system comprises a crocus sativus stylus recognition method, a stylus posture modeling method and a manipulator posture control method. The crocus sativus stylus identification method comprises the following steps: respectively carrying out image identification processing on stylus contours which can be used as medicines and petals which cannot be used as medicines in a single-plant crocus sativus flower, constructing three-dimensional point cloud data, and numbering each stylus; the stylus attitude modeling method comprises the following steps: constructing an attitude vector of each stylus for each identified stylus in a global coordinate system; according to the manipulator posture control method, coordinates of a shearing point of a manipulator and a reference plane posture of a shearing blade are adjusted according to a posture vector of each stylus, so that the stylus is sheared from a lower position. The control device is used for controlling the motion trail of the automatic saffron picking manipulator, the stylus part capable of being used as a medicine can be quickly recognized, the posture of the manipulator is adjusted, the bottom of the stylus is cut off, and maximum cutting is achieved.

Description

[Technical Field]

[0001] This invention relates to the technical field of automatic saffron harvesting, and in particular to an automatic saffron harvesting system. [Background Technology]

[0002] Saffron, also known as Tibetan saffron, has extremely low domestic and international production. In my country, it is listed as a rare and precious Chinese medicinal material. At present, saffron cultivation mainly adopts the traditional natural planting method. Correspondingly, the harvesting also requires the traditional manual method. Therefore, during the saffron harvest month each year, there will be a situation of concentrated labor but low efficiency. It takes about 100,000 to 200,000 flowers to produce one kilogram of medicinal saffron, so the labor cost is high.

[0003] Existing technologies have proposed some refined and mechanized methods for saffron cultivation and harvesting, such as the cultivation box for indoor saffron cultivation proposed in ZL201710332577.6, and the saffron reverse docking air suction fully open self-positioning flexible stigma stripping robot proposed in CN201910725093.7. These technologies have improved the production efficiency of saffron to a certain extent, but there is currently no fully automated harvesting technology solution.

[0004] The key technical challenge in achieving fully automated saffron harvesting lies in the identification and maximizing of the stigma, necessitating the development of an automated saffron harvesting system. [Summary of the Invention]

[0005] The purpose of this invention is to solve the problems in the prior art and propose an automatic saffron harvesting system. This system controls the movement trajectory of the automatic saffron harvesting robot, which can quickly identify the pistil part that can be used in medicine and adjust the posture of the robot to cut it from the bottom of the pistil, thereby maximizing the cutting.

[0006] To achieve the above objectives, this invention proposes an automatic saffron harvesting system for controlling the motion trajectory of an automatic saffron harvesting robot. The system includes a saffron pistil recognition method, a pistil posture modeling method, and a robot posture control method. The saffron pistil recognition method performs image recognition processing on the outlines of medicinal pistils and non-medicinal petals within a single saffron flower, constructing three-dimensional point cloud data and numbering each pistil. The pistil posture modeling method constructs the posture vector of each identified pistil in a global coordinate system. The robot posture control method adjusts the coordinates of the robot's cutting point and the posture of the cutting blade reference plane based on the posture vector of each pistil, thereby cutting the pistil from a lower position.

[0007] Preferably, the saffron style identification method includes the following steps:

[0008] 1) Construct a global Cartesian coordinate system;

[0009] 2) Images are acquired from the top of the target saffron using a depth camera to generate 3D point cloud data. The coordinate values ​​of each node in the point cloud data include three dimensions: X, Y, and Z.

[0010] 3) Extract the outline nodes of each medicinal flower column from the point cloud data, construct a node group for the outline nodes of each flower column, and assign a number to each node group.

[0011] Preferably, the flower column posture modeling method includes the following steps:

[0012] a) Extract the coordinates of the nodes containing the maximum and minimum Z-axis coordinates from the node group representing each flower column, i.e., the top and bottom coordinates;

[0013] b) The attitude vector of each flower column in the global coordinate system is formed by the top and bottom coordinates of each flower column.

[0014] Preferably, the robotic arm posture control method includes the following steps:

[0015] A) Derive the coordinates of the shearing point from the coordinates of the end of the flower column;

[0016] B) Derive the vertical plane of the flower column attitude vector from the flower column attitude vector, and make the reference plane of the robot shearing blade parallel to the vertical plane of the flower column attitude vector.

[0017] The beneficial effects of this invention are:

[0018] The automated saffron harvesting system provides fast and highly reliable robotic arm trajectory control logic for automated saffron harvesting machinery. Since the saffron stigma exhibits a relatively upright posture during the flower's opening period, the system models the stigma using posture vectors. This approach involves minimal computational data, high reliability, and significant practicality. By controlling the shearing blade's reference plane to be perpendicular to the stigma's posture vector and determining the shearing point coordinates based on this vector, the system can cut off the majority of the stigma without miscutting, thereby increasing yield.

[0019] The features and advantages of the present invention will be described in detail through embodiments.

Detailed Implementation Methods

[0020] Example 1

[0021] An automatic saffron harvesting system is provided for controlling the motion trajectory of an automatic saffron harvesting robot, including a saffron pistil recognition method, a pistil posture modeling method, and a robot posture control method.

[0022] The method for identifying saffron stigmas includes the following steps:

[0023] 1) Construct a global Cartesian coordinate system O-xyz;

[0024] The reachable space of saffron culture medium and automated harvesting robot uses the same global coordinate system. This global coordinate system has a fixed origin O (0, 0, 0). The calibration of the origin is a mature technology in the industry and will not be elaborated further.

[0025] 2) Images are acquired from the top of the target saffron using a depth camera, generating 3D point cloud data. The coordinates of each node N in the point cloud data contain three dimensions: X, Y, and Z. n (X n ,Y n Z n ), where n represents the nth node;

[0026] 3) Extract the outline nodes of each medicinal flower column from the point cloud data. The outline node must contain at least two points: the top and the bottom. Construct a node group G for the outline nodes of each flower column and assign a number to each node group to form a node group library [G].

[0027] The method for modeling the posture of the flower column includes the following steps:

[0028] a) From the node group representing each flower column, extract the coordinates of the node containing the maximum and minimum Z-axis coordinates, i.e., the top coordinate N. T (X t ,Y t Z t ) and end coordinate N D (X d ,Y d Z d );

[0029] b) The coordinates of the top of each flower column N T and the end coordinate N D This forms the attitude vector of each flower column in the global coordinate system.

[0030] The robotic arm posture control method includes the following steps:

[0031] A) From the coordinate N of the end of the flower column D (X d ,Y d Z d ), and derive the coordinates N of the shear point. C (X c ,Y c Z c );

[0032] The shearing point is set 0.5mm above the end node of the flower column, i.e.

[0033] N c =N d +(0,0,0.5)=(X d ,Y d Z d +0.5)

[0034] B) From the flower column posture vector Derive the vertical plane of the flower column attitude vector and make the reference plane of the robot shearing blade parallel to the vertical plane of the flower column attitude vector.

[0035] Example 2

[0036] The saffron automatic harvesting system controls an automatic saffron harvesting robot, which includes two L-shaped shearing blades. A support shaft is fixedly connected to the top of each shearing blade. The support shaft consists of a fixed shaft and a moving shaft. The bottom of the shearing blade, fixed to the fixed shaft, serves as the reference plane for the shearing blade. A drive unit is connected to the top of the moving shaft to control its rotation, causing the shearing blade connected to the moving shaft to rotate clockwise and counterclockwise relative to the other fixed shearing blade, thus achieving the shearing action. The L-shaped shearing blades need to have their shearing blades as short as possible to minimize interference with non-medicinal parts during the robot's movement from the top of the saffron down to the shearing point.

[0037] The above embodiments are illustrative of the present invention and are not intended to limit the present invention. Any simple modifications to the present invention are within the scope of protection of the present invention.

Claims

1. An automatic saffron harvesting system for controlling the motion trajectory of an automatic saffron harvesting robot, characterized in that: The invention includes a saffron style recognition method, a style posture modeling method, and a robotic arm posture control method. The saffron style recognition method performs image recognition processing on the outlines of medicinal styles and non-medicinal petals within a single saffron flower, constructing three-dimensional point cloud data and numbering each style. The style posture modeling method constructs the posture vector of each identified style in a global coordinate system. The robotic arm posture control method adjusts the coordinates of the robotic arm's cutting point and the posture of the cutting blade reference plane based on the posture vector of each style, thereby cutting the style from a lower position.

2. The saffron automatic picking system as claimed in claim 1, characterized in that: The method for identifying saffron stigmas includes the following steps: 1) Construct a global Cartesian coordinate system; 2) Images are acquired from the top of the target saffron using a depth camera to generate 3D point cloud data. The coordinate values ​​of each node in the point cloud data include three dimensions: X, Y, and Z. 3) Extract the outline nodes of each medicinal flower column from the point cloud data, construct a node group for the outline nodes of each flower column, and assign a number to each node group.

3. The saffron automatic picking system as claimed in claim 1, characterized in that: The method for modeling the posture of the flower column includes the following steps: a) Extract the coordinates of the nodes containing the maximum and minimum Z-axis coordinates from the node group representing each flower column, i.e., the top and bottom coordinates; b) The attitude vector of each flower column in the global coordinate system is formed by the top and bottom coordinates of each flower column.

4. The saffron automatic picking system as claimed in claim 1, characterized in that: The robotic arm posture control method includes the following steps: A) Derive the coordinates of the shearing point from the coordinates of the end of the flower column; B) Derive the vertical plane of the flower column attitude vector from the flower column attitude vector, and make the reference plane of the robot shearing blade parallel to the vertical plane of the flower column attitude vector.