CN-121970612-A - Self-propelled red pepper picking system

CN121970612ACN 121970612 ACN121970612 ACN 121970612ACN-121970612-A

Abstract

The invention relates to the technical field of red pepper picking, in particular to a self-propelled red pepper picking system which comprises a red pepper picking device and a central processor, wherein the central processor is arranged on the red pepper picking device, the red pepper picking device comprises a self-propelled chassis, a first mechanical arm, a second mechanical arm, an image acquisition module, a flexible end effector, a control box and a branch pulling module, and the first mechanical arm, the second mechanical arm and the control box are all arranged on the self-propelled chassis. According to the invention, the picking path planning is determined by the central processing unit according to the actual current position of the flexible end effector and the actual target position of the red pepper cluster, the flexible end effector is moved to the position of the red pepper cluster to be picked according to the picking path planning, and then the red pepper cluster is picked by the flexible end effector, so that the red pepper is automatically picked from near to far in sequence, the dependence of manual picking is reduced, and the labor intensity of picking personnel is relieved.

Inventors

CHEN LIN
LI YUANFU
ZHAO HANQI
HE JIMEI
LIU YANG
Chen Zongran
XIAO LIANGHUI

Assignees

四川农业大学

Dates

Publication Date: 20260505
Application Date: 20260326

Claims (10)

1. The self-propelled red pepper picking system is characterized by comprising a red pepper picking device (100) and a central processor, wherein the central processor is arranged on the red pepper picking device (100); The red pepper picking device (100) comprises a self-propelled chassis (110), a first mechanical arm (120), a second mechanical arm (130), an image acquisition module (140), a flexible end effector (150), a control box (170) and a branch pulling module (180), wherein the first mechanical arm (120), the second mechanical arm (130) and the control box (170) are all arranged on the self-propelled chassis (110), the image acquisition module (140) comprises a depth camera, the depth camera is arranged on the output end of the first mechanical arm (120), and the flexible end effector (150) is arranged on the output end of the second mechanical arm (130); The image acquisition module (140) can acquire red pepper image information, the image acquisition module (140) can transmit the red pepper image information to the central processing unit, the central processing unit can analyze the image information and identify red pepper clusters in the image, the central processing unit can reconstruct a three-dimensional model of the red pepper clusters to be picked according to position information acquired by the depth camera and determine the corresponding actual target position of the red pepper clusters to be picked in an actual coordinate system according to the image, and the central processing unit can determine the execution action of the mechanical arm II (130) and the picking path planning of the flexible end effector (150) according to the actual current position of the flexible end effector (150) in the actual coordinate system, the actual target position of the red pepper clusters and the action model of the mechanical arm II (130).
2. The self-propelled red pepper picking system as claimed in claim 1, wherein the red pepper picking device (100) further comprises an inflator pump (160), the inflator pump (160) is used for providing an air source for the flexible end effector (150), an air pump controller is arranged inside the control box (170), the air pump controller is communicated with the flexible end effector (150) through an air pipe, and the air pump controller can adjust the driving pressure of the flexible end effector (150).
3. The self-propelled red pepper picking system according to claim 1, wherein when determining an actual target position corresponding to the red pepper cluster to be picked in the actual coordinate system according to the image information, the actual target position is determined by using a preset conversion model and a virtual target position, wherein the conversion model is a conversion relationship between the virtual target position of the red pepper cluster in the image and the actual position in the actual coordinate system.
4. The self-propelled red pepper picking system as claimed in claim 1, wherein said image acquisition module (140) further comprises a mounting plate (141), said depth camera being fixedly arranged on the mounting plate (141), said mounting plate (141) being fixedly arranged on the output end of the first mechanical arm (120); The pruning module (180) includes: The base (181) is fixedly arranged on the mounting plate (141), and an adjusting motor is arranged in the base (181); the transmission shaft (182) is rotatably arranged on the base (181), and the output end of the adjusting motor is in transmission connection with the transmission shaft (182); The toggle rod (183) is fixedly arranged on the transmission shaft (182).
5. The self-propelled red pepper picking system as claimed in claim 1, wherein when identifying red pepper clusters in the image, coordinates are marked in the image for image information which cannot be completely confirmed as red pepper clusters, i.e. image information of uncertain red pepper clusters whose characteristic parameters do not meet a threshold, triggering a discard picking instruction; Transmitting the marked real-time image data to a handheld control panel through wireless communication, synchronously displaying the position mark, morphological parameters and peripheral branch tip environments of the fruit clusters on a panel end interface, manually issuing a picking or confirming a picking-discarding instruction by a user according to actual planting requirements, selecting a three-dimensional coordinate point corresponding to a target fruit cluster in a manual clicking model according to a three-dimensional space model of the pepper plant, manually selecting a picking position, and automatically planning a picking path according to the picking position by a central processor.
6. The self-propelled red pepper picking system according to any one of claims 1 to 5, wherein when the central processing unit recognizes red pepper clusters in the image, the collected image information is input into the YOLO neural network after the preset training is completed, and a recognition result is output, wherein the recognition result includes the area information of the red pepper clusters in the image, and fruits of the red pepper clusters to be picked are classified into dispersed clusters and clustered clusters.
7. The self-propelled red pepper picking system of claim 6, wherein the pre-trained neural network is trained by using a training sample set to obtain a trained YOLO neural network, wherein an initial image sample set is firstly obtained, the initial image sample set comprises a plurality of sample images, each sample image in the initial image sample set contains red pepper clusters, the sample images in the initial image sample set are marked, the areas of the red pepper clusters in the sample images are selected in a frame, the types of the red pepper clusters are marked to obtain a training sample set, the training sample set is input into the pre-trained neural network to train, and the trained YOLO neural network is formed after the training is completed.
8. The self-propelled red pepper picking system according to any one of claims 1-5, wherein when the mechanical arm two (130) drives the flexible end effector (150) to move, collision detection of fruits and branches is performed during picking action, firstly, according to the current joint angle, the length of a connecting rod and a connecting mode of the mechanical arm two (130), the real-time gesture of each joint and connecting rod of the mechanical arm two (130) in a three-dimensional space is obtained through positive kinematic calculation, and an integral model comprising the joint sphere, the enveloping body of the connecting rod and the appearance of the end effector is constructed based on the structural parameters of the mechanical arm two (130); Meanwhile, the dynamic environment point cloud is collected and updated through the image collection module (140), and the electronic fence boundaries set according to the operation area are fused together to generate a unified environment grid representation, the central processor utilizes a space minimum distance calculation method to carry out periodic solution on the distance between any connecting rod surface of the mechanical arm II (130) and all obstacle surfaces in the environment, and synchronous detection is carried out on possible self-collision among the connecting rods, and when any minimum distance is lower than a set safety critical threshold, the potential collision risk in the current state is judged.
9. The self-propelled red pepper picking system according to claim 8, wherein in the collision detection process, the theoretical moment of each joint is calculated in real time based on the kinetic model of the mechanical arm two (130), and compared with the actual current signal fed back by the driving motor, when an abnormal stress condition inconsistent with the normal movement trend occurs, the system regards the condition as an advance sign of contact collision, so as to perform secondary verification on the geometric detection result.
10. The self-propelled red pepper picking system of claim 9, wherein the limiting mode is entered when any one of the geometric distance determination or the force check satisfies a collision risk condition, and wherein the mechanical arm two (130) is prevented from continuing to move to the high risk area by adjusting the joint speed, limiting the direction of movement, or suspending the trajectory execution.

Description

Self-propelled red pepper picking system Technical Field The invention relates to the technical field of red pepper picking, in particular to a self-propelled red pepper picking system. Background The red pepper is a special spice and Chinese medicinal material in China, is widely planted in areas such as Sichuan, northwest and the like, is mainly manually picked in the process of picking the red pepper, has high labor intensity and low operation efficiency, and the surfaces of plant branches are densely covered with sharp thorns, so that picking staff are easily scratched by the thorns on the branches in the operation process, the operation environment is severe, the labor cost is high, and although the hand-held picking equipment can be adopted for assisting in picking, such as saw teeth, scissors, rotation and the like, manual hand-held operation is still needed, and the labor intensity of the picking staff cannot be effectively relieved. Disclosure of Invention The invention aims to provide a self-propelled red pepper picking system for solving the problems in the background technology. In order to achieve the above purpose, the present invention provides the following technical solutions: The self-propelled red pepper picking system comprises a red pepper picking device and a central processor, wherein the central processor is arranged on the red pepper picking device; The red pepper picking device comprises a self-propelled chassis, a first mechanical arm, a second mechanical arm, an image acquisition module, a flexible end effector, a control box and a branch pulling module, wherein the first mechanical arm, the second mechanical arm and the control box are all arranged on the self-propelled chassis, the image acquisition module comprises a depth camera, the depth camera is arranged at the output end of the first mechanical arm, the flexible end effector is arranged at the output end of the second mechanical arm, the central processor is arranged inside the control box, and the self-propelled chassis is fixedly provided with a collecting frame; The image acquisition module can acquire image information of red pepper, the image acquisition module can transmit the image information of the red pepper to the central processing unit, the central processing unit can analyze the image information, identify red pepper clusters in the image, the central processing unit can reconstruct a three-dimensional model of the red pepper clusters to be picked according to the position information acquired by the depth camera, determine the corresponding actual target position of the red pepper clusters to be picked in an actual coordinate system according to the image, and the central processing unit can determine the execution action of the mechanical arm II and the picking path planning of the mechanical arm II according to the actual current position of the flexible end effector in the actual coordinate system, the actual target position of the red pepper clusters and the action model of the mechanical arm II, and execute the picking action through the flexible end effector after the mechanical arm II moves the flexible end effector to the actual target position of the red pepper clusters to be picked. Further in that, the device is picked to red pepper still includes the pump, and the pump is used for providing the air supply to flexible end effector, and the control box is inside to be provided with the air pump controller, and the air pump controller passes through the trachea intercommunication with flexible end effector, and the driving pressure of flexible end effector can be adjusted to the air pump controller, realizes opening, centre gripping, multiple action modes such as release. When the corresponding actual target position of the red pepper cluster to be picked in the actual coordinate system is determined according to the image information, the actual target position is determined by utilizing a preset conversion model and a virtual target position, wherein the conversion model is a conversion relation between the virtual target position of the red pepper cluster in the image and the actual position in the actual coordinate system, the virtual target position is the cluster center position of the red pepper cluster to be picked in the image, and the actual target position comprises a plane position and a height position. Further, when the red pepper cluster in the image is identified, the image information of the red pepper cluster which cannot be completely confirmed, namely the image information of the uncertain red pepper cluster with the characteristic parameters not meeting the threshold value, is subjected to coordinate marking in the image, and a picking instruction is triggered to be abandoned; Transmitting the marked real-time image data to a handheld control panel through wireless communication, synchronously displaying the position mark, the morphological parameters and the surrounding branc