Automatic broussonetia ramification device and control method

Abstract

The present application provides a broussonetia automatic seedling device and control method, the device contains mounting plate, pit unit, conveying unit, filling unit, control unit, irrigation unit, level, mobile wheel and image recognition unit. The pit, conveying, filling unit are all set in the mounting plate, the irrigation unit is equipped with a solenoid valve, and is connected with the control unit. The level is also connected with the control unit. The image recognition unit is integrated in the control unit, which identifies the position of the tree planting pit and controls the conveying unit to accurately send seedlings. The present application uses the level to detect the state of the device in real time. When the inclination angle of the device is greater than the preset threshold, the level sends a signal to the control unit. After receiving the signal, the control unit controls the telescopic rod to adjust the device, so that the device returns to the horizontal state. At the same time, the device makes the broussonetia always perpendicular to the horizontal plane when planting, preventing the broussonetia from growing obliquely, thereby improving the growth quality and survival rate of the broussonetia.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of seedling planting machines, in particular to a poplar automatic seedling planting device and control method. BACKGROUND

[0002] Poplar is an important tree species for wind prevention and sand fixation, soil and water conservation, and ecological environment improvement, and is usually planted in areas such as river banks and lake banks that are easily affected by soil and water loss.

[0003] The traditional poplar planting method is usually through artificial digging, planting, filling and watering processes, which not only has high labor intensity, but also has low efficiency. With the development of science and technology, automation technology is more and more widely used in the field of agriculture, especially in the planting link, the appearance of automatic seedling planting device greatly improves the planting efficiency and accuracy. However, the existing automatic seedling planting device still has some problems in actual application. For example, due to the complex and changeable terrain, the existing device is easy to misjudge when identifying the tree planting hole, resulting in inaccurate planting position; in addition, due to the complex terrain in the areas such as river banks and lake banks, the existing device is easy to tilt during operation, affecting the planting effect, and the survival rate of planted poplar is low. SUMMARY

[0004] The present application aims to at least solve the technical problem of low survival rate of planted poplar in the prior art, and particularly innovatively provides a poplar automatic seedling planting device and control method.

[0005] In order to achieve the above-mentioned purpose of the present application, the present application provides a poplar automatic seedling planting device, which comprises:

[0006] a mounting plate;

[0007] a digging unit arranged on the mounting plate and used for digging a tree planting hole;

[0008] a conveying unit arranged on the mounting plate and used for conveying a seedling into the tree planting hole;

[0009] a filling unit arranged on the mounting plate and used for filling soil to the root of the seedling;

[0010] a power supply arranged on the mounting plate and used for providing electric energy for the device;

[0011] a control unit connected with the digging unit, the conveying unit and the filling unit;

[0012] a watering unit arranged on the mounting plate and used for watering the seedling filled with soil, wherein an electromagnetic valve is arranged on the watering unit, and the electromagnetic valve is connected with the control unit;

[0013] a level gauge arranged on the mounting plate and connected with the control unit.

[0014] At least four mobile wheels are arranged at the bottom of the mounting plate through telescopic rods, each of the mobile wheels is provided with a wheel driver connected with the control unit, the telescopic rods are connected with the control unit, and the control unit adjusts the length of the telescopic rods in real time based on the signals sent by the level meter;

[0015] An image recognition unit is arranged on the mounting plate and integrated on the control unit, used for recognizing the position information of the tree planting hole and sending the position information to the control unit, and the control unit controls the conveying unit to convey the sapling into the tree planting hole.

[0016] As an optional embodiment of the present application, the hole digging unit comprises:

[0017] At least two air cylinders are arranged on the mounting plate and connected with the control unit;

[0018] A moving plate is arranged on the air cylinder and used for moving under the driving of the air cylinder;

[0019] An auger driver is arranged at the bottom of the moving plate and connected with the control unit;

[0020] An auger is arranged on the output shaft of the auger driver and used for digging soil to form a tree planting hole.

[0021] As an optional embodiment of the present application, the conveying unit comprises:

[0022] A reversing groove is arranged on the mounting plate, the reversing groove comprises a first bent plate, a rotating shaft, a straight plate and a second bent plate, the first bent plate is rotationally connected with one end of the straight plate through the rotating shaft, the second bent plate is rotationally connected with the other end of the straight plate through the rotating shaft, electromagnets are arranged on the top of the first bent plate and the second bent plate, the electromagnets are connected with the control unit, an axle fixing sleeve is arranged at the bottom of the straight plate, and a protrusion is arranged in the axle fixing sleeve;

[0023] A reversing groove driver is arranged on the mounting plate and used for rotating one end of the reversing groove to the tree planting hole;

[0024] A conveyor belt is arranged on the mounting plate and inclined to the reversing groove, and at least one baffle is arranged on the conveyor belt;

[0025] A conveyor belt driver is arranged on the mounting plate and used for driving the conveyor belt.

[0026] As an optional embodiment of the present application, the filling unit comprises:

[0027] At least two stretchers are arranged at the bottom of the mounting plate and connected with the control unit;

[0028] At least two arc-shaped filling boards are arranged on the output shaft of the stretchers and used for filling soil to the saplings.

[0029] At least two push plates are symmetrically arranged at the bottom of the mounting plate.

[0030] As an optional embodiment of the present application, the image recognition unit comprises an image acquisition module and an image recognition model.

[0031] The image acquisition module sends the real-time collected image data of the tree-planting hole to the image recognition model, and the image recognition model identifies the image data of the tree-planting hole to determine whether the image contains the tree-planting hole.

[0032] If the image data of the tree-planting hole does not contain the tree-planting hole, the control unit controls the wheel driver to make the device continue to move forward.

[0033] If the image data of the tree-planting hole contains the tree-planting hole, the control unit controls the wheel driver to make the device stop moving forward.

[0034] As an optional embodiment of the present application, the image recognition model is trained by a deep learning algorithm.

[0035] The step of training the image recognition model comprises:

[0036] S1, obtaining historical image data of a tree-planting hole and pre-processing the historical image data of the tree-planting hole;

[0037] S2, dividing the pre-processed image data of the tree-planting hole into a training data set and a test data set;

[0038] S3, obtaining an initial model and performing feature decomposition on the initial model to obtain heavy particle parameters and feature sub-parameters;

[0039] S4, iteratively training the initial model using the training data set and adjusting the heavy particle parameters and the feature sub-parameters through a loss function and an incremental learning algorithm;

[0040] S5, verifying the trained initial model using the test data set;

[0041] S6, adjusting the heavy particle parameters and the feature sub-parameters of the initial model based on the verification result, repeating steps S4 and S5 until the initial model reaches a predetermined recognition accuracy, and obtaining the image recognition model.

[0042] As an optional embodiment of the present application, the expression of the loss function is:

[0043]

[0044] in, Indicates the loss value. Indicates the number of samples. Indicates the first One sample, Indicates an indicator variable. Indicates taking the logarithm. Indicates the first The predicted probability value of each sample.

[0045] As an optional embodiment of the present invention, the image recognition model may optionally include an input layer, a convolutional layer, an activation layer, a pooling layer, a fully connected layer, a Dropout layer, and a batch normalization layer. The input layer is used to receive tree pit image data acquired by the image acquisition module; the convolutional layer is used to extract image features from the tree pit image data; the activation layer is used to introduce nonlinear transformations; the pooling layer is used to reduce feature dimensionality; the fully connected layer is used to integrate and classify the extracted features; the Dropout layer is used to prevent the image recognition model from overfitting; and the batch normalization layer is used to accelerate the training process of the image recognition model.

[0046] On the other hand, the present invention also provides a control method for an automatic mulberry seedling planting device, including the device; the method further includes:

[0047] A1. Start the control unit and use the control unit to initialize the device;

[0048] A2. The control unit sends a control signal to the digging unit to start the digging unit to perform digging operations;

[0049] A3. After the hole is dug, the control unit receives the completion signal from the hole-digging unit and sends a control signal to the wheel drive and image recognition unit. The device begins to move forward. The image acquisition module collects image data in real time and sends it to the image recognition model. When the image recognition model recognizes the planting hole, the wheel drive stops working.

[0050] A4. After receiving the stop signal from the wheel drive, the control unit sends a control signal to the conveying unit to start the conveying unit to carry out the seedling conveying operation.

[0051] A5. After the saplings are transported to the planting pit, the control unit receives the completion signal from the transport unit and sends a control signal to the filling unit to start the filling operation.

[0052] A6, after the filling is completed, the control unit receives the completion signal of the filling unit, sends a start signal to the wheel driver, the device drives forward to a preset distance, stops the wheel driver, the control unit sends a watering command to the watering unit, and starts the watering unit to perform watering work;

[0053] A7, after the watering is completed, the control unit receives the completion signal of the watering unit;

[0054] A8, after the control unit receives the completion signal of the watering unit, it is judged whether there is a tree seedling to be planted or a tree planting hole to be completed, if yes, steps A1 to A7 are repeated, and if no, the device enters a maintenance mode and waits for the next planting task.

[0055] As another optional embodiment of the application, optionally, the A3 further comprises:

[0056] If the image recognition model recognizes the tree planting hole within the preset distance, the control unit controls the wheel driver to drive the device backward to the position of the last tree planting hole, and controls the hole digging unit to perform hole digging work again;

[0057] The A4 further comprises:

[0058] A401, when the conveying belt conveys the tree seedling to the reversing groove, the control unit sends a control command to the electromagnet at the top of the first bending plate and the second bending plate, so that the top of the first bending plate and the top of the second bending plate are closed under the action of the two electromagnets, and the tree seedling is fixed;

[0059] A402, after the control unit receives the tree seedling fixing information, the control unit sends a control information to the reversing groove driver, and the reversing groove driver drives the reversing groove to rotate above the tree planting hole;

[0060] A403, after the control unit receives the reversing groove rotation completion signal, the control unit sends a control command to the two electromagnets, so that the first bending plate and the second bending plate are separated, the tree seedling is released from the reversing groove, and the tree seedling falls into the tree planting hole;

[0061] The control method further comprises:

[0062] A9, before digging the tree planting hole, when the level detector detects that the inclination angle of the device exceeds a preset threshold, the control unit sends a control signal to the corresponding telescopic rod to adjust the length of the telescopic rod, so that the device restores to be horizontal, and the device continues to work.

[0063] The beneficial effects of the present application are that the present application automatically plants broussonetia by installing a hole digging unit, a conveying unit, a soil filling unit and a control unit, reduces labor costs, and uses a level real-time detection device, when the inclination angle of the device is greater than a preset threshold, the level sends a signal to the control unit, and the control unit controls the telescopic rod to adjust the device to restore it to a horizontal state, so that the device always grows vertically to the horizontal plane when planting broussonetia, preventing broussonetia from growing obliquely, thereby improving the growth quality and survival rate of broussonetia. In addition, the image recognition model is used to identify the tree planting hole in real time, so that the seedling can be accurately planted into the tree planting hole, improving the accuracy and efficiency of planting.

[0064] Additional aspects and advantages of the present application will be in part apparent and in part pointed out hereinafter in the description of the application. BRIEF DESCRIPTION OF DRAWINGS

[0065] The above and / or additional aspects and advantages of the present application will become apparent and be readily appreciated from the description of the embodiments, given in conjunction with the accompanying drawings, in which:

[0066] Figure 1 is the control logic diagram of the automatic seedling planting device of the present application.

[0067] Figure 2 is a structural schematic diagram of the automatic seedling planting device of the present application.

[0068] Figure 3 is a partial structural schematic diagram of the automatic seedling planting device of the present application.

[0069] Figure 4 is a structural schematic diagram of the reversing groove of the automatic seedling planting device of the present application.

[0070] Figure 5 is a structural schematic diagram of the reversing groove of the automatic seedling planting device of the present application.

[0071] Figure 6 is a partial structural schematic diagram of the automatic seedling planting device of the present application.

[0072] In the figure: 1, mounting plate, 2, control chamber, 3, hole digging opening, 4, moving plate, 5, reversing groove, 51, first bent plate, 52, rotating shaft, 53, straight plate, 54, second bent plate, 55, electromagnet, 56, shaft fixing sleeve, 6, reversing groove driver, 7, planting opening, 8, water filling unit, 9, conveying belt, 10, baffle, 11, image acquisition module, 12, electromagnetic valve, 13, auger, 14, telescopic device, 15, arc-shaped soil filling plate, 16, push plate, 17, moving wheel, 18, wheel driver, 19, air cylinder, 20, auger driver. DETAILED DESCRIPTION

[0073] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0074] Example 1

[0075] like Figure 1 As shown, an automatic seedling planting device for paper mulberry trees includes:

[0076] Mounting plate 1 is made of chromium alloy to prevent it from being attracted by electromagnet 55 and to improve the stability of electromagnet 55 during use.

[0077] The digging unit is mounted on the mounting plate 1 and is used to dig planting pits; the digging unit digs pits in the ground through the digging opening 3;

[0078] A conveying unit, mounted on the mounting plate 1, is used to convey saplings into the planting pit;

[0079] A soil-filling unit, mounted on the mounting plate 1, is used to fill soil around the roots of the sapling;

[0080] A power supply, located on the mounting plate 1, is used to provide electrical energy to the device. In this embodiment, the power supply is a storage battery, installed in the control room 2. The power supply provides the necessary electrical energy to the different drives of the device. When the power is too low, an alarm is automatically triggered to remind the operator to charge it.

[0081] The control unit is connected to the digging unit, conveying unit, and filling unit. In this embodiment, the control unit includes a PLC controller and an industrial computer. The industrial computer is integrated with the PLC controller and communicates via RS-232 or RS-485 to achieve precise control of the device. The image recognition model in the image recognition unit is set in the industrial computer to identify the location of the planting pit in real time. The PLC controller is equipped with a communication module for communicating with the outside world, sending alarm signals, or receiving control signals, etc.

[0082] The irrigation unit 8 is arranged on the mounting plate 1 and used for irrigating the seedling filled with soil. The irrigation unit 8 is arranged with an electromagnetic valve 12 connected with the control unit. The irrigation unit 8 is fixed on the tail of the mounting plate 1 by welding. In this embodiment, two irrigation units 8 are arranged. Each irrigation unit 8 is arranged with an electromagnetic valve 12 near the planting hole 7 at the outer side of the bottom. The electromagnetic valve 12 is connected with the output port of the PLC controller. The opening and closing of the electromagnetic valve is controlled by the PLC controller to realize the accurate control of irrigation. After the seedling is planted in the planting hole, the control unit sends a signal to the irrigation unit. The electromagnetic valve is opened. Water flows into the planting hole through the pipeline of the irrigation unit to provide sufficient water for the seedling.

[0083] The level gauge is arranged on the mounting plate 1 and connected with the control unit. The level gauge is arranged in the control chamber 2 of the mounting plate 1 by screwing. The output end of the level gauge is connected with the input end of the PLC controller to realize the real-time monitoring of the inclination angle of the device. When the level gauge detects that the included angle (inclination angle) between the mounting plate and the horizontal direction exceeds the preset threshold value, a signal is sent to the PLC controller. The PLC controller immediately sends a control signal to the corresponding telescopic rod to adjust the length of the telescopic rod to restore the horizontal state of the device to ensure the working accuracy and the growth quality of the tree.

[0084] The at least four moving wheels 17 are arranged at the bottom of the mounting plate 1 by telescopic rods. Each moving wheel 17 is arranged with a wheel driver 18 connected with the control unit. The telescopic rod is connected with the control unit. The control unit adjusts the length of the telescopic rod based on the signal sent by the level gauge. Figure 3 and 6 As shown in the figure, in this embodiment, four telescopic rods are arranged with one moving wheel 17 at each corner of the bottom of the mounting plate 1. Each moving wheel 17 is arranged with a wheel driver 18 on the rotating shaft. Each wheel driver 18 is connected with the output end of the PLC controller. The PLC controller and the wheel driver 18 drive the moving wheel 17 to rotate to move the device forward, backward, left and right. The wheel driver 18 is a direct current motor. The rotating speed and direction of the motor are controlled to realize the flexible movement and positioning of the device. In this embodiment, the telescopic rod is a hydraulic telescopic rod. The telescopic length of the hydraulic telescopic rod is controlled by the hydraulic system to adapt to different terrains. The telescopic length of the hydraulic telescopic rod is accurately controlled by the PLC controller according to the signal of the level gauge to ensure that the device can be kept stable and horizontal on different terrains. In addition, in order to further improve the adaptability of the device, the wheel driver 18 of each moving wheel 17 can be independently controlled. The device can be flexibly moved in complex terrain to avoid the problem of inaccurate planting caused by uneven terrain.

[0085] An image recognition unit is disposed on the mounting plate 1 and integrated on the control unit. It is used to identify the location information of the planting pit and send the location information to the control unit. The control unit controls the conveying unit to convey the seedling into the planting pit.

[0086] like Figure 1 As shown, during use, the operator inputs relevant parameters through an industrial computer, such as the type of seedling, planting spacing, and depth, and the control unit automatically adjusts based on these parameters. The device is then placed at the digging location. The PLC controller controls the digging unit to begin digging according to a preset program. After digging is completed, the PLC controller sends control commands to the four wheel drives 18, causing the device to move forward. During this movement, the image acquisition module 11 in the image recognition unit acquires real-time image information of the planting pit and transmits it to the image recognition model via a wired connection. When the image recognition model in the industrial computer recognizes the planting pit in the image information, the industrial computer sends information to the PLC controller. Upon receiving the information, the PLC controller stops the device and controls the moving wheels 17 to align the planting inlet 7 with the planting pit. Then, the PLC controller starts the conveying unit to transport the mulberry seedling to the planting pit. At this time, the PLC controller controls the filling unit to fill the soil according to a preset program. After filling, the PLC controller controls the device to continue moving forward to a preset distance and then stops. At this point, the water outlet of the irrigation unit is aligned with the roots of the mulberry seedling, and the PLC controller sends a control command to the solenoid valve 12 to activate the valve and begin irrigation. After irrigation is completed, the PLC controller controls the device to continue moving forward until it reaches the next planting pit. In practical applications, the automatic paper mulberry planting device of this embodiment can significantly improve planting efficiency and quality, reduce manual labor intensity, and lower production costs. Furthermore, the device of this embodiment has good adaptability, capable of adapting to different terrains and environments to meet various planting needs.

[0087] As an optional embodiment of the present invention, the digging unit may optionally include:

[0088] At least two cylinders 19 are mounted on the mounting plate 1 and connected to the control unit;

[0089] The movable plate 4 is mounted on the cylinder 19 and is used to move under the drive of the cylinder 19;

[0090] The auger driver 20 is located at the bottom of the movable plate 4 and is connected to the control unit;

[0091] Screw 13, located on the output shaft of the screw drive 20, is used to excavate soil to form planting pits.

[0092] like Figure 3As shown, in the embodiment, two air cylinders 19 are bolted on both sides of the pit opening 3 of the mounting plate 1, the moving plate 4 is fixed on the output shafts of the two air cylinders 19 by screws, the two air cylinders 19 are connected with the output ends of the PLC controller, the two air cylinders 19 are controlled to move up and down by the PLC controller, so as to drive the moving plate 4 to move up and down, the auger driver 20 is a DC motor, which is fixed on the bottom of the moving plate 4 by screws and connected with the output end of the PLC controller, the PLC controller controls the rotating speed and direction of the motor to realize the rotating digging action of the auger 13, the auger 13 is fixed on the output shaft of the auger driver 20 by welding and rotates with the output shaft of the auger driver 20; the tip part of the auger 13 is designed in a spiral shape, which can effectively cut into the soil and discharge it to form a tree planting pit. When the pit digging operation is needed, the PLC controller controls the two air cylinders 19 and the auger driver 20 to work simultaneously according to the preset program, the air cylinder 19 pushes the moving plate 4 to move up and down, and the auger driver 20 drives the auger 13 to rotate, so as to realize the fast and efficient digging of the tree planting pit. The spiral tip part of the auger 13 discharges the soil during the rotating process to form a neat tree planting pit. In order to ensure the accuracy of the pit digging, the rotating speed of the auger 13 and the moving speed of the air cylinder 19 are accurately controlled by the PLC controller to adapt to different soil conditions and depth requirements.

[0093] As an optional embodiment of the application, optionally, the conveying unit comprises:

[0094] The reversing groove 5 is arranged on the mounting plate 1, and comprises a first bent plate 51, a rotating shaft 52, a straight plate 53 and a second bent plate 54. The first bent plate 51 is rotationally connected with one end of the straight plate 53 through the rotating shaft 52, and the second bent plate 54 is rotationally connected with the other end of the straight plate 53 through the rotating shaft 52. An electromagnet 55 is arranged on the top of each of the first bent plate 51 and the second bent plate 54, and the electromagnet 55 is connected with the control unit. An axle fixing sleeve 56 is arranged on the bottom of the straight plate 53, and a protrusion is arranged in the axle fixing sleeve 56.

[0095] As Figure 4 and 5As shown, the electromagnet 55 on the top of the first bending plate 51 and the second bending plate 54 is an electromagnet 55 formed by connecting a plurality of small electromagnets with the same polarity in parallel, and the small electromagnets are fixed on the top of the first bending plate 51 and the second bending plate 54 by embedding; the electromagnet 55 on the top of the first bending plate 51 and the second bending plate 54 is connected with the output end of the PLC controller through the commutator, and the PLC controller changes the polarity of the electromagnet 55 on the top of the first bending plate 51 or the second bending plate 54 by controlling the commutator to change the current direction; in the initial state, the electromagnet 55 on the top of the first bending plate 51 and the electromagnet 55 on the top of the second bending plate 54 have the same polarity, so they repel each other, and the first bending plate 51 and the second bending plate 54 are open; when the conveying belt 9 conveys the tree seedling into the commutating groove 5, the PLC controller sends a signal to the commutator, the commutator changes the current direction of the electromagnet 55 on the top of the first bending plate 51 or the second bending plate 54, and the polarity of the electromagnet 55 changes, so the electromagnet 55 on the top of the first bending plate 51 and the electromagnet 55 on the top of the second bending plate 54 attract each other, the first bending plate 51 and the second bending plate 54 are closed, and the tree seedling is fixed in the commutating groove 5. The material near the electromagnet 55 is made of chromium alloy, which can improve the stability of the electromagnet 55 and ensure the long-term stable operation of the electromagnet 55, and it will not attract other parts during use.

[0096] The commutating groove driver 6 is arranged on the mounting plate 1 and used for rotating one end of the commutating groove 5 to the tree planting hole;

[0097] As shown in Figure 2 , 4 and 5, the commutating groove driver 6 is a speed reducer, which is fixed on the mounting plate 1 by screws, the output shaft of the commutating groove driver 6 is fixed in the shaft fixing sleeve 56 of the commutating groove 5, and the groove on the output shaft of the commutating groove driver 6 matches the protrusion in the shaft fixing sleeve 56 to realize the rotary motion of the commutating groove 5. The shaft fixing sleeve 56 is fixed on the bottom of the commutating groove 5 by welding; the commutating groove driver 6 is electrically connected with the output end of the PLC controller, when the PLC controller receives the signal that the tree seedling is fixed in the commutating groove 5, the PLC controller controls the commutating groove driver 6 to drive the commutating groove 5 to rotate in a predetermined direction and to a predetermined angle.

[0098] The conveying belt 9 is arranged on the mounting plate 1 and inclined to the commutating groove 5, and at least one baffle 10 is arranged on the conveying belt 9;

[0099] The conveying belt driver is arranged on the mounting plate 1 and used for driving the conveying belt 9, and the conveying belt driver is connected with the output end of the PLC controller.

[0100] As shown in Figure 1 , 2, 4 and 5, when the conveyor drive receives the control signal of the PLC, the bauhinia tree seedling on the conveyor belt 9 is conveyed into the reversing groove 5. The inclined design of the conveyor belt 9 enables the bauhinia tree seedling to smoothly slide into the reversing groove 5, and the baffle 10 ensures that the bauhinia tree seedling does not fall during the conveying process. After the bauhinia tree seedling is successfully conveyed into the reversing groove 5, the PLC controller sends a signal to the reverser to change the polarity of the electromagnet 55 at the top of the first curved plate 51 or the electromagnet 55 at the top of the second curved plate 54, so that the electromagnet 55 at the top of the first curved plate 51 and the electromagnet 55 at the top of the second curved plate 54 attract each other, the first curved plate 51 and the second curved plate 54 are closed, the bauhinia tree seedling is fixed in the reversing groove 5, and the controller sends control information to the reversing groove drive 6 to start working. The reversing groove drive 6 starts to drive the reversing groove 5, so that the reversing groove 5 rotates to a predetermined angle, and the bauhinia tree seedling is accurately sent into the planting hole.

[0101] As an optional embodiment of the present application, optionally, the filling unit comprises:

[0102] At least two stretchers 14 are arranged at the bottom of the mounting plate 1 and connected with the control unit;

[0103] As shown in Figure 6 In this embodiment, two stretchers 14 are welded by welding plates, and the stretchers 14 are air cylinders controlled by the PLC controller.

[0104] At least two arc-shaped filling plates 15 are arranged on the output shafts of the stretchers 14 for filling soil to the seedling;

[0105] As shown in Figure 6 In this embodiment, one arc-shaped filling plate 15 is welded on the output shaft of each of the two stretchers 14, which fills the soil to the bauhinia tree seedling placed in the planting hole and also straightens the bauhinia tree seedling.

[0106] At least two push plates 16 are symmetrically arranged at the bottom of the mounting plate 1, and both of the push plates 16 are welded at the bottom of the mounting plate 1 by welding plates.

[0107] As shown in Figure 6 In use, after the bauhinia tree seedling is placed in the planting hole, the PLC controller sends a control command to the two stretchers 14, and the two stretchers 14 push the soil into the planting hole according to the preset program, fill the planting hole and straighten the bauhinia tree seedling, and then the PLC control device moves forward, and the two inclined and symmetrically arranged push plates 16 scrape away the excess soil on both sides of the bauhinia tree seedling during the movement, ensuring the stable growth of the bauhinia tree seedling in the planting hole.

[0108] As an optional embodiment of the present application, the image recognition unit comprises an image acquisition module 11 and an image recognition model;

[0109] The image acquisition module 11 sends the real-time acquired tree planting pit image data to the image recognition model, and the image recognition model identifies the tree planting pit image data to determine whether the image contains the tree planting pit;

[0110] If the tree planting pit image data does not contain the tree planting pit, the control unit controls the wheel driver 18 to make the device continue to move forward;

[0111] If the tree planting pit image data contains the tree planting pit, the control unit controls the wheel driver 18 to make the device stop moving forward.

[0112] As shown in Figure 2 The image acquisition module 11 is an industrial camera fixed on the planting opening 7 by screws, which is used to take pictures of the land in real time when the digging unit finishes work and the device starts to move forward, and send the pictures to the image recognition model to identify whether there is a tree planting pit in the pictures.

[0113] As an optional embodiment of the present application, the image recognition model is trained by a deep learning algorithm;

[0114] The step of training the image recognition model comprises:

[0115] S1, obtaining historical tree planting pit image data and pre-processing the historical tree planting pit image data;

[0116] It should be noted that the historical tree planting pit image data is the tree planting pit image taken by the image acquisition module 11 under different illuminations and different angles after the digging unit finishes work. The pre-processing step includes image cropping, scaling, normalization, contrast enhancement, denoising, etc. to improve the accuracy of subsequent identification.

[0117] S2, dividing the pre-processed tree planting pit image data into a training data set and a test data set;

[0118] It should be noted that the ratio of the training data set and the test data set in this embodiment is 8:2.

[0119] S3, obtaining an initial model and performing feature decomposition on the initial model to obtain heavy sub-parameters and feature sub-parameters;

[0120] It should be noted that the initial model is a deep learning-based convolutional neural network model with multiple layers capable of extracting features from image data. Feature decomposition refers to the decomposition of weights and biases in the model to optimize the performance of the model. The method of feature decomposition is SVD (singular value decomposition) or PCA (principal component analysis) to extract important feature sub-parameters. Sub-parameters refer to the weight and bias parameters of the fully connected layer in the model, and feature sub-parameters refer to the weight and bias parameters of the feature extraction in the model, that is, the weight and bias of the convolutional layer. Through feature decomposition, the performance of the model can be further optimized to improve the recognition accuracy.

[0121] S4, iteratively train the initial model using the training data set, and adjust the heavy sub-parameters and feature sub-parameters through the loss function and incremental learning algorithm;

[0122] It should be noted that the loss function is used to measure the difference between the predicted value of the model and the true value, and the incremental learning algorithm is to gradually optimize the performance of the model (adjust the heavy sub-parameters and feature sub-parameters) based on the existing model by continuously learning new data. In the iterative training process, the loss function is used to guide the optimization direction of the model, while the incremental learning algorithm ensures that the model does not forget the knowledge learned before while learning new data. In this way, the model can gradually improve the recognition accuracy of the tree planting pit image. In this embodiment, the incremental learning algorithm uses the stochastic gradient descent (SGD) method to update the model parameters. Through multiple iterations of training, the model can gradually adapt to the characteristics in the training data set and improve the recognition accuracy of the tree planting pit image.

[0123] S5, verify the trained initial model using the test data set;

[0124] It should be noted that during the training process, the test data set is used regularly to evaluate the performance of the model. This helps to detect whether the model has overfitting (i.e. the model performs well on training data but poorly on unseen data) or other problems.

[0125] S6, adjust the heavy sub-parameters and feature sub-parameters of the initial model based on the verification result, repeat steps S4 and S5 until the initial model reaches the predetermined recognition accuracy, and obtain the image recognition model.

[0126] It should be noted that when it is found that the model has low accuracy in predicting tree planting pit images, the heavy sub-parameters can be adjusted. These parameters are automatically adjusted by gradient descent and other optimization algorithms to minimize the loss function. When it is found that the model cannot extract effective features from the input image (such as edges, textures, etc. in the image), the feature sub-parameters can be adjusted. This includes changing the filter size, step size, padding method, etc. of the convolutional layer, as well as adjusting the activation function and pooling layer, etc.

[0127] As an optional embodiment of the present invention, the expression of the loss function is optionally:

[0128]

[0129] in, Indicates the loss value;

[0130] Indicates the number of samples;

[0131] Indicates the first One sample;

[0132] Indicates an indicator variable;

[0133] Indicates taking the logarithm;

[0134] Indicates the first The predicted probability value of each sample.

[0135] As an optional embodiment of the present invention, the image recognition model may optionally include an input layer, a convolutional layer, an activation layer, a pooling layer, a fully connected layer, a Dropout layer, and a batch normalization layer. The input layer is used to receive the tree planting pit image data acquired by the image acquisition module 11; the convolutional layer is used to extract image features from the tree planting pit image data; the activation layer is used to introduce nonlinear transformations; the pooling layer is used to reduce the feature dimension; the fully connected layer is used to integrate and classify the extracted features; the Dropout layer is used to prevent the image recognition model from overfitting; and the batch normalization layer is used to accelerate the training process of the image recognition model.

[0136] It should be noted that when the image data of the tree planting pits is sent to the image recognition model, it is first received by the input layer, and then features are extracted through convolutional layers. The convolutional layers perform convolution operations on the input image using a series of convolutional kernels to extract local features from the image. These features are then transformed nonlinearly through activation layers, enabling the model to capture more complex patterns. Pooling layers further reduce the feature dimensionality, decreasing computation while retaining important feature information. Fully connected layers integrate and classify the features extracted by the preceding layers, outputting the final prediction result.

[0137] To prevent overfitting, a Dropout layer is introduced, which reduces the model's dependence on training data by randomly dropping a portion of neurons. Batch normalization layers are used to accelerate the model training process by normalizing the input to each layer, making the model more likely to converge.

[0138] After the model training is completed, the tree planting pit image data collected by the image acquisition module 11 will be sent to the image recognition model for processing. The model will recognize the input image according to the features and classification rules learned during the training process, determine whether there is a tree planting pit, and output the recognition result.

[0139] In order to further improve the accuracy of recognition, the embodiment introduces an attention mechanism, so that the model can pay more attention to the area related to tree planting pit recognition in the image, thereby improving the accuracy of recognition. The attention mechanism learns a weight distribution to highlight important features in the image and suppress unimportant information, so that the model focuses more on the key area.

[0140] Embodiment 2

[0141] A control method of a tree automatic seedling planting device, comprising the device; the method further comprises:

[0142] A1, starting the control unit, initializing the device by using the control unit;

[0143] A2, the control unit sends a control signal to the hole digging unit to start the hole digging unit for hole digging operation;

[0144] A3, after the hole digging is completed, the control unit receives the completion signal of the hole digging unit, sends control signals to the wheel driver 18 and the image recognition unit, and the device starts to move forward. The image acquisition module 11 collects image data in real time and sends it to the image recognition model. When the image recognition model recognizes the tree planting pit, the wheel driver 18 stops working;

[0145] A4, after the control unit receives the signal that the wheel driver 18 stops working, it sends a control signal to the conveying unit to start the conveying unit for tree seedling conveying operation;

[0146] A5, after the tree seedling is conveyed to the tree planting pit, the control unit receives the completion signal of the conveying unit, sends a control signal to the soil filling unit, and starts the soil filling unit for soil filling operation;

[0147] A6, after the soil filling is completed, the control unit receives the completion signal of the soil filling unit, sends a start signal to the wheel driver 18, and the device drives forward to a preset distance. The wheel driver 18 stops working, the control unit sends a watering command to the watering unit 8, and the watering unit starts watering operation;

[0148] A7, after watering is completed, the control unit receives the completion signal of the watering unit;

[0149] A8、the control unit receives the completion signal of the irrigation unit, judges whether there are still un-planting seedlings or un-completed planting holes, if yes, repeats steps A1 to A7, if no, the device enters maintenance mode, and waits for the next planting task.

[0150] As another optional embodiment of the application, optionally, the A3 further comprises:

[0151] If the image recognition model identifies planting holes within the preset distance, the control unit controls the wheel driver 18 to drive the device to move backward to the position of the previous planting hole, and controls the hole digging unit to re-dig holes;

[0152] The A4 further comprises:

[0153] A401、When the conveying belt 9 conveys the seedling to the reversing groove 5, the control unit sends a control command to the electromagnet 55 on the top of the first bending plate 51 and the second bending plate 54, so that the top of the first bending plate 51 and the top of the second bending plate 54 are closed under the action of the two electromagnets 55, and the seedling is fixed;

[0154] A402、After the control unit receives the seedling fixing information, the control unit sends a control information to the reversing groove driver 6, and the reversing groove driver 6 drives the reversing groove 5 to rotate above the planting hole;

[0155] A403、After the control unit receives the rotating completion signal of the reversing groove 5, the control unit sends a control command to the two electromagnets 55, so that the first bending plate 51 and the second bending plate 54 are separated, the seedling is released from the reversing groove 5, and the seedling falls into the planting hole;

[0156] The control method further comprises:

[0157] A9、Before digging the planting hole, when the level detector detects that the inclination angle of the device exceeds the preset threshold, the control unit sends a control signal to the corresponding telescopic rod to adjust the length of the telescopic rod, so that the device restores to the horizontal state, and the device continues to work.

[0158] Although the embodiments of the application have been shown and described, those skilled in the art can understand that various changes, modifications, replacements and variations can be made to the embodiments without departing from the principles and purposes of the application, and the scope of the application is defined by the claims and their equivalents.

Claims

1. An automatic seedling planting device for paper mulberry, characterized in that, The device includes: Mounting plate (1); A pit-digging unit is installed on the mounting plate (1) and is used to dig planting pits; A conveying unit is installed on the mounting plate (1) and is used to convey the seedlings to the planting pit; A soil-filling unit is installed on the mounting plate (1) and is used to fill soil into the roots of the seedling; A power supply, located on the mounting plate (1), is used to provide electrical energy to the device; The control unit is connected to the excavation unit, the conveying unit, and the filling unit; The irrigation unit (8) is installed on the mounting plate (1) and is used to irrigate the seedlings that have been filled with soil. The irrigation unit (8) is equipped with a solenoid valve (12) which is connected to the control unit. A level is mounted on the mounting plate (1) and connected to the control unit; At least four movable wheels (17) are set at the bottom of the mounting plate (1) via telescopic rods. Each movable wheel (17) is equipped with a wheel drive (18), which is connected to the control unit. The telescopic rod is connected to the control unit, which adjusts the length of the telescopic rod in real time based on the signal sent by the level. An image recognition unit is disposed on the mounting plate (1) and integrated on the control unit. It is used to identify the location information of the planting pit and send the location information to the control unit. The control unit controls the conveying unit to convey the seedling into the planting pit. The conveying unit includes: A reversing groove (5) is provided on the mounting plate (1). The reversing groove (5) includes a first bent plate (51), a rotating shaft (52), a straight plate (53), and a second bent plate (54). The first bent plate (51) is rotatably connected to one end of the straight plate (53) through the rotating shaft (52). The second bent plate (54) is rotatably connected to the other end of the straight plate (53) through the rotating shaft (52). Electromagnets (55) are provided on the top of both the first bent plate (51) and the second bent plate (54). The electromagnets (55) are connected to the control unit. A shaft fixing sleeve (56) is provided at the bottom of the straight plate (53), and a protrusion is provided inside the shaft fixing sleeve (56). A commutator drive (6) is mounted on the mounting plate (1) and is used to rotate one end of the commutator (5) onto the planting pit. The conveyor belt (9) is inclined to the reversing groove (5) and is mounted on the mounting plate (1). At least one baffle (10) is provided on the conveyor belt (9). A conveyor belt driver is mounted on the mounting plate (1) and is used to drive the conveyor belt (9). The image recognition unit includes an image acquisition module (11) and an image recognition model; The image acquisition module (11) sends the real-time acquired tree pit image data to the image recognition model. The image recognition model identifies the tree pit image data and determines whether the tree pit is in the image. If the tree planting pit is not found in the image data of the tree planting pit, the control unit controls the wheel drive (18) to make the device continue to move forward; If the tree planting pit is present in the image data of the tree planting pit, the control unit controls the wheel drive (18) to stop the device from moving forward; The image recognition model is trained using a deep learning algorithm; The steps for training the image recognition model include: S1. Obtain historical tree planting pit image data and preprocess the historical tree planting pit image data; S2. Divide the preprocessed tree pit image data into a training dataset and a test dataset; S3. Obtain the initial model and perform feature decomposition on the initial model to obtain the bary parameters and feature parameters; S4. Iteratively train the initial model using the training dataset, and adjust the sub-parameters and feature sub-parameters using the loss function and incremental learning algorithm; S5. Validate the trained initial model using the test dataset. S6. Based on the verification results, adjust the initial model sub-parameters and feature sub-parameters, and repeat steps S4 and S5 until the initial model reaches the predetermined recognition accuracy to obtain the image recognition model. The expression for the loss function is: in, Indicates the loss value. Indicates the number of samples. Indicates the first One sample, Indicates an indicator variable. Indicates taking the logarithm. Indicates the first The predicted probability value of each sample; The image recognition model includes an input layer, a convolutional layer, an activation layer, a pooling layer, a fully connected layer, a Dropout layer, and a batch normalization layer. The input layer is used to receive the tree planting pit image data collected by the image acquisition module (11). The convolutional layer is used to extract the image features of the tree planting pit image data. The activation layer is used to introduce nonlinear transformations. The pooling layer is used to reduce the feature dimension. The fully connected layer is used to integrate and classify the extracted features. The Dropout layer is used to prevent the image recognition model from overfitting. The batch normalization layer is used to accelerate the training process of the image recognition model.

2. The automatic mulberry seedling planting device as described in claim 1, characterized in that, The pit-digging unit includes: At least two cylinders (19) are mounted on the mounting plate (1) and connected to the control unit; A movable plate (4) is mounted on the cylinder (19) and is used to move under the drive of the cylinder (19); The auger driver (20) is located at the bottom of the movable plate (4) and is connected to the control unit; The auger (13), located on the output shaft of the auger driver (20), is used to excavate soil to form planting pits.

3. The automatic mulberry seedling planting device as described in claim 1, characterized in that, The backfill unit includes: At least two telescopic joints (14) are disposed at the bottom of the mounting plate (1) and connected to the control unit; At least two arc-shaped filler plates (15) are set on the output shaft of the expansion joint (14) for filling soil into the sapling; At least two push plates (16) are symmetrically arranged at the bottom of the mounting plate (1).

4. A control method for an automatic mulberry seedling planting device, characterized in that, The method includes the apparatus as described in any one of claims 1 to 3; the method further includes: A1. Start the control unit and use the control unit to initialize the device; A2. The control unit sends a control signal to the digging unit to start the digging unit to perform digging operations; A3. After the hole is dug, the control unit receives the completion signal from the hole digging unit and sends control signals to the wheel drive (18) and the image recognition unit. The device starts to move forward. The image acquisition module (11) collects image data in real time and sends it to the image recognition model. When the image recognition model recognizes the tree planting hole, the wheel drive (18) stops working. A4. After receiving the stop signal from the wheel drive (18), the control unit sends a control signal to the conveying unit to start the conveying unit to carry out the seedling conveying operation. A5. After the saplings are transported to the planting pit, the control unit receives the completion signal from the transport unit and sends a control signal to the filling unit to start the filling operation. A6. After the backfilling is completed, the control unit receives the completion signal of the backfilling unit and sends a start signal to the wheel drive (18). The device moves forward to the preset distance and stops working to the wheel drive (18). The control unit sends a watering command to the watering unit (8) and starts the watering unit to carry out watering operations. A7. After watering is completed, the control unit receives the completion signal from the watering unit; A8. After receiving the completion signal from the irrigation unit, the control unit determines whether there are any unplanted seedlings or unfinished planting pits. If so, steps A1 to A7 are repeated; if not, the device enters maintenance mode and waits for the next planting task.

5. The control method for the automatic mulberry seedling planting device as described in claim 4, characterized in that, A3 also includes: If the image recognition model fails to recognize the planting pit within the preset distance, the control unit controls the wheel drive (18) to make the device move backward and return to the position of the previous planting pit, and controls the digging unit to start the digging operation again. The A4 also includes: A401. When the conveyor belt (9) transports the seedling to the reversing groove (5), the control unit sends a control command to the electromagnets (55) at the top of the first bending plate (51) and the second bending plate (54), so that the top of the first bending plate (51) and the top of the second bending plate (54) are closed under the action of the two electromagnets (55) to fix the seedling. A402. After receiving the seedling fixing information, the control unit sends control information to the reversing slot driver (6), and the reversing slot driver (6) drives the reversing slot (5) to rotate above the planting pit. A403. After receiving the signal that the reversing slot (5) has completed rotation, the control unit sends a control command to the two electromagnets (55) to separate the first bending plate (51) and the second bending plate (54), and the sapling is released from the reversing slot (5) and falls into the planting pit. The control method further includes: A9. Before digging the planting pit, when the level instrument detects that the tilt angle of the device exceeds the preset threshold, the control unit sends a control signal to the corresponding telescopic rod to adjust the length of the telescopic rod so that the device is restored to a horizontal position and the device continues to operate.

Images