Inter-plant mechanical weeding system and method for transplanted crops

Abstract

The invention discloses an inter-plant mechanical weeding system and method for transplanted crops, and the system comprises an imaging system which comprises an imaging darkroom; the illumination system is located in the imaging darkroom and comprises a plurality of LED lamps; the weeding knife system is located on the rear portion of the imaging darkroom and comprises a servo electric cylinder, a rolling sliding rail and a weeding blade, and the servo electric cylinder drives the weeding blade to horizontally move on the rolling sliding rail when executing actions so as to achieve avoiding of crop plants and inter-plant weeding; the GNSS positioning system is used for acquiring the position information of the current device in real time; the crawler-type chassis system is used for enabling the system to advance according to a planned route; the power distribution control box comprises a power supply, a single-chip microcomputer and a microcomputer which are arranged in the power distribution control box. According to the system and the method, crops can be accurately positioned, the weeding efficiency is improved, the system and the method can be popularized to crop transplanting, and sustainable agriculture can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent agriculture, and specifically to an inter-row mechanical weeding system and method for transplanted crops based on systematic crop signal technology and multi-view imaging. Background Art

[0002] The identification of crops is mainly carried out in two ways, including using machine vision and studying the spectral characteristics of weeds and main crops. The crop / weed identification algorithm established by traditional machine vision based on color, texture and shape features combined with machine learning is greatly affected by the environment; the identification scheme based on deep learning requires a large amount of data sets for pre-training, and most of them are identification models for specific crop weeds, with poor generalization ability; using hyperspectral imaging technology will generate a large amount of redundant spectral information, resulting in waste of resources and extended processing time. Therefore, the main challenge restricting the wide application of intelligent weeding robots in the agricultural field is the lack of a general, accurate and fast crop / weed identification method.

[0003] The traditional weeding method in crop fields relies on manual weeding, which is costly and has low work efficiency, and cannot meet the needs of large-scale farmland planting. Spraying pesticides is a widely used weeding method. However, the extensive use of herbicides not only causes soil and groundwater pollution, but also easily leads to problems such as increased weed resistance, crop phytotoxicity and poor quality of agricultural products. In addition, using plastic films to control weeds is also a widely used method. However, the main disadvantage of this method is the generation of plastic waste that is not easily degraded. Although biodegradable plastics can be used, these materials are expensive, unaffordable for farmers, and unstable. The existing technical methods are not applicable to in-row weeding of crops and cannot accurately and timely remove weeds close to crops. Thus, there is currently a lack of a general, fast and accurate identification method in the field of weeding robots, as well as a weeding technology applicable to in-row weeds.

[0004] Regarding the problems in the related technology, no effective solution has been proposed yet. Summary of the Invention

[0005] In view of the above deficiencies, the present invention provides an inter-row mechanical weeding system and method for transplanted crops.

[0006] The technical solution adopted by the present invention to solve its technical problems is as follows: An inter-row mechanical weeding system for transplanted crops, comprising an imaging system, including an imaging darkroom, inside which a camera for taking images of crops and weeds is fixed; a lighting system, located inside the imaging darkroom, including a plurality of LED lights for providing light source or supplementary light for the crops; The weeding knife system, located at the rear of the imaging darkroom, includes a servo electric cylinder, a rolling slide rail, and a weeding blade. When the servo electric cylinder performs an action, it drives the weeding blade to translate on the rolling slide rail to avoid crop plants and weed between plants. The GNSS positioning system, located above the imaging system, is used to obtain the position information of the current device in real time. The crawler chassis system, located below the imaging system, includes crawlers, a crawler drive motor, and a crawler drive wheel set. It realizes differential motion through PID control to make the system move forward along the planned route. The power distribution control box, located in front of the imaging system, includes a power supply, a single-chip microcomputer, and a microcomputer arranged therein. The microcomputer is used to process the image data captured by the camera, calculate the opening and closing time of the weeding blade, run the PID control algorithm, and transmit the information to the single-chip microcomputer.

[0007] As an improvement of the above technical solution, the camera is a high-frame-rate grayscale camera, which is installed in the middle of the inner top of the imaging darkroom.

[0008] As an improvement of the above technical solution, the grayscale camera is equipped with a 16 mm focal length lens, and a 520 nm filter is placed in front of its lens.

[0009] As an improvement of the above technical solution, a mirror bracket is fixed at the inner bottom end of the imaging darkroom, and a plane reflector is placed on the mirror bracket to form a multi-view system to provide a side view of the crop stalks.

[0010] As an improvement of the above technical solution, the mirror bracket is arranged parallel to the system moving direction. Each mirror bracket can place three plane reflectors. The bottom edge of the plane reflector in the middle is parallel to the traveling direction and forms a 45 ° angle with the horizontal ground. The angles between the plane reflectors on both sides and the middle mirror are 120 °.

[0011] As an improvement of the above technical solution, the imaging darkroom is built with metal profiles and wrapped with black matte sheet metal on the outside to prevent external light from interfering with the inside of the system.

[0012] As an improvement of the above technical solution, the lighting system includes eight LED lights. Four of them are located under the mirror brackets inside the imaging darkroom to provide sufficient light sources for the crop stalks, and the other four are fixed in the middle of the inside of the imaging darkroom to supplement light for the crop plants.

[0013] As an improvement of the above technical solution, the LED lights are equipped with 10W heat sinks to ensure that the LED lights will not be burned out due to excessive heat generation.

[0014] As an improvement of the above technical solution, a condenser cup is placed at the front end of the LED lamp to converge the light emitted by the LED lamp so that it converges on the plant to be identified.

[0015] As an improvement of the above technical solution, two sets of rolling slide rails are arranged up and down, and two rolling sliders are slidably arranged thereon. An electric cylinder controller is installed on the rolling slider located above, and a rotary motor is installed on the rolling slider located below. The weeding blade is installed at the bottom of the rotary motor. The rolling sliders on the same side are connected by a weeding knife connecting rod. Among them, the electric cylinder controller is used to control the telescopic movement of the telescopic rod of the servo electric cylinder. Its telescopic rod is connected to the rolling slider. When its telescopic movement occurs, it can drive the weeding blade to translate on the rolling slide rail to achieve the avoidance of crop plants and weeding between plants.

[0016] As an improvement of the above technical solution, the weeding blade is designed as an exponential blade, and its blade curve is obtained by the following formula: 。

[0017] As an improvement of the above technical solution, the crawler drive wheel set includes a crawler drive pulley, a crawler driven pulley, and a crawler moving pulley. The crawler drive motor drives the crawler to rotate by driving the crawler drive pulley, the crawler driven pulley, the crawler moving pulley, and the crawler to realize field walking.

[0018] As an improvement of the above technical solution, an angle sensor is further arranged in the power distribution control box. The angle sensor is used to obtain the angle at which the system deviates from a predetermined straight line and transmit the information to the microcomputer.

[0019] The present invention also adopts the following technical solutions while solving its technical problems: A mechanical weeding method between plants of transplanted crops, including Step 1: Root-treat the crops to be transplanted with an effective dose of sodium fluorescein solution, and transplant the treated crops into the field; Step 2: When the crops to be transplanted need to be weeded, use the imaging system and the lighting system to collect images of crop plants; adjust the positions of the plane reflector and the LED lamp so that the plane reflector does not reflect light and the LED light is smoothly emitted towards the plants; Step 3: Adjust the position of the camera so that its imaging field of view includes the target crops and weeds, and set appropriate camera exposure values and gain values through the microcomputer; Step 4: pre-set the coordinates of the starting and ending positions of the crop rows where the system is traveling. When the system is traveling in a straight line along the crop rows in the field according to the travel control algorithm of the microcomputer, the system takes real-time photos through the camera, extracts the region of interest through the image processing algorithm of the microcomputer, and determines the coordinates of the unearthed point of the stem of the crop seedling using the intersection of the lines formed by different viewing angles; Step 5: Use the data of the emergence point of the crop seedling stem to control the opening and closing action of the weeding blade.

[0020] As an improvement of the above technical solution, in step 1, the crop roots are immersed in a 200 ppm sodium fluorescein solution in a dark room for 48 hours.

[0021] As an improvement of the above technical solution, step 4 specifically includes: S41, using a microcomputer to obtain the position information returned by the GNSS positioning system and the angle deviation returned by the angle sensor, and moving along the crop row according to the preset start and end point coordinates. When the system deviates from the predetermined straight line, the direction is adjusted through the differential rotation of the crawler chassis system based on the PID control algorithm; S42, performing median filtering and image sharpening operations on the image taken by the camera to remove noise; S43, binarizing the image after the noise is removed based on the difference in grayscale values, and extracting the fluorescence signal of the crop plant; S44, performing an opening and closing operation on the binary image to remove image noise and smooth the boundaries, and remove areas with weak fluorescence signals, that is, areas with too small connected domain areas; S45. Define seven regions of interest in the binary image, and name them as: upper left (LR), left middle (LM), lower left (LB), middle (M), upper right (RM), right middle (RM), and lower right (RM); S46, finding the point of the six surrounding interest regions closest to the middle interest region; S47. Draw a corresponding straight line based on the nearest point found in the region of interest, find the point where the stem of the crop plant emerges from the soil, and use it as the actual position point of the crop.

[0022] As an improvement of the above technical solution, in step S41, the angle obtained by the angle sensor is used as the input of the PID control algorithm, and the differential rotation of the crawler chassis system is realized according to the incremental PID expression, and its expression is as follows:

[0023] in, , To control the increment, , , They are the input angle information at the current moment, the input angle information at the previous moment, and the change in the angle, respectively. , , .

[0024] As an improvement to the above technical solution, in step S43, the fluorescence signal of the crop plant is extracted, and the fluorescence signal of the weed is removed. The thresholding rule is shown as follows:

[0025] Among them, represents the original grayscale image captured by the camera; is the segmentation threshold, and the specific size of the threshold depends on the on-site lighting conditions, treatment concentration and time, soil color and other situations; is the binary image after thresholding.

[0026] As an improvement to the above technical solution, in step S47, the method for identifying the emergence point of the crop plant stem includes: Ideally, six nearest points can be captured; When the weed is located between the mirror and the crop plant, it may block the mirror imaging, causing the loss of the fluorescence signal in the region of interest, thus affecting the solution of the nearest point; According to the nearest point obtained from the region of interest, the corresponding straight line is made, and the following situations may occur: (1) Three straight lines can be formed; (2) Two straight lines can be formed; (3) One straight line can be formed; (4) No straight line can be formed and the nearest points are on both sides; (5) The nearest points that cannot form a straight line are on one side and are the nearest points of the upper and lower regions; (6) The nearest points that cannot form a straight line are on one side and are the nearest points of the middle region; (7) Only the centroid point; The calculation rules for various situations are as follows:

[0027] ;

[0028] ;

[0029] ;

[0030]

[0031] ;

[0032]

[0033]

[0034] ;

[0035]

[0036]

[0037] ;

[0038]

[0039]

[0040]

[0041]

[0042]

[0043]

[0044]

[0045] ; Among them, represents the number of straight lines formed by the nearest points; represents the emergence point of the stem of the tomato plant; represents the intersection point of two straight lines formed by the opposite nearest points; represents the number of sides where the nearest points are located. LT, LM, LB, M, RT, RM, RB respectively represent the ROI regions of the upper left, left middle, lower left, middle, upper right, right middle, and lower right; represents the abscissa of the nearest point in the left ROI region; represents the abscissa of the nearest point in the right ROI region; represents the ordinate of the center point of the picture; represents the abscissa of the center point of the picture; represents the ordinate of the nearest point in the upper ROI region; represents the ordinate of the nearest point in the lower ROI region; Represents the ordinate of the ROI region in the middle of the left and right sides; Represents the centroid coordinates of the middle ROI region.

[0046] As an improvement to the above technical solution, in step 5, using the data of the stem emergence points of crop seedlings, by calculating the delay time for the weeding blade to open, the opening and closing actions of the weeding blade are controlled and transmitted to the single-chip microcomputer. The calculation formula is as follows:

[0047] Among them, Is the delay time for the weeding blade to open after detecting the position of the stem emergence point of the crop plant, Is the pixel value of the image along the forward direction, Is the pixel value of the distance from the stem emergence point of the crop plant along the traveling direction to the boundary where the crop plant enters the device, Is the conversion ratio between the image and the actual distance, Is the distance between the weeding blade and the imaging darkroom, Is the length of the weeding safety area, Is the traveling speed of the system, Is the average processing time of the program.

[0048] As an improvement to the above technical solution, after the single-chip microcomputer obtains the delay time for the weeding blade to open, when the established opening time is reached, the single-chip microcomputer transmits the information to the electric cylinder controller. The electric cylinder controller controls the servo electric cylinder to be energized, and its telescopic rod pushes the weeding blade to make a uniform opening movement along the rolling slide rail at the same speed as the traveling speed of the device. After the weeding blade passes through the set safety area, then the single-chip microcomputer controls the telescopic rod of the servo electric cylinder to retract, and the weeding blade closes. In this way, the cycle is repeated to achieve the avoidance of the weeding blade for the crop plant and the weeding work between plants; Among them, the safety area is set according to the actual situation, and the formula for determining the opening time of the weeding blade is as follows:

[0049] Among them, Is the opening time of the weeding blade; Is the length of the weeding safety area; Is the traveling speed of the device.

[0050] Compared with the prior art, the present invention can bring the following beneficial effects: (1) The identification and inter-plant mechanical weeding system for transplanted crops provided by the present invention can be extended to various transplanted crops. After treating the crops with an effective dose of sodium fluorescein solution, under the excitation light of 365 nm, it exhibits a fluorescence signal significantly different from that of weeds, and this signal can be captured by a grayscale camera equipped with a 520 nm filter. After sodium fluorescein is applied to the crop plants, it is absorbed by the crop plants and systematically transported within the crop plants, avoiding the situation where the marker is removed by natural factors such as wind and rain or irrigation. This method can be applied before crop transplantation, can be processed in large batches, is easy to operate, and sodium fluorescein is economical and affordable, which can bring benefits to farmers.

[0051] (2) The dose of sodium fluorescein solution applied will not affect the growth of transplanted crops, will not transfer to crop fruits, and will not cause harm to the human body.

[0052] (3) Relying solely on the top view of the camera cannot completely capture the fluorescence of all parts of the crop plants. The six mirrors provide six side views for the camera, and multiple views avoid the problem of occlusion caused by high weed density. By combining the side views with the top view, the emergence point of the stem of the crop plant can be calculated through an algorithm, thereby accurately determining the specific position of the crop plant.

[0053] (4) After obtaining the information of the GNSS positioning system, the inter-plant mechanical weeding system can be controlled by a computer to move forward along the crop row. The electric cylinder controller controls the telescopic movement of the servo electric cylinder telescopic rod to achieve a weeding solution with adjustable row spacing. The crawler chassis ensures that the device can operate on various terrains, accurately remove the weeds around the crops and avoid damage to the crops.

[0054] (5) The present invention reduces the high-intensity labor of operators, saves labor costs, improves weeding efficiency, and does not require chemical agents, avoiding environmental damage and the generation of weed resistance, meeting the requirements of sustainable agriculture. Description of the Drawings

[0055] In order to more clearly illustrate the embodiments in the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required in the embodiments. The drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative labor.

[0056] Figure 1 It is a schematic diagram of the overall structure of an embodiment of the present invention; Figure 2 It is a front view schematic diagram of an embodiment of the present invention; Figure 3 It is a top-side view schematic diagram of an embodiment of the present invention; Figure 4 Rear schematic view of the embodiment of the present invention; Figure 5 Internal enlarged schematic view of the power distribution control box of the embodiment of the present invention; Figure 6 Schematic diagram of the LED lamp integrated structure of the embodiment of the present invention; Figure 7 Schematic flow chart of the weeding method of the embodiment of the present invention; Figure 8 Schematic diagram of the step of dividing the region of interest of the embodiment of the present invention; Figure 9 Schematic diagram of the navigation controller of the embodiment of the present invention; Figure 10 Schematic diagram of the working principle of the weeding blade of the embodiment of the present invention; Figure 11 Schematic diagram of the recognition result of the emergence point of the tomato stem of the embodiment of the present invention.

[0057] Explanation of reference numerals: 1 - Profile, 2 - Power distribution control box, 3 - Crawler chassis system, 4 - Crawler drive motor, 5 - GNSS positioning system, 6 - Sheet metal, 7 - Crawler drive sprocket, 8 - Crawler driven sprocket, 9 - Crawler moving pulley, 10 - Camera, 11 - Imaging darkroom, 12 - LED lamp, 13 - Mirror bracket, 14 - Heat sink, 15 - Crawler, 16 - Plane mirror, 17 - LED lamp fixing part, 18 - Rolling slide rail, 19 - Servo electric cylinder, 20 - Weeding knife system, 21 - Electric cylinder controller, 22 - Rolling slider, 23 - Weeding knife connecting rod, 24 - Rotating motor, 25 - Weeding blade, 26 - LED lamp integration above the plane mirror, 27 - LED lamp integration below the plane mirror, 28 - Microcomputer, 29 - Power supply, 30 - Angle sensor, 31 - Single-chip microcomputer, 32 - Constant current module, 33 - UV LED, 34 - Condensing cup, 35 - Safety area, 36 - Weeds, 37 - Tomato plants. Detailed implementation manners

[0058] Next, the technical solutions in the embodiments of the present invention will be described clearly and completely in conjunction with the accompanying drawings in the embodiments of the present invention.

[0059] It should be noted that the described embodiments are only some of the embodiments of the present invention, not all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the protection scope of the present invention.

[0060] To make the above objects, features, and advantages of the present invention clearer and easier to understand, the present invention will be further explained in detail below in conjunction with the accompanying drawings and specific embodiments, taking tomatoes as the representative crop.

[0061] As Figures 1 to 6 shown, the first embodiment of the present invention provides an identification and inter-plant mechanical weeding system for transplanted tomatoes, including: an imaging darkroom 11, a multi-view system, a lighting system, a weeding knife system 20, a GNSS positioning system 5, a crawler chassis system 3, a power distribution control box 2, a power supply 29, a single-chip microcomputer 31, an angle sensor 30, and a microcomputer 28.

[0062] The imaging darkroom 11 is built by metal profiles 1, and the outside of the system is wrapped with black matte sheet metal 6 to prevent external light sources from interfering with the inside. In the middle of the top end inside the imaging darkroom 11, a high-frame-rate grayscale camera 10 equipped with a 16-mm focal length lens and a 520 nm filter is fixed in front of the lens. The high-frame-rate grayscale camera 10 is connected to the microcomputer 28. Parallel to the moving direction at the bottom end inside the imaging darkroom 11, two mirror brackets 13 are fixed on each side. Plane reflectors 16 are placed on the mirror brackets 13, and each mirror bracket 13 can hold three plane reflectors 16.

[0063] The lighting system includes eight LED lights 12 equipped with heat sinks 14 and condenser cups 34, including four LED light assemblies 26 above the plane reflectors 16 and four LED light assemblies 27 below the plane reflectors 16, which are fixed in corresponding positions by LED light fixing parts 17.

[0064] The weeding knife system 20 is located at the rear of the imaging darkroom 11 and is composed of rolling sliders 22, rolling rails 18, an electric cylinder controller 21, a servo electric cylinder 19, a rotary motor 24, weeding blades 25, and a weeding knife connecting rod 23. Two sets of rolling rails 18 are arranged up and down, and two rolling sliders 22 are slidably arranged on each of them. The electric cylinder controller 21 is installed on the rolling slider 22 above, and the rotary motor 24 is installed on the rolling slider 22 below. The weeding blades 25 are installed at the bottom of the rotary motor 24. The rolling sliders 22 on the same side are connected by the weeding knife connecting rod 23. The weeding blades 25 are driven to rotate by the rotary motor 24, and the electric cylinder controller 21 controls the telescopic rod of the servo electric cylinder 19. Its telescopic rod is connected to the corresponding rolling slider 22. When its telescopic action occurs, it can drive the weeding blades 25 to translate on the rolling rails 18 to avoid crop plants and perform inter-plant weeding, realizing a weeding scheme with adjustable row spacing.

[0065] The GNSS positioning system 5 is located above the imaging darkroom 11 and above the weeding knife system 20. By obtaining the device position information in real time, it completes unmanned operations using precise autonomous navigation.

[0066] The crawler chassis system 3 is located below the imaging darkroom 11. The crawler driving motor 4 drives the crawler drive pulley 7, the crawler driven pulley 8, the crawler moving pulley 9, and the crawler 15 to rotate, realizing walking in the field. Differential rotation can be achieved through PID control to ensure that the device moves forward along the planned route.

[0067] The control distribution box 2 is located in front of the imaging darkroom 11 and above the crawler chassis system 20. Inside, there are the microcomputer 28, the power supply 29, the angle sensor 30, and the single-chip microcomputer 31. The power supply 29 is used to supply power to the LED lamp 12, the servo electric cylinder 19, and the rotary motor 24, ensuring that the device walks and performs weeding operations. The single-chip microcomputer 31 is used to control the electric cylinder controller 21 to control the telescopic movement of the expansion rod of the servo electric cylinder 19, the rotation of the rotary motor 24, and the differential movement of the crawler chassis system 3. The angle sensor 30 is used to obtain the angle at which the device deviates from the predetermined straight line and transmit the information to the microcomputer 28. The microcomputer 28 is also used to process the image information captured by the high-frame-rate grayscale camera 10, calculate the opening and closing time of the weeding blade 25, run the PID control algorithm, and transmit the relevant information to the single-chip microcomputer 31.

[0068] In this embodiment, the wavelength of the LED lamp 12 is 365 nm. The LED lamp integration 26 located above the plane mirror 16 is responsible for supplementary lighting to ensure that the camera 10 can capture the image of the tomato plant 37. The LED lamp integration 27 located below the plane mirror 16 is responsible for irradiating the root of the tomato plant 37, and this LED lamp integration 27 will not be captured by the high-frame-rate grayscale camera 10.

[0069] In another embodiment, the ultraviolet LED 33 is powered by the DC power supply 29, and a constant current module 32 can be added between the power supply and the ultraviolet LED 33 to ensure stable current and no overload. The ultraviolet LED 33 is equipped with a 10W heat sink 14 to ensure that it will not be burned out due to excessive heat generation. A condenser cup 34 is placed in front of the ultraviolet LED 33 to converge the emitted light so that the emitted light converges on the tomato plant 37 to be recognized.

[0070] In this embodiment, the bottom edge of the plane mirror 16 in the middle of the mirror bracket 13 at the bottom of the imaging darkroom 11 is parallel to the traveling direction. The plane mirror 16 forms a 45 ° angle with the horizontal ground, and the included angle between the other two plane mirrors 16 and the middle plane mirror 16 is 120 °. The plane mirror 16 is in the shape of an isosceles trapezoid, facing each other in pairs, so that the images in the plane mirror 16 can display the images of the tomato plant 37 and the weed 36.

[0071] When the weeding system of the first embodiment is operating in the field, the camera 10 performs real-time shooting, and calculates the position of the emergence point of the tomato stem in real time to achieve intelligent inter-plant mechanical weeding. The process is as follows Figure 7 shown.

[0072] The second embodiment of the present invention provides a method for mechanical weeding between plants of transplanted tomatoes, including Step 1: Root-treat the crops to be transplanted with an effective dose of sodium fluorescein solution, and transplant the treated crops into the field; The tomato plants 37 are soaked in sodium fluorescein solution before transplantation. Under laboratory conditions, soaking the plants in a 200 ppm sodium fluorescein solution for 48 hours will not affect the growth of the tomato plants 37, and can produce a stable and unique fluorescence signal within one month after transplantation. Sodium fluorescein mainly exists on the stems of the tomato plants 37. After being irradiated by the LED lamp 12 equipped with the condenser cup 34, it will emit emission light with a wavelength peak of 520 nm. For the untreated tomato plants 37 and weeds 36 under this excitation light, chlorophyll fluorescence is the main body, and the wavelength peaks are approximately 680 nm and 720 nm. The 520 nm filter equipped on the high-frame-rate grayscale camera 10 can filter out the stray light other than the emission light of sodium fluorescein, so the tomato plants 37 can be identified.

[0073] Step 2: When the crops to be transplanted need to be weeded, use the imaging system and the lighting system to collect images of the crop plants; adjust the positions of the plane mirror 16 and the LED lamp 12 so that the plane mirror 16 does not reflect light and the light of the LED lamp 12 shines smoothly on the tomato plants 37.

[0074] Step 3: Adjust the position of the camera so that its imaging field of view includes the target crops and weeds, and set appropriate camera exposure values and gain values through the microcomputer 28.

[0075] Step 4: Preset the starting and ending position coordinates of the crop rows for the system to travel, and automatically travel in a straight line along the crop rows in the field according to the travel control algorithm of the microcomputer 28; Excite fluorescence through the LED lamp 12, reflect the tomato plants 37 by the mirror, perform real-time shooting through the camera 10, extract the region of interest through the image processing algorithm of the microcomputer 28, and determine the coordinates of the emergence point of the stem of the crop seedling by using the intersection of the lines formed by different perspectives; Step 5: Control the opening and closing actions of the weeding blade 25 by using the data of the emergence point of the stem of the crop seedling.

[0076] In this embodiment, step 4 specifically includes: S41. Use the microcomputer 28 to obtain the position information returned by the GNSS positioning system 5 and the angle deviation returned by the angle sensor 30, move forward along the crop row according to the preset starting and ending point coordinates, and when the system deviates from the predetermined straight-line driving, based on the PID control algorithm, realize the direction adjustment through the differential rotation of the crawler chassis system 3; S42. Perform median filtering and image sharpening operations on the pictures taken by the camera 10 to remove noise; S43. Binarize the image after removing noise based on the difference in gray values, and extract the fluorescence signal of the crop plants; S44. Perform opening and closing operations on the binarized image to remove image noise and smooth the boundary, and remove the areas with weak fluorescence signals, that is, the areas with too small connected domain areas; S45. Define seven regions of interest in the binarized image, and name them respectively: upper left (LR), left middle (LM), lower left (LB), middle (M), upper right (RM), right middle (RM), lower right (RM); S46. Solve the points in the six surrounding regions of interest that are closest to the region of interest in the middle; S47. According to the closest points obtained from the regions of interest, make corresponding straight lines, solve the emergence points of the stems of the crop plants, and use them as the actual position points of the crops.

[0077] Refer to Figure 8 In step S41, the angle obtained by using the angle sensor 30 is used as the input of the PID control algorithm, and the differential rotation of the crawler chassis system 3 is realized according to the incremental PID expression, and its expression is as follows:

[0078] Among them, , is the control increment, , , are the input angle information at the current moment, the input angle information at the previous moment and the change in angle respectively, , , .

[0079] In step S43, the microcomputer 28 is used to perform median filtering and image sharpening operations on the image obtained by the camera 10 based on Python to remove noise. Binarize the image after removing noise based on the difference in gray values, extract the fluorescence signal of the tomato plants 37, and remove the fluorescence signal of the weeds 36. The thresholding rule is as follows:

[0080] Among them, Represents the original grayscale image captured by camera 10; is the segmentation threshold, and the specific size of the threshold depends on the on-site lighting conditions, processing concentration and time, soil color, etc.; is the binary image after thresholding.

[0081] Perform opening and closing operations on the binary image to remove image noise and smooth the boundaries, and remove areas with weak fluorescence signals, that is, areas with too small connected domain areas.

[0082] Refer to Figure 9 , define seven regions of interest (ROIs) in the binary image, named: upper left (LR), left middle (LM), lower left (LB), middle (M), upper right (RT), right middle (RM), lower right (RB). Solve the points in the six surrounding regions of interest that are closest to the region of interest in the middle and the centroid point of the connected domain in the middle region. According to the closest points obtained from the regions of interest, draw corresponding straight lines, solve the emergence point of the stem of tomato plant 37, and use it as the actual position point of tomato plant 37.

[0083] According to the closest points obtained from the regions of interest, draw corresponding straight lines, and the following situations may occur: (1) Three straight lines can be formed; (2) Two straight lines can be formed; (3) One straight line can be formed; (4) Straight lines cannot be formed and the closest points are on both sides; (5) The closest points that cannot form a straight line are on one side and are the closest points of the upper and lower regions; (6) The closest points that cannot form a straight line are on one side and are the closest points of the middle region; (7) There is only the centroid point; The calculation rules for various situations are as follows:

[0084] ;

[0085] ;

[0086] ;

[0087]

[0088] ;

[0089]

[0090]

[0091] ;

[0092]

[0093]

[0094] ;

[0095]

[0096]

[0097]

[0098]

[0099]

[0100]

[0101]

[0102] ; Among them, represents the number of straight lines formed by the nearest points; represents the emergence point of the stem of tomato plant 37; represents the intersection point of two straight lines formed by the opposite nearest points; represents the number of sides where the nearest points are located. LT, LM, LB, M, RT, RM, RB respectively represent the ROI regions of the upper left, left middle, lower left, middle, upper right, right middle, and lower right; represents the abscissa of the nearest point in the left ROI region; represents the abscissa of the nearest point in the right ROI region; represents the ordinate of the center point of the picture; represents the abscissa of the center point of the picture; represents the ordinate of the nearest point in the upper ROI region; represents the ordinate of the nearest point in the lower ROI region; represents the ordinate of the middle ROI region between the left and right sides; represents the centroid coordinates of the middle ROI region.

[0103] After obtaining the stem emergence point of the tomato plant 37, the delay time for the weeding blade 25 to open can be calculated using the formula:

[0104] Where, is the delay time for the weeding blade 25 to open after detecting the stem emergence point position of the crop plant, is the pixel value of the image along the forward direction, is the pixel value of the distance from the stem emergence point of the crop plant along the traveling direction to the boundary of the crop plant entry device, is the conversion ratio of the image to the actual distance, is the distance between the weeding blade 25 and the imaging darkroom 11, is the length of the weeding safety area 35, is the traveling speed of the system, is the average processing time of the program.

[0105] After completing the conversion from the position information of the stem emergence point of the tomato plant 37 to the opening and closing time information of the weeding blade 25, the single-chip microcomputer 31 starts timing. When the timing reaches zero, the single-chip microcomputer 31 controls the servo electric cylinder 19 to be energized through the electric cylinder controller 21. The telescopic rod of the servo electric cylinder 19 pushes the weeding blade 25 to move uniformly along the rolling slide rail 18 at the same speed as the traveling speed of the device to open. After the weeding blade 25 passes through the set safety area 35, then the single-chip microcomputer 31 controls the telescopic rod of the servo electric cylinder 19 to retract, and the weeding blade 25 closes. Such a cycle is used to achieve the avoidance of the weeding blade for the tomato plant 37 and the inter-row weeding work.

[0106] Refer to Figure 10 , where the safety area 35 is set according to the actual situation, and the formula for determining the opening time of the weeding blade 25 is as follows:

[0107] Where, is the opening time of the weeding blade 25; is the length of the weeding safety area 35; is the traveling speed of the device.

[0108] During the traveling process of the device, control the rotation motor 24 to rotate, so that the weeding blade 25 rotates to remove the weeds 36. The weeding blade 25 is designed as an exponential blade, and the specific blade curve is obtained by the following formula: .

[0109] Refer to Figure 11, the white circle in the picture is the emergence point of the stem of tomato plant 37. There will be deviations in the positioning of the emergence point of the stem of tomato plant 37, especially when a straight line cannot be formed. However, due to the existence of the safety zone 35, even if there are deviations, it can still be protected. It should be noted that the above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or perform equivalent replacements on some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for mechanical weeding between transplanted crops, characterized in that: The method comprises Step 1: Treat the roots of the crops to be transplanted with an effective dose of sodium fluorescein solution, and transplant the treated crops into the field; Step 2: When weeding is required for the crops to be transplanted, the imaging system and the lighting system are used to collect images of the crop plants; the positions of the plane reflector and the LED lamp are adjusted so that the plane reflector does not reflect light and the LED lamp light is steadily directed toward the plants; Step 3: Adjust the camera position so that the target crop and weeds are included in the imaging field of view, and set the appropriate camera exposure value and gain value through the microcomputer; Step 4: pre-set the coordinates of the starting and ending positions of the crop rows where the system is traveling. When the system is traveling in a straight line along the crop rows in the field according to the travel control algorithm of the microcomputer, the system takes real-time photos through the camera, extracts the region of interest through the image processing algorithm of the microcomputer, and determines the coordinates of the unearthed point of the stem of the crop seedling using the intersection of the lines formed by different viewing angles; Step 5: Using the data of the emergence point of the stem of the crop seedling, the opening and closing action of the weeding blade is controlled; Wherein, step 4 specifically includes: S41, using a microcomputer to obtain the position information returned by the GNSS positioning system and the angle deviation returned by the angle sensor, and moving along the crop row according to the preset start and end point coordinates. When the system deviates from the predetermined straight line, the direction is adjusted through the differential rotation of the crawler chassis system based on the PID control algorithm; S42, performing median filtering and image sharpening operations on the image taken by the camera to remove noise; S43, binarizing the image after the noise is removed based on the difference in grayscale values, and extracting the fluorescence signal of the crop plant; S44, performing an opening and closing operation on the binary image to remove image noise and smooth the boundaries, and remove areas with weak fluorescence signals, that is, areas with too small connected domain areas; S45, defining seven regions of interest in the binary image, and naming them as: upper left LR, left middle LM, lower left LB, middle M, upper right RM, right middle RM, and lower right RM; S46, finding the point of the six surrounding interest regions closest to the middle interest region; S47. Draw a corresponding straight line based on the nearest point found in the region of interest, find the point where the stem of the crop plant emerges from the soil, and use it as the actual position point of the crop.

2. The method for mechanical weeding between transplanted crops according to claim 1, characterized in that: In step S41, the angle obtained by the angle sensor is used as the input of the PID control algorithm, and the differential rotation of the crawler chassis system is realized according to the incremental PID expression, and the expression is as follows: in, , To control the increment, , , They are the input angle information at the current moment, the input angle information at the previous moment, and the angle change.

3. The method for mechanical weeding between transplanted crops according to claim 1, characterized in that: In step S43, the fluorescence signals of the crop plants are extracted and the fluorescence signals of the weeds are removed. The thresholding rule is as follows: in, Represents the original grayscale image captured by the camera; The specific value of the threshold is determined by the actual lighting conditions, treatment concentration and time, soil color, etc. is the binary image after thresholding.

4. The method for mechanical weeding between transplanted crops according to claim 1, characterized in that: In step S47, the method for identifying the emergence point of the crop plant stem includes: Ideally, the six closest points can be captured; When weeds are located between the mirror and the crop plants, they may block the mirror imaging, causing the fluorescence signal to be lost in the area of ​​interest, thus affecting the solution of the nearest point; Based on the nearest point found in the region of interest, draw the corresponding straight line. The following situations may occur: (1) Can form three straight lines; (2) Can form two straight lines; (3) Can form one straight line; (4) Cannot form a straight line and the closest points are on both sides; (5) Cannot form a straight line and the closest point is on one side and is the closest point between the upper and lower areas; (6) Cannot form a straight line and the closest point is on one side and is the closest point in the middle area; (7) Only the centroid point; The calculation rules for each case are as follows: ； ； ； ； ； ； ； in, Represents the number of straight lines formed by the closest points; Represents the point where the stem of a tomato plant emerges from the soil; Represents the intersection of two straight lines formed by the closest points relative to each other; Represents the number of edges where the nearest point is located. LT, LM, LB, M, RT, RM, and RB represent the ROI areas of the upper left, middle left, lower left, middle, upper right, middle right, and lower right, respectively. Represents the horizontal coordinate of the nearest point in the ROI area on the left; Represents the horizontal coordinate of the nearest point in the ROI area on the right; The vertical coordinate representing the center point of the image; The horizontal coordinate representing the center point of the image; Represents the ordinate of the nearest point in the upper ROI area; Represents the ordinate of the nearest point in the lower ROI area; Represents the vertical coordinate of the ROI area in the middle of the left and right sides; Represents the centroid coordinates of the middle ROI area.

5. The method for mechanical weeding between transplanted crops according to claim 1, characterized in that: Step 5 uses the data of the stem emergence point of the crop seedlings to control the opening and closing of the weeding blade by calculating the delay time of the weeding blade opening, and transmits it to the single-chip microcomputer. The calculation formula is as follows: in, To detect the delay time of weeding blade opening after the stem of crop plant emerges from soil. is the pixel value of the image along the forward direction, is the pixel value of the distance from the point where the crop plant stem emerges from the soil to the boundary where the crop plant enters the device along the travel direction, is the conversion ratio between image and actual distance, is the distance between the weeding blade and the imaging darkroom, The length of the safe area for weeding, is the system travel speed, is the average processing time of the program; After the single-chip microcomputer obtains the delay time for the weeding blade to open, when the predetermined opening time is reached, the single-chip microcomputer transmits the information to the electric cylinder controller, and the electric cylinder controller controls the servo electric cylinder to be energized, and its telescopic rod pushes the weeding blade to open along the rolling slide rail at a uniform speed equal to the travel speed of the device, and waits for the weeding blade to pass through the set safety area, and then the single-chip microcomputer controls the telescopic rod of the servo electric cylinder to be retracted, and the weeding blade is closed, and this cycle is repeated to achieve the weeding blade avoiding the crop plants and weeding between the plants; The formula for setting the safety area according to the actual situation and determining the time when the weeding blade is turned on is as follows: in, time to open the blade for weeding; Length of safe area for weeding; is the device travel speed.

6. A mechanical weeding system for transplanted crops, characterized in that: The system is used to implement the method for mechanical weeding between transplanted crops according to any one of claims 1 to 5, and comprises An imaging system, including an imaging darkroom in which a camera for capturing images of crops and weeds is fixed; The lighting system is located inside the imaging darkroom and includes a number of LED lights used to provide light or supplementary light for the crops; The weeding blade system is located at the rear of the imaging darkroom and includes a servo electric cylinder, a rolling slide rail and a weeding blade. When the servo electric cylinder performs an action, it drives the weeding blade to move horizontally on the rolling slide rail to avoid crop plants and weed between plants. GNSS positioning system, located above the imaging system, is used to obtain the current device location information in real time; The crawler chassis system is located below the imaging system and includes crawlers, crawler drive motors, and crawler drive wheel sets. It realizes differential motion through PID control so that the system moves along the planned route. The power distribution control box is located in front of the imaging system and includes a power supply, a single-chip microcomputer and a microcomputer arranged therein, wherein the microcomputer is used to process image data taken by the camera, calculate the opening and closing time of the weeding blade, run the PID control algorithm, and transmit information to the single-chip microcomputer.

7. The inter-row mechanical weeding system for transplanted crops according to claim 6, characterized in that: A mirror bracket is fixed at the inner bottom end of the imaging darkroom, and a plane reflector is placed on the mirror bracket to form a multi-view system to provide a side view of the crop stem; The mirror bracket is arranged parallel to the moving direction of the system, and each mirror bracket can hold three plane reflectors, wherein the bottom edge of the plane reflector in the middle is parallel to the moving direction and forms an angle of 45° with the horizontal ground, and the plane reflectors on both sides form an angle of 120° with the middle mirror.

8. The inter-row mechanical weeding system for transplanted crops according to claim 6, characterized in that: The lighting system includes eight LED lamps, four of which are located below the mirror bracket inside the imaging darkroom to provide sufficient light source for the crop stems, and the remaining four are fixed in the middle of the imaging darkroom to provide supplementary light for the crop plants.

9. The inter-row mechanical weeding system for transplanted crops according to claim 6, characterized in that: The rolling slide rail is provided with two groups up and down, and two rolling sliders are slidably provided on the two groups. The rolling slider located at the upper side is provided with an electric cylinder controller, and the rolling slider located at the lower side is provided with a rotating motor. The weeding blade is provided at the bottom of the rotating motor, and the rolling sliders located at the same side are connected by a weeding blade connecting rod, wherein the electric cylinder controller is used to control the telescopic action of the telescopic rod of the servo electric cylinder, and the telescopic rod is connected to the rolling slider, and when the telescopic action is extended and retracted, the weeding blade can be driven to translate on the rolling slide rail, so as to achieve the avoidance of crop plants and weeding between plants.

10. The inter-row mechanical weeding system for transplanted crops according to claim 6, characterized in that: An angle sensor is also provided in the power distribution control box, and the angle sensor is used to obtain the angle of the system deviation from a predetermined straight line and transmit the information to the microcomputer.

Images