Bulbus fritillariae ussuriensis topsoil stripping depth automatic control system based on machine vision

Abstract

A fritillary bulb topsoil stripping depth automatic control system based on machine vision relates to the technical field of agricultural mechanical equipment, and comprises an image acquisition module, an image processing module, a control module and an execution feedback module, the image processing module runs a lightweight YOLOv5s target detection algorithm on an upper computer, detects and counts the fritillary bulb scraped out of the soil in real time, and calculates a density value as feedback information; the control module adopts a single-chip microcomputer, according to the deviation between the real-time density and the preset target density, a pulse control signal is sent to a servo motor through a servo driver, and an electric push rod is driven to dynamically adjust the depth of a soil scraping plate. The feedback module monitors the position of the executing mechanism in real time through a high-precision displacement sensor and a limiting switch, and closed-loop control is achieved. The problem of high dependence on manual continuous observation in a traditional operation mode is effectively solved, and mechanical damage to the fritillary bulb caused by unstable stripping depth is effectively reduced.

Description

[0001] The present invention relates to the technical field of agricultural machinery and equipment, and in particular to an automatic control system for the stripping depth of Fritillaria cirrhosa surface soil based on machine vision. Background Art

[0002] Fritillaria cirrhosa is a perennial herbaceous plant of the genus Fritillaria in the Liliaceae family. Its underground bulbs are the primary medicinal and edible part. Traditionally, Fritillaria cirrhosa harvesting requires operators to continuously visually monitor the depth of topsoil removal and frequently adjust mechanical equipment. This results in slow response, low control accuracy, and a high damage rate, severely limiting harvesting efficiency and product quality.

[0003] In the existing technology, the control of the stripping depth of the Fritillaria cirrhosa topsoil mainly relies on the operator's manual adjustment of the harvester's mechanical hydraulic suspension system and the screw depth limiter. During the operation, the operator needs to constantly look back to observe the relative position of the scraper, the soil, and the bulbs, change the suspension height by adjusting the hydraulic valve, and rotate the screw to change the depth of the depth-limiting scraper into the soil. The response speed is slow and easily affected by experience and fatigue, resulting in unstable stripping depth and a high Fritillaria cirrhosa damage rate. In addition, the existing depth-limiting device lacks real-time perception of Fritillaria cirrhosa distribution information and cannot adaptively adjust the operating depth, making it difficult to meet the needs of efficient mechanized harvesting in different plots and complex environments. Summary of the Invention

[0004] The purpose of the present invention is to overcome the problems of slow response, low precision and lack of real-time perception of crop distribution in the existing technology in stripping depth control, and to provide an automatic control system for the surface stripping depth of Fritillaria cirrhosa based on machine vision. It can detect the target density of Fritillaria cirrhosa in the stripping area in real time during mechanized movement, and adaptively adjust the depth of the scraper to achieve precise control of the stripping depth, significantly reduce the damage rate of Fritillaria cirrhosa and improve harvesting efficiency.

[0005] In order to solve the problems existing in the background technology, the present invention adopts the following technical solutions: including an integral frame, a motor mounting plate, an equipment mounting panel, a front wheel, a universal wheel, an image acquisition module, an image processing module, a control module, an execution module and a feedback module. A motor mounting plate is provided at one end of the integral frame, a DC drive motor is fixed on the motor mounting plate, an output shaft of the DC drive motor is connected to a sprocket through a chain, the sprocket is fixed on the transmission shaft, a rotating shaft is horizontally mounted on the integral frame, front wheels are mounted at both ends of the rotating shaft, a universal wheel is mounted at the other end of the integral frame, an equipment mounting panel is provided on the integral frame, an image acquisition module, an image processing module, a control module, an execution module and a feedback module are mounted on the equipment mounting panel; the image acquisition module includes a camera adjustment guide rail and an industrial camera, the longitudinal section of the camera adjustment guide rail is T-shaped, and the camera adjustment guide rail is arranged It is installed at the bottom of the equipment installation panel, and an industrial camera is installed on the camera adjustment rail; the image processing module includes a host computer, and the image processing module is electrically connected to the image acquisition module; the control module includes a single-chip microcomputer, and the control module is electrically connected to the image processing module; the execution module includes a servo driver, a servo motor, an electric push rod and a scraper board, the servo driver is installed at the top of the equipment installation panel, and an electric cylinder mounting seat is provided at the bottom of the equipment installation panel, and a servo motor and an electric push rod are installed on the electric cylinder mounting seat, the servo motor is connected to the servo driver and the electric push rod respectively, and the electric push rod is connected to the scraper board; the execution module is electrically connected to the control module; the feedback module includes a displacement sensor and a limit switch, the displacement sensor is installed along the movement direction of the electric push rod, and the limit switch is set at the upper and lower limit positions of the scraper board; the feedback module is electrically connected to the execution module.

[0006] The single chip microcomputer is configured with a prediction delay calculation unit for calculating the control delay time according to the horizontal distance between the camera and the scraper and the current travel speed, and using the delay time for timing compensation to ensure that the depth adjustment is synchronized with the working area.

[0007] A Hall sensor is installed on the axle of the front wheel, and the Hall sensor is electrically connected to the control module, that is, the Hall sensor is electrically connected to the single-chip microcomputer.

[0008] A mobile power supply is provided on the equipment installation panel, and the mobile power supply is electrically connected to the DC drive motor, the image acquisition module, the image processing module, the control module, the execution module and the feedback module respectively.

[0009] The control module is configured with a step limiter unit for limiting the absolute value of each depth adjustment to no more than 5 mm, so as to avoid excessive response due to target density detection error and ensure smooth operation of the system.

[0010] The human-machine interaction module is installed on the device installation panel. The human-machine interaction module is electrically connected to the mobile power supply. The human-machine interaction module includes a touch screen. The human-machine interaction module is electrically connected to the control module, that is, the touch screen is electrically connected to the single-chip microcomputer.

[0011] The control method of the system comprises the following steps: S1. Image acquisition: The industrial camera in the image acquisition module collects surface images of the operated area and the scraper blade area at a preset frequency, transmits the collected image information to the image processing module, and synchronizes the operating speed v of the acquisition device with the distance L from the image acquisition point to the scraper blade; S2. Image recognition and target number extraction: The host computer in the image processing module processes the collected scraping area image based on the YOLOv5s model, detects and counts the Fritillaria cirrhosa, and outputs the number of Fritillaria cirrhosa targets. Based on the number of Fritillaria cirrhosa targets output after image processing and the field of view area of ​​the industrial camera, the Fritillaria cirrhosa number density per unit area ρ is converted t , and the density ρ t The information is transmitted to the control module for depth limit determination; S3, depth control strategy judgment: the control module will obtain the currently identified Fritillaria number density ρ t Compared with the preset target density ρ0, if ρ t <ρ0, it is inferred that the plants in the front area are sparse, and a "deepen peeling" control command is issued; if ρ t >ρ0, then a "shallowing peeling" or "maintaining depth" control instruction is issued; to ensure the continuity and stability of the adjustment, the control module sets the maximum depth adjustment step Δd max , limit the depth change to no more than this value each time the command is adjusted; S4, prediction compensation: The single chip microcomputer in the control module calculates the current speed v m And the horizontal distance L from the industrial camera to the scraper, calculate the predicted delay time T=L / v m , and compensate for the depth adjustment timing; S5, Depth Adjustment: Based on the predictive control strategy, the control module calculates the forward depth adjustment increment Δd (t + T) = K p ×(ρ0-ρ t ), where K p is the proportional adjustment coefficient, which reflects the influence of density deviation on depth adjustment amplitude; Δd(t+T) is divided into the limit condition Δd max Compare, if exceeded, cut off to the maximum allowable adjustment amount to achieve control stability and protection; S6. Execution control: The control module uses the corrected depth adjustment value as a control signal and sends it to the execution module in real time through the control output module. After the execution module receives the control signal, the servo driver in the execution module converts the adjustment value into a pulse signal, drives the electric push rod to extend and retract, and adjusts the height position of the scraper blade. The scraper blade dynamically strips the soil layer in the target area according to the adjusted depth, completing a depth control closed-loop process. S7. Closed-loop feedback: During the entire operation process, the system continuously executes the image acquisition module for image acquisition, the image processing module for recognition, the control module for control decision-making, the execution module for execution, and the feedback module for feedback, forming a continuous closed-loop intelligent control process to achieve precise control of the stripping depth of the surface soil of the Fritillaria cirrhosa. Among them, the displacement sensor and limit switch detect the position of the scraper in real time, and input the feedback information into the control module to correct the depth calculation of the next cycle.

[0012] In step S1, the industrial camera captures images at a frequency of 1 frame per second, and the distance between frames does not exceed the length of the field of view of the industrial camera to ensure that adjacent images have an overlapping area of ​​at least 20%.

[0013] In step S2, the YOLOv5s model is lightweight and improved and integrated with the attention mechanism to improve the detection accuracy of the Fritillaria target in a complex background.

[0014] In step S7, when the displacement sensor or limit switch detects that the scraper blade reaches the preset limit position, the control module immediately stops or reverses the movement of the execution module to ensure safe and reliable operation of the system.

[0015] The beneficial effects of the present invention are: 1. Use the image acquisition device to obtain images of the working area in real time, combine the target detection model to identify the number of Fritillaria cirrhosa plants in the image and convert their density, and use the control module to implement the prediction compensation and depth limit adjustment strategy to drive the actuator to achieve automatic control of the scraping depth, thus building an integrated intelligent control closed loop of "identification-judgment-control-feedback".

[0016] 2. It can realize adaptive adjustment of scraping depth under different density distribution conditions, effectively solving the problems of uneven soil coverage, high crop damage rate, and adjustment lag under traditional manual judgment or fixed depth stripping methods. It has significant advantages such as high stripping accuracy, rapid control response, and strong operational adaptability, and can provide technical support for the intelligent and mechanized harvesting of Fritillaria cirrhosa. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the overall structure of the present invention; Figure 2 It is a side view of the present invention; Figure 3 It is a front view of the present invention; Figure 4 It is a bottom view of the present invention; Figure 5 It is a schematic structural diagram of the surface soil stripping device of the present invention; Figure 6 It is a block diagram of the system connection structure of the present invention; Figure 7 It is a functional module block diagram of the control system of the present invention; Figure 8 It is a flowchart of the system work flow of the present invention; Figure 9 It is the electrical connection principle diagram of the present invention. DETAILED DESCRIPTION

[0018] With reference to the figures, the present invention specifically adopts the following implementation mode: comprising an integral vehicle frame 1, a motor mounting plate 2, an equipment mounting panel 3, a front wheel 4, a universal wheel 5, an image acquisition module, an image processing module, a control module, an execution module and a feedback module, a motor mounting plate 2 is provided at one end of the integral vehicle frame 1, a DC drive motor 10 is fixed on the motor mounting plate 2, an output shaft of the DC drive motor 10 is connected to a sprocket 11 through a chain 12, the sprocket 11 is fixed on a transmission shaft 13, a rotating shaft 13 is horizontally mounted on the integral vehicle frame 1, and front wheels 4 are mounted at both ends of the rotating shaft 13, the integral vehicle frame 1 is provided with a motor mounting plate 2, a DC drive motor 10 is fixed on the motor mounting plate 2, an output shaft of the DC drive motor 10 is connected to a sprocket 11 through a chain 12, the sprocket 11 is fixed on a transmission shaft 13, a rotating shaft 13 is horizontally mounted on the integral vehicle frame 1, and front wheels 4 are mounted at both ends of the rotating shaft 13, A universal wheel 5 is installed at the other end of the frame 1; an equipment installation panel 3 is provided on the overall frame 1, on which an image acquisition module, an image processing module, a control module, an execution module and a feedback module are installed; the image acquisition module includes a camera adjustment rail 8 and an industrial camera 21, the longitudinal section of the camera adjustment rail 8 is T-shaped, the camera adjustment rail 8 is installed at the bottom of the equipment installation panel 3, and the industrial camera 21 is installed on the camera adjustment rail 8, which is used to adjust the installation position of the industrial camera 21 relative to the scraper 6 in the horizontal and vertical directions to adapt to different working conditions The image processing module includes a host computer 20, which is electrically connected to the image acquisition module. The image processing module adopts an improved lightweight YOLOv5s target detection model. The optimal weight of the model is about 13.6M, and the single-frame detection time is less than 30ms to meet the real-time recognition performance requirements of the system; the control module includes a single-chip microcomputer 30, which is electrically connected to the image processing module; the execution module includes a servo driver 31, a servo motor 33, an electric push rod 32 and a scraper 6, and the servo driver 31 is installed on An electric cylinder mounting base 7 is provided at the top and bottom of the equipment installation panel 3, and a servo motor 33 and an electric push rod 32 are installed on the electric cylinder mounting base 7. The servo motor 33 is connected to the servo driver 31 and the electric push rod 32 respectively, and the electric push rod 32 is connected to the scraper blade 6; the execution module is electrically connected to the control module; the feedback module includes a displacement sensor 34 and a limit switch 35, the displacement sensor 34 is installed along the movement direction of the electric push rod 32, and the limit switch 35 is set at the upper and lower limit positions of the stroke of the scraper blade 6; the feedback module is electrically connected to the execution module.

[0019] The single chip microcomputer 30 is configured with a prediction delay calculation unit for calculating the control delay time according to the horizontal distance between the camera and the scraper and the current travel speed, and using the delay time for timing compensation to ensure that the depth adjustment is synchronized with the working area.

[0020] A Hall sensor 9 is installed on the axle of the front wheel 4. The Hall sensor 9 is electrically connected to the control module, that is, the Hall sensor 9 is electrically connected to the single-chip microcomputer 30, and is used to calculate the spatial coverage of adjacent frame images based on the travel speed and the camera field of view length. If the coverage is lower than the preset threshold, a control instruction to reduce the travel speed or increase the image acquisition frequency is sent to the control module to ensure continuous image coverage and ensure the real-time and integrity of the peeling depth adjustment.

[0021] A mobile power supply 40 is provided on the device installation panel 3, and the mobile power supply 40 is electrically connected to the DC drive motor 10, the image acquisition module, the image processing module, the control module, the execution module and the feedback module respectively. The mobile power supply 40 provides 24V DC power to the DC drive motor 10, the image acquisition module, the image processing module, the control module and the execution module respectively to ensure stable power supply to each module during long-term operation.

[0022] The control module is configured with a step limiter unit for limiting the absolute value of each depth adjustment to no more than 5 mm, so as to avoid excessive response due to target density detection error and ensure smooth operation of the system.

[0023] A human-machine interaction module is installed on the equipment installation panel 3. The human-machine interaction module is electrically connected to the mobile power supply 40. The human-machine interaction module includes a touch screen 22. The human-machine interaction module is electrically connected to the control module, that is, the touch screen 22 is electrically connected to the single-chip computer 30, and is used to display the density of the Fritillaria cirrhosa, the scraping depth and the system status in real time when the system is running, and allows the operator to adjust parameters such as the preset target density and the initial stripping depth through the HMI touch screen to realize online monitoring and parameter configuration of the system.

[0024] The control method of the system comprises the following steps: S1. Image acquisition: The industrial camera 21 in the image acquisition module acquires surface images of the operated area and the area around the scraper blade 6 at a preset frequency, transmits the acquired image information to the image processing module, and synchronizes the operating speed v of the acquisition device with the distance L from the image acquisition point to the scraper blade 6; S2. Image recognition and target number extraction: The host computer 20 in the image processing module performs image processing on the collected scraping area image based on the YOLOv5s model, detects and counts the Fritillaria cirrhosa, and outputs the number of Fritillaria cirrhosa targets. Based on the number of Fritillaria cirrhosa targets output after image processing and the field of view area of ​​the industrial camera 21, the Fritillaria cirrhosa number density per unit area ρ is converted t , and the density ρ t The information is transmitted to the control module for depth limit determination; S3, depth control strategy judgment: the control module will obtain the currently identified Fritillaria number density ρt Compared with the preset target density ρ0, if ρ t <ρ0, it is inferred that the plants in the front area are sparse, and a "deepen peeling" control command is issued; if ρ t >ρ0, then a "shallowing peeling" or "maintaining depth" control instruction is issued; to ensure the continuity and stability of the adjustment, the control module sets the maximum depth adjustment step Δd max , limit the depth change to no more than this value each time the command is adjusted; S4, prediction compensation: the single chip microcomputer 30 in the control module calculates the current travel speed v m And the horizontal distance L from the industrial camera 21 to the scraper 6, calculate the predicted delay time T=L / v m , and compensate for the depth adjustment timing; S5, Depth Adjustment: Based on the predictive control strategy, the control module calculates the forward depth adjustment increment Δd (t + T) = K p ×(ρ0-ρ t ), where K p is the proportional adjustment coefficient, which reflects the influence of density deviation on depth adjustment amplitude; Δd(t+T) is divided into the limit condition Δd max Compare, if exceeded, cut off to the maximum allowable adjustment amount to achieve control stability and protection; S6. Execution control: The control module uses the corrected depth adjustment value as a control signal and sends it to the execution module in real time through the control output module. After the execution module receives the control signal, the servo driver 31 in the execution module converts the adjustment value into a pulse signal, drives the electric push rod 32 to extend and retract, and adjusts the height position of the scraper blade 6. The scraper blade 6 dynamically strips the soil layer in the target area according to the adjusted depth, completing a depth control closed-loop process. S7. Closed-loop feedback: During the entire operation process, the system continuously executes the image acquisition module for image acquisition, the image processing module for recognition, the control module for control decision-making, the execution module for execution, and the feedback module for feedback, forming a continuous closed-loop intelligent control process to achieve precise regulation of the stripping depth of the surface soil of the Fritillaria cirrhosa. Among them, the displacement sensor 34 and the limit switch 35 detect the position of the scraper plate 6 in real time, and input the feedback information into the control module to correct the depth calculation of the next cycle.

[0025] In step S1, the industrial camera 21 captures images at a frequency of 1 frame per second, and the distance between frames does not exceed the field of view length of the industrial camera 21 to ensure that adjacent images have at least 20% overlapping areas.

[0026] In step S2, the YOLOv5s model is lightweight and improved and integrated with the attention mechanism to improve the detection accuracy of the Fritillaria target in a complex background.

[0027] In step S7, when the displacement sensor 34 or the limit switch 35 detects that the scraper blade 6 has reached the preset limit position, the control module immediately stops or reverses the movement of the execution module to ensure safe and reliable operation of the system.

[0028] The present invention discloses an automatic control system for the stripping depth of the surface soil of the Fritillaria thunbergii based on machine vision, comprising a movable front wheel 4 and a universal wheel 5, an image acquisition module, an image processing module, a control module, an execution module and a scraper 6; the image acquisition module is installed at the rear end of the overall frame 1, behind the scraper 5, and is used to collect surface image information of the stripped area in real time; the upper computer 20 in the image processing module is electrically connected to the industrial camera 21 in the image acquisition module, and is used to identify and analyze the collected image to obtain the distribution density of the Fritillaria thunbergii target; the control module is connected to the image processing module, and is used to generate a control instruction according to the recognition result and the driving state of the overall frame 1; the execution module is connected to the control module, and is used to drive the scraper 6 to adjust the stripping depth according to the control instruction, so as to realize dynamic regulation of the surface soil stripping process.

[0029] The actuator module includes an electric push rod 32, a servo driver 31, an electric cylinder mounting bracket 7, a displacement sensor 34, and limit switches 5 and 35. The electric push rod 32 is fixed to the lower portion of the equipment mounting panel via the electric cylinder mounting bracket 7. Its cylinder is connected to the scraper blade 6. Driven by the servo driver 31, the piston rod can accurately extend and retract, driving the scraper blade 6 up and down to achieve depth adjustment. The displacement sensor 34 is installed along the movement direction of the electric push rod 32 to detect the actual displacement of the scraper blade 6 in real time and provide feedback to the control module. The limit switches 35 are set at the upper and lower limits of the scraper blade 6's travel to prevent overshoot. By combining displacement feedback and limit protection, this actuator module achieves high-precision, fast-response, and stable control of the scraper blade 6, making it suitable for real-time depth adjustment in complex terrain.

[0030] The image processing module includes a Jetson Nano host computer 20 and a data interface module. The industrial camera 21 is electrically connected to the Jetson Nano host computer 20 and is used to acquire high-definition images of the operating area behind the equipment. The image processing unit performs denoising, correction, and feature extraction on the acquired image data, using a target detection algorithm to identify the number and distribution of Fritillaria cirrhosa plants in the image. The data interface module is used to transmit the recognition results to the control module in real time for subsequent control command generation. This module can accurately identify and extract information about Fritillaria cirrhosa targets in the operating area, providing data support for dynamic depth control.

[0031] The control module includes an information receiving module, a decision-making and calculation module, and a control module. The information receiving module is used to receive the Fritillaria cirrhosa density information and vehicle driving status parameters sent by the image recognition and processing module. The decision-making and calculation module comprehensively judges the input information based on a preset depth limit strategy model and generates corresponding depth adjustment instructions. The control output module transmits the control instructions to the actuator, achieving real-time control of the scraper blade 6. This control module adopts a distributed functional structure, featuring high data processing efficiency and timely control response, and can adapt to changing working terrain and crop distribution environments.

[0032] During operation, the system of the present invention operates collaboratively according to the following process: As the equipment travels along the operating path, the image acquisition device acquires real-time surface image information from behind and transmits the image to the image recognition processing module; the image recognition processing module analyzes the image, extracts the distribution density and location information of the Fritillaria cirrhosa, and transmits the recognition results to the control module; the control module combines the vehicle's driving state parameters with the depth control strategy to calculate the required depth adjustment of the scraper blade 6, generates a control instruction, and sends it to the actuator; the actuator drives the electric cylinder to achieve dynamic adjustment of the position of the scraper blade 6, thereby achieving real-time control of the topsoil stripping depth. Through this process, the various functional modules work together to achieve rapid perception and precise response of the system to the operating environment.

[0033] Reference Figure 1 The overall structure of the machine vision-based automatic control system for the stripping depth of a flat fritillaria includes a vehicle frame 1, front wheels 4, universal wheels 5, a motor mounting plate 2, a device mounting panel 3, an electric push rod 32, a scraper 6, a host computer 20, an industrial camera 21, a touch screen 22, a mobile power supply 40, a single-chip microcomputer 30, a servo driver 31, a displacement sensor 34, a limit switch 35, and other auxiliary structures. The universal wheels 5 enhance the steering flexibility of the device during travel.

[0034] Reference Figure 2 The side view clearly illustrates the mounting relationship between the scraper blade 6, the actuator module, and the overall frame 1. The electric push rod 32 is connected to the electric cylinder mounting base 7 and, through a screw structure, drives the scraper blade 6 to achieve vertical linear adjustment, thereby controlling the scraping depth.

[0035] Reference Figure 3 , The front view shows the front structure of the scraper blade 6 and its relative wheelbase position. Reference Figure 4 The bottom view shows the bottom structure of the overall frame 1 and the layout relationship between the control module and the front wheel 4 and the universal wheel 5.

[0036] Reference Figure 5The actuator module includes components such as an electric push rod 32, a servo motor 33, a limit switch 35, a displacement sensor 34, a scraper blade 6, and an electric cylinder mounting base 7. The electric push rod 32 drives the scraper blade 6 up and down to adjust the working depth. The servo motor 33 is built into the push rod structure, and its output shaft is connected to a screw mechanism, providing high-precision positioning capabilities.

[0037] The electric cylinder mounting bracket 7 is fixed to the lower crossbeam of the equipment mounting panel 3, providing support and limiting functions. A displacement sensor 34 is installed along the travel direction of the electric push rod 32 to monitor the travel of the scraper blade in real time. Limit switches 35 are installed at the upper and lower limit positions of the servo cylinder, forming a travel protection mechanism to prevent damage to the electric cylinder due to overshoot.

[0038] The scraper plate 6 is an L-shaped steel plate structure, and its operating edge is set close to the ground surface. It cooperates with the electric push rod 32 to peel the topsoil at a fixed depth to ensure that the tissue structure of the flat fritillaria is not damaged.

[0039] The above structures together constitute the topsoil stripping mechanism of the present invention, which realizes precise and controllable adjustment of the stripping depth and provides a reliable execution basis for subsequent image recognition and control strategies.

[0040] Reference Figure 6 The control system's connectivity structure primarily includes an image acquisition module, an image processing module, a host computer 20, a single-chip microcomputer 30, a servo driver 31, an electric actuator 32, and a human-computer interaction module. The industrial camera 21 is connected to the host computer 20 via a wired data interface and is used to capture images of the scraped area. After the host computer 20 processes the image, it transmits the resulting Fritillaria cirrhosa population density information to the single-chip microcomputer 30.

[0041] The microcontroller 30 processes the received identification data in conjunction with the control model and outputs a corresponding depth adjustment command signal to the servo driver 31, which drives the electric actuator 32 to adjust the height of the scraper blade 6. The displacement sensor 34 and limit switch 35 provide stroke position feedback and limit position signals, respectively, forming a complete closed-loop control system. The touch screen 22 in the human-computer interaction module can be used to set parameters, display current operation status, and provide system feedback.

[0042] The above connection structure realizes the intelligent control closed-loop process of "image perception - control calculation - execution adjustment - feedback correction", providing system guarantee for the stable operation and high-precision peeling of the present invention.

[0043] Reference Figure 7The control system functional modules of the present invention are mainly composed of an image processing module, a control module, an execution control module, and a human-computer interaction module. The image recognition processing module is deployed in the host computer 20 and is responsible for receiving image data collected by the industrial camera 21, performing preprocessing, model reasoning, and number density recognition, and outputting the number index of flat fritillaria per unit area ρ t .

[0044] The control module is set in the single chip microcomputer 30, and is used to t The deviation from the preset threshold ρ0 is combined with the current working height and the response model to calculate the target depth adjustment amount Δd, generating an adjustment instruction. After the execution module interprets this instruction, it controls the servo driver 31 to execute the extension and retraction of the electric push rod 32, achieving dynamic adjustment of the peeling depth.

[0045] The human-computer interaction module primarily consists of a touch screen 22, which can be used to set ρ0, adjust control sensitivity parameters, and display the current recognition density value and execution status, thereby improving operational convenience and system controllability. The entire functional module works in coordination via a data bus and control lines, forming a complete recognition-calculation-execution control path.

[0046] Reference Figure 8 The system workflow of the present invention includes the steps of image acquisition, density recognition, density judgment, depth calculation, control execution, and feedback update. First, the industrial camera 21 captures the image of the target area after scraping in real time and transmits the image to the host computer 20.

[0047] The image recognition module inside the host computer 20 processes the image and identifies the number of Fritillaria cirrhosa per unit area ρ t , and transmits the density information to the single chip microcomputer 30. The single chip microcomputer 30 compares the current recognition density ρ t With the preset threshold ρ0, if ρ t <ρ0, the system determines that the current scraping depth is insufficient and needs to be deepened; if ρ t >ρ0, the depth is judged to be too large and needs to be reduced; if the two are close, the current depth is maintained.

[0048] According to the density deviation and the response curve model, the target depth adjustment amount Δd is calculated, and a pulse instruction is generated and sent to the servo driver 31 to control the lifting movement of the electric push rod 32 to complete the peeling depth adjustment in real time.

[0049] During adjustment, a displacement sensor 34 monitors the displacement of the electric push rod 32 in real time, and a limit switch 35 prevents overtravel. The recognition results and adjustment feedback together form a closed-loop control mechanism. Control status and operation information are simultaneously displayed on the HMI touch screen 22 for user review and adjustment.

[0050] Reference Figure 9 The electrical connection principle diagram of the present invention shows in detail the wiring relationship of each control component. The mobile power supply 40 provides stable power for the entire machine, and the positive and negative poles of the power supply are connected to the main power input terminals of the microcontroller 30 and the servo driver 31 respectively.

[0051] The industrial camera 21 is connected to the host computer 20 via a USB interface for high-speed transmission of image data. The host computer 20 and the single-chip microcomputer 30 communicate via serial ports to achieve real-time transmission of recognition results, using the UART communication protocol.

[0052] The microcontroller 30 outputs pulse signals to the control port of the servo driver 31, which controls the movement of the electric actuator 32. Simultaneously, it collects analog or coded pulse signals from the displacement sensor 34 to determine the displacement. A limit switch 35 is connected to the interrupt input port of the microcontroller 30, forcing the actuator to stop when the limit position is reached.

[0053] The touch screen 22 communicates with the microcontroller 30 via the HDMI / USB bus, enabling command issuance and data display. Shielded cables and standard terminal connectors connect the various components to ensure anti-interference performance and modular maintenance requirements.

[0054] The present invention addresses the problems of slow response, low precision and high reliance on manual adjustment in the existing control of the stripping depth of the Fritillaria cirrhosa, which leads to low mechanized harvesting efficiency and high damage rate. The present invention provides a Fritillaria cirrhosa surface soil stripping depth automatic control system based on machine vision and its control method. The Fritillaria cirrhosa targets in the stripping area are detected and density statistics are performed in real time through industrial cameras, edge computing and deep learning algorithms. Data analysis and intelligent decision-making are performed in combination with the driving speed and prediction delay model. Key operating parameters such as the scraper stripping depth and displacement range are intelligently adjusted, thereby significantly improving the operating quality and efficiency of the mechanized harvesting of the Fritillaria cirrhosa, and solving the problems of large stripping depth fluctuations, delayed response and high mechanical damage rate in the prior art.

[0055] In summary, this machine vision-based automatic control system for the topsoil stripping depth of Fritillaria truncatum utilizes an image acquisition device to acquire images of the work area in real time. It then uses a target detection model to identify the number of Fritillaria truncatum plants in the image and calculate their density. The control module then implements a predictive compensation and depth-limiting adjustment strategy, driving the actuator to automatically control the scraping depth. This creates an integrated intelligent control closed loop of "identification-judgment-control-feedback." This system can adaptively adjust the scraping depth under varying density distribution conditions, effectively resolving the problems of uneven soil coverage, high crop damage, and delayed adjustment that exist with traditional manual judgment or fixed-depth stripping methods. It boasts significant advantages such as high stripping accuracy, rapid control response, and strong operational adaptability, providing technical support for the intelligent and mechanized harvesting of Fritillaria truncatum.

Claims

1. A machine vision-based automatic control system for the depth of surface soil stripping of Fritillaria cirrhosa, characterized by: The vehicle comprises an integral vehicle frame (1), a motor mounting plate (2), a device mounting panel (3), a front wheel (4), a universal wheel (5), an image acquisition module, an image processing module, a control module, an execution module and a feedback module. The integral vehicle frame (1) is provided with a motor mounting plate (2) at one end, a DC drive motor (10) is fixed on the motor mounting plate (2), an output shaft of the DC drive motor (10) is connected to a sprocket (11) via a chain (12), the sprocket (11) is fixed on a transmission shaft (13), a rotating shaft (13) is horizontally mounted on the integral vehicle frame (1), front wheels (4) are mounted at both ends of the rotating shaft (13), a universal wheel (5) is mounted on the other end of the integral vehicle frame (1), a device mounting panel (3) is provided on the integral vehicle frame (1), and an image acquisition module, an image processing module, a control module, an execution module and a feedback module are mounted on the device mounting panel (3); The image acquisition module includes a camera adjustment rail (8) and an industrial camera (21), wherein the longitudinal section of the camera adjustment rail (8) is T-shaped, the camera adjustment rail (8) is mounted on the bottom of the equipment mounting panel (3), and the industrial camera (21) is mounted on the camera adjustment rail (8); The image processing module includes a host computer (20), and the image processing module is electrically connected to the image acquisition module; The control module includes a single chip microcomputer (30), and the control module is electrically connected to the image processing module; The execution module includes a servo driver (31), a servo motor (33), an electric push rod (32) and a scraper (6), wherein the servo driver (31) is mounted on the top of the equipment installation panel (3), an electric cylinder mounting seat (7) is provided at the bottom of the equipment installation panel (3), the servo motor (33) and the electric push rod (32) are mounted on the electric cylinder mounting seat (7), the servo motor (33) is connected to the servo driver (31) and the electric push rod (32) respectively, and the electric push rod (32) is connected to the scraper (6); the execution module is electrically connected to the control module; The feedback module includes a displacement sensor (34) and a limit switch (35), wherein the displacement sensor (34) is installed along the movement direction of the electric push rod (32), and the limit switch (35) is set at the upper and lower limit positions of the travel of the scraper blade (6).

2. The automatic control system for the surface soil stripping depth of Fritillaria cirrhosa based on machine vision according to claim 1, characterized in that: The single chip microcomputer (30) is configured with a prediction delay calculation unit for calculating a control delay time according to the horizontal distance between the camera and the scraper blade and the current travel speed, and using the delay time for timing compensation to ensure that the depth adjustment is synchronized with the working area.

3. The automatic control system for the surface soil stripping depth of Fritillaria cirrhosa based on machine vision according to claim 1, characterized in that: A Hall sensor (9) is installed on the axle of the front wheel (4), and the Hall sensor (9) is electrically connected to the control module, that is, the Hall sensor (9) is electrically connected to the single-chip microcomputer (30).

4. The automatic control system for the surface soil stripping depth of Fritillaria cirrhosa based on machine vision according to claim 1, characterized in that: A mobile power supply (40) is provided on the surface of the equipment installation panel (3), and the mobile power supply (40) is electrically connected to the DC drive motor (10), the image acquisition module, the image processing module, the control module, the execution module and the feedback module respectively.

5. The automatic control system for the surface soil stripping depth of Fritillaria cirrhosa based on machine vision according to claim 1, characterized in that: The control module is configured with a step limiter unit for limiting the absolute value of each depth adjustment to no more than 5 mm, so as to avoid excessive response due to target density detection error and ensure smooth operation of the system.

6. The automatic control system for the surface soil stripping depth of Fritillaria cirrhosa based on machine vision according to claim 1 or 4, characterized in that: A human-machine interaction module is installed on the device installation panel (3), the human-machine interaction module is electrically connected to the mobile power supply (40), the human-machine interaction module includes a touch screen (22), and the human-machine interaction module is electrically connected to the control module, that is, the touch screen (22) is electrically connected to the single-chip computer (30).

7. A control method for the automatic control system for the surface soil stripping depth of Fritillaria cirrhosa based on machine vision according to claim 1, characterized in that: The following steps are involved: S1, image acquisition: the industrial camera (21) in the image acquisition module acquires the surface image of the operating area at a preset frequency, acquires the image of the scraper (6) area, transmits the acquired image information to the image processing module, and synchronizes the operating speed v of the acquisition device and the distance L from the image acquisition point to the scraper (6); S2. Image recognition and target number extraction: The host computer (20) in the image processing module processes the collected scraping area image based on the YOLOv5s model, detects and counts the Fritillaria cirrhosa, and outputs the number of Fritillaria cirrhosa targets. Based on the number of Fritillaria cirrhosa targets output after image processing and the field of view area of ​​the industrial camera (21), the Fritillaria cirrhosa number density per unit area ρ is converted. t , and the density ρ t The information is transmitted to the control module for depth limit determination; S3, depth control strategy judgment: the control module will obtain the currently identified Fritillaria number density ρ t Compared with the preset target density ρ0, if ρ t <ρ0, it is inferred that the plants in the front area are sparse, and a "deepen peeling" control command is issued; if ρ t >ρ0, then a "shallow peeling" or "maintain depth" control command is issued; to ensure the continuity and stability of the adjustment, the control module sets the maximum depth adjustment step Δd max , limit the depth change to no more than this value each time the command is adjusted; S4, prediction compensation: the single chip microcomputer (30) in the control module calculates the prediction compensation according to the current travel speed v m And the horizontal distance L from the industrial camera (21) to the scraper (6), calculate the predicted delay time T = L / v m , and compensate for the depth adjustment timing; S5, Depth Adjustment: Based on the predictive control strategy, the control module calculates the forward depth adjustment increment Δd (t + T) = K p ×(ρ0-ρ t ), where K p is the proportional adjustment coefficient, which reflects the influence of density deviation on depth adjustment amplitude; Δd(t+T) is divided into the limit condition Δd max Compare, if exceeded, cut off to the maximum allowable adjustment amount to achieve control stability and protection; S6, execution control: the control module uses the corrected depth adjustment value as a control signal and sends it to the execution module in real time through the control output module. After the execution module receives the control signal, the servo driver (31) in the execution module converts the adjustment amount into a pulse signal, drives the electric push rod (32) to extend and retract, and adjusts the height position of the scraper (6). The scraper (6) realizes dynamic stripping of the soil layer in the target area according to the adjusted depth, completing a depth control closed loop process; S7, closed-loop feedback: During the entire operation process, the system continuously executes the image acquisition module for image acquisition, the image processing module for recognition, the control module for control decision-making, the execution module for execution, and the feedback module for feedback, forming a continuous closed-loop intelligent control process to achieve precise control of the stripping depth of the flat fritillaria surface soil. Among them, the displacement sensor (34) and the limit switch (35) detect the position of the scraper (6) in real time, and input the feedback information into the control module to correct the depth calculation of the next cycle.

8. The automatic control system for the surface soil stripping depth of Fritillaria cirrhosa based on machine vision according to claim 7, characterized in that: In step S1, the industrial camera (21) captures images at a frequency of 1 frame per second, and the distance between frames does not exceed the field of view length of the industrial camera (21) to ensure that adjacent images have an overlapping area of ​​at least 20%.

9. The automatic control system for the surface soil stripping depth of Fritillaria cirrhosa based on machine vision according to claim 1, characterized in that: In step S2, the YOLOv5s model is lightweight and improved and integrated with the attention mechanism to improve the detection accuracy of the Fritillaria target in a complex background.

10. The automatic control system for the surface soil stripping depth of Fritillaria cirrhosa based on machine vision according to claim 1, characterized in that: In step S7, when the displacement sensor (34) or the limit switch (35) detects that the scraper blade (6) has reached a preset limit position, the control module immediately stops or reverses the movement of the execution module to ensure safe and reliable operation of the system.

Images