YOLO-based garlic normal bud recognition intelligent seeding device and method

Abstract

The invention relates to the technical field of garlic seeding, in particular to a YOLO-based garlic positive bud recognition intelligent seeding device and method. The method comprises the following steps: S1, acquiring a garlic clove image through an image acquisition module; s2, performing real-time reasoning on the image by an AI reasoning module, and outputting a garlic clove position p, a bulbil direction and a confidence degree c; s3, position verification, buffer area updating and decision making are carried out according to a reasoning result; s4, the motor control module drives a stepping motor to adjust the postures of the garlic cloves according to the decision result. According to the device, accurate control over garlic normal bud sowing is achieved through the AI intelligent recognition technology, dependence of traditional mechanical sowing on the physical characteristics of garlic cloves is thoroughly eliminated, the YOLO reasoning model is specially used for optimizing garlic bulbil recognition, the biological bud point direction can be directly and accurately recognized, the influence of the factors such as the size, shape and malformation of the garlic cloves is avoided, and the method is suitable for large-scale popularization and application. The core problems of low bud setting rate and instability of a traditional seeding mode are effectively solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of garlic seeding, in particular to a garlic bud orientation recognition intelligent seeding method based on YOLO. BACKGROUND

[0002] Garlic is an important economic crop and seasoning in China, occupying an indispensable position in agricultural production. The scientificity and accuracy of the seeding process directly affect the subsequent growth trend and final output. In the growth cycle of garlic, seeding is a critical step to lay the foundation for high yield, and the orientation of the scale bud (bud tip) is a key factor affecting the quality of emergence. The growth characteristics of garlic determine that the scale bud must be seeded upward, and this "right bud" state is the premise for ensuring the smooth germination and healthy growth of garlic cloves. Only when the scale bud is upward, can the garlic clove break through the seed coat in the most reasonable posture, with the root system rapidly extending downward to absorb water and nutrients, and the leaf blade growing upward to receive light for photosynthesis, thereby forming a neat and healthy seedling population, which accumulates sufficient nutrients for the subsequent bulb enlargement.

[0003] For a long time, the traditional garlic seeding method has faced insurmountable technical bottlenecks. Traditional mechanized seeding mainly relies on mechanical structures such as spoon chains and slides to achieve directional seeding by using the physical size or center of gravity distribution of garlic cloves. However, the physiological growth center of gravity of garlic cloves does not completely coincide with the physical center of gravity, and this inherent characteristic greatly reduces the accuracy of mechanical orientation. At the same time, the physical form of garlic cloves has natural irregularities, and the size and shape of different garlic cloves differ, and some garlic cloves are deformed, which further interferes with the directional effect of mechanical structures on garlic cloves, making the right bud rate of traditional seeding methods always at a low level, which cannot meet the needs of modern agriculture for high yield and high efficiency.

[0004] In actual production, the chain reaction caused by low right bud rate has a serious impact on garlic yield and quality. A large number of garlic cloves with improper scale bud orientation after seeding consume their own stored nutrients to adjust the growth direction, which not only delays the emergence time, but also makes the seedlings weak and reduces their disease resistance. The growth cycle of seedlings in the field is not the same, and the height is not uniform, which causes an imbalance in the competition for resources such as light, heat, water, and fertilizer. The healthy plants occupy the dominant resources, while the weak plants gradually decline due to lack of sufficient growth space and nutrient supply, ultimately leading to a decrease in the number of effective garlic heads per unit area, and affecting the weight and quality of individual garlic heads. In addition, low right bud rate also increases the additional cost of agricultural production, farmers need to invest more seeds to make up for the lack of emergence, and also spend a lot of manpower for reseeding and spacing, which not only wastes resources, but also reduces production efficiency. SUMMARY

[0005] The YOLO-based intelligent garlic sprout recognition planting device includes: A regulating hopper is connected to a drive motor, and the rotation direction of the regulating hopper is controlled by the drive motor. The seeding and feeding channel is located below the control hopper; The image acquisition module is used to acquire images of garlic cloves on the regulating hopper; The AI ​​inference module includes an AI computing chip, a storage module, and a pre-installed YOLO inference model. The AI ​​inference module uses the YOLO inference model to perform real-time inference on the image, outputting the garlic clove position p and the direction of the bud. And the confidence level c, based on the inference results, performs location verification, buffer update and decision-making; The motor control module drives the motor to control the rotation direction of the hopper based on the decision results, thereby controlling the feeding direction of the garlic cloves.

[0006] According to one embodiment of the present invention, the image acquisition module consists of an industrial camera and PWM-controlled LED fill lights. Several LED fill lights surround the industrial camera, and by adapting to ambient light, the camera can capture high-quality images with consistent exposure and clear features at different times.

[0007] According to one embodiment of the present invention, a status feedback module is further included, which consists of a buzzer and an alarm indicator light.

[0008] According to one embodiment of the present invention, the motor control module sends instructions to the driver of the drive motor through a 485 communication interface. The instruction packet contains the target angle and the movement speed. After the drive motor shaft rotates to the specified position or exceeds the rated time, it returns to zero and waits for the next action instruction.

[0009] According to one embodiment of the present invention, a feeding mechanism is further included. The feeding mechanism includes a housing and a flexible guide. The upper end of the housing has a guide opening. The lower end of the guide opening is connected to the flexible guide. The inner diameter of the guide opening gradually decreases from top to bottom. The guide opening and the flexible guide form the seeding feeding channel. The inner diameter of the flexible guide gradually increases from the middle to both ends.

[0010] A YOLO-based intelligent planting method for garlic sprout recognition, employing the aforementioned YOLO-based intelligent planting device for garlic sprout recognition, includes the following steps: S1. Acquire garlic clove images using the image acquisition module; S2, the AI ​​inference module performs real-time inference on the image and outputs the garlic clove position p, the bud direction d (d∈{0,1}) and the confidence level c; S3. Perform location verification, buffer update, and decision-making based on the reasoning results; S4. The motor control module drives the stepper motor to adjust the garlic clove posture based on the decision result.

[0011] According to one embodiment of the present invention, S2 defines The specified location area for the garlic cloves is defined only if the garlic clove position p falls within the specified area. The system only considers the reasoning result valid if the garlic clove is located within a certain time frame. Then clear buffer B (i.e. ), and the motor does not operate; if If so, continue with the subsequent steps.

[0012] According to one embodiment of the present invention, the bud direction d in S2 is a binary variable used to control the rotation angle and direction of the motor. and Corresponding to the motor rotation angles ( (For a preset fixed angle); Confidence To indicate the reliability of the inference results, the system can set a confidence threshold. Only when Time, direction It was then added to the buffer.

[0013] According to one embodiment of the present invention, in step three, buffer B is of a fixed size. When the buffer is full (i.e. When the decision-making process is triggered, the mode of the bud direction in the buffer is calculated. The motor maps to the corresponding preset angle according to the mode direction, satisfying: if ;like After the motor starts rotating, clear the buffer zone B (i.e., (in order to collect the next set of reasoning results).

[0014] According to one embodiment of the present invention, the mode calculation of the buffer in step four satisfies: Let Then the mode , in The indicator function n is the size of the buffer and is an odd number.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention achieves precise control over garlic bud planting using AI intelligent recognition technology, completely eliminating the reliance on the physical characteristics of garlic cloves in traditional mechanical planting. The YOLO inference model is specifically optimized for garlic bud recognition, enabling direct and accurate identification of the direction of biological buds, unaffected by factors such as clove size, shape, or deformity. This significantly improves the accuracy and adaptability of bud recognition, effectively solving the core problem of low and unstable bud rates in traditional planting methods, and laying a solid foundation for high garlic yields.

[0016] 2. This invention significantly improves the automation and intelligence level of garlic planting, reducing labor costs and resource waste in agricultural production. The system can achieve fully automated operation of image acquisition, intelligent recognition, and posture adjustment, eliminating the need for manual intervention in the orientation of garlic cloves and reducing the labor intensity of manual operations. Simultaneously, the high positive germination rate avoids seed waste due to insufficient seedling emergence, eliminating the need for additional seed input for replanting and saving the tedious manual replanting and thinning work, thus improving planting efficiency, reducing total production costs, and aligning with the development trend of modern agriculture towards large-scale and high-efficiency production.

[0017] 3. This invention possesses excellent stability, reliability, and scalability, providing a feasible technical solution for intelligent agricultural sowing. The adaptive supplementary lighting design of the image acquisition module ensures image quality in complex environments, the lightweight design of the AI ​​inference module guarantees real-time processing capabilities, the mode decision-making mechanism in the buffer improves decision stability, and the status feedback module provides timely fault warnings, comprehensively ensuring the smooth operation of the sowing process. Furthermore, the hardware and software decoupling design of the storage module allows the system to be upgraded in hardware or optimized in software algorithms according to actual production needs. It is not only suitable for garlic sowing but can also be extended to other direction-sensitive crop sowing scenarios through model training, demonstrating broad application prospects and promotional value. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the overall process of the present invention; Figure 2 This is a schematic diagram of the overall structure of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Example

[0020] Please refer to Figure 2The YOLO-based intelligent garlic sprout recognition planting device includes: A regulating hopper 300 is connected to a drive motor 200, and the rotation direction of the regulating hopper 300 is controlled by the drive motor 200. Preferably, the regulating hopper 300 has a triangular groove, and the length of the triangular groove is greater than its width, that is, both sides are inclined surfaces and both ends are open, so that the garlic cloves falling on the regulating hopper 300 tend to be laid flat to the left and right. The sowing and feeding channel 430 is located below the regulating hopper 300. The sowing and feeding channel 430 receives garlic cloves and maintains the falling direction of the garlic cloves, ensuring that the direction in which the garlic cloves enter the soil is consistent with the direction after falling from the regulating hopper 300, that is, the scales and buds face upwards. The image acquisition module is used to acquire images of garlic cloves on the regulating hopper 300; The AI ​​inference module includes an AI computing chip, a storage module, and a pre-installed YOLO inference model. The AI ​​inference module uses the YOLO inference model to perform real-time inference on the image, outputting the garlic clove position p and the direction of the bud. And the confidence level c, based on the inference results, performs position verification, buffer update and decision-making; the AI ​​computing chip provides powerful hardware computing power support for the operation of the YOLO inference model, ensuring that the model can complete the inference processing of a single frame image within 10 milliseconds, meeting the real-time requirements of the sowing process; after the inference processing of the YOLO inference model, it can accurately identify the position of garlic cloves and the direction of the scales and output the confidence level, providing core data for subsequent decision-making; The motor control module, based on the decision result, drives the motor 200 to control the rotation direction of the hopper 300, thereby controlling the feeding direction of the garlic cloves. The motor control module drives the motor 200 to adjust the hopper 300 to turn left or right. Understandably, when the scales tend to the right, the hopper 300 is adjusted to turn left (counter-clockwise); similarly, when the scales tend to the left, the hopper 300 is adjusted to turn right (clockwise).

[0021] According to one embodiment of the present invention, the image acquisition module consists of an industrial camera 100 and PWM-controlled LED fill lights 110. Several LED fill lights 110 surround the industrial camera 100, ensuring that the camera captures high-quality images with consistent exposure and clear features at different times by adapting to ambient light. Composed of the industrial camera 100 and the PWM-controlled LED fill lights 110, this module is responsible for stably capturing high-quality garlic clove images under complex and variable ambient lighting conditions, providing a reliable image data source for the entire system.

[0022] According to one embodiment of the present invention, a status feedback module is further included, which consists of a buzzer and an alarm indicator light. During normal system operation, the status feedback module remains silent; when the system detects abnormalities such as motor stall, abnormal image acquisition, or unreliable inference results, the GPIO pin outputs a high / low level signal to drive the buzzer to emit alarm sounds of different modes, and simultaneously controls the alarm indicator light to illuminate or flash, promptly reminding operators to troubleshoot and ensuring the smooth progress of the seeding process.

[0023] According to one embodiment of the present invention, the motor control module sends a command to the driver of the drive motor 200 via a 485 communication interface. The command packet contains the target angle and movement speed. After the shaft of the drive motor 200 rotates to the specified position or exceeds the rated time, it returns to zero to wait for the next action command. The driver converts the received command into an executable electrical signal for the drive motor 200, and the drive motor 200 rotates according to the preset angle and speed to achieve the adjustment of the garlic clove posture.

[0024] According to one embodiment of the present invention, a feeding mechanism is further included. The feeding mechanism includes a housing 410 and a flexible guide 420. The upper end of the housing 410 has a guide opening 411, and the lower end of the guide opening 411 is connected to the flexible guide 420. The inner diameter of the guide opening 411 gradually decreases from top to bottom. The guide opening 411 and the flexible guide 420 form a sowing and feeding channel 430. The inner diameter of the flexible guide 420 gradually increases from the middle to both ends. The housing 410 facilitates the insertion of the feeding mechanism into the soil. The flexible guide 420 maintains the garlic cloves with their scales facing upwards and can accommodate garlic cloves of various outer diameters, enhancing versatility. Preferably, the housing 410 is made of iron or steel. Preferably, the flexible guide 420 is an elastic mesh cylinder, which is continuous from top to bottom. Using an elastic mesh cylinder enhances versatility, allowing garlic cloves of different sizes to pass through while maintaining their scales facing upwards. In particular, the lower inner diameter of the flexible guide 420 is smaller than the upper inner diameter of the flexible guide 420.

[0025] The storage module stores critical data during system operation, including collected garlic clove images, AI inference results, and motor action logs. This data can be used for subsequent system optimization analysis and provides a basis for troubleshooting. Furthermore, the independent design of the storage module separates the hardware and software components, facilitating future hardware upgrades or software algorithm optimizations, thus improving system maintainability and scalability. Typically, the storage module is an SD card.

[0026] The device acquires images of garlic cloves through an image acquisition module, identifies relevant information about the bulb scales through an AI inference module, adjusts the posture of the garlic cloves through a motor control module, provides real-time feedback on the system's operating status through a status feedback module, and achieves data storage and decoupling between hardware and software through a storage module.

[0027] Example 2 Please refer to Figure 1 A YOLO-based intelligent planting method for garlic sprout recognition, employing any of the YOLO-based intelligent planting devices for garlic sprout recognition described in Example 1, includes the following steps: S1. Acquire garlic clove images using the image acquisition module. The core function of this step is to provide high-quality garlic clove image data support for subsequent AI inference. The image acquisition module consists of an industrial camera 100 and a PWM-controlled LED supplementary light 110. The industrial camera 100 is responsible for directly capturing the image information of the garlic cloves, while the PWM-controlled LED supplementary light 110 plays a role in adapting to ambient light. In actual planting scenarios, ambient lighting conditions are complex and variable, with differences in light intensity and angle at different times, which directly affects image clarity and feature recognition. Through PWM control technology, the LED supplementary light 110 can automatically adjust the supplementary light power according to the real-time ambient light intensity, ensuring that the industrial camera 100 can capture high-quality images with consistent exposure and clear garlic clove features at different times, such as early morning, noon, or evening. This lays a solid foundation for the subsequent AI inference module to accurately identify the position and direction of the garlic bulbs.

[0028] S2, the AI ​​inference module performs real-time inference on the image, outputting the garlic clove position p, the bud direction d (d∈{0,1}), and the confidence score c. This step is the core of the entire method, its function being to perform real-time analysis and processing of the acquired garlic clove image and output key decision information. The YOLO inference model pre-installed in the AI ​​inference module has been specifically trained and optimized for garlic bud recognition, possessing the characteristics of lightweight, high accuracy, and high speed. The model performs real-time inference on the input garlic clove image, first determining the specific position p of the garlic clove in the image, clarifying the imaging area of ​​the garlic clove; secondly, determining the direction d of the bud, this direction parameter is a binary variable, taking only two values, 0 or 1, corresponding to two preset fixed angles for subsequent motor adjustment; simultaneously, it also outputs the confidence score c of the inference result, this parameter taking a value between 0 and 1, used to characterize the reliability of the model's recognition result.

[0029] S3. Based on the inference results, perform position verification, buffer update, and decision-making. This step integrates and analyzes the valid information output by the AI ​​inference to form the final motor control decision. The buffer is a fixed-size storage unit, which temporarily stores the bud orientation data after position verification and confidence filtering.

[0030] S4. The motor control module drives the stepper motor to adjust the garlic clove posture based on the decision result. This step's function is to drive the stepper motor to precisely adjust the garlic clove posture based on the decision result. After receiving the decision command output from step three, the motor control module drives the stepper motor to rotate at the corresponding preset angle, thereby adjusting the garlic clove's placement posture to ensure the scales and buds face upwards. The stepper motor features high control precision and fast response speed, accurately executing rotation commands to adjust the garlic clove to the posture required for planting. After the motor completes its rotation, the system automatically clears buffer B, allowing it to collect the next set of valid direction data output by AI inference, preparing for the next garlic clove posture adjustment, forming a cyclical operation process to ensure the continuity of the planting process.

[0031] According to one embodiment of the present invention, S2 defines The specified location area for the garlic cloves is defined only if the garlic clove position p falls within the specified area. The system only considers the reasoning result valid if the garlic clove is located within a certain time frame. Then clear buffer B (i.e. ), and the motor does not operate; if If so, continue with the subsequent steps.

[0032] According to one embodiment of the present invention, the bud direction d in S2 is a binary variable used to control the rotation angle and direction of the motor. and Corresponding to the motor rotation angles ( (For a preset fixed angle); preferred One is greater than 0 and the other is less than 0.

[0033] Confidence To indicate the reliability of the inference results, the system can set a confidence threshold. Only when Time, direction It was then added to the buffer.

[0034] According to one embodiment of the present invention, in step three, buffer B is of a fixed size. When the buffer is full (i.e. When the decision-making process is triggered, the mode of the bud direction in the buffer is calculated. The motor maps to the corresponding preset angle according to the mode direction, satisfying: if ;like After the motor starts rotating, clear the buffer zone B (i.e., This is to allow for the collection of the next set of inference results. According to one embodiment of the invention, the mode calculation in step four satisfies: Let... Then the mode ,in The indicator function n is the size of the buffer and is an odd number.

[0035] The YOLO inference model is specifically trained and optimized for garlic sprout recognition. The model is lightweight and highly accurate, with a recognition speed of less than 10 milliseconds, ensuring the real-time performance and accuracy of the planting process.

[0036] The working principle of this device is as follows: After the sowing operation begins, the image acquisition module operates continuously. The industrial camera 100, in conjunction with the PWM-controlled LED supplemental lighting 110, continuously captures images of garlic cloves passing through the conveyor channel of the sowing device. The acquired image data is transmitted in real time to the AI ​​inference module. The AI ​​computing chip drives the pre-installed YOLO inference model to process the images quickly, completing the identification and calculation of garlic clove position, bud direction, and confidence level within 10 milliseconds, and feeding the results back to the system's main control unit.

[0037] The system's main control unit first verifies the validity of the inference results, determining whether the garlic cloves are located within a pre-defined area and whether the confidence level reaches a set threshold. If the verification passes, the bud orientation data is stored in a buffer; if the verification fails, the buffer is cleared, and the motor does not perform any action. As the sowing operation progresses, the buffer continuously receives and stores valid bud orientation data. When the number of stored data reaches the buffer's preset fixed size n (where n is an odd number), the system triggers a decision-making process.

[0038] The main control unit, following the mode calculation rule, counts the number of 0s and 1s in the garlic clove direction in the buffer, determines the more frequent mode direction, and maps this direction to the corresponding preset rotation angle of the motor. Subsequently, the main control unit sends control commands containing the target angle and movement speed to the motor control module via the 485 communication interface. The driver converts the commands into electrical signals to drive the stepper motor to rotate, causing the garlic cloves to adjust their posture, ensuring that the garlic cloves fall into the planting and feeding channel 430 with the garlic cloves facing upwards. After the motor completes its rotation, it automatically returns to zero, and the buffer is cleared, ready to receive the next set of valid data and enter the next cycle.

[0039] Throughout the operation, the status feedback module continuously monitors the system's operating status, receiving status signals from the main control unit in real time via GPIO pins. When abnormal situations occur, such as motor stall, image acquisition failure, or inference confidence consistently falling below a threshold, the GPIO pins output corresponding level signals, driving the buzzer to emit a specific alarm sound and simultaneously controlling the alarm indicator light to illuminate or flash, providing timely feedback of fault information to the operator. The storage module synchronously records various data during system operation, including image data, inference results, and motor motion records, achieving long-term data preservation. This provides data support for subsequent system maintenance and optimization, and through hardware-software decoupling design, improves the system's flexibility and scalability.

[0040] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A YOLO-based intelligent planting device for garlic sprout recognition, characterized in that, include: A regulating hopper is connected to a drive motor, and the rotation direction of the regulating hopper is controlled by the drive motor. The seeding and feeding channel is located below the control hopper; The image acquisition module is used to acquire images of garlic cloves on the regulating hopper; The AI ​​inference module includes an AI computing chip, a storage module, and a pre-installed YOLO inference model. The AI ​​inference module uses the YOLO inference model to perform real-time inference on the image, outputting the garlic clove position p and the direction of the bud. And confidence level c; and confidence level c, based on the inference results, perform position verification, buffer update and decision-making; The motor control module drives the motor to control the rotation direction of the hopper based on the decision results, thereby controlling the feeding direction of the garlic cloves.

2. The intelligent planting device for garlic sprout recognition based on YOLO according to claim 1, characterized in that, The image acquisition module consists of an industrial camera and PWM-controlled LED fill lights. Several LED fill lights surround the industrial camera, and by adapting to ambient light, the camera can capture high-quality images with consistent exposure and clear features at different times.

3. The intelligent garlic sprout recognition planting device based on YOLO according to claim 1, characterized in that, It also includes a status feedback module, which consists of a buzzer and an alarm indicator light.

4. The intelligent garlic sprout recognition planting device based on YOLO according to claim 1, characterized in that, The motor control module sends commands to the driver of the drive motor via the 485 communication interface. The command packet contains the target angle and movement speed. After the drive motor shaft rotates to the specified position or exceeds the rated time, it returns to zero and waits for the next action command.

5. The intelligent garlic sprout recognition planting device based on YOLO according to claim 1, characterized in that, It also includes a feeding mechanism, which includes a housing and a flexible guide. The upper part of the housing has a guide opening, and the lower end of the guide opening is connected to the flexible guide. The inner diameter of the guide opening gradually decreases from top to bottom. The guide opening and the flexible guide form the seeding feeding channel. The inner diameter of the flexible guide gradually increases from the middle to both ends.

6. A smart planting method for garlic based on YOLO for identifying correct garlic buds, characterized in that, The method employs the YOLO-based intelligent planting device for garlic sprout recognition as described in any one of claims 1-5, and includes the following steps: S1. Acquire garlic clove images using the image acquisition module; S2, the AI ​​inference module performs real-time inference on the image and outputs the garlic clove position p, the bud direction d (d∈{0,1}) and the confidence level c; S3. Perform location verification, buffer update, and decision-making based on the reasoning results; S4. The motor control module drives the stepper motor to adjust the garlic clove posture based on the decision result.

7. The intelligent sowing method for garlic sprout recognition based on YOLO according to claim 6, characterized in that: Defined in S2 The specified location area for the garlic cloves is defined; only when the garlic clove's position p falls within this area... The system only considers the reasoning result valid if the garlic clove is located within a certain time frame. Then clear buffer B (i.e. ), and the motor does not operate; if If so, continue with the subsequent steps.

8. The intelligent planting method for garlic sprout recognition based on YOLO according to claim 7, characterized in that: In S2, the direction d of the scale bud is a binary variable used to control the rotation angle and direction of the motor. and Corresponding to the motor rotation angles and ( and (For a preset fixed angle); Confidence To indicate the reliability of the inference results, the system can set a confidence threshold. Only when Time, direction It was then added to the buffer.

9. The intelligent planting method for garlic sprout recognition based on YOLO according to claim 6, characterized in that: In step three, buffer B is of a fixed size. When the buffer is full (i.e. When the decision-making process is triggered, the mode of the bud direction in the buffer is calculated. The motor maps to the corresponding preset angle according to the mode direction, satisfying: if ;like After the motor starts rotating, clear the buffer zone B (i.e., (in order to collect the next set of reasoning results).

10. The intelligent sowing method for garlic sprout recognition based on YOLO according to claim 6, characterized in that: In step four, the mode calculation of the buffer satisfies: Let Then the mode , in The indicator function n is the size of the buffer and is an odd number.

Images