A smart monitoring system for seeding under all working conditions

Abstract

A comprehensive intelligent seeding monitoring system includes an NI data acquisition card, a finger-clamp precision seed metering device, a wireless WIFI transmitter / receiver unit, and an anomaly alarm unit. The wireless WIFI transmitter / receiver unit sends various anomaly alarm information to the anomaly alarm unit, which then emits an audible alarm. The finger-clamp precision seed metering device comprises a signal acquisition module, a signal and image processing module, and a high-precision feature recognition module, all electrically connected in sequence. This system uses real-time image reading for feature recognition, enabling rapid and accurate identification of various seeding conditions. The system can store single or even multiple seed metering operations in real time, facilitating verification of the system's accuracy after seeding. During testing, the system achieved an accuracy rate of 99.7%. The system's alarm function responds extremely quickly, immediately alerting the operator upon detecting an anomaly, reducing unnecessary seeding obstacles and improving seeding efficiency.

Description

Technical Field

[0001] This invention relates to the field of agricultural machinery technology, specifically to an intelligent monitoring system for sowing under all working conditions. Background Technology

[0002] Precision seeding technology is one of the key technologies for reducing energy consumption and increasing yield in agricultural production. Precision seed metering technology requires the seed metering device to provide a uniform and consistent flow of corn seeds. However, complex field operation conditions and malfunctions can lead to double seeding, missed seeding, broken seeds, and seed metering device jamming. This not only reduces the efficiency of seed metering but also affects the quality of seedling emergence. Double seeding can result in slow crop growth, stunted seedlings, and deformed ears; missed seeding requires replanting or supplementary planting and carries a high risk of missing the optimal sowing time, affecting crop growth and harvest quality. Therefore, improving the quality of seeding operations is crucial for ensuring high and stable crop yields. Effective and precise seed metering monitoring is the foundation and key to achieving high-quality and precise seeding.

[0003] Currently, in the research field, technologies for directly monitoring the seeding effect of seeders are still in the initial exploratory stage. Most widely used seeding quality assessment systems rely on indirect detection strategies, typically using pressure sensors or piezoelectric sensors to detect potential missed seeding. However, in field seeding operations, variable factors, such as seed breakage, can generate signal waveforms on the sensors, affecting detection accuracy. Furthermore, the uncertainty of field operating conditions and the vibration of agricultural equipment itself can induce interference in signal waveforms, potentially leading to chaotic signal waveforms recorded by the sensors. Therefore, traditional seeding quality monitoring methods are prone to misjudgment. In experiments conducted under these circumstances, these methods also fail to provide effective and real-time monitoring data on seeding quality.

[0004] Performance monitoring of finger-clamp seed meters during the sowing process is crucial, as their impact on sowing quality is significant. The finger-clamp spring is a core component of the seed metering device, and its tension is critical for sowing accuracy. Improper spring tension—too loose or too tight—can lead to missed sowing or broken seeds, directly affecting germination rate and subsequent growth. Furthermore, internal structural malfunctions in the finger-clamp seed metering device can cause persistent missed sowing, negatively impacting overall crop growth and yield. This not only reduces operational efficiency but may also require additional manpower and resources for remedial measures, increasing complexity and cost. Therefore, meticulous and comprehensive performance monitoring of finger-clamp seed meters to ensure efficient and accurate sowing is key to improving agricultural productivity and crop yield. This necessitates the development of more precise monitoring technologies and methods to promptly identify and correct potential problems, ensuring optimal operating conditions for the sowing machinery. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the existing systems and provide a smart monitoring system for sowing under all working conditions.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a full-condition intelligent seeding monitoring system, including an NI data acquisition card, a finger clip precision seed metering device, a wireless WIFI transmitting and receiving unit, and an abnormal alarm unit. The wireless WIFI transmitting and receiving unit and the abnormal alarm unit transmit data wirelessly. The wireless WIFI transmitting and receiving unit sends each abnormal alarm status to the abnormal alarm unit, and the abnormal alarm unit emits an audible alarm.

[0007] The finger clip precision seed metering device includes a signal acquisition module, a signal and image processing module, and a high-precision feature recognition module that are connected in sequence.

[0008] Preferably, the signal acquisition module includes a pressure sensor installed at the seeding port of the finger-clamp precision seed metering device, which is connected in series with a 12V DC regulated power supply and a DC signal isolation transmitter to form a closed loop. An NI USB-6003 acquisition card is connected in parallel to the DC signal isolation transmitter. The NI USB-6003 acquisition card is connected to an industrial control computer. A program is written using the NI-DAQmxAPI library in Python to acquire voltage data. The program is then used in Python's SciPy library to smooth and filter the messy signal data points, preserving the main characteristics of the signal while reducing noise.

[0009] Preferably, the signal and image processing module is used to perform high-order nonlinear fitting on the signal after smoothing and filtering in the signal acquisition module using the powerful numerical computing capabilities of Python's NumPy library, Matplotlib library, and Pillow library, and automatically convert the continuous waveform signal into a set of continuous "V"-shaped signal images with identifiable features and store and record them.

[0010] Preferably, the high-precision feature recognition module includes an image reading unit, an image feature recognition unit, and a full-condition seeding anomaly alarm module that are electrically connected in sequence.

[0011] Preferably, the image reading unit starts timing from the moment the first smooth signal inflection point is detected, and performs real-time continuous reading based on the time interval of a single seeding cycle, which is the single seeding image.

[0012] Preferably, the image feature recognition unit detects extreme points and image contours in the image based on the improved Laplacian-Gaussian method and contour detection method in the scale space method, uses Hough transform to identify straight lines in the image, especially straight lines perpendicular to the X-axis, and performs real-time feature recognition on the signal image. By identifying the number of extreme points and image height in different images in real time, the sowing situation is accurately distinguished, thereby realizing the monitoring of normal sowing, reseeding, missed sowing, single-seed fragmentation, and reseeding fragmentation phenomena.

[0013] Preferably, the normal sowing phenomenon is characterized by the existence of a single extreme point;

[0014] The replay phenomenon is characterized by the existence of at least two extreme points, resulting in two consecutive waveforms.

[0015] The aforementioned missed broadcast phenomenon refers to the existence of only a single straight line parallel to the X-axis;

[0016] The single-seed fragmentation phenomenon is characterized by the existence of a single extreme point, the absence of an upward sloping edge, the presence of a downward sloping edge, and the existence of a straight line perpendicular to the X-axis.

[0017] The re-seeding fragmentation phenomenon is characterized by the presence of at least two upward sloping edges, at least two downward sloping edges, and a straight line perpendicular to the X-axis, while the overall image height is significantly higher.

[0018] Preferably, the full-condition seed metering abnormality alarm module includes a finger clamp positioning alarm unit for the finger clamp seed meterer, an internal structure alarm unit for the finger clamp seed meterer, a finger clamp seeding abnormality alarm unit, and a wireless WIFI transmitting and receiving unit.

[0019] The finger clamp positioning alarm unit of the finger clamp seed metering device uses the Laplace-Gaussian method and contour detection method to identify the height of the "V"-shaped signal image. It uses a fuzzy control method to set a certain threshold range. If the height value of a certain signal is always lower than the minimum threshold in a single consecutive seeding cycle, the spring of the finger clamp will loosen.

[0020] The internal alarm unit of the finger clip seed metering device identifies an internal malfunction if the height value of all signals is higher than the maximum threshold throughout the entire seed metering cycle.

[0021] The finger clip type abnormality alarm unit identifies that if a certain signal recognition height value is always higher than the highest threshold in a single consecutive planting cycle, and there are only two extreme points, two consecutive image waveforms, and the two image waveforms are one wide and one narrow, then the finger clip has a type abnormality.

[0022] The wireless WIFI transmitting and receiving unit sends various abnormal alarm situations to the external mobile receiving unit, and the mobile receiving unit emits an audible alarm.

[0023] The beneficial effects of this invention are as follows:

[0024] 1. The system is written on the Python software platform, and has a certain degree of stability and high reliability, and is easy to modify for new requirements in the future;

[0025] 2. The system has low power supply voltage requirements and does not require a large power supply device. For example, it can be easily powered by a few batteries, making it portable.

[0026] 3. Faced with complex field sowing environments and a series of complex sowing conditions, traditional detection methods that rely on pulse signals or radio wave signals to determine sowing conditions are prone to misjudgment. This system uses a real-time image reading mode for feature recognition, which can still quickly and accurately identify various sowing conditions.

[0027] 4. This system can store single or even multiple seeding operations in real time, making it convenient to check the accuracy of the system's identification after sowing is completed. The system's accuracy rate can reach 99.6% during testing.

[0028] 5. The system's alarm function reacts extremely quickly. Once an anomaly is detected, it will immediately alert the operator, reducing unnecessary sowing obstacles and improving sowing efficiency. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the structure of this monitoring system.

[0030] Figure 2 This is a schematic diagram of the middle finger clamp precision seed metering device of this monitoring system.

[0031] Figure 3 These are examples of images captured during experiments showing the sowing process. Detailed Implementation

[0032] 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.

[0033] Please see Figure 1-3 The present invention provides a technical solution: a full-condition intelligent seeding monitoring system, including an NI acquisition card, a finger clip precision seed metering device, a wireless WIFI transmitting and receiving unit and an abnormal alarm unit. The wireless WIFI transmitting and receiving unit and the abnormal alarm unit transmit data wirelessly. The wireless WIFI transmitting and receiving unit sends each abnormal alarm to the abnormal alarm unit, and the abnormal alarm unit emits an audible alarm.

[0034] The finger clip precision seed metering device includes a signal acquisition module, a signal and image processing module, and a high-precision feature recognition module that are connected in sequence.

[0035] The signal acquisition module includes a pressure sensor installed at the seeding port of a finger-clamp precision seed meterer, which is connected in series with a 12V DC regulated power supply and a DC signal isolation transmitter to form a closed loop. An NI USB-6003 acquisition card is connected in parallel to the DC signal isolation transmitter. The NI USB-6003 acquisition card is connected to an industrial control computer. A program is written using the NI-DAQmx API library in Python to acquire voltage data. The program is then used in Python's SciPy library to smooth and filter the messy signal data points, preserving the main characteristics of the signal while reducing noise.

[0036] A pressure sensor is installed on the seed guide ring between the seed outlet of the finger-type seed meterer and the seed cleaning brush. Each finger of the finger-type seed meterer passes through the pressure sensor. During sowing, the pressure sensor is subjected to the pressure of the seed, causing a change in resistance. At the same time, a Hall effect speed sensor is installed on the tractor tire to detect the forward speed of the tractor. The pressure sensor, 12V DC regulated power supply, and DC signal isolation transmitter are connected in series to form a closed loop. The NI data acquisition card is connected in parallel to the DC signal isolation transmitter and is connected to the industrial control computer. Pressure causes a change in resistance, which in turn causes a change in the current in the circuit. The NI data acquisition card 5 is used to acquire the voltage signal and send the acquired voltage signal to the industrial control computer.

[0037] The signal acquisition module includes a pressure sensor installed at the seed dispensing port of a finger-clamp precision seed meterer. During sowing, the seeds are held tightly to the seed metering disc by the clamping force of the finger clamps, causing a change in the resistance of the pressure sensor. This change, along with a 12V DC regulated power supply and a DC signal isolation transmitter, forms a closed loop. The change in resistance results in a change in the current of the entire circuit. Simultaneously, a Hall effect speed sensor is installed on the tractor tire to detect the tractor's forward speed. An NI USB-6003 acquisition card is connected in parallel to the DC signal isolation transmitter. The Hall effect speed sensor is connected to the acquisition card, which is then connected to an industrial control computer. The industrial control computer is equipped with Python and its user-friendly interface. A program is written using the NI-DAQmx API library in Python to acquire voltage data. SciPy in Python is used to smooth and filter the messy signal data points, preserving the main signal characteristics while reducing noise. The acquired voltage data is monitored in real time to determine if it exceeds a preset threshold. When the current or voltage exceeds the preset threshold, an alarm mechanism is automatically triggered to prevent short circuits or damage to the equipment. These thresholds can be customized by the user within the system.

[0038] The signal and image processing module is used to perform high-order nonlinear fitting on the signal after smoothing and filtering in the signal acquisition module using the powerful numerical computing capabilities of Python's NumPy, Matplotlib, and Pillow libraries as piecewise functions. This automatically converts the continuous waveform signal into a set of continuous "V"-shaped signal images with identifiable features and stores them.

[0039] The high-precision feature recognition module includes an image reading unit, an image feature recognition unit, and a full-condition seeding anomaly alarm module that are connected in sequence.

[0040] The image reading unit includes a timing function that starts timing from the inflection point of the first smooth signal to ensure accurate detection of a complete seeding cycle image. It performs real-time continuous reading based on the time interval (t) of a single seeding cycle. During the experiment, it can ensure fast and accurate reading for different seeding speeds (20km / h, 30km / h, 40km / h). Here, t is also the time it takes for the two finger clips to pass through the pressure sensor, which can be changed in real time according to different seeding speeds (the speed collected by the Hall sensor).

[0041] In image processing, because the pressure of the seed on the sensor is extremely high in the case of seed fragmentation, the sensor resistance is extremely low. Therefore, after exceeding a certain threshold, the unilateral oblique signal in the image is transformed into a straight line signal using edge detection algorithms and Hough transform.

[0042] The image feature recognition unit detects extreme points and image contours (including image height, width, and shape) in the image based on the improved Laplacian-Gaussian method and contour detection method in the scale space method. It uses Hough transform to identify straight lines in the image, especially straight lines perpendicular to the X-axis, and performs real-time feature recognition on the signal image. By identifying the number of extreme points and image height in different images in real time, it accurately distinguishes the sowing situation and realizes the monitoring of normal sowing, reseeding, missed sowing, single-seed fragmentation, and reseeding fragmentation.

[0043] The normal sowing phenomenon is characterized by the existence of a single extreme point;

[0044] The replay phenomenon is characterized by the existence of at least two extreme points, resulting in two consecutive waveforms.

[0045] The aforementioned missed broadcast phenomenon refers to the existence of only a single straight line parallel to the X-axis;

[0046] The single-seed fragmentation phenomenon is characterized by the existence of a single extreme point, the absence of an upward sloping edge, the presence of a downward sloping edge, and the existence of a straight line perpendicular to the X-axis.

[0047] The re-seeding fragmentation phenomenon is characterized by the presence of at least two extreme points, at least two sloping descending edges, and at least one straight line perpendicular to the X-axis, while the overall image height is significantly higher.

[0048] The improved Laplacian-Gaussian method involves selecting Gaussian kernels with different standard deviations to construct a scale-space pyramid and calculating the Laplacian pyramid: for each scale level, the Laplacian-Gaussian operator is applied to the image to compute the response.

[0049] The response is expressed by the following formula:

[0050]

[0051] When σ = 0.5 is optimal, for each scale level of the Laplace-Gaussian response, local extrema are sought, and an appropriate threshold is used to filter the extrema, retaining the unique and most significant extrema on one side.

[0052] The normal sowing phenomenon has a single extreme point; the reseeding phenomenon has multiple extreme points (at least two), with two consecutive waveforms, which may be continuous or have an interval between them; the missed sowing phenomenon has only a single straight line parallel to the X-axis; the single-seed fragmentation phenomenon has a single extreme point, no upward sloping edge, a downward sloping edge, and a straight line perpendicular to the X-axis; the reseeding fragmentation phenomenon has multiple (at least two) upward sloping edges, multiple (at least two) downward sloping edges, and a straight line perpendicular to the X-axis, while the overall image height is significantly higher;

[0053] By accumulating all seeding scenarios and the total number of seedings, we can obtain the seeding qualification rate, seeding uniformity, and seeding area.

[0054]

[0055] Where E is the sowing qualification efficiency, and P is the actual sowing efficiency. 实际 This is the actual seeding quantity, P. 理论 This is the theoretical seeding quantity;

[0056] Sowing uniformity can also be obtained through analysis:

[0057]

[0058] Where U is the uniformity of sowing, σ is the standard deviation of sowing amount, and μ is the average sowing amount;

[0059] A = L × W

[0060] Where A is the sowing area, L is the sowing length, and W is the sowing width;

[0061] Collect data on seeding rate, speed, and area, and conduct in-depth analysis to optimize sowing plans and improve sowing efficiency and crop yield.

[0062] The full-condition seed metering abnormality alarm module includes a finger clamp positioning alarm unit for the finger clamp seed meterer, an internal structure alarm unit for the finger clamp seed meterer, a finger clamp seeding abnormality alarm unit, and a wireless WIFI transmitting and receiving unit.

[0063] The finger clamp positioning alarm unit of the finger clamp seed metering device uses the Laplace-Gaussian method and contour detection method to identify the height of the "V"-shaped signal image. It uses a fuzzy control method to set a certain threshold range. If the height value of a certain signal is always lower than the minimum threshold in a single consecutive seeding cycle, the spring of the finger clamp will loosen.

[0064] The internal alarm unit of the finger clip seed metering device identifies an internal malfunction if the height value of all signals is higher than the maximum threshold throughout the entire seed metering cycle.

[0065] The finger clip type abnormality alarm unit identifies that if a certain signal recognition height value is always higher than the highest threshold in a single consecutive planting cycle, and there are only two extreme points, two consecutive image waveforms, and the two image waveforms are one wide and one narrow, then the finger clip has a type abnormality.

[0066] The wireless WIFI transmitting and receiving unit sends various abnormal alarm situations to the external mobile receiving unit, and the mobile receiving unit emits an audible alarm.

[0067] The system is written on the Python software platform, and it has a certain degree of stability and high reliability, and is easy to modify for new requirements in the future.

[0068] The system has low power supply voltage requirements and does not require a large power supply device. For example, it can be easily powered by a few batteries, making it portable.

[0069] Faced with complex field sowing environments and a series of complex sowing conditions, traditional detection methods that rely on pulse or radio wave signals to assess sowing conditions are prone to misjudgment. This system, however, uses real-time image reading for feature recognition, enabling it to quickly and accurately identify various sowing conditions.

[0070] This system can store single or even multiple seeding operations in real time, making it convenient to check the accuracy of the system's identification after seeding is completed. The system achieved an accuracy rate of 99.6% during testing.

[0071] The system's alarm function reacts extremely quickly; once an anomaly is detected, it will immediately alert the operator, reducing unnecessary sowing obstacles and improving sowing efficiency.

[0072] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A full-condition intelligent seeding monitoring system, comprising an NI data acquisition card, a finger-clamp precision seed metering device, and an anomaly alarm unit, characterized in that, The finger clip precision seed metering device includes a signal acquisition module, a signal and image processing module, and a high-precision feature recognition module that are electrically connected in sequence. The high-precision feature recognition module includes an image reading unit, an image feature recognition unit, and a full-condition seeding anomaly alarm module that are electrically connected in sequence. The full-condition seed metering abnormal alarm module includes a finger clamp positioning alarm unit for the finger clamp seed meterer, an internal structure alarm unit for the finger clamp seed meterer, a finger clamp seeding abnormal alarm unit, and a wireless WIFI transmitting and receiving unit. The wireless WIFI transmitting and receiving unit and the abnormal alarm unit transmit data wirelessly. The wireless WIFI transmitting and receiving unit sends each abnormal alarm to the abnormal alarm unit, and the abnormal alarm unit emits an sound to sound an alarm. The image reading unit starts timing from the moment the first smooth signal inflection point is detected, and performs real-time continuous reading based on the time interval of a single seeding as one cycle, which is the single seeding image; The image feature recognition unit uses Hough transform to identify straight lines in the image based on the extreme points and image contours in the detected image, and performs real-time feature recognition on the signal image. By recognizing the number of extreme points and image height in different images in real time, it can accurately distinguish the sowing situation and realize the monitoring of normal sowing, reseeding, missed sowing, single-seed fragmentation, and reseeding fragmentation. The normal sowing phenomenon is characterized by the existence of a single extreme point; The replay phenomenon is characterized by the existence of at least two extreme points, resulting in two consecutive waveforms. The aforementioned missed broadcast phenomenon refers to the existence of only a single straight line parallel to the X-axis; The single-seed fragmentation phenomenon is characterized by the existence of a single extreme point, the absence of an upward sloping edge, the presence of a downward sloping edge, and the existence of a straight line perpendicular to the X-axis. The re-seeding fragmentation phenomenon is characterized by the presence of at least two extreme points, at least two sloping descending edges, and at least one straight line perpendicular to the X-axis, while the overall image height is significantly higher.

2. The intelligent monitoring system for sowing under all working conditions according to claim 1, characterized in that: The signal acquisition module includes a pressure sensor installed at the seeding port of a finger-clamp precision seed meterer, which is connected in series with a 12V DC regulated power supply and a DC signal isolation transmitter to form a closed loop. An NI acquisition card is connected in parallel to the DC signal isolation transmitter. The NI acquisition card is connected to an industrial control computer. A program is written using the NI-DAQmx API library in Python to acquire voltage data. The program is then used in Python's SciPy library to smooth and filter the messy signal data points, preserving the main characteristics of the signal while reducing noise.

3. The intelligent monitoring system for sowing under all working conditions according to claim 2, characterized in that: The signal and image processing module is used to perform high-order nonlinear fitting on the signal after smoothing and filtering in the signal acquisition module using the powerful numerical computing capabilities of Python's NumPy, Matplotlib, and Pillow libraries as piecewise functions. This automatically converts the continuous waveform signal into a set of continuous "V"-shaped signal images with identifiable features and stores them.

4. The intelligent monitoring system for sowing under all working conditions according to claim 3, characterized in that: The finger clamp positioning alarm unit of the finger clamp seed metering device uses the Laplace-Gaussian method and contour detection method to identify the height of the "V"-shaped signal image, and uses fuzzy control method to set a certain threshold range. If the height value of a certain signal is always lower than the minimum threshold in a single consecutive seeding cycle, the spring of the finger clamp will loosen. The internal alarm unit of the finger clip seed metering device identifies an internal malfunction if the height value of all signals is higher than the maximum threshold throughout the entire seed metering cycle. The finger clip type abnormality alarm unit identifies that if a certain signal recognition height value is always higher than the highest threshold in a single consecutive planting cycle, and there are only two extreme points, two consecutive image waveforms, and the two image waveforms are one wide and one narrow, then the finger clip has a type abnormality. The wireless WIFI transmitting and receiving unit sends various abnormal alarm situations to the external mobile receiving unit, and the mobile receiving unit emits an audible alarm.

Images