Device and method for monitoring loss of mechanically harvested grains of crops
By using the programmable logic gate array FPGA module in the crop machine-coated grain loss monitoring device for signal processing, the problem that the monitoring device in the prior art cannot quickly adapt to different varieties and moisture content crops is solved, and efficient and accurate grain loss monitoring is achieved, reducing costs and maintenance burdens.
Patent Information
- Application Number
- CN202510241164.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-03
- Publication Date
- 2025-06-03
AI Technical Summary
The existing crop machine-harvested grain loss monitoring device cannot quickly adapt to different varieties and moisture content crops, and the signal processing speed is not enough to accurately monitor dense grain shocks, resulting in a decrease in loss monitoring accuracy and increased cost.
The programmable logic gate array FPGA module and sensor sensitive board are combined, and the band-pass filtering and identification of vibration signals are performed through the FPGA module, and the low-frequency machine vibration interference waves and impurity shock waves are eliminated to obtain the number of impacted grains per unit time.
It realizes flexible monitoring of different varieties of crops, improves monitoring accuracy and efficiency, reduces hardware costs and maintenance burdens, and improves the intelligence level of harvesting equipment.
Smart Images
Figure CN120087401A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of intelligent agricultural machinery, and in particular relates to a device and method for monitoring the loss of mechanically harvested crops. Background Art
[0002] In recent years, the level of mechanized harvesting of major grains and cash crops in my country has steadily improved. During the harvesting process, due to the large differences in the density of field crops, the different operating speeds of the operators, the unstable feeding amount of the combine harvester, and the nonlinear changes in the loads of the machine's cutting, threshing, cleaning, and collecting functional systems, the grain loss rate of crops is relatively high. In response to this problem, domestic researchers are exploring and developing grain loss monitoring principles and devices. The research route is: using piezoelectric ceramics to convert grain impact vibration signals into electrical signals, and then designing conditioning circuits to convert electrical signals into square wave pulses, and finally using a single-chip microcomputer to count the square waves. From the research route, the technical defects in the development of the device are mainly concentrated in the conditioning circuit design, which is reflected in the following aspects:
[0003] First, due to the limitation of the hardware design of the conditioning circuit, the device cannot quickly meet the loss monitoring needs of crops of different varieties and different moisture contents. For example, the traditional rice grain loss monitoring device based on the conditioning circuit is not compatible with the integrated grain loss monitoring of harvesters such as rice, wheat, soybeans, and corn. In other words, when changing a crop variety during field harvesting, the entire set of grain loss sensors must be replaced; in addition, even if the same variety of crops is harvested, due to changes in moisture content in different plots and at different times, the traditional grain loss monitoring device needs to be shut down and spend a long time adjusting the hardware parameters. The adjustment work is impossible for the operator to complete, which reduces the intelligence level of the harvesting machine.
[0004] Second, due to the signal processing speed limitation of the conditioning circuit, the number of impact vibration signals that can be collected per unit time is relatively small, which forms a monitoring bottleneck when the number of grain impacts is large (more than 100 grains per second). The traditional grain loss monitoring device based on the single-chip microcomputer and conditioning circuit is technically unable to complete the accurate monitoring of the vibration electrical signals of the dense grain impact situation. For example, when rapeseed is harvested, many grains will impact the sensor panel almost at the same time. In fact, the time interval between the impacts of two grains is extremely short. At this time, the traditional conditioning circuit is difficult to cope with, and the monitoring accuracy will be greatly reduced.
[0005] Third, if we insist on using the traditional grain loss monitoring method based on single-chip microcomputer and conditioning circuit technology, to solve the above two technical problems, researchers will be forced to develop loss monitoring conditioning circuits for different crops, which will bring two prominent problems: cost and reliability: adding conditioning circuits will inevitably increase hardware and R&D costs, and the increase of hardware conditioning circuits will reduce the reliability of the device. At the same time, too many conditioning circuits will reduce the user experience and increase the maintenance burden.
[0006] Therefore, from the perspective of scientific research and market demand, developing a monitoring device for the loss of machine-harvested crop seeds with high speed, adjustable parameters and low price is of great significance for reducing the loss of machine-harvested seeds, meeting market demand and improving the intelligent level of harvesting machines. Summary of the Invention
[0007] In view of the above technical problems, the present invention provides a monitoring device and method for the loss of machine-harvested crop seeds, which is used for automatically monitoring the loss of machine-harvested crop seeds and is beneficial to accurately, generally and efficiently online monitoring the number of lost seeds.
[0008] Note that the recording of these objectives does not prevent the existence of other objectives. One embodiment of the present invention does not need to achieve all the above objectives. Objectives other than the above can be extracted from the descriptions of the specification, drawings and claims.
[0009] The present invention achieves the above technical objectives through the following technical means.
[0010] A monitoring device for the loss of machine-harvested crop seeds includes a secondary instrument and a sensor sensitive plate; the secondary instrument includes an AD conversion module and a field programmable gate array (FPGA) module; a piezoelectric sheet is provided behind the sensor sensitive plate; the sensor sensitive plate is used to collect grain impact vibration waves, and the vibration waves are transmitted from the sensor sensitive plate to the piezoelectric sheet, and the piezoelectric sheet is used to convert the vibration signal into an electrical signal, and the electrical signal is transmitted to the AD conversion module of the secondary instrument, and the AD conversion module is used to collect the electrical signal of the piezoelectric sheet; the field programmable gate array (FPGA) module is used to receive the data stream of the AD conversion module and perform band-pass filtering and identification processing by applying a band-pass filter and an identification model for grain impact signals to eliminate low-frequency machine vibration interference waves and impurity shock waves and obtain the number of impacted grains per unit time.
[0011] In the above solution, the secondary instrument further includes a human-computer interaction module; the human-computer interaction module is used to select the current harvested crop variety, select channels and acquisition cycle parameters, call the band-pass filter and identification model for grain impact signals, and is used to display the number of impacted grains per unit time.
[0012] In the above solution, the secondary instrument further includes a UART serial communication module; the UART serial communication module is used to transmit the data of the impacted grains per unit time to the host computer.
[0013] In the above solution, the secondary instrument further includes an SD card storage module; the SD card storage module saves the data.
[0014] A method for the monitoring device for the loss of machine-harvested crop seeds as described above includes the following steps:
[0015] The sensor sensitive plate collects the impact vibration waves of the grains. The vibration waves are transmitted from the sensor sensitive plate to the piezoelectric sheet. The piezoelectric sheet converts the vibration signal into an electrical signal, and the electrical signal is transmitted to the AD conversion module of the secondary instrument. The AD conversion module collects the electrical signal of the piezoelectric sheet and transmits it to the Field Programmable Gate Array (FPGA) module; the FPGA module receives the data stream from the AD conversion module and performs band-pass filtering and identification processing using the band-pass filtering and identification model for grain impact signals to eliminate low-frequency machine vibration interference waves and impurity shock waves, and obtain the number of grains impacted per unit time.
[0016] In the above solution, the FPGA module performs band-pass filtering on the data received from the AD conversion module, including the following steps:
[0017] Set the soft trigger threshold voltage range [-K, K], set the width W of the vibration waveform sampling points, collect W consecutive data points exceeding the threshold range [-K, K], calculate the root mean square (RMS), number of peaks (Peak), amplitude (Amplitude), and frequency (Frequency) of this section of the impact signal data stream in the secondary instrument, and determine whether the characteristic data of this section of the impact signal conforms to the set band-pass filtering and identification model for crop grains. If it conforms, perform counting processing on the impact signal and count the identified crop grains per unit time.
[0018] Furthermore, the calculation method of the width W value of the vibration waveform sampling points is as follows:
[0019] W = KST
[0020] Where T is the time of the grain impact sensor sensitive plate;
[0021] S is the sampling rate of the AD conversion module;
[0022] K is a coefficient, and its value ranges from 0.6 to 1.0.
[0023] Furthermore, the calculation method of the root mean square (RMS) is: Where M n is the nth sampling data, and W is the width of the sampling points of a single impact vibration waveform, that is, the number of sampling points exceeding the soft trigger threshold voltage K value;
[0024] The calculation method of the number of peaks (Peak) is: if the previous value and the subsequent value are less than the current value, it is recorded as 1 Peak, and the total number of Peaks in W sampling data points is recorded as the number of peaks;
[0025] The calculation methods for the amplitude and frequency are as follows: perform a fast Fourier transform (FFT) on W sampling data, record the maximum value as the amplitude, and record the fundamental frequency as the frequency.
[0026] In the above solution, the band-pass filtering and recognition model for the grain impact signal is constructed after learning from the training data set:
[0027] Before the harvesting operation, a training data set is collected. The training data set contains a large amount of material impact feature data, and the features include the root mean square (RMS), the number of peaks (Peak), the amplitude, and the fundamental frequency. The grain and impurity categories are manually identified to form the training data set. A model is constructed based on the data mining algorithm of the machine learning training data set. The data mining algorithms include decision trees, random forests, or neural networks.
[0028] In the above solution, the data mining algorithm is the C45 decision tree algorithm.
[0029] In the above solution, for the band-pass filtering and recognition model of the grain impact signal, the software filtering and recognition process includes the following steps:
[0030] First, set the sampling speed and the data reading speed;
[0031] Perform band-pass filtering to allow waveforms within a certain impact frequency range to pass through;
[0032] Set the soft trigger threshold K and the waveform width W;
[0033] Extract K features of the waveform data: {X 0 , X 1 , X 2 , …, X k , …} T
[0034] X k = {RMS k , Amplitude k , Frequency k , PeaksNumber k}
[0035] Whether the waveform feature values conform to the recognition model: If the waveform feature values conform to the recognition model, it is recognized as a grain, the grain count is incremented by 1, and then it is displayed; if not, it is judged as a non-grain and directly enters the display process;
[0036] If the Stop button is pressed, then stop; otherwise, return to judging whether the waveform feature values conform to the recognition model.
[0037] Compared with the prior art, the beneficial effects of the present invention are as follows:
[0038] 1. The present invention combines a Field Programmable Gate Array (FPGA) module and a sensor sensitive board, and uses the nanosecond-level high-speed characteristics of the FPGA module to analyze, process, and store the vibration signal data stream. In the device, characteristic parameters of various crop grains such as rice, wheat, corn, and soybeans can be integrated. During field harvesting, parameters can be conveniently selected to call the band-pass filtering and recognition models for impact signals of different crop grains, breaking through the processing speed and variety adaptability limitations of the hardware conditioning circuit in the traditional piezoelectric monitoring method, which is beneficial for accurately, generally, and efficiently monitoring the number of lost grains online.
[0039] 2. The piezoelectric sheet in the present invention converts the vibration signal into an electrical signal, which is different from the traditional piezoelectric monitoring method. In this device, the signal is directly transmitted to the FPGA module through the AD conversion module, eliminating the processes of amplifying, shaping, and filtering the analog signal by the conditioning circuit.
[0040] 3. The FPGA module in the present invention receives the data stream from the AD conversion module and applies different band-pass filtering and recognition models for impact signals of grains to perform band-pass filtering and recognition processing on it to eliminate low-frequency machine vibration interference waves and impurity shock waves, and obtain the number of impact grains per unit time.
[0041] 4. The FLASH storage space occupied by the band-pass filtering and recognition model for impact signals of grains in the present invention is extremely small. Theoretically, the number of crop varieties that can be written into the software is much larger than the number of existing crop varieties in China. Compared with the traditional piezoelectric monitoring system, this device has a great technical advantage in terms of the general adaptability to crop varieties.
[0042] Note that the recording of these effects does not prevent the existence of other effects. One embodiment of the present invention does not necessarily have all the above effects. Effects other than the above can be obviously seen and extracted from the descriptions in the specification, drawings, claims, etc. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] Figure 1 is a schematic structural diagram of a sensor sensitive board according to an embodiment of the present invention.
[0044] Figure 2 is a schematic diagram of the installation position of a sensor sensitive board according to an embodiment of the present invention.
[0045] Figure 3 is a schematic diagram of the signal acquisition / processing flow according to an embodiment of the present invention.
[0046] Figure 4 is a schematic diagram of the physical object of the device hardware according to an embodiment of the present invention.
[0047] Figure 5 It is a table for rice grain feature extraction and classification according to an embodiment of the present invention.
[0048] Figure 6 It is a schematic diagram of a machine learning C45 decision tree recognition model for grains and impurities according to an embodiment of the present invention.
[0049] Figure 7 It is a schematic diagram of a software filtering and recognition process according to an embodiment of the present invention.
[0050] In the figure: 1. Sensor sensitive board; 2. Piezoelectric sheet; 3. Installation position. Detailed implementation manners
[0051] The embodiments of the present invention will be described in detail below. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to explain the present invention, and should not be construed as limiting the present invention.
[0052] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "transverse", "length", "width", "thickness", "front", "rear", "left", "right", "upper", "lower", "axial", "radial", "vertical", "horizontal", "inner", "outer", etc. indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the present invention. In addition, the terms "first" and "second" are only used for descriptive purposes and should not be construed as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present invention, "a plurality" means two or more unless otherwise specifically defined.
[0053] In the present invention, unless otherwise clearly defined and limited, the terms "installed", "connected", "connected to", "fixed", etc. should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the internal communication of two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0054] A monitoring device for the loss of machine-harvested grains of crops, comprising a secondary instrument and a sensor sensitive plate 1; the secondary instrument includes an AD conversion module and a Field Programmable Gate Array (FPGA) module; a piezoelectric sheet 2 is provided behind the sensor sensitive plate 1; the sensor sensitive plate 1 is used to collect grain impact vibration waves, and the vibration waves are transmitted from the sensor sensitive plate 1 to the piezoelectric sheet 2, and the piezoelectric sheet 2 is used to convert the vibration signal into an electrical signal, and the electrical signal is transmitted to the AD conversion module of the secondary instrument through a wire, and the AD conversion module is used to collect the electrical signal of the piezoelectric sheet 2 at a high speed greater than 100KS / s; the Field Programmable Gate Array (FPGA) module is used to receive the data stream of the AD conversion module and perform band-pass filtering and identification processing on it using different grain impact signal band-pass filtering and identification models to eliminate low-frequency machine vibration interference waves and impurity shock waves, and obtain the number of impact grains per unit time.
[0055] Preferably, the piezoelectric sheet 2 is a piezoelectric ceramic sheet, which converts the vibration signal into an electrical signal. Different from the traditional piezoelectric monitoring method, in this device, the signal is directly transmitted to the Field Programmable Gate Array (FPGA) module through the AD conversion module, eliminating the processes of amplifying, shaping, and filtering the analog signal by the conditioning circuit.
[0056] Preferably, the AD conversion module is connected to the Field Programmable Gate Array (FPGA) module through a socket. The AD conversion module converts the analog signal into a signed I16 digital signal with a sampling rate greater than 100kS / s to form an initial data stream, which flows into the Field Programmable Gate Array (FPGA) module.
[0057] Preferably, the Field Programmable Gate Array (FPGA) module uses FIFO data transmission technology to perform operations such as band-pass filtering, grain identification, and storage counting between threads on the initial data stream of the impact signal, and obtain the number of impact grains per unit time.
[0058] The secondary instrument further includes a human-computer interaction module; the human-computer interaction module is provided with a button and a liquid crystal display screen. The button is used to select parameters such as the variety of the currently harvested crop, the selected channel, and the acquisition period, and call different grain impact signal band-pass filtering and identification models, and the liquid crystal display screen is used to display the number of impact grains per unit time.
[0059] The secondary instrument further includes a UART serial communication module; the UART serial communication module is used to transmit the data of the impact grains per unit time to the upper computer.
[0060] The secondary instrument further includes an SD card storage module; the SD card storage module saves the data locally.
[0061] A method according to the above-mentioned monitoring device for the loss of machine-harvested grains of crops, comprising the following steps:
[0062] The sensor sensitive plate 1 collects the impact vibration waves of the grains. The vibration waves are transmitted from the sensor sensitive plate 1 to the piezoelectric sheet 2. The piezoelectric sheet 2 converts the vibration signal into an electrical signal, and the electrical signal is transmitted through a wire to the AD conversion module of the secondary instrument. The AD conversion module high-speed collects (greater than 100KS / s) the electrical signal of the piezoelectric sheet 2 and transmits it to the programmable logic gate array FPGA module; the programmable logic gate array FPGA module receives the data stream of the AD conversion module and applies different band-pass filtering and identification models for grain impact signals to perform band-pass filtering and identification processing on it to eliminate low-frequency machine vibration interference waves and impurity shock waves, and obtain the number of grains impacted per unit time.
[0063] According to this embodiment, specifically, the present invention converts the internal mechanism of the grain impact vibration electrical signal characteristics into a mathematical model (i.e., the band-pass filtering and identification model for grain impact signals), embeds the mathematical model into the FPGA microprocessor, and adds software to filter the low-frequency machine vibration signals to complete the vibration signal identification task on the sensitive plate. The sensor sensitive plate 1 includes a square steel plate and a piezoelectric sheet 2; the square steel plate collects the impact vibration waves of the grains, and the vibration waves are transmitted on the square steel plate to the piezoelectric sheet 2. The piezoelectric sheet 2 converts the vibration wave energy into an electrical signal, and the electrical signal is transmitted through a wire to the secondary instrument; the AD conversion module of the secondary instrument high-speed collects (greater than 100KS / s) the electrical signal of the sensor sensitive plate 1. The programmable logic gate array FPGA module receives the data stream of the AD conversion module and applies different mathematical models to perform software filtering and identification processing on it to eliminate low-frequency machine vibration interference waves and impurity shock waves, and obtain the grain impact data per unit time; the buttons of the human-computer interaction module select parameters such as the current harvested crop variety, selected channel, and acquisition period, call different mathematical models, the liquid crystal screen displays the grain impact data per unit time, and transmits this data to the upper computer and stores it locally through the UART serial communication module and the SD card storage module. The sensor sensitive plate 1 of the grain loss monitoring device is as Figure 1 shown, and the installation position 3 of the sensor sensitive plate 1 is as Figure 2 shown, the signal acquisition / processing flow is as Figure 3 shown, and the physical object is as Figure 4 shown. Preferably, there are 8 groups in total through 16-pin cable terminals, which are respectively connected to the signals of the 8-channel sensor sensitive plate 1. One end of the USB wire harness is connected to the port on the side of the secondary instrument, and the other port is connected to the battery for power supply of the instrument.
[0064] The present invention uses the piezoelectric principle to collect the impact vibration electrical signals of the material on the sensor sensitive plate 1, then converts the electrical signals into a data stream through a high-speed AD conversion module, and finally completes the acquisition, filtering, and identification of the data stream on the FPGA microprocessor.
[0065] From the perspective of waveform characteristics, the electrical signals of material impact vibrations are different: when different material grains impact the sensitive plate 1 of the sensor, the characteristics of the electrical signals correspond to those of the vibration waveforms and are different. For example, compared with corn grains, the sound of rice grains impacting the sensitive plate is small and shrill because the amplitude of the vibration waveform is small and the frequency is high; the sounds of corn grains and corn cobs impacting the sensitive plate are also different, and the reason behind this is also the difference in signal characteristics.
[0066] The steps for the programmable logic gate array FPGA module to perform band-pass filtering on the data received by the AD conversion module are as follows:
[0067] The data stream for band-pass filtering comes from the AD conversion module and is transmitted into the FIFO memory of the programmable logic gate array FPGA module. The band-pass software filter allows waveforms in a certain set frequency band to pass through, while other waveforms are significantly attenuated.
[0068] A certain set frequency band is the frequency range of material impact vibrations. After experimental verification, materials have their own unique and relatively stable impact vibration frequency bands, which generally do not change with the impact force and harvesting environment. Different materials have relatively stable impact vibration frequency bands at a stable moisture content. For crop grains such as rice, wheat, corn, and soybeans, the lower limit of the band-pass frequency is generally set to 3 kHz, and the upper limit is generally between 5 - 10 kHz. Preferably, the frequency bands of crops such as rice, wheat, soybeans, corn, and red beans are set to 3 - 8 kHz.
[0069] For the data acquisition of the band-pass filtering and recognition model of grain impact signals, an identification event is triggered when the value in the data stream is greater than the soft trigger threshold K. The value of the soft trigger threshold voltage K is related to parameters such as the thousand-grain weight of crop grains, impact speed, and thickness of the sensitive plate. After triggering the identification event, the programmable logic gate array FPGA module transfers the subsequent W data through the FIFO to another thread, performs band-pass software filtering on the data stream collected at high speed by the AD, and filters out low-frequency vibrations of machine vibrations and other low-frequency interference waveforms; set the soft trigger threshold voltage range [-K, K], set the width W of the waveform sampling points, that is, the number of sampling points exceeding the soft trigger threshold voltage K value, collect W consecutive data points exceeding the threshold range [-K, K], calculate features (including but not limited to) such as the root mean square RMS, number of peaks Peak, amplitude Amplitude, and frequency Frequency of this section of the impact signal data stream in the secondary instrument, determine whether the characteristic data of this section of the impact signal conforms to the set band-pass filtering and recognition model of crop grains. If it conforms, perform counting processing on the impact signal, perform unit time counting on the identified and counted crop grains, and display and store the values.
[0070] The value of the soft trigger threshold voltage K is determined according to the crop variety, moisture content, and the material and thickness of the sensor sensitive plate 1. After setting the crop variety, moisture content, and the material and thickness of the sensor sensitive plate 1 before the harvesting operation, the value of K can be automatically called in the software.
[0071] The width W of the waveform sampling points is obtained by automatically calling the formula for calculation after setting the harvested crop variety, moisture content, and the material and thickness of the sensitive plate. The calculation method is as follows:
[0072] W = KST
[0073] Where T is the time when the grain impacts the sensor sensitive plate 1, with the unit of second (s);
[0074] S is the sampling rate of the AD conversion module, with the unit of point / second (S / s);
[0075] K is a coefficient, with a value ranging from 0.6 to 1.0, and the usual value is 0.8.
[0076] The calculation method of the root mean square RMS is as follows: Where M n is the nth sampling data, and W is the width of the sampling points of a single impact vibration waveform;
[0077] The calculation method of the number of wave peaks Peak is as follows: When the previous value and the subsequent value are less than the current value, it is recorded as 1 Peak (the amplitude threshold for less than is automatically called according to different crops), and the total number of Peaks in W sampling data points is recorded as the number of wave peaks;
[0078] The calculation methods of the amplitude Amplitude and frequency Frequency are as follows: Perform a fast Fourier transform FFT on W sampling data, and the maximum value is recorded as the amplitude Amplitude, and the fundamental frequency is recorded as the frequency Frequency.
[0079] The band-pass filtering and recognition model of the grain impact signal is constructed after learning from the training data set:
[0080] Before the harvesting operation, a training data set is collected. The training data set contains a large amount of material impact characteristic data, and the characteristics include but are not limited to the root mean square RMS, the number of wave peaks Peak, the amplitude Amplitude, and the fundamental frequency Frequency. The grain and impurity categories Class are manually identified to form a training data set, and a model is constructed based on the data mining algorithm of the machine learning training data set. The data mining algorithms include supervised algorithms such as decision trees, random forests, or neural networks.
[0081] According to this embodiment, preferably, Figure 5Shown are but not limited to rice grain feature extraction and classification. Kernel represents grains, and MOG represents impurities (measured values). The formed training dataset is mined through machine learning algorithms to construct an identification model.
[0082] The data mining algorithm is the C45 decision tree algorithm. Figure 6 Shown is the C45 decision tree identification model for grains and impurities in machine learning.
[0083] The band-pass filtering and identification model for grain impact signals is a classification model established based on waveform feature data such as root mean square (RMS), number of peaks (Peak), amplitude (Amplitude), and frequency (Frequency). The modeling method is a supervised data mining algorithm based on the signal feature test dataset.
[0084] Specifically, for the band-pass filtering and identification model of grain impact signals, the software filtering and identification process is as Figure 7 shown, including the following steps:
[0085] First, set the sampling speed and data reading speed;
[0086] Perform band-pass filtering to allow waveforms within a certain impact frequency range to pass through;
[0087] Set the soft trigger threshold K and waveform width W;
[0088] Extract 4 features of the waveform data: {X 0 , X 1 , X 2 , …, X k , …} T
[0089] X k = {RMS k , Amplitude k , Frequency k , PeaksNumber k}
[0090] Whether the waveform feature values conform to the identification model: If the waveform feature values conform to the identification model, it is identified as a grain, the grain count is incremented by 1, and then it is displayed; if not, it is judged as a non-grain and directly enters the display link.
[0091] If the Stop button is pressed, it stops; otherwise, it returns to judging whether the waveform feature values conform to the identification model.
[0092] The present invention adapts to the impact characteristics of grains of various crop varieties. After selecting the current harvested crop variety through the key in the human-computer interaction module, the software automatically calls the corresponding band-pass filtering and identification model for grain impact signals.
[0093] The FLASH storage space occupied by the grain impact signal band-pass filtering and recognition model is extremely small, and the number of crop varieties that can be written into the software is theoretically much larger than the existing crop varieties in China. Compared with the traditional piezoelectric monitoring system, this device has great technical advantages in the universality of crop varieties.
[0094] The present invention combines a programmable logic gate array FPGA module with a sensor sensitive board 1, and uses the nanosecond-level high-speed characteristics of the programmable logic gate array FPGA module to analyze, process, and store the vibration signal data stream. Preferably, the device can integrate the grain characteristic parameters of various crops such as rice, wheat, corn, and soybeans. During field harvesting, parameters can be conveniently selected to call the grain impact signal band-pass filtering and recognition models of different crop varieties, breaking through the processing speed and variety adaptability limitations of the hardware conditioning circuit in the traditional piezoelectric monitoring method, which is beneficial to accurately, universally, and efficiently online monitor the number of lost grains.
[0095] It should be understood that although this specification is described according to each embodiment, not every embodiment only contains an independent technical solution. This narrative manner of the specification is only for clarity. Those skilled in the art should regard the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.
[0096] The series of detailed descriptions listed above are only specific descriptions of the feasible embodiments of the present invention, and they are not intended to limit the protection scope of the present invention. Any equivalent embodiments or changes made without departing from the technical spirit of the present invention should be included within the protection scope of the present invention.
Claims
1. A device for monitoring the loss of mechanically harvested crops, characterized in that: It includes a secondary instrument and a sensor sensitive plate (1); The secondary instrument includes an AD conversion module and a programmable logic gate array FPGA module; A piezoelectric sheet (2) is provided behind the sensor sensitive plate (1); the sensor sensitive plate (1) is used to collect seed impact vibration waves, the vibration waves are transmitted from the sensor sensitive plate (1) to the piezoelectric sheet (2), the piezoelectric sheet (2) is used to convert the vibration signal into an electrical signal, the electrical signal is transmitted to the AD conversion module of the secondary instrument, the AD conversion module is used to collect the electrical signal of the piezoelectric sheet (2); the programmable logic gate array FPGA module is used to receive the data stream of the AD conversion module and apply the seed impact signal bandpass filtering and recognition model to perform bandpass filtering and recognition processing to eliminate low-frequency machine vibration interference waves and impurity shock waves, and obtain the number of impacted seeds per unit time.
2. The device for monitoring the loss of mechanically harvested crops according to claim 1, characterized in that: The secondary instrument also includes a human-computer interaction module; The human-computer interaction module is used to select the current crop variety, select the channel and acquisition cycle parameters, call the grain impact signal bandpass filter and recognition model, and display the number of impacted grains per unit time.
3. The device for monitoring the loss of mechanically harvested crops according to claim 1, characterized in that: The secondary instrument also includes a UART serial communication module; the UART serial communication module is used to transmit the data of impacting the grains within a unit time to the host computer.
4. The device for monitoring the loss of mechanically harvested crops according to claim 1, characterized in that: The secondary instrument also includes an SD card storage module; the SD card storage module stores data.
5. A method for monitoring the amount of grain loss in mechanical harvesting of crops according to any one of claims 1 to 4, characterized in that: The following steps are involved: The sensor sensitive plate (1) collects the grain impact vibration wave, and the vibration wave is transmitted from the sensor sensitive plate (1) to the piezoelectric plate (2). The piezoelectric plate (2) converts the vibration signal into an electrical signal, and the electrical signal is transmitted to the AD conversion module of the secondary instrument. The AD conversion module collects the electrical signal of the piezoelectric plate (2) and transmits it to the programmable logic gate array FPGA module; the programmable logic gate array FPGA module receives the data stream of the AD conversion module and applies the grain impact signal bandpass filtering and recognition model to perform bandpass filtering and recognition processing to eliminate low-frequency machine vibration interference waves and impurity impact waves, and obtain the number of impacted grains per unit time.
6. The method of the device for monitoring the loss of mechanically harvested crops according to claim 5, characterized in that: The programmable logic gate array FPGA module performs bandpass filtering on the received AD conversion module data, including the following steps: Set the soft trigger threshold voltage range [-K, K], set the waveform sampling point width W, collect W consecutive data points exceeding the threshold range [-K, K], calculate the RMS, Peak, Amplitude and Frequency of the impulse signal data stream in the secondary instrument, and determine whether the characteristic data of the impulse signal conforms to the set crop grain bandpass filtering and recognition model. If so, count the impulse signal and count the identified and counted crop grains per unit time.
7. The method of the device for monitoring the loss of mechanically harvested crops according to claim 6, characterized in that: The waveform sampling point width W value is calculated as follows: W=KST Wherein, T is the time when the seed impacts the sensitive plate (1) of the sensor; S is the sampling rate of the AD conversion module; K is a coefficient, and its value is between 0.6 and 1.
0.
8. The method of the device for monitoring the loss of mechanically harvested crops according to claim 6, characterized in that: The root mean square RMS calculation method is: Among them, M n is the nth sampling data, W is the width of a single impact vibration waveform sampling point; The peak number Peak calculation method is: the previous value and the next value less than the current value is recorded as 1 Peak, and the total number of Peaks in W sampling data points is recorded as the peak number; The amplitude Amplitude and frequency Frequency are calculated by performing fast Fourier transform FFT on W sampled data, recording the maximum value as the amplitude Amplitude and the base frequency as the frequency Frequency.
9. The method of the device for monitoring the loss of mechanically harvested crops according to claim 6, characterized in that: The grain impact signal bandpass filtering and recognition model is constructed after learning from the training data set: A training data set is collected before the harvesting operation. The training data set contains a large amount of material impact feature data. The features include root mean square RMS, peak number Peak, amplitude Amplitude and fundamental frequency Frequency. The grain and impurity categories Class are manually identified to form a training data set. A model is constructed based on the data mining algorithm of the machine learning training data set. The data mining algorithm includes decision tree, random forest or neural network.
10. The method of the device for monitoring the loss of mechanically harvested crops according to claim 9, characterized in that: The grain impact signal bandpass filtering and recognition model, software filtering and recognition process comprises the following steps: First, set the sampling speed and data reading speed; Bandpass filtering allows waveforms within a certain impulse frequency range to pass; Set the soft trigger threshold K and waveform width W; Extract K features of waveform data: {X0,X1,X2,…,X k ,…} T X k ={RMS k ,Amplitude k ,Frequency k ,PeaksNumber k } Whether the waveform characteristic value conforms to the recognition model: If the waveform characteristic value conforms to the recognition model, it is recognized as a grain, the number of grains is increased by 1, and then displayed; if it does not conform, it is judged as not a grain and directly enters the display stage; If the Stop button is pressed, the process stops; otherwise, the process returns to judging whether the waveform feature value conforms to the recognition model.