Seed falling number identification method and system for seed metering flow sensor

Abstract

The invention provides a seed falling number identification method and system for a seed metering flow sensor. The method comprises the following steps: forming a training sequence; calculating a core point Ci of an ith period; determining the Euclidean distance EDxi between each sample point Dx in the training sequence and the core point Ci in the ith core period; determining the core distance CDi of the core point Ci; determining the reachable distance RDxi between the sample point Dx and the core point Ci in the ith core calculation cycle training sequence; selecting a sample point serial number srd which can reach the minimum distance from the core point Ci; and finding out all seed falling measurement sample points and the seed falling number in the sampling measurement period T, and accumulating the seed falling measurement sample points into the seed metering flow in the measurement period T. According to the invention, the seed metering flow detection accuracy can be improved, and the seed metering flow precision and the seeding uniformity are ensured.

Description

Technical Field

[0001] This invention relates to the field of agricultural machinery automation, and in particular to a method and system for identifying the number of seeds dropped using a seed flow sensor. Background Art

[0002] To achieve the desired wheat yield, wheat seeds need to be sown evenly and accurately in the designated fields during wheat sowing operations. Therefore, it is necessary to monitor the seed flow rate at the seed metering outlet of the wheat seeder in real time. If the seed flow rate of a certain seed metering device deviates, it can be promptly fed back to the seed flow rate control system for adjustment, or displayed on the monitoring terminal in the cab, allowing the operator to stop the machine for inspection.

[0003] To obtain accurate seed flow rate, researchers have developed a seed flow rate monitoring sensor. The sensor's internal cavity contains a through-beam photoelectric detection unit. When a seed enters the channel and falls along the channel, it blocks the photoelectric detection unit. At this time, the output of the photoelectric detection unit changes from low to high level. When the seed completely passes through the photoelectric detection unit, the output changes from high to low level again. Therefore, traditional seed drop identification methods assume that one change in output indicates one seed has fallen. This does not take into account the situation where wheat and rice seeds fall continuously in dense rows, or the continuous detection of light in the channel when multiple seeds are planted in corn and soybean replanting conditions. In this case, the traditional seed drop identification method will lead to missed detections, reducing the accuracy of online seed flow rate monitoring.

[0004] If the missed measurement value obtained by the traditional method is used to feed back the seed flow rate, the control system will think that the seed flow rate is less than the theoretical value. Therefore, it will speed up the seed flow rate, resulting in a seed flow rate that is greater than the theoretical value required. This wastes seeds and causes the seed density in the unit plot to be too high, resulting in water and fertilizer competition among seeds and reducing yield. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method and system for identifying the number of seeds dropped using a seed flow sensor, solving the problem of real-time and accurate monitoring of seed flow.

[0006] The present invention achieves the above-mentioned technical objectives through the following technical means.

[0007] A method for identifying the number of seeds dropped using a seed flow sensor includes the following steps:

[0008] S0: Set the core distance confirmation bit CEB, and set both the sorting sequence O and the result sequence R to be empty;

[0009] S1: m groups of consecutive seeds with 1, 2, and 3 seeds respectively pass through the photoelectric detection unit of the seed flow sensor. The microcontroller collects the analog signal output from the photoelectric detection unit and converts it into a digital sequence via an ADC function. This digital sequence is then regularized into a DIM-dimensional digital sequence SS. x d The sample point ID number, the DIM-dimensional digital sequence, and the actual number of seeds landed are combined into a structure to form m sample points. The digital signal SS of the number of seeds landed within the measurement period T is then used to calculate the number of seeds landed. d The sample point ID number x and the number of occurrences are concatenated in the form of a structure after the m sample points to form the training sequence D.

[0010] S2: Based on the ID number srd of the minimum reachable distance sample point output in step S8, extract the (i-1)th core point from the training sequence D and calculate the minimum reachable distance sample point D within the calculation period. srd It is regarded as the core point C of the i-th cycle. i The sample point D will be used as the core point. srd Remove the sample point IDs srd and reachability distances RD from the training sequence D and the sorted sequence O, and add them as key-value pairs to the result sequence R;

[0011] S3: Check if there are any remaining training sample points in the training sequence D. If there are, proceed to step S4; if there are no sample points in the training sequence D, proceed to step S9.

[0012] S4: Determine the sample points D in the training sequence D. x With the core point C of the i-th core cycle i European distance ED x i The sample point index x and the Euclidean distance ED are calculated. x i Place them into an increasing sequence E in the form of key-value pairs;

[0013] S5: According to each sample point D in the training sequence x With core point C i The Euclidean distance is calculated by comparing the sample point ID x with the Euclidean distance ED. x i The key-value pairs are sorted in ascending order, and the increasing sequence E is updated according to the ascending order.

[0014] S6: Locate the sample point of the core distance confirmation bit CEB in the increasing sequence E, and compare the sample point of the core distance confirmation bit CEB in the increasing sequence E with the core point C. i Euclidean distance e i CEB As the core point C i Core distance CDi :

[0015]

[0016] In the formula:

[0017] CD i For the i-th core point, calculate the core point C in the cycle. i Core distance;

[0018] e i CEB Calculate the element of the core distance confirmation position CEB in the increasing sequence E during the period of the i-th core point;

[0019] S7: Based on Euclidean distance ED x i And core point C i Core distance CD i Determine the sample point D in the training sequence of the i-th core computation cycle. x With core point C i Reachable distance RD x i Compare the sample point x with the reachability distance RD x i The sample point x is placed into the sorting sequence O as a key-value pair. If the sample point x key-value pair is already in the sorting sequence O, then the reachable distance RD corresponding to the i-th core point is calculated. x i renew;

[0020] S8: Arrange the key-value pairs of each sample point in the sorted sequence O according to the reachability distance RD x i Sort in ascending order and select the core point C. i The index srd of the sample point with the smallest reachable distance is output to step S2;

[0021] S9: If all sample points in the training sequence have been deleted, perform a forward first-order difference operation on the reachability distances of adjacent elements in the result sequence R. The resulting first-order difference operation forms a difference sequence Y. Find the element index k1 with the largest absolute difference value and the element index k2 with the second largest absolute difference value in the difference sequence Y. The relatively smaller value k2 is the index segmentation threshold TL for the number of occurrences equal to 1 and 2. The relatively larger value k1 is the index segmentation threshold TH for the number of occurrences equal to 2 and 3.

[0022] S10: In the result sequence R, the sample points with an index less than TL are considered as single-grain droplets, the sample points with an index greater than TL and less than TH are considered as double-grain droplets, and the sample points with an index greater than TH are considered as triple-grain droplets.

[0023] S11: Traverse the result sequence R, find all seeding measurement sample points and their seeding counts in this sampling measurement period T, and accumulate them as the seeding flow rate within the measurement period T.

[0024] Furthermore, according to the Euclidean distance ED x i And core point C i Core distance CD i Determine the sample point D in the training sequence of the i-th core computation cycle. x With core point C i Reachable distance RD x i Specifically:

[0025] D of each training sample point x To the core point C i European distance ED x i With core point C i Core distance CD i The larger of the two values ​​is used as the sample point D. x To the core point C i Reachable Distance Reference Value RRD x i Compare sample points D in the training sequence x For core point C i Reachable Distance Reference Value RRD x i With sample point D in the training sequence x For the (i-1)th core point C i-1 Reachable distance RD x i-1 The smaller of the two values ​​is used as the sample point D in the training sequence of the i-th core calculation cycle. x With core point C i Reachable distance RD x i , can be represented as:

[0026]

[0027]

[0028] In the formula:

[0029] RRD x i For the training sample point with index x, the distance from the core point C to the i-th core point is... i Reference value for reachable distance;

[0030] ED x i For the x-th training sample point Dx To the core point C i The Euclidean distance;

[0031] CD i Core point C i Core distance;

[0032] RD x i For sample point D with ID number x x To the i-th core point C i The reachable distance;

[0033] RD x i-1 For sample point D with ID number x x To the (i-1)th core point C i-1 The reachable distance.

[0034] Furthermore, determine the sample points D in the training sequence D. x With the core point C of the i-th core cycle i European distance ED x i The formula is as follows:

[0035]

[0036] In the formula:

[0037] ED x i For training sample point D with index x x To the core point C i The Euclidean distance;

[0038] j is the coordinate dimension index, and its value is... DIM represents the spatial dimension of a digital signal.

[0039] x j x For training sample point D with ID number x x The j-th dimension coordinate of the number of seeds;

[0040] x j ci Core point C i The j-th dimension coordinate of the number of seeds.

[0041] Furthermore, the training sequence D is represented as:

[0042]

[0043] In the formula:

[0044] D is the training sequence;

[0045] x is the sample point ID number. Z is an integer, m is the total number of sample points in the initial training sample library when the number of falles is 1, 2 or 3, and n is the number of sample points measured in the measurement period T.

[0046] SS x d Let x be the seed digital signal with ID number x, and let x be a DIM-dimensional vector.

[0047] N x This represents the number of possible landings for the sample point with ID number x.

[0048] Furthermore, the core point C of the i-th cycle i Represented as:

[0049]

[0050] In the formula:

[0051] C i The core point of the cycle is calculated for the i-th core point;

[0052] i is the core point calculation cycle number. Z is an integer, m is the total number of sample points in the initial training sample library when the number of falles is 1, 2 or 3, and n is the number of sample points measured in the measurement period T.

[0053] srd represents the Cth i-1 The ID number of the sample point with the smallest reachable distance output by the reachable distance solution module within the core point calculation cycle;

[0054] D srd For the Cth i-1 The sample point in the training sequence corresponding to the ID number srd of the minimum reachable sample point output by the reachable distance solution module within the core point calculation cycle;

[0055] rand represents a random number between 0 and m+n-1;

[0056] D rand These are the sample points corresponding to the random number rand;

[0057] When i=0, C0 does not need to be removed from the sorting sequence O during the C0 calculation cycle.

[0058] Furthermore, step S9 is specifically calculated as follows:

[0059]

[0060]

[0061]

[0062]

[0063]

[0064]

[0065] In the formula:

[0066] This is the forward first-order difference between the k-th element and the (k-1)-th element in the resulting sequence R;

[0067] r k This refers to the k-th element in the resulting sequence R;

[0068] r k-1 This refers to the (k-1)th element in the resulting sequence R;

[0069] k is the element index of the resulting sequence R. Z is an integer;

[0070] Y max The maximum absolute value of each element in the difference sequence Y;

[0071] k1 is the index k of the element with the largest absolute value in the difference sequence Y;

[0072] k2 is the index k of the second largest element in the difference sequence Y;

[0073] TL is the index threshold for separating the number of occurrences of a species into 1 and 2 in the resulting sequence R;

[0074] TH is the index threshold for separating the number of occurrences of 2 and 3 in the resulting sequence R.

[0075] A system for identifying the number of seeds dropped using a seeding flow sensor includes a seeding signal acquisition module, an initial training sample library, a training sequence generation module, a state determination module, a Euclidean distance calculation module, an Euclidean distance ascending order sorting module, a core distance calculation module, a reachable distance calculation module, a reachable distance ascending order sorting module, a segmentation threshold calculation module, a seeding number determination module, and a seeding flow accumulation module.

[0076] The initial training sample library contains a total of m initial training sample points. Each initial training sample point is stored in the form of a structure, which includes a sample point ID number, a seeding digital signal, and a seeding number.

[0077] The seed-drop signal acquisition module is connected to the microprocessor input port and the training sequence generation module, and is used to acquire the analog voltage signal SS output by the photoelectric detection unit. aThe analog signal of seeding is converted into a digital signal SS. d ;

[0078] The training sequence generation module is connected to the core point processing module. The seeding signal acquisition module and the initial training sample library serve as the input terminals of the training sequence generation module. The training sequence generation module is used to form the training sequence D.

[0079] The core point processing module is connected to the reachability distance ascending sorting module. The reachability distance ascending sorting module inputs the ID number srd of the sample point with the smallest reachability distance into the core point processing module. The core point processing module outputs the core point C of the i-th period based on the input training sequence D and srd. i The core point C of the i-th cycle i The input state determination module; the core point processing module deletes the sample points that serve as core points from the training sequence D and the sorting sequence O, and adds the sample point ID number srd and reachability distance RD as key-value pairs to the result sequence R, while inputting the result sequence R into the segmentation threshold solution module;

[0080] The state determination module is used to detect whether there are any remaining training sample points in the training sequence D. If so, the core point C of the i-th core calculation cycle is set. i Output to the Euclidean distance calculation module; if there are no sample points in the training sequence D, send the execution instruction to the segmentation threshold calculation module;

[0081] The Euclidean distance calculation module calculates the core point C of the i-th core calculation cycle based on the input. i Find the sample points D with ID number x in the training sequence D. x With the core point C of the i-th core cycle i European distance ED x i And the Euclidean distance ED of each sample point ID number x is calculated. x i The Euclidean distance solution module inputs the increasing sequence E into the Euclidean distance ascending order sorting module, which then places the sequence into an increasing sequence E in the form of key-value pairs.

[0082] The Euclidean distance ascending sorting module sorts samples D in the training sequence according to their respective sample points. x With core point C i The Euclidean distance is calculated by comparing the sample point ID x with the Euclidean distance ED. x i The key-value pairs are sorted in ascending order, and the increasing sequence E is updated according to the ascending order; the Euclidean distance ascending order sorting module inputs the increasing sequence E in ascending order into the core distance solving module;

[0083] The core distance calculation module is used to find the sample point of the core distance confirmation position CEB in the increasing sequence E, and to compare the sample point of the core distance confirmation position CEB in the increasing sequence E with the core point C. i Euclidean distance e i CEB As the core point C i Core distance CD i The core distance solving module will calculate CD. i Input the reachable distance calculation module;

[0084] The reachability distance calculation module is used to obtain the sample point D in the training sequence of the i-th core computing cycle. x With core point C i Reachable distance RD x i Compare the sample point x with the reachability distance RD x i The reachability distance calculation module updates the sorted sequence O in the form of key-value pairs; the reachability distance calculation module inputs the sorted sequence O into the reachability distance ascending order sorting module.

[0085] The ascending order sorting module based on reachability distance is used to sort the key-value pairs of each sample point in the sorted sequence O according to the reachability distance RD. x i Sort in ascending order and select the core point C. i The index srd of the sample point with the smallest reachable distance is output to the core point processing module;

[0086] The segmentation threshold solving module is used to obtain the sequence segmentation threshold TL for the number of falling species equal to 1 and the number of falling species equal to 2, and the sequence segmentation threshold TH for the number of falling species equal to 2 and the number of falling species equal to 3; the segmentation threshold solving module inputs TL and TH into the number of falling species determining module;

[0087] The seed count determination module is used to treat sample points with indices less than TL in the result sequence R as single-seed drops, sample points with indices greater than TL and less than TH as double-seed drops, and sample points with indices greater than TH as triple-seed drops. The seed count determination module will confirm the result N. x Input the seeding flow accumulation module;

[0088] The seed flow accumulation module calculates the seed flow rate within the measurement period T based on the average thousand-seed weight and all seed-falling measurement sample points and their seed-falling counts within the measurement period T.

[0089] The beneficial effects of the present invention are:

[0090] 1. The seed count identification method for seed flow sensors described in this invention addresses the different seed counts in single-seed, double-seed, and triple-seed seeding scenarios. It treats each seed count signal as a data point in a high-dimensional space. The seed count signals obtained within the measurement period are sorted with those in the pre-training sample library according to their reachability. If the seed counts are the same, the reachability distance is relatively short; if the seed counts are different, the reachability distance of the seed count signal points is relatively far. Therefore, three types of spatial points with similar reachability distances are identified. Since the seed counts corresponding to the pre-training sample points in these three types of spatial points are pre-defined, the seed count corresponding to the measured seed count signal can be determined.

[0091] 2. The seed count identification method for seed metering flow sensor of the present invention improves the ability to distinguish continuous falling of more than one seed compared with traditional photoelectric sensors. It can detect reseeding problems of precision-sown crops such as soybeans and corn in a timely manner and prompt the farm machinery operator to adjust the seed metering device. It can also be used in high-flow-rate row sowing occasions such as wheat and rice sowing to improve the accuracy of seed metering flow detection, ensure seed metering flow accuracy and sowing uniformity, and ensure that the yield is increased while reducing the amount of seed used. Attached Figure Description

[0092] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. The drawings described below are some embodiments of the present invention. For those skilled in the art, it is obvious that other drawings can be obtained from these drawings without creative effort.

[0093] Figure 1 This is a flowchart of the seed count identification method for a seed flow sensor according to the present invention.

[0094] Figure 2 This is a system block diagram of the seed count identification method for a seed flow sensor according to the present invention. Detailed Implementation

[0095] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein 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 intended to explain the present invention, and should not be construed as limiting the present invention.

[0096] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "axial," "radial," "vertical," "horizontal," "inner," and "outer," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0097] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0098] like Figure 1 As shown, the seed count identification method for a seed flow sensor according to the present invention includes the following steps:

[0099] Initialization steps: Set the core distance confirmation bit CEB, and set both the sorting sequence O and the result sequence R to empty. Since the seed outlet space of the seed metering device is limited, it was found that generally only a maximum of 3 seeds will continuously pass through the photoelectric detection of the seed flow sensor. Therefore, 1, 2, and 3 consecutive seeds passing through the seed flow sensor are used as the data collection objects of this invention.

[0100] Step 1: m groups of consecutive seeds with 1, 2, and 3 seeds respectively are passed through the photoelectric detection unit of the seed flow sensor. The microcontroller collects the analog signal output from the photoelectric detection unit and converts it into a digital sequence using an ADC function. This digital sequence is then regularized into a DIM-dimensional digital sequence SS. x d The sample point ID number, the DIM-dimensional digital sequence, and the actual number of seeds landed are combined into a structure to form m sample points. The digital signal SS of the number of seeds landed within the measurement period T is then used to calculate the number of seeds landed. dThe sample point ID number x and the number of occurrences (initialized to 0) are concatenated as a structure after the m sample points to form the training sequence D.

[0101]

[0102] In the formula:

[0103] D is the training sequence;

[0104] x is the sample point ID number. Z is an integer, m is the total number of sample points in the initial training sample library when the number of falles is 1, 2 or 3, and n is the number of sample points measured in the measurement period T.

[0105] SS x d Let x be the seed digital signal with ID number x, and let x be a DIM-dimensional vector.

[0106] N x This represents the number of possible landings for the sample point with ID number x.

[0107] Step 2: Based on the ID number srd of the minimum reachable sample point output in Step 8, extract the (i-1)th core point from the training sequence D and calculate the minimum reachable sample point D within the calculation period. srd It is regarded as the core point C of the i-th cycle. i The sample point D will be used as the core point. srd Remove the sample point ID srd and reachability distance RD from the training sequence D and the sorting sequence O, and add them as key-value pairs to the result sequence R; when i=0, when the core point C0 is selected for the first time, a sample point in the training sequence D is randomly selected, and C0 does not need to be removed from the sorting sequence O during the C0 calculation cycle.

[0108] Core point C of the i-th cycle i , can be represented as:

[0109]

[0110] In the formula:

[0111] C i The core point of the cycle is calculated for the i-th core point;

[0112] i is the core point calculation cycle number. Z is an integer, m is the total number of sample points in the initial training sample library when the number of falles is 1, 2 or 3, and n is the number of sample points measured in the measurement period T.

[0113] srd represents the Cth i-1 The ID number of the sample point with the smallest reachable distance output by the reachable distance solution module within the core point calculation cycle;

[0114] D srd For the Cth i-1 The sample point in the training sequence corresponding to the ID number srd of the minimum reachable sample point output by the reachable distance solution module within the core point calculation cycle;

[0115] rand represents a random number between 0 and m+n-1;

[0116] D rand These are the sample points corresponding to the random number rand;

[0117] Step 3: Check if there are any remaining training sample points in the training sequence D. If so, then add the core point C of the i-th core calculation cycle. i Output to the Euclidean distance calculation module, i.e., step 4; if there are no sample points in the training sequence D, then proceed to the segmentation threshold calculation module, i.e., step 9.

[0118] Step 4: Solve for each sample point D in the training sequence D. x With the core point C of the i-th core cycle i European distance ED x i The sample point index x and the Euclidean distance ED are calculated. x i Place them into an increasing sequence E in the form of key-value pairs;

[0119]

[0120] In the formula:

[0121] ED x i For training sample point D with index x x To the core point C i The Euclidean distance;

[0122] j is the coordinate dimension index, and its value is... DIM is the spatial dimension of the digital signal, which is set to 7 in this patent, but it can also be selected as 9 or 11.

[0123] x j x For training sample point D with ID number x x The j-th dimension coordinate of the number of seeds;

[0124] x j ci Core point C i The j-th dimension coordinate of the number of seeds;

[0125] Step 5: According to each sample point D in the training sequence x With core point Ci The Euclidean distance is calculated by comparing the sample point ID x with the Euclidean distance ED. x i The key-value pairs are sorted in ascending order, and the increasing sequence E is updated according to the ascending order.

[0126] Step 6: Locate the sample point of the core distance confirmation bit CEB in the increasing sequence E, and compare the sample point of the core distance confirmation bit CEB in the increasing sequence E with the core point C. i Euclidean distance e i CEB As the core point C i Core distance CD i :

[0127]

[0128] In the formula:

[0129] CD i For the i-th core point, calculate the core point C in the cycle. i Core distance;

[0130] e i CEB Calculate the element of the core distance confirmation position CEB in the increasing sequence E during the period of the i-th core point;

[0131] Step 7: Transfer each training sample point D x To the core point C i European distance ED x i With core point C i Core distance CD i The larger of the two values ​​is used as the sample point D. x To the core point C i Reachable Distance Reference Value RRD x i Compare sample points D in the training sequence x For core point C i Reachable Distance Reference Value RRD x i With sample point D in the training sequence x For the (i-1)th core point C i-1 Reachable distance RD x i-1 The smaller of the two values ​​is used as the sample point D in the training sequence of the i-th core calculation cycle. x With core point C i Reachable distance RD x i Compare the sample point x with the reachability distance RD x iThe sample point x is placed into the sorting sequence O as a key-value pair. If the sample point x key-value pair is already in the sorting sequence O, then the reachable distance RD corresponding to the i-th core point is calculated. x i renew:

[0132]

[0133]

[0134] In the formula:

[0135] RRD x i For the training sample point with index x, the distance from the core point C to the i-th core point is... i Reference value for reachable distance;

[0136] ED x i For the x-th training sample point D x To the core point C i The Euclidean distance;

[0137] CD i Core point C i Core distance;

[0138] RD x i For sample point D with ID number x x To the i-th core point C i The reachable distance;

[0139] RD x i-1 For sample point D with ID number x x To the (i-1)th core point C i-1 The reachable distance;

[0140] Step 8: Arrange the key-value pairs of each sample point in the sorted sequence O according to the reachability distance RD x i Sort in ascending order and select the core point C. i Output the index srd of the sample point with the smallest reachable distance to step 2;

[0141]

[0142] In the formula:

[0143] srd represents the core point C in the i-th core calculation cycle. i The ID number of the sample point with the smallest reachable distance;

[0144] o0 idTo sort out the ID numbers of the elements with index 0 in sequence O, we can find the ID number of the element with the smallest distance.

[0145] Step 9: If all sample points in the training sequence have been deleted, perform a forward first-order difference operation on the reachability distances between adjacent elements in the result sequence R. The resulting first-order difference operation forms a difference sequence Y. Find the element index k1 with the largest absolute difference value and the element index k2 with the second largest absolute difference value in the difference sequence Y. The relatively smaller value k2 is the index segmentation threshold TL for the number of occurrences equal to 1 and 2. The relatively larger value k1 is the index segmentation threshold TH for the number of occurrences equal to 2 and 3.

[0146]

[0147]

[0148]

[0149]

[0150]

[0151]

[0152] In the formula:

[0153] This is the forward first-order difference between the k-th element and the (k-1)-th element in the resulting sequence R;

[0154] r k This refers to the k-th element in the resulting sequence R;

[0155] r k-1 This refers to the (k-1)th element in the resulting sequence R;

[0156] k is the element index of the resulting sequence R. Z is an integer;

[0157] Y max The maximum absolute value of each element in the difference sequence Y;

[0158] k1 is the index k of the element with the largest absolute value in the difference sequence Y;

[0159] k2 is the index k of the second largest element in the difference sequence Y;

[0160] TL is the index threshold for separating the number of occurrences of a species into 1 and 2 in the resulting sequence R;

[0161] TH is the threshold for separating the sequence number of occurrences in the result sequence R from 2 to 3.

[0162] Step 10: In the result sequence R, the sample points with an index less than TL are considered as single-seed landings, the sample points with an index greater than TL and less than TH are considered as double-seed landings, and the sample points with an index greater than TH are considered as triple-seed landings.

[0163]

[0164] In the formula:

[0165] N x This represents the number of possible landings corresponding to training sample point ID;

[0166] RD x The reachability distance corresponding to the training sample point with ID number;

[0167] x is the sample point ID number;

[0168] TL is the index threshold for separating the number of occurrences of a species into 1 and 2 in the resulting sequence R;

[0169] TH is the threshold for separating the sequence number of occurrences in the result sequence R from 2 to 3.

[0170] Step 11: Traverse the result sequence R, find all seeding measurement sample points and their seeding counts in this sampling measurement period T, and accumulate them as the seeding flow rate within the measurement period T;

[0171]

[0172] In the formula:

[0173] SFR T To measure the seeding flow rate within the measurement period T;

[0174] M is the weight of 1000 seeds, which is set as a constant before the measurement begins;

[0175] N x The number of possible landings corresponding to the x-th training sample point;

[0176] x is the sample point ID number.

[0177] like Figure 2 As shown, the self-evolving seed drop count identification system of the present invention includes a seed drop signal acquisition module, an initial training sample library, a training sequence generation module, a state determination module, a Euclidean distance calculation module, an Euclidean distance ascending order sorting module, a core distance calculation module, a reachable distance calculation module, a reachable distance ascending order sorting module, a segmentation threshold calculation module, a seed drop count determination module, and a seeding flow accumulation module.

[0178] The initial training sample library contains a total of m initial training sample points. Each initial training sample point is stored in the form of a structure, which includes a sample point ID number, a seeding digital signal, and a seeding number.

[0179] The seed-drop signal acquisition module is connected to the microprocessor input port and the training sequence generation module, and is used to acquire the analog voltage signal SS output by the photoelectric detection unit. a The analog signal of seeding is converted into a digital signal SS. d ;

[0180] The training sequence generation module is connected to the core point processing module. The seeding signal acquisition module and the initial training sample library serve as the input terminals of the training sequence generation module, which processes the seeding digital signal SS. d Converting the data into DIM-dimensional digital signals, and converting the ID number x of each measurement sample point and the seed digital signal SS obtained within the measurement period T. x d The number of occurrences Nx (initialized to 0) is represented as a structure and concatenated after the initial training sample points to form the training sequence D.

[0181]

[0182] In the formula:

[0183] D is the training sequence;

[0184] x is the sample point ID number. Z is an integer, m is the total number of sample points in the initial training sample library when the number of falles is 1, 2 or 3, and n is the number of sample points measured in the measurement period T.

[0185] SS x d Let x be the seed digital signal with ID number x, and let x be a DIM-dimensional vector.

[0186] N x This represents the number of possible landings for the sample point with ID number x.

[0187] The core point processing module is connected to the reachability distance ascending sorting module. The reachability distance ascending sorting module inputs the ID number srd of the sample point with the smallest reachability distance into the core point processing module. The core point processing module outputs the core point C of the i-th period based on the input training sequence D and srd. i The core point C of the i-th cycle i The input state determination module; the core point processing module deletes the sample points that serve as core points from the training sequence D and the sorting sequence O, and adds the sample point ID number srd and reachability distance RD as key-value pairs to the result sequence R. At the same time, the result sequence R is input to the segmentation threshold solution module.

[0188] The state determination module is used to detect whether there are any remaining training sample points in the training sequence D. If so, the core point C of the i-th core calculation cycle is set. i Output to the Euclidean distance calculation module; if there are no sample points in the training sequence D, then proceed to the segmentation threshold calculation module.

[0189] The Euclidean distance calculation module calculates the core point C of the i-th core calculation cycle based on the input. i Find the sample points D with ID number x in the training sequence D. x With the core point C of the i-th core cycle i European distance ED x i And the Euclidean distance ED of each sample point ID number x is calculated. x i The Euclidean distance solution module inputs the increasing sequence E into the Euclidean distance ascending order sorting module, which then places the sequence into an increasing sequence E in the form of key-value pairs.

[0190] The Euclidean distance ascending sorting module sorts samples D in the training sequence according to their respective sample points. x With core point C i The Euclidean distance is calculated by comparing the sample point ID x with the Euclidean distance ED. x i The key-value pairs are sorted in ascending order, and the increasing sequence E is updated according to the ascending order; the Euclidean distance ascending order sorting module inputs the increasing sequence E in ascending order into the core distance solving module;

[0191] The core distance calculation module is used to find the sample point of the core distance confirmation position CEB in the increasing sequence E, and to compare the sample point of the core distance confirmation position CEB in the increasing sequence E with the core point C. i Euclidean distance e i CEB As the core point C i Core distance CD i The core distance solving module will calculate CD. i Input the reachable distance calculation module;

[0192] The reachability distance calculation module is used to calculate the reachability distance of each training sample point D. x To the core point C i European distance ED x i With core point C i Core distance CD i The larger of the two values ​​is used as the sample point D. x To the core point C i Reachable Distance Reference Value RRD x iCompare sample points D in the training sequence x For core point C i Reachable Distance Reference Value RRD x i With sample point D in the training sequence x For the (i-1)th core point C i-1 Reachable distance RD x i-1 The smaller of the two values ​​is used as the sample point D in the training sequence of the i-th core calculation cycle. x With core point C i Reachable distance RD x i Compare the sample point x with the reachability distance RD x i The reachability distance calculation module updates the sorted sequence O in the form of key-value pairs; the reachability distance calculation module inputs the sorted sequence O into the reachability distance ascending order sorting module.

[0193] The ascending order sorting module based on reachability distance is used to sort the key-value pairs of each sample point in the sorted sequence O according to the reachability distance RD. x i Sort in ascending order and select the core point C. i The index srd of the sample point with the smallest reachable distance is output to the core point processing module;

[0194] The segmentation threshold solving module is used to obtain the sequence segmentation threshold TL for the number of falling species equal to 1 and the number of falling species equal to 2, and the sequence segmentation threshold TH for the number of falling species equal to 2 and the number of falling species equal to 3; the segmentation threshold solving module inputs TL and TH into the number of falling species determining module;

[0195] The seed count determination module is used to treat sample points with indices less than TL in the result sequence R as single-seed drops, sample points with indices greater than TL and less than TH as double-seed drops, and sample points with indices greater than TH as triple-seed drops. The seed count determination module will confirm the result N. x Input the seeding flow accumulation module;

[0196] The seed flow accumulation module calculates the seed flow rate within the measurement period T based on the average thousand-seed weight and all seed-falling measurement sample points and their seed-falling counts within the measurement period T.

[0197] It should be understood that although this specification is described according to various embodiments, not every embodiment contains only one independent technical solution. This way of describing the specification is only for clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

[0198] The detailed descriptions listed above are merely specific illustrations of feasible embodiments of the present invention and are not intended to limit the scope of protection of the present invention. All equivalent embodiments or modifications made without departing from the spirit of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for identifying the number of seeds dropped using a seed flow sensor, characterized in that, Includes the following steps: S0: Set the core distance confirmation bit CEB, and set both the sorting sequence O and the result sequence R to be empty; S1: m groups of consecutive seeds with 1, 2, and 3 seeds respectively pass through the photoelectric detection unit of the seed flow sensor. The microcontroller collects the analog signal output from the photoelectric detection unit and converts it into a digital sequence via an ADC function. This digital sequence is then regularized into a DIM-dimensional digital sequence SS. x d The sample point ID number, the DIM-dimensional digital sequence, and the actual number of seeds landed are combined into a structure to form m sample points. The digital signal SS of the number of seeds landed within the measurement period T is then used to calculate the number of seeds landed. d The sample point ID number x and the number of occurrences are concatenated in the form of a structure after the m sample points to form the training sequence D. S2: Based on the ID number srd of the minimum reachable distance sample point output in step S8, extract the (i-1)th core point from the training sequence D and calculate the minimum reachable distance sample point D within the calculation period. srd It is regarded as the core point C of the i-th cycle. i The sample point D will be used as the core point. srd Remove the sample point IDs srd and reachability distances RD from the training sequence D and the sorted sequence O, and add them as key-value pairs to the result sequence R; S3: Check if there are any remaining training sample points in the training sequence D. If there are, proceed to step S4; if there are no sample points in the training sequence D, proceed to step S9. S4: Determine the sample points D in the training sequence D. x With the core point C of the i-th core cycle i European distance ED x i The sample point index x and the Euclidean distance ED are calculated. x i Place them into an increasing sequence E in the form of key-value pairs; S5: According to each sample point D in the training sequence x With core point C i The Euclidean distance is calculated by comparing the sample point ID x with the Euclidean distance ED. x i The key-value pairs are sorted in ascending order, and the increasing sequence E is updated according to the ascending order. S6: Locate the sample point of the core distance confirmation bit CEB in the increasing sequence E, and compare the sample point of the core distance confirmation bit CEB in the increasing sequence E with the core point C. i Euclidean distance e i CEB As the core point C i Core distance CD i : ； In the formula: CD i For the i-th core point, calculate the core point C in the cycle. i Core distance; e i CEB Calculate the element of the core distance confirmation position CEB in the increasing sequence E during the period of the i-th core point; S7: Based on Euclidean distance ED x i And core point C i Core distance CD i Determine the sample point D in the training sequence of the i-th core computation cycle. x With core point C i Reachable distance RD x i Compare the sample point x with the reachability distance RD x i The sample point x is placed into the sorting sequence O as a key-value pair. If the sample point x key-value pair is already in the sorting sequence O, then the reachable distance RD corresponding to the i-th core point is calculated. x i renew; S8: Arrange the key-value pairs of each sample point in the sorted sequence O according to the reachability distance RD x i Sort in ascending order and select the core point C. i The index srd of the sample point with the smallest reachable distance is output to step S2; S9: If all sample points in the training sequence have been deleted, perform a forward first-order difference operation on the reachability distances of adjacent elements in the result sequence R. The resulting first-order difference operation forms a difference sequence Y. Find the element index k1 with the largest absolute difference value and the element index k2 with the second largest absolute difference value in the difference sequence Y. The relatively smaller value k2 is the index segmentation threshold TL for the number of occurrences equal to 1 and 2. The relatively larger value k1 is the index segmentation threshold TH for the number of occurrences equal to 2 and 3. S10: In the result sequence R, the sample points with an index less than TL are considered as single-grain droplets, the sample points with an index greater than TL and less than TH are considered as double-grain droplets, and the sample points with an index greater than TH are considered as triple-grain droplets. S11: Traverse the result sequence R, find all seeding measurement sample points and their seeding counts in this sampling measurement period T, and accumulate them as the seeding flow rate within the measurement period T.

2. The method for identifying the number of seeds dropped using a seed flow sensor according to claim 1, characterized in that, According to Euclidean distance ED x i And core point C i Core distance CD i Determine the sample point D in the training sequence of the i-th core computation cycle. x With core point C i Reachable distance RD x i Specifically: D of each training sample point x To the core point C i European distance ED x i With core point C i Core distance CD i The larger of the two values ​​is used as the sample point D. x To the core point C i Reachable Distance Reference Value RRD x i ; Compare sample points D in the training sequence x For core point C i Reachable Distance Reference Value RRD x i With sample point D in the training sequence x For the (i-1)th core point C i-1 Reachable distance RD x i-1 The smaller of the two values ​​is used as the sample point D in the training sequence of the i-th core calculation cycle. x With core point C i Reachable distance RD x i , can be represented as: ， ， In the formula: RRD x i For the training sample point with index x, the distance from the core point C to the i-th core point is... i Reference value for reachable distance; ED x i For the x-th training sample point D x To the core point C i The Euclidean distance; CD i Core point C i Core distance; RD x i For sample point D with ID number x x To the i-th core point C i The reachable distance; RD x i-1 For sample point D with ID number x x To the (i-1)th core point C i-1 The reachable distance.

3. The method for identifying the number of seeds dropped using a seed flow sensor according to claim 1, characterized in that, Determine each sample point D in the training sequence D. x With the core point C of the i-th core cycle i European distance ED x i The formula is as follows: ， In the formula: ED x i For training sample point D with index x x To the core point C i The Euclidean distance; j is the coordinate dimension index, and its value is... DIM represents the spatial dimension of a digital signal. x j x For training sample point D with ID number x x The j-th dimension coordinate of the number of seeds; x j ci Core point C i The j-th dimension coordinate of the number of seeds.

4. The method for identifying the number of seeds dropped using a seed flow sensor according to claim 1, characterized in that, The training sequence D is represented as: ， In the formula: D is the training sequence; x is the sample point ID number. Z is an integer, m is the total number of sample points in the initial training sample library when the number of falles is 1, 2 or 3, and n is the number of sample points measured in the measurement period T. SS x d Let x be the seed digital signal with ID number x, and let x be a DIM-dimensional vector. N x This represents the number of possible landings for the sample point with ID number x.

5. The method for identifying the number of seeds dropped using a seed flow sensor according to claim 1, characterized in that, Core point C of the i-th cycle i Represented as: ， In the formula: C i The core point of the cycle is calculated for the i-th core point; i is the core point calculation cycle number. Z is an integer, m is the total number of sample points in the initial training sample library when the number of falles is 1, 2 or 3, and n is the number of sample points measured in the measurement period T. srd represents the Cth i-1 The ID number of the sample point with the smallest reachable distance output by the reachable distance solution module within the core point calculation cycle; D srd For the Cth i-1 The sample point in the training sequence corresponding to the ID number srd of the minimum reachable sample point output by the reachable distance solution module within the core point calculation cycle; rand represents a random number between 0 and m+n-1; D rand These are the sample points corresponding to the random number rand; When i=0, C0 does not need to be removed from the sorting sequence O during the C0 calculation cycle.

6. The method for identifying the number of seeds dropped using a seed flow sensor according to claim 1, characterized in that, Step S9 is calculated as follows: ， ， ， ， ， ， In the formula: This is the forward first-order difference between the k-th element and the (k-1)-th element in the resulting sequence R; r k This refers to the k-th element in the resulting sequence R; r k-1 This refers to the (k-1)th element in the resulting sequence R; k is the element index of the resulting sequence R. Z is an integer; Y max The maximum absolute value of each element in the difference sequence Y; k1 is the index k of the element with the largest absolute value in the difference sequence Y; k2 is the index k of the second largest element in the difference sequence Y; TL is the index threshold for separating the number of occurrences of a species into 1 and 2 in the resulting sequence R; TH is the index threshold for separating the number of occurrences of 2 and 3 in the resulting sequence R.

7. A system for identifying the number of seeds dropped using a seed flow sensor according to any one of claims 1-6, characterized in that, It includes a seeding signal acquisition module, an initial training sample library, a training sequence generation module, a state determination module, a Euclidean distance calculation module, an Euclidean distance ascending order sorting module, a core distance calculation module, a reachable distance calculation module, a reachable distance ascending order sorting module, a segmentation threshold calculation module, a seeding number determination module, and a seeding flow accumulation module; The initial training sample library contains a total of m initial training sample points. Each initial training sample point is stored in the form of a structure, which includes a sample point ID number, a seeding digital signal, and a seeding number. The seed-drop signal acquisition module is connected to the microprocessor input port and the training sequence generation module, and is used to acquire the analog voltage signal SS output by the photoelectric detection unit. a The analog signal of seeding is converted into a digital signal SS. d ; The training sequence generation module is connected to the core point processing module. The seeding signal acquisition module and the initial training sample library serve as the input terminals of the training sequence generation module. The training sequence generation module is used to form the training sequence D. The core point processing module is connected to the reachability distance ascending sorting module. The reachability distance ascending sorting module inputs the ID number srd of the sample point with the smallest reachability distance into the core point processing module. The core point processing module outputs the core point C of the i-th period based on the input training sequence D and srd. i The core point C of the i-th cycle i Input status determination module; The core point processing module removes the sample points that serve as core points from the training sequence D and the sorting sequence O, and adds the sample point ID number srd and reachability distance RD as key-value pairs to the result sequence R. At the same time, the result sequence R is input into the segmentation threshold solving module. The state determination module is used to detect whether there are any remaining training sample points in the training sequence D. If so, the core point C of the i-th core calculation cycle is set. i Output to the Euclidean distance calculation module; If there are no sample points in the training sequence D, then send the execution instruction to the segmentation threshold calculation module; The Euclidean distance calculation module calculates the core point C of the i-th core calculation cycle based on the input. i Find the sample points D with ID number x in the training sequence D. x With the core point C of the i-th core cycle i European distance ED x i And the Euclidean distance ED of each sample point ID number x is calculated. x i The Euclidean distance solution module inputs the increasing sequence E into the Euclidean distance ascending order sorting module, which then places the sequence E into the key-value pair in the increasing sequence E. The Euclidean distance ascending sorting module sorts samples D in the training sequence according to their respective sample points. x With core point C i The Euclidean distance is calculated by comparing the sample point ID x with the Euclidean distance ED. x i The key-value pairs are sorted in ascending order, and the increasing sequence E is updated according to the ascending order; the Euclidean distance ascending order sorting module inputs the increasing sequence E in ascending order into the core distance solving module; The core distance calculation module is used to find the sample point of the core distance confirmation position CEB in the increasing sequence E, and to compare the sample point of the core distance confirmation position CEB in the increasing sequence E with the core point C. i Euclidean distance e i CEB As the core point C i Core distance CD i The core distance solving module will calculate CD. i Input the reachable distance calculation module; The reachability distance calculation module is used to obtain the sample point D in the training sequence of the i-th core computing cycle. x With core point C i Reachable distance RD x i Compare the sample point x with the reachability distance RD x i The reachability distance calculation module updates the sorted sequence O in the form of key-value pairs; the reachability distance calculation module inputs the sorted sequence O into the reachability distance ascending order sorting module. The ascending order sorting module based on reachability distance is used to sort the key-value pairs of each sample point in the sorted sequence O according to the reachability distance RD. x i Sort in ascending order and select the core point C. i The index srd of the sample point with the smallest reachable distance is output to the core point processing module; The segmentation threshold solving module is used to obtain the sequence segmentation threshold TL for the number of falling species equal to 1 and the number of falling species equal to 2, and the sequence segmentation threshold TH for the number of falling species equal to 2 and the number of falling species equal to 3; the segmentation threshold solving module inputs TL and TH into the number of falling species determining module; The seed count determination module is used to treat sample points with indices less than TL in the result sequence R as single-seed drops, sample points with indices greater than TL and less than TH as double-seed drops, and sample points with indices greater than TH as triple-seed drops. The seed count determination module will confirm the result N. x Input the seeding flow accumulation module; The seed flow accumulation module calculates the seed flow rate within the measurement period T based on the average thousand-seed weight and all seed-falling measurement sample points and their seed-falling counts within the measurement period T.

Images