Seed metering flow detection system and detection method based on linear array charge-coupled device

Abstract

The invention provides a seed discharge flow detection system and detection method based on a linear array charge-coupled device, and the method comprises the following steps: scanning a falling seed flow through the linear array charge-coupled device, converting an analog signal into a digital signal, and splicing the digital signal to form a seed shadow image; processing the seed shadow image, and marking all seed shadow connected domains; calculating the light volume of each seed shadow connected domain; dividing each light volume into a corresponding fuzzy interval; according to the current light volume fuzzy interval, querying a likelihood probability table and calculating posterior probabilities of different seed falling number predicted values; judging the predicted value corresponding to the maximum posterior probability as the seed falling number of the connected domain; and accumulating the seed falling numbers of all the connected domains to obtain the total seed falling number. Pixel-level monitoring in the seed falling process can be achieved, the seed form can be effectively distinguished according to a seed shadow image, the overlapping state of the seeds is judged according to the pixel value of each pixel point, and seed discharging flow online detection is conducted by integrating the shape, posture and overlapping information of the falling seeds.

Description

Technical Field

[0001] This invention relates to the field of agricultural machinery automation, and in particular to a seed flow rate detection system and method based on a linear array charge-coupled device. Background Technology

[0002] Seeders require a seed flow rate detection system to monitor the seed flow rate in real time, which is used to provide feedback and control the seeding rate of the work area. In addition, if a seed metering device repeatedly re-seeds or misses seeding for a long time, the seed flow rate detection system can issue an alarm, and the agricultural machinery operator can then check and troubleshoot the corresponding seed metering device.

[0003] Traditional seed flow sensors for seeders are mainly classified into photoelectric, piezoelectric, and capacitive types. Photoelectric sensors offer fast response and high sensitivity, enabling non-contact measurement of rapidly falling seeds. However, their mechanism relies on the degree of seed occlusion of emitted light to identify the seed quantity, making them susceptible to changes in ambient light in the field. Furthermore, dust mixed in with the seeds can obscure the optical sensing element, reducing the amount of modulated light received by the receiver and causing detection errors. Piezoelectric sensors operate based on the piezoelectric effect, where the pressure-sensitive material generates an electric charge when impacted by a seed (e.g., by pressure or vibration). Unlike photoelectric sensors, piezoelectric sensors do not rely on visual characteristics to detect the number of fallen seeds. Regardless of seed color or shape, a physical impact on the pressure-sensitive material generates a piezoelectric signal, solving the problem of photoelectric sensors misjudging seeds with complex colors and shapes. However, if two or more seeds simultaneously or consecutively impact the sensor surface, they may only generate a single, combined vibration signal. The circuit cannot recognize this as two independent events, leading to missed counts. Capacitive sensors are not sensitive to the color or shape of seeds. When a seed passes between two capacitor plates, the dielectric constant of the capacitor changes, which can detect whether a seed has fallen.

[0004] However, due to their low resolution, the aforementioned sensors are unable to effectively extract the shape and degree of occlusion of falling seeds, resulting in insufficient detection accuracy for overlapping and irregularly shaped seeds. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a seed flow rate detection system and method based on a linear array charge-coupled device (CCD). By using a CCD as a sensor, pixel-level monitoring of the seed falling process can be achieved. Seed morphology can be effectively distinguished based on seed shadow images, and the overlap state of seeds can be determined by the pixel values ​​of each pixel. The seed flow rate is detected online by comprehensively considering the shape, posture, and overlap information of the falling seeds, thus solving the problem of real-time and accurate monitoring of seed flow rate.

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

[0007] A seed flow detection method based on a linear charge-coupled device includes the following steps:

[0008] The falling seed stream is scanned by a linear array charge-coupled device, the analog signal is converted into a digital signal and stitched together to form a seed image;

[0009] Process the seed shadow image to identify all seed shadow connected components;

[0010] Calculate the optical volume of each species-shadow connected region;

[0011] Each optical volume is divided into a corresponding fuzzy region;

[0012] Based on the current optical volume ambiguity interval, query the likelihood probability table and calculate the posterior probability of different predicted drop numbers;

[0013] The number of possible outcomes corresponding to the maximum posterior probability is determined as the number of possible outcomes in the shadow connected component of that type.

[0014] The total number of possible outcomes is obtained by summing the number of outcomes in all connected components.

[0015] Furthermore, the seed shadow image is processed to identify all connected components of the seed shadow, specifically including the following steps:

[0016] Traverse the image pixels in the scanning order and compare the pixel values ​​with the light intensity threshold;

[0017] The first pixel with a light intensity greater than the light intensity threshold is taken as the germination point and assigned a new label;

[0018] Check the neighboring pixels of the germination point and push the coordinates of neighboring pixels whose pixel values ​​are greater than the light intensity threshold onto the stack;

[0019] Pop the top pixel coordinates from the stack, assign the current label, check its neighboring pixels, and push the coordinates of unlabeled pixels with pixel values ​​greater than the light intensity threshold onto the stack.

[0020] Repeat the process of popping, marking, checking neighbors, and pushing onto the stack until the stack is empty, thus completing the marking of a connected component;

[0021] Continue scanning until the entire image has been processed and all connected components have been identified.

[0022] Furthermore, the neighborhood is a four-neighborhood or an eight-neighborhood.

[0023] Furthermore, the optical volume of each species-shadow connected region is calculated using the following formula:

[0024]

[0025] In the formula,

[0026] V l The light volume of the connected region of the species labeled l;

[0027] I l i (x,y) is the pixel value of the i-th pixel in the shaded connected domain of label l;

[0028] i is the pixel index in the shaded connected region of label l;

[0029] n is the maximum value of the pixel index;

[0030] x is the x-coordinate of the pixel;

[0031] y is the ordinate value of the pixel;

[0032] l represents the label number of the connected domain.

[0033] Furthermore, each optical volume is divided into corresponding fuzzy regions, as follows:

[0034] The light volume of the seed-shadow connected region is determined into one of five fuzzy intervals based on a segmentation threshold. These fuzzy intervals are: minimal, small, medium, large, and maximal.

[0035]

[0036] In the formula:

[0037] FV l The fuzzy interval corresponding to the light volume of the connected region of the shadow of label number l;

[0038] V l The light volume of the connected region of the species labeled l;

[0039] ES represents a fuzzy region where the optical volume is extremely small;

[0040] S represents a fuzzy region with a small light volume;

[0041] M represents the fuzzy region in the optical volume;

[0042] L represents the fuzzy region where the light volume is large;

[0043] EL represents a fuzzy region where the light volume is extremely large;

[0044] T ES The threshold for determining when the light volume is minimal;

[0045] T S The threshold for determining if a light volume is small;

[0046] T M To determine the segmentation threshold in the light volume;

[0047] T L The segmentation threshold for determining whether the light volume is large;

[0048] T EL The segmentation threshold is used to determine when the light volume is extremely large.

[0049] Furthermore, based on the current optical volume ambiguity interval, the likelihood probability table is consulted and the posterior probability of different predicted landing numbers is calculated, specifically:

[0050] Based on the fuzzy interval of optical volume, look up the table to find the likelihood probability P(OV=FV) of the fuzzy interval of optical volume corresponding to label l, given that the predicted number of landing species is j. l |N=j),

[0051] Multiplying the obtained likelihood probability by the prior probability with j as the number of occurrences, we obtain the optical volume ambiguity interval as FV. l Under the condition that the predicted number of species is j, the posterior probability is...

[0052] In the formula:

[0053] N is a random variable representing the number of seeds that fall;

[0054] j is the predicted number of seeds to fall;

[0055] OV is a random variable in the fuzzy interval of the optical volume of the connected domain of the seed shadow;

[0056] FV l The optical volume fuzzy region of the seed-shadow connected region of label l;

[0057] P(N=j|OV=FV l The image volume blur value is FV. l Under the condition that the number of seeds falls is the predicted value j;

[0058] P(OV=FV l |N=j) is the case where the number of seeds is j and the seed shadow volume is FV l The likelihood probability;

[0059] P(N=j) is the prior probability of the number of possible outcomes being j.

[0060] Furthermore, the number of possible outcomes corresponding to the maximum posterior probability is determined as the number of possible outcomes in the shadow connected region of that type, as follows:

[0061] The maximum posterior probability P(N=j|OV=FV) l The predicted number of species corresponding to label l is considered as the species count detection result of the connected component of the species shadow:

[0062]

[0063] In the formula:

[0064] n l The result of detecting the number of species in the connected domain of the species with label number l;

[0065] P(N=j|OV=FV l The image volume blur range is FV. l Under the condition that the predicted number of seeds is j, the posterior probability is given.

[0066] N is a random variable representing the number of seeds that fall;

[0067] j is the predicted number of seeds to fall;

[0068] OV is a random variable in the fuzzy interval of the optical volume of the connected domain of the seed shadow;

[0069] FV l The light volume ambiguity interval of the species-connected domain of label l is represented.

[0070] Furthermore, by summing the number of possible outcomes for all connected components, the total number of possible outcomes is obtained, specifically:

[0071]

[0072] In the formula,

[0073] n' represents the seed count detection result for the entire seed image;

[0074] n l The result of detecting the number of species in the connected component of the shadow domain labeled l;

[0075] a is the maximum value of the shaded connected region;

[0076] l represents the label number of the connected domain.

[0077] A detection system for a detection method includes a linear charge-coupled device, an LED array, and a signal processing system;

[0078] The linear charge-coupled device (CCD) is mounted on one side of the seed flow falling channel of the seed metering device to scan the falling seed flow and convert the optical signal into an analog electrical signal; the LED array is symmetrically arranged with the linear CCCD on the other side of the seed flow falling channel to provide a light source; the signal processing system includes peripheral circuitry and an embedded processor to receive and process the analog electrical signal, the embedded processor including:

[0079] The seed image acquisition module is used to convert the analog signal output by the linear charge-coupled device into a digital signal according to the scanning cycle, and stitch them together through multiple scanning cycles to form a seed image;

[0080] The seed image connected component search module is used to perform binarization processing on the seed image and mark all connected components;

[0081] The light volume calculation module is used to accumulate the pixel values ​​of all pixels in each species' shadow connected region to obtain the light volume of that connected region;

[0082] The optical volume blurring module is used to divide the optical volume of each connected component into the corresponding blur interval according to multiple preset segmentation thresholds.

[0083] The seed drop probability inference module is used to determine the posterior probability of different seed drop numbers under the optical volume ambiguity interval of the current connected domain.

[0084] The number of occurrences determination module is used to select the number of occurrences corresponding to the maximum posterior probability as the detection result of the connected component;

[0085] The seed count accumulation module accumulates the seed count detection results of all seed shadow connected regions in the seed shadow image to obtain the seed count detection result of the entire seed shadow image.

[0086] The beneficial effects of this invention are as follows:

[0087] The seeding flow detection system and method based on linear array charge-coupled devices described in this invention obtain pixel-level images of the seeding process through linear array charge-coupled devices. Compared with traditional photoelectric, piezoelectric, and capacitive sensors, this improves the resolution of the seeding state image. Therefore, the number of seeds in each seed shadow connected region can be distinguished by utilizing the shape of the seed shadow image and the light intensity of each pixel. In terms of the detection method, the pixel value of each pixel is essentially the received light intensity. This invention accumulates the pixel values ​​of each pixel in each connected region, which essentially combines the shape of the connected region and the light intensity of each pixel. The custom connected region light volume can comprehensively characterize the shape, size, and light intensity amplitude of the seed shadow connected region. By using the calibrated fuzzy intervals of each light volume to obtain the likelihood probability of the number of seeds and the prior probability of each number of seeds, a Bayesian classifier is used to obtain the probability of the number of seeds under any light volume, improving the accuracy of seed number discrimination. In addition, the use of fuzzy intervals can effectively reduce the number of light volumes that need to be judged, improve the detection speed, and ensure real-time detection. Therefore, this invention can accurately and quickly detect the seeding flow in real time. Attached Figure Description

[0088] 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.

[0089] Figure 1 This is a flowchart of the seed flow detection method based on linear array charge-coupled devices described in this invention.

[0090] Figure 2 This is a block diagram of the seed flow detection system based on a linear array charge-coupled device as described in this invention. Detailed Implementation

[0091] 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.

[0092] 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.

[0093] 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.

[0094] like Figure 1 As shown, the seed flow detection method based on linear array charge-coupled devices of the present invention includes the following steps:

[0095] S01: Linear charge-coupled devices and LED arrays are installed on both sides of the seed flow channel to form a detection optical path; the analog signal output by the linear charge-coupled devices is converted into a digital signal according to the scanning cycle, and the seed image is formed by splicing multiple cycles.

[0096] S02: Scan the seed image and identify all connected components based on the region growing algorithm, specifically including the following sub-steps:

[0097] S2.1: Initialize the label number l and the current pixel index i;

[0098] S2.2: First scan each pixel of the image from left to right, then from top to bottom, and compare the pixel value I(x,y) at the current pixel coordinate position (x,y) with the light intensity threshold T. i Compare;

[0099] S2.3: When the first condition I(x, y) > T is found... i When a pixel is found, that point is taken as the germination point, and a new connected component label l and pixel index i+1 are assigned to it.

[0100] S2.4: Check the four neighboring pixels of the germination point (i.e., the pixels corresponding to coordinates (x,y-1), (x-1,y), (x,y+1), (x+1,y)). If its pixel value is greater than the light intensity threshold T, i If so, then push the coordinates of that pixel onto the stack;

[0101] S2.5: Pop a pixel coordinate from the stack and assign it the current connected component label l and pixel index i+1;

[0102] S2.6: Check the four neighboring pixels of this pixel, and identify the unmarked pixels with a value greater than T. i The pixel coordinates are pushed onto the stack;

[0103] S2.7: Repeat steps S2.5 and S2.6 until the stack is empty. At this point, the labeling of a connected region with the label 1 is complete.

[0104] S2.8: Increment the label number l + 1 and return to step S2.2 to continue scanning the untraversed areas in the image until the entire image is scanned, thereby identifying all connected regions in the image.

[0105] S03: Light volume calculation, summing the pixel values ​​of each connected component.

[0106] The sum of the pixel values ​​of the seed-shadow connected region for label l is defined as the light volume of the seed-shadow connected region. The light volume of each seed-shadow connected region is calculated as follows:

[0107]

[0108] In the formula,

[0109] V lThe light volume of the connected region of the species labeled l;

[0110] I l i (x,y) is the pixel value of the i-th pixel in the shaded connected domain of label l;

[0111] i is the pixel index in the shaded connected region of label l;

[0112] n is the maximum value of the pixel index;

[0113] x is the x-coordinate of the pixel;

[0114] y is the ordinate value of the pixel;

[0115] l is the label number of the connected domain of the shadow;

[0116] S4: Fuzzy region division, classifying the obtained light volume into one of five fuzzy regions, as follows:

[0117] The light volume of the seed-shadow connected region is determined into one of five fuzzy intervals based on a segmentation threshold. These fuzzy intervals are: minimal, small, medium, large, and maximal.

[0118]

[0119] In the formula:

[0120] FV l The fuzzy interval corresponding to the light volume of the connected region of the shadow of label number l;

[0121] V l The light volume of the connected region of the species labeled l;

[0122] ES represents a fuzzy region where the optical volume is extremely small;

[0123] S represents a fuzzy region with a small light volume;

[0124] M represents the fuzzy region in the optical volume;

[0125] L represents the fuzzy region where the light volume is large;

[0126] EL represents a fuzzy region where the light volume is extremely large;

[0127] T ES The threshold for determining when the light volume is minimal;

[0128] T S The threshold for determining if a light volume is small;

[0129] T M To determine the segmentation threshold in the light volume;

[0130] T L The segmentation threshold for determining whether the light volume is large;

[0131] T EL The threshold for determining when the light volume is maximized;

[0132] S05: Consult the likelihood probability table and calculate the posterior probability based on the prior probability; details are as follows:

[0133] Based on the fuzzy interval of optical volume, look up the table to find the likelihood probability P(OV=FV) of the fuzzy interval of optical volume corresponding to label l, given that the predicted number of landing species is j. l |N=j), then multiply the obtained likelihood probability by the prior probability with the number of occurrences j, to obtain the optical volume ambiguity interval as FV. l Under the condition that the predicted number of seeds is j, the posterior probability is given.

[0134]

[0135] In the formula:

[0136] N is a random variable representing the number of seeds that fall;

[0137] j is the predicted number of seeds to fall;

[0138] OV is a random variable in the fuzzy interval of the optical volume of the connected domain of the seed shadow;

[0139] FV l The optical volume fuzzy region of the seed-shadow connected region of label l;

[0140] P(N=j|OV=FV l The image volume blur value is FV. l Under the condition that the number of seeds falls is the predicted value j;

[0141] P(OV=FV l |N=j) is the case where the number of seeds is j and the seed shadow volume is FV l The likelihood probability;

[0142] P(N=j) is the prior probability of having j possible outcomes;

[0143] S06: The number of possible outcomes corresponding to the maximum posterior probability value is taken as the number of possible outcomes for the shadow connected component of that type, as follows:

[0144] The maximum posterior probability P(N=j|OV=FV) l The predicted number of species corresponding to label l is considered as the species count detection result of the connected component of the species shadow:

[0145]

[0146] In the formula:

[0147] n l The result of detecting the number of species in the connected domain of the species with label number l;

[0148] P(N=j|OV=FV l The image volume blur range is FV. l Under the condition that the predicted number of seeds is j, the posterior probability is given.

[0149] N is a random variable representing the number of seeds that fall;

[0150] j is the predicted number of seeds to fall;

[0151] OV is a random variable in the fuzzy interval of the optical volume of the connected domain of the seed shadow;

[0152] FV l The optical volume fuzzy region of the seed-shadow connected region of label l;

[0153] S07: Accumulate the total number of landing types across all regions and output the total number of landing types. Specifically:

[0154] The seed count detection results of all connected components in the seed shadow image are summed to obtain the seed count detection result of the entire seed shadow image:

[0155]

[0156] In the formula,

[0157] n' represents the seed count detection result for the entire seed image;

[0158] n l The result of detecting the number of species in the connected component of the shadow domain labeled l;

[0159] a is the maximum value of the shaded connected region;

[0160] l represents the label number of the connected domain.

[0161] like Figure 2 As shown, the seed flow detection system based on linear charge-coupled devices of the present invention includes linear charge-coupled devices, an LED array, and a signal processing system;

[0162] The linear array charge-coupled device is installed inside the seed flow sensor and is used to scan the cross-section of the seed flow falling channel of the seed flow sensor of the seeder and convert the optical signal into an analog electrical signal.

[0163] The LED array is symmetrically arranged with the linear charge-coupled device (CCD) and is used to provide a light source for the CCCD.

[0164] The signal processing system includes peripheral circuits and an embedded processor. The embedded processor includes a seed image acquisition module, a seed image connected component search module, an optical volume calculation module, an optical volume blurring module, a seed landing probability inference module, a seed landing number determination module, and a seed landing number accumulation module, which are used to process the image information output by the linear array charge-coupled device to obtain the number of seeds landing within the sampling time.

[0165] The seed image acquisition module is used to convert the analog signal output by the linear charge-coupled device into a digital signal according to the scan cycle, and then stitch them together after multiple cycles to form a seed image. Each scan cycle T... s The image acquisition module converts the analog signal output by the linear array charge-coupled device into a digital signal, places the linear array scanning result in a new row, and completes Y scanning cycles T. s This forms a seed image; the seed image sampling time T = Y × T s Y is the set number of image rows, T s The scan cycle.

[0166] The seed shadow connected region search module is used to binarize the image and mark the connected regions, and input all connected regions in the seed shadow image into the light volume calculation module.

[0167] The optical volume calculation module is used to accumulate the pixel values ​​of the seed-shadow connected region of label number l, and input the optical volume of the seed-shadow connected region of label number l into the optical volume blurring module.

[0168] The optical volume fuzzification module is used to divide the optical volume into five fuzzy intervals: "extremely small, small, medium, large, and extremely large," based on a preset threshold. These fuzzy intervals are then input into the seeding probability inference module.

[0169] The seed drop probability inference module is used to calculate the posterior probability of different seed drop numbers under a given optical volume ambiguity interval, and input it into the seed drop number determination module;

[0170] The seed count determination module is used to select the seed count corresponding to the maximum posterior probability as the detection result of the connected component. The detection result is then input into the seed count accumulation module.

[0171] The seed count accumulation module is used to accumulate the seed count detection results of all seed shadow connected regions in the seed shadow image to obtain the seed count detection result of the entire seed shadow image.

[0172] 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.

[0173] 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 seed flow rate detection method based on a linear array charge-coupled device, characterized in that, Includes the following steps: The falling seed stream is scanned by a linear array charge-coupled device, the analog signal is converted into a digital signal and stitched together to form a seed image; Process the seed shadow image to identify all seed shadow connected components; Calculate the optical volume of each species-shadow connected region; Each optical volume is divided into a corresponding fuzzy region; Based on the current optical volume ambiguity interval, query the likelihood probability table and calculate the posterior probability of different predicted drop numbers; The number of possible outcomes corresponding to the maximum posterior probability is determined as the number of possible outcomes in the shadow connected component of that type. The total number of possible outcomes is obtained by summing the number of outcomes in all connected components.

2. The seed flow detection method based on a linear array charge-coupled device according to claim 1, characterized in that, The seed shadow image is processed to identify all connected components of the seed shadow, specifically including the following steps: Traverse the image pixels in the scanning order and compare the pixel values ​​with the light intensity threshold; The first pixel with a light intensity greater than the light intensity threshold is taken as the germination point and assigned a new label; Check the neighboring pixels of the germination point and push the coordinates of neighboring pixels whose pixel values ​​are greater than the light intensity threshold onto the stack; Pop the top pixel coordinates from the stack, assign the current label, check its neighboring pixels, and push the coordinates of unlabeled pixels with pixel values ​​greater than the light intensity threshold onto the stack. Repeat the process of popping, marking, checking neighbors, and pushing onto the stack until the stack is empty, thus completing the marking of a connected component; Continue scanning until the entire image has been processed and all connected components have been identified.

3. The seed flow detection method based on a linear array charge-coupled device according to claim 2, characterized in that, The neighborhood is either a four-neighbor or an eight-neighbor.

4. The seed flow detection method based on a linear array charge-coupled device according to claim 1, characterized in that, The optical volume of each species-shading connected region is calculated using the following formula: ； In the formula, V l The light volume of the connected region of the species labeled l; I l i (x,y) is the pixel value of the i-th pixel in the shaded connected domain of label l; i is the pixel index in the shaded connected region of label l; n is the maximum value of the pixel index; x is the x-coordinate of the pixel; y is the ordinate value of the pixel; l represents the label number of the connected domain.

5. The seed flow detection method based on a linear array charge-coupled device according to claim 1, characterized in that, Each optical volume is divided into a corresponding fuzzy region, as follows: The light volume of the seed-shadow connected region is determined into one of five fuzzy intervals based on a segmentation threshold. These fuzzy intervals are: minimal, small, medium, large, and maximal. ； In the formula: FV l The fuzzy interval corresponding to the light volume of the connected region of the shadow of label number l; V l The light volume of the connected region of the species labeled l; ES represents a fuzzy region where the optical volume is extremely small; S represents a fuzzy region with a small light volume; M represents the fuzzy region in the optical volume; L represents the fuzzy region where the light volume is large; EL represents a fuzzy region where the light volume is extremely large; T ES The threshold for determining when the light volume is minimal; T S The threshold for determining if a light volume is small; T M To determine the segmentation threshold in the light volume; T L The segmentation threshold for determining whether the light volume is large; T EL The segmentation threshold is used to determine when the light volume is extremely large.

6. The seed flow detection method based on a linear array charge-coupled device according to claim 1, characterized in that, Based on the current optical volume ambiguity interval, the likelihood probability table is consulted and the posterior probability of different predicted landing numbers is calculated, specifically: Based on the fuzzy interval of optical volume, look up the table to find the likelihood probability P(OV=FV) of the fuzzy interval of optical volume corresponding to label l, given that the predicted number of landing species is j. l |N=j), Multiplying the obtained likelihood probability by the prior probability with j as the number of occurrences, we obtain the optical volume ambiguity interval as FV. l Under the condition that the predicted number of species is j, the posterior probability is... ; In the formula: N is a random variable representing the number of seeds that fall; j is the predicted number of seeds to fall; OV is a random variable in the fuzzy interval of the optical volume of the connected domain of the seed shadow; FV l The optical volume fuzzy region of the seed-shadow connected region of label l; P(N=j|OV=FV l The image volume blur value is FV. l Under the condition that the number of seeds falls is the predicted value j; P(OV=FV l |N=j) is the case where the number of seeds is j and the seed shadow volume is FV l The likelihood probability; P(N=j) is the prior probability of the number of possible outcomes being j.

7. The seed flow detection method based on a linear array charge-coupled device according to claim 1, characterized in that, The number of possible outcomes corresponding to the maximum posterior probability is determined as the number of possible outcomes in the shadow connected region of that type, as follows: The maximum posterior probability P(N=j|OV=FV) l The predicted number of species corresponding to label l is considered as the species count detection result of the connected component of the species shadow: ； In the formula: n l The result of detecting the number of species in the connected domain of the species with label number l; P(N=j|OV=FV l The image volume blur range is FV. l Under the condition that the predicted number of seeds is j, the posterior probability is given. N is a random variable representing the number of seeds that fall; j is the predicted number of seeds to fall; OV is a random variable in the fuzzy interval of the optical volume of the connected domain of the seed shadow; FV l The light volume ambiguity interval of the species-connected domain of label l is represented.

8. The seed flow detection method based on a linear array charge-coupled device according to claim 1, characterized in that, The total number of possible outcomes is obtained by summing the number of outcomes in all connected components, specifically: ； In the formula, n' represents the seed count detection result for the entire seed image; n l The result of detecting the number of species in the connected component of the shadow domain labeled l; a is the maximum value of the shaded connected region; l represents the label number of the connected domain.

9. A detection system according to any one of claims 1-8, characterized in that, This includes linear charge-coupled devices, LED arrays, and signal processing systems; The linear array charge-coupled device is installed on one side of the seed flow falling channel of the seed metering device to scan the falling seed flow and convert the optical signal into an analog electrical signal; the LED array is symmetrically arranged with the linear array charge-coupled device on the other side of the seed flow falling channel to provide a light source; The signal processing system includes peripheral circuitry and an embedded processor for receiving and processing the analog electrical signal. The embedded processor includes: The seed image acquisition module is used to convert the analog signal output by the linear charge-coupled device into a digital signal according to the scanning cycle, and stitch them together through multiple scanning cycles to form a seed image; The seed image connected component search module is used to perform binarization processing on the seed image and mark all connected components; The light volume calculation module is used to accumulate the pixel values ​​of all pixels in each species' shadow connected region to obtain the light volume of that connected region; The optical volume blurring module is used to divide the optical volume of each connected component into the corresponding blur interval according to multiple preset segmentation thresholds. The seed drop probability inference module is used to determine the posterior probability of different seed drop numbers under the optical volume ambiguity interval of the current connected domain. The number of occurrences determination module is used to select the number of occurrences corresponding to the maximum posterior probability as the detection result of the connected component; The seed count accumulation module accumulates the seed count detection results of all seed shadow connected regions in the seed shadow image to obtain the seed count detection result of the entire seed shadow image.

Images