Method and system for collecting pesticide application data for intelligent atomizer

Abstract

The present invention relates to the field of image acquisition technology, and in particular to a method and system for collecting drug application data for an intelligent atomizer. Firstly, the droplet diffusion direction of the overall velocity distribution diagram of a fog flow field is determined in combination with the distribution of droplet velocity values ​​in each droplet velocity cluster in the dimension of local analysis; then, according to the characteristics that the injection direction is usually unchanged and the droplet velocity distribution is usually consistent, an initial abnormal velocity distribution diagram is screened out according to the standard direction deviation of the droplet diffusion direction and the droplet velocity value distribution deviation; and on this basis, according to the characteristic that the droplet velocity value under the influence of spatial obstacles will present corresponding characteristic changes, a spatial occlusion velocity distribution diagram that has no effect on the disinfection effect is screened out, so that a more accurate real abnormal velocity distribution diagram is obtained after removing the spatial occlusion velocity distribution diagram, so that the effect of collecting drug application data of the atomizer according to the real abnormal velocity distribution diagram is better.

Description

Technical Field

[0001] The present invention relates to the technical field of image acquisition, and in particular to a method and system for acquiring drug application data for an intelligent atomizer. Background Art

[0002] The atomizer atomizes the drug solution and exerts the drug effect by means of droplet suspension and diffusion, thereby achieving the function of disinfection and sterilization. It is usually used in agriculture, medical care, and carbon neutrality. In the agricultural field, atomizers are usually used to disinfect farms or sheds to effectively disinfect pathogenic microorganisms in the sheds, thereby reducing the occurrence and spread of animal diseases. At present, when using atomizers to disinfect the interior of sheds, the animals are usually moved out of the sheds, and after closing the doors and windows of the sheds, the internal space of the sheds is sprayed with disinfectants through atomizers. In order to strictly control the quality of shed disinfection and screen out areas with unqualified disinfection quality, relevant data on drug spraying at different locations in the sheds are usually collected and stored.

[0003] Traditional methods for testing the disinfection effect of aquaculture environments usually use drug concentration testing or microbial testing, but both methods require manual assistance. In planned aquaculture environments, the site is usually large, resulting in high time costs and low efficiency for methods of manually testing the disinfection effect of aquaculture environments.

[0004] Considering that the particle image velocimetry method can measure the particle flow velocity, and in a breeding environment with good disinfection effect, the droplet velocity values ​​of the droplet particles sprayed by the atomizer at different positions usually show a similar distribution; therefore, the prior art usually uses the particle image velocimetry method to collect the fog flow field velocity distribution map at different positions in the breeding environment, and then performs template matching based on the collected fog flow field velocity distribution map and the fog flow field velocity distribution map under ideal conditions, and determines the abnormal situation of each fog flow field velocity distribution map according to the matching degree to measure the corresponding disinfection effect; that is, the prior art usually directly collects the fog flow field velocity distribution map at various positions in the breeding environment through the particle image velocimetry method, and marks the abnormal velocity distribution map with a lower matching degree, that is, the poor disinfection effect.

[0005] However, in a breeding environment that is greatly affected by the environment, the overall droplet velocity of the droplet particles will generally be affected; and the presence of obstacles usually does not affect the disinfection effect, but will significantly affect the droplet velocity distribution in the fog flow field velocity distribution diagram; this results in the accuracy of template matching being greatly affected, whether in the fog flow field velocity distribution diagram that is greatly affected by the outside world or in the fog flow field velocity distribution diagram that is affected by obstacles, that is, the prior art is not accurate enough in measuring the abnormal conditions of each fog flow field velocity distribution diagram, resulting in low accuracy of the abnormal velocity distribution diagram marked after directly collecting the fog flow field velocity distribution diagram through the particle image velocimetry method, resulting in poor data collection effect of the atomizer for applying pesticides. Summary of the invention

[0006] The present application provides a method and system for collecting drug application data for an intelligent atomizer. First, the overall droplet diffusion direction of the fog flow field velocity distribution diagram is determined in combination with the distribution of droplet velocity values ​​in each droplet velocity cluster in the dimension of local analysis; then, based on the characteristics that the injection direction is usually unchanged and the droplet velocity distribution is usually consistent, an initial abnormal velocity distribution diagram is screened out according to the standard direction deviation of the droplet diffusion direction and the deviation of the droplet velocity value distribution; and on this basis, based on the characteristics that the droplet velocity values ​​under the influence of spatial obstacles will show corresponding characteristic changes, a spatial occlusion velocity distribution diagram that has no effect on the disinfection effect is screened out, so that a more accurate real abnormal velocity distribution diagram is obtained after removing the spatial occlusion velocity distribution diagram, which solves the problem that the prior art does not accurately measure the abnormal situation of each fog flow field velocity distribution diagram, resulting in low accuracy of the abnormal velocity distribution diagram marked after directly collecting the fog flow field velocity distribution diagram through the particle image velocimetry method, so that the effect of collecting drug application data for the atomizer according to the real abnormal velocity distribution diagram is better.

[0007] The first aspect of the present application provides a method for collecting drug application data for an intelligent atomizer, comprising:

[0008] When the atomizer moves along the top track of the breeding house and sprays disinfectants, a fog flow field velocity distribution diagram of the internal image of the breeding house at each sampling time is collected; each pixel point in the fog flow field velocity distribution diagram corresponds to a droplet velocity value;

[0009] Cluster analysis is performed on each pixel point in the fog flow field velocity distribution map in combination with the droplet velocity value, and all the pixel points are divided into at least two droplet velocity clusters; the cluster diffusion direction of each droplet velocity cluster is determined according to the change trend of the droplet velocity value in each neighborhood direction in each droplet velocity cluster and the overall size of the change rate; the droplet diffusion direction of the fog flow field velocity distribution map at each sampling time is determined according to the cluster diffusion direction of each droplet velocity cluster;

[0010] Determine the reference abnormality degree of the velocity distribution diagram of the fog flow field at each sampling time according to the deviation of the standard direction of the droplet diffusion direction and the deviation of the droplet velocity value distribution; and select the initial abnormal velocity distribution diagram according to the reference abnormality degree;

[0011] According to the symmetric situation of the temporal variation of the reference anomaly degree of each initial anomaly velocity distribution diagram that is adjacent in time sequence, all the spatial obstruction velocity distribution diagrams are screened out;

[0012] The spatially blocked velocity distribution maps in all initial abnormal velocity distribution maps are removed to determine the real abnormal velocity distribution map; and the spraying data of the atomizer is collected according to the real abnormal velocity distribution map.

[0013] Furthermore, the process of obtaining the droplet velocity clustering clusters includes:

[0014] In the image coordinate system of the fog flow field velocity distribution map, a three-dimensional coordinate system of the fog flow field velocity distribution map is constructed with the droplet velocity value of the pixel point as the z-axis; in the three-dimensional coordinate system, the three-dimensional coordinate points of each pixel point in the fog flow field velocity distribution map are counted; and a cluster analysis is performed using the DBSCAN algorithm based on the three-dimensional coordinate points of all the pixel points to obtain at least two droplet velocity clusters.

[0015] Furthermore, the process of obtaining the cluster diffusion direction includes:

[0016] In each droplet velocity cluster, the diffusion starting possibility of each pixel is determined according to the overall size and fluctuation of the droplet velocity in the neighborhood of each pixel; the pixel with the largest diffusion starting possibility in each droplet velocity cluster is taken as the droplet diffusion starting point;

[0017] All pixel points in each neighborhood direction of the droplet diffusion starting point are arranged in the order of the corresponding neighborhood directions, and a reference pixel point sequence in each neighborhood direction including the droplet diffusion starting point is determined; in the reference pixel point sequence, the difference in droplet velocity value between each pixel point and the next pixel point is used as the reference difference value of each pixel point; and the pixel points with reference difference values ​​greater than 0 are used as valid pixel points of each reference pixel point sequence;

[0018] In each neighborhood direction, the product of the diffusion starting possibility extreme difference of all pixels in the corresponding reference pixel sequence and the total number of valid pixels is used as the diffusion direction possibility of each neighborhood direction;

[0019] The neighborhood direction with the greatest probability of corresponding diffusion direction is taken as the cluster diffusion direction of each droplet velocity cluster.

[0020] Furthermore, the process of obtaining the diffusion starting possibility includes:

[0021] In each droplet velocity cluster, all pixels within a preset neighborhood window of each pixel are taken as corresponding neighborhood pixels; the mean droplet velocity values ​​and the extreme value of the droplet velocity values ​​of all neighborhood pixels are counted; and the product between the negative correlation mapping value of the extreme value of the droplet velocity value and the mean droplet velocity value is taken as the diffusion starting possibility of each pixel in each droplet velocity cluster.

[0022] Furthermore, the process of obtaining the droplet diffusion direction includes:

[0023] The unit vector of the cluster diffusion direction of each droplet velocity cluster is determined with unit length as the modulus value;

[0024] In the fog flow field velocity distribution diagram at each sampling moment, the unit vectors of the cluster diffusion directions of all droplet velocity clusters are vector-added to determine the corresponding reference vector; the vector direction corresponding to the reference vector is used as the droplet diffusion direction of the fog flow field velocity distribution diagram at each sampling moment.

[0025] Furthermore, the process of obtaining the reference abnormality degree includes:

[0026] After setting the modulus of the reference vector of the fog flow field velocity distribution diagram at each sampling moment to 1, the corresponding comparison unit vector is determined;

[0027] After vector addition of the comparison unit vectors of the fog flow field velocity distribution diagram at all sampling moments, a standard direction vector is determined; a normalized value of the angle between the direction of the comparison unit vector corresponding to the fog flow field velocity distribution diagram at each sampling moment and the direction of the standard direction vector is used as the corresponding droplet diffusion direction deviation;

[0028] The average of the droplet velocity values ​​of all pixels in the fog flow field velocity distribution map at each sampling time is used as the corresponding droplet velocity characteristic value; the average of the droplet velocity characteristic values ​​corresponding to all sampling times is used as the reference velocity characteristic value; the difference between the droplet velocity characteristic value and the reference velocity characteristic value is used as the droplet velocity distribution deviation of the fog flow field velocity distribution map at each sampling time;

[0029] The product of the droplet diffusion direction deviation and the droplet velocity distribution deviation is normalized to determine the reference abnormality degree of the fog flow field velocity distribution diagram at each sampling moment.

[0030] Furthermore, the process of acquiring the initial abnormal velocity distribution map includes:

[0031] The fog flow field velocity distribution map corresponding to the reference abnormality degree greater than the preset abnormality threshold is used as the initial abnormal velocity distribution map.

[0032] Furthermore, the process of acquiring the spatial occlusion velocity distribution map includes:

[0033] The sampling time corresponding to the initial abnormal velocity distribution diagram is used as the initial abnormal time; each initial abnormal time in a continuous time series is merged to determine at least one initial abnormal time period;

[0034] In each initial abnormal time period, the difference between the reference abnormality level corresponding to each sampling moment and the reference abnormality level corresponding to the next sampling moment is taken as the abnormality change value at each sampling moment; the sampling moment when the corresponding abnormality change value is greater than 0 is taken as the abnormality rising moment; the sampling moment when the corresponding abnormality change value is less than 0 is taken as the abnormality falling moment;

[0035] Obtain the abnormal rise end time and abnormal decline start time of each initial abnormal time period; wherein the abnormal change values ​​of all sampling moments between the first sampling moment of each initial abnormal time period and the abnormal rise end time are less than 0, and the abnormal rise end time is not the abnormal degree rise time; the abnormal change values ​​of all sampling moments between the last sampling moment of each initial abnormal time period and the abnormal decline start time are greater than 0, and the abnormal decline start time is not the abnormal degree decline time;

[0036] The number of all sampling moments between the first sampling moment of each initial abnormal time period and the abnormal rise end moment is taken as the first reference number; the number of all sampling moments between the last sampling moment of each initial abnormal time period and the abnormal drop start moment is taken as the second reference number; the difference between the first reference number and the second reference number is negatively correlated and mapped to determine the occlusion time symmetry of each initial abnormal time period;

[0037] The accumulated value of the absolute values ​​of the abnormal change values ​​at all sampling moments in each initial abnormal time period is taken as the abnormal change degree of each initial abnormal time period;

[0038] The product of the abnormal change degree and the occlusion time symmetry is normalized to determine the spatial occlusion possibility of each initial abnormal time period; the initial abnormal velocity distribution map of all sampling moments in the initial abnormal time period corresponding to the spatial occlusion possibility greater than the preset occlusion threshold is used as the spatial occlusion velocity distribution map.

[0039] Furthermore, the process of collecting the drug application data of the atomizer according to the real abnormal velocity distribution diagram includes:

[0040] The internal image of the breeding house corresponding to the real abnormal velocity distribution map is marked as abnormal, compressed by lossless compression and stored in the data acquisition module; other fog flow field velocity distribution maps outside the real abnormal velocity distribution map are compressed by lossy compression and stored in the data acquisition module.

[0041] In a second aspect, the present application provides a drug application data collection system for an intelligent atomizer, the system comprising:

[0042] The mist flow field velocity distribution map acquisition module is used to collect the mist flow field velocity distribution map of the internal image of the breeding house at each sampling time when the atomizer moves along the top track of the breeding house and sprays disinfectants; each pixel point in the mist flow field velocity distribution map corresponds to a droplet velocity value;

[0043] A droplet diffusion direction determination module is used to perform cluster analysis on each pixel point in the fog flow field velocity distribution map in combination with the droplet velocity value, and divide all the pixel points into at least two droplet velocity clusters; determine the cluster diffusion direction of each droplet velocity cluster according to the change trend of the droplet velocity value in each neighborhood direction in each droplet velocity cluster and the overall size of the change rate; determine the droplet diffusion direction of the fog flow field velocity distribution map at each sampling time according to the cluster diffusion direction of each droplet velocity cluster;

[0044] The first screening module is used to determine the reference abnormality degree of the fog flow field velocity distribution diagram at each sampling time according to the standard direction deviation of the droplet diffusion direction and the distribution deviation of the droplet velocity value; and screen out the initial abnormal velocity distribution diagram according to the reference abnormality degree;

[0045] The second screening module is used to screen out all the spatial blocking velocity distribution diagrams according to the symmetry of the temporal changes of the reference abnormality degrees of each initial abnormal velocity distribution diagram that is adjacent in time sequence;

[0046] The atomizer spraying data acquisition module is used to remove the spatially blocked velocity distribution maps in all initial abnormal velocity distribution maps to determine the real abnormal velocity distribution map; and to collect the atomizer spraying data according to the real abnormal velocity distribution map.

[0047] In a third aspect, the present application provides a computer device, comprising a memory and a processor. The memory is used to store computer program code, and the processor is used to call and run the computer program code from the memory to execute the method of the first aspect or any embodiment of the first aspect of the present application.

[0048] In a fourth aspect, the present application provides a computer program product, comprising a computer program code, which, when executed, performs the method of the first aspect of the present application or any embodiment of the first aspect.

[0049] In a fifth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program code. When the computer program code is executed, the method of the first aspect or any embodiment of the first aspect of the present application is performed.

[0050] This application has the following beneficial effects:

[0051] First, the liquid film sprayed by the nozzle will be affected by air resistance during the spraying process, resulting in a decreasing trend in the droplet velocity, so that the corresponding fog flow field velocity distribution diagram always shows the characteristic that the farther the pixel point is from the injection point, the smaller the droplet velocity value is under normal circumstances; however, since the direction of the nozzle spraying at different sampling moments is unchanged, when there is no abnormality, the overall diffusion direction of the droplets at each sampling moment should be consistent or close; if the droplet diffusion direction at a certain sampling moment is more discrete than the droplet diffusion direction at other sampling moments, the fog flow field velocity distribution diagram at the corresponding sampling moment is more abnormal; therefore, it is first necessary to determine the droplet diffusion direction of the fog flow field velocity distribution diagram at each sampling moment.

[0052] When determining the direction of droplet diffusion, it is necessary to take into account that for the overall fog flow field velocity distribution diagram, the overall figure will always show that the farther the pixel point is from the injection point, the smaller the droplet velocity value corresponding to the pixel point is. This results in that if the fog flow field velocity distribution diagram as a whole is directly analyzed, some local abnormal interference factors will be ignored; therefore, the present application first divides different areas according to the size of the droplet velocity value, that is, different droplet velocity clusters; thus, from the local to the overall analysis perspective, the cluster diffusion direction of each local area, that is, the droplet velocity cluster, is first analyzed, and then the diffusion directions of each cluster are integrated to determine the overall droplet diffusion direction, so that the influence of local interference factors can be taken into account, thereby making the subsequent screened initial abnormal velocity distribution diagram more accurate.

[0053] It is further necessary to consider that the properties of the liquid film sprayed by the nozzle at different sampling moments are usually consistent. Therefore, under normal circumstances, the distribution of droplet velocity values ​​in the fog flow field velocity distribution diagram at different sampling moments should be similar. On this basis, combined with the fact that the more discrete the droplet diffusion direction at a certain sampling moment is relative to the droplet diffusion direction at other sampling moments, the more abnormal the fog flow field velocity distribution diagram at the corresponding sampling moment is. The degree of abnormality is determined by reference based on the deviation of the standard direction of the droplet diffusion direction and the deviation of the droplet velocity value distribution, thereby preliminarily screening out the initial abnormal velocity distribution diagram that may have abnormalities.

[0054] It is further necessary to consider that the presence of obstacles usually does not affect the disinfection effect, but it will significantly affect the characteristics of the droplet velocity distribution and the droplet diffusion direction in the fog flow field velocity distribution diagram, resulting in the normal fog flow field velocity distribution diagram with obstacles being likely to be divided into the initial abnormal velocity distribution diagram; therefore, it is further necessary to screen out the spatial occlusion velocity distribution diagram according to the characteristics of the fog flow field velocity distribution diagram with obstacles. First, it is necessary to consider that when there are obstacles, the obstacles will continuously affect the fog flow field velocity distribution diagram at multiple sampling moments; secondly, when the obstacles are large, they usually show the characteristics of the reference abnormal degree gradually increasing in time sequence and gradually decreasing after a period of time; therefore, this application screens out all the spatial occlusion velocity distribution diagrams according to the symmetric time series changes of the reference abnormal degree of each initial abnormal velocity distribution diagram that is adjacent in time sequence; thus, after removing the spatial occlusion velocity distribution diagrams in all initial abnormal velocity distribution diagrams, a more accurate real abnormal velocity distribution diagram required by this application is obtained, so that the effect of collecting the drug application data of the atomizer according to the real abnormal velocity distribution diagram is better. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0056] Figure 1 A flow chart of a method for collecting drug application data for an intelligent atomizer provided by an embodiment of the present invention;

[0057] Figure 2 A schematic diagram of a data collection scene for a fog flow field velocity distribution diagram provided by an embodiment of the present invention;

[0058] Figure 3 A schematic diagram of a velocity distribution diagram of a fog flow field provided by an embodiment of the present invention;

[0059] Figure 4 A structural diagram of a drug application data acquisition system for an intelligent atomizer provided by an embodiment of the present invention;

[0060] Figure 5 A schematic diagram of the structure of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0061] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of a method and system for collecting drug administration data for an intelligent atomizer proposed in accordance with the present invention, its specific implementation method, structure, features and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment, and specific features, structures or characteristics in one or more embodiments may be combined in any suitable form. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as implying or suggesting relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features.

[0062] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0063] The following is a detailed description of a specific solution of a drug application data collection method and system for an intelligent atomizer provided by the present invention in conjunction with the accompanying drawings.

[0064] The present application embodiment provides a method for collecting drug application data for an intelligent atomizer. Figure 1 , which shows a flow chart of a method for collecting drug application data for an intelligent atomizer provided by an embodiment of the present invention, the method comprising:

[0065] Step S101: When the atomizer moves along the top track of the breeding house and sprays disinfectants, a fog flow field velocity distribution diagram of the internal image of the breeding house at each sampling time is collected; each pixel point in the fog flow field velocity distribution diagram corresponds to a droplet velocity value.

[0066] See also Figure 2 , which shows a schematic diagram of a data collection scene of a fog flow field velocity distribution diagram provided by an embodiment of the present invention, Figure 2In the figure, 1 is the track of the atomizer, 2 is the atomizer, 3 is the nozzle of the atomizer, 4 is the connecting rod, 5 is the laser emitting device, and 6 is the camera. In a specific implementation of the embodiment of the present invention, after the animals in the breeding house are removed, the internal space of the breeding house is disinfected by using the atomizer. In order to evenly distribute the droplets to evenly disinfect the space, the disinfection path, i.e., the track of the atomizer, is set in the center of the aisle of the breeding house, and the central axis of the atomizer is aligned with the track of the atomizer, so that the atomizer moves at a constant speed on the track and the moving speed is set to 0.1 meters per second; the pressure of the atomizer nozzle is fixed and it is a fan-shaped nozzle, and in front of the nozzle according to Figure 2 As shown, a camera is installed on the track of the atomizer, and an image of the interior of the breeding house is taken once a second. The velocity distribution diagram of the fog flow field corresponding to the image of the interior of the breeding house is determined by particle image velocimetry with the help of a laser emitting device on the track of the atomizer. The number of sampling moments is determined by the length of the time period during which the atomizer moves along the track of the atomizer; wherein the laser emitting device, the camera and the atomizer move synchronously, that is, the laser emitting device, the camera and the atomizer are relatively stationary. It should be noted that the velocity distribution diagram of the fog flow field can also be determined by other methods. For example, after obtaining the images of the interior of the breeding house at each sampling moment under the influence of the laser emitting device, the Tecplot image processing software is used to process all the images of the interior of the breeding house to determine the velocity distribution diagram of the fog flow field at each sampling moment, or the optical velocimetry is used to determine the velocity distribution diagram of the fog flow field. Please refer to Figure 3 , which shows a schematic diagram of a velocity distribution diagram of a fog flow field provided by an embodiment of the present invention, Figure 3 In the grayscale representation, the smaller the grayscale value of the pixel point, the larger the droplet velocity value corresponding to it. Therefore, it can be clearly observed that the liquid film sprayed by the fan-shaped nozzle presents an approximately symmetrical fan shape along the nozzle spray direction, and Figure 3 The black areas corresponding to the upper left and upper right corners are backgrounds, not pixels of specific droplet velocity values. It should be noted that Tecplot image processing software and particle image velocimetry are both well-known technical means to those skilled in the art, and will not be further described here.

[0067] Step S102: performing cluster analysis on each pixel point in the fog flow field velocity distribution diagram in combination with the droplet velocity value, and dividing all pixel points into at least two droplet velocity clusters; determining the cluster diffusion direction of each droplet velocity cluster according to the change trend of the droplet velocity value in each neighborhood direction in each droplet velocity cluster and the overall size of the change rate; determining the droplet diffusion direction of the fog flow field velocity distribution diagram at each sampling time according to the cluster diffusion direction of each droplet velocity cluster.

[0068] First, the liquid film sprayed by the nozzle will be affected by air resistance during the spraying process, resulting in a decreasing trend in the droplet velocity, so that the corresponding fog flow field velocity distribution diagram always shows the characteristic that the farther the pixel point is from the injection point, the smaller the droplet velocity value is under normal circumstances; however, since the direction of the nozzle spraying at different sampling moments is unchanged, when there is no abnormality, the overall diffusion direction of the droplets at each sampling moment should be consistent or close; if the droplet diffusion direction at a certain sampling moment is more discrete than the droplet diffusion direction at other sampling moments, the fog flow field velocity distribution diagram at the corresponding sampling moment is more abnormal; therefore, it is first necessary to determine the droplet diffusion direction of the fog flow field velocity distribution diagram at each sampling moment.

[0069] When determining the direction of droplet diffusion, it is necessary to take into account that for the fog flow field velocity distribution diagram as a whole, it will always show that the farther the pixel point is from the injection point, the smaller the droplet velocity value corresponding to the pixel point is. This results in that if the fog flow field velocity distribution diagram as a whole is directly analyzed, the overall velocity change area will erase the influence of local anomalies, so that the interference factors of local anomalies will be ignored; therefore, the present application first divides different local areas according to the size of the droplet velocity value, that is, different droplet velocity clusters; further, from the perspective of analysis from the local to the overall, first analyze the cluster diffusion direction of each local area, that is, the droplet velocity cluster, and then integrate the diffusion directions of each cluster to determine the overall droplet diffusion direction, so that the influence of local interference factors can be taken into account, so that the initial abnormal velocity distribution diagram screened out later is more accurate.

[0070] Preferably, in some possible implementations of the embodiments of the present invention, the process of acquiring the droplet velocity clustering clusters includes:

[0071] In the image coordinate system of the fog flow field velocity distribution map, the three-dimensional coordinate system of the fog flow field velocity distribution map is constructed with the droplet velocity value of the pixel point as the z-axis; in the three-dimensional coordinate system, the three-dimensional coordinate points of each pixel point in the fog flow field velocity distribution map are counted; cluster analysis is performed using the DBSCAN algorithm based on the three-dimensional coordinate points of all pixel points to obtain at least two droplet velocity clustering clusters. In a specific implementation of an embodiment of the present invention, the minimum number of samples of the DBSCAN algorithm is set to 5, and the minimum Euclidean distance between each three-dimensional coordinate point and all three-dimensional coordinate points is calculated; the mean of the minimum Euclidean distance of all three-dimensional coordinate points is used as the neighborhood radius of the DBSCAN algorithm, and the minimum number of samples and the neighborhood radius can be set according to the specific implementation environment, which will not be further elaborated here. In other possible implementation methods of the embodiment of the present invention, the droplet velocity value itself can also be directly clustered and analyzed, so that each droplet velocity cluster corresponds to a corresponding droplet velocity value range, thereby determining each pixel point corresponding to each droplet velocity cluster according to the droplet velocity value range.

[0072] Preferably, in some possible implementations of the embodiments of the present invention, the process of acquiring the cluster diffusion direction includes:

[0073] In each droplet velocity cluster, the diffusion start possibility of each pixel is determined according to the overall size and fluctuation of the neighborhood droplet velocity of each pixel. The process of obtaining the diffusion start possibility includes: in each droplet velocity cluster, all pixels in the preset neighborhood window of each pixel are used as corresponding neighborhood pixels; the mean droplet velocity value and the extreme difference of the droplet velocity value of all neighborhood pixels are counted; the product between the negative correlation mapping value of the extreme difference of the droplet velocity value and the mean droplet velocity value is used as the diffusion start possibility of each pixel in each droplet velocity cluster; further, the pixel with the largest diffusion start possibility in each droplet velocity cluster is used as the droplet diffusion start point. In a specific implementation of an embodiment of the present invention, the preset neighborhood window is set to an eight-neighborhood window, which can be adjusted at will.

[0074] After the atomizer nozzle sprays droplets, the force on the droplets is relatively concentrated when they just start to diffuse, and the corresponding droplet velocity is relatively large and close to the velocity of adjacent droplets; therefore, for each pixel point in each droplet velocity cluster, the larger the droplet velocity value is and the closer it is to the droplet velocity value of the adjacent pixel point, the greater the corresponding possibility of diffusion initiation; therefore, within the preset neighborhood window of each pixel point, the smaller the range of the droplet velocity values ​​of all neighboring pixel points, the closer the droplet velocity of the corresponding pixel point is to the droplet velocity of the adjacent pixel point, and the greater the corresponding possibility of diffusion initiation; therefore, the range of the droplet velocity value is negatively correlated with the possibility of diffusion initiation; and the larger the droplet velocity value is, the greater the corresponding possibility of diffusion initiation is, so the larger the overall mean of the droplet velocity values ​​of all neighboring pixel points is, the greater the corresponding possibility of diffusion initiation is.

[0075] In a specific implementation of the embodiment of the present invention, the process of obtaining the diffusion start possibility is expressed by the formula: ;in, For the The velocity distribution diagram of the fog flow field at the sampling time The first The diffusion starting probability of each pixel; For the The velocity distribution diagram of the fog flow field at the sampling time The first The maximum value of the droplet velocity of all neighboring pixels corresponding to the pixel; For the The velocity distribution diagram of the fog flow field at the sampling time The first The minimum value of the droplet velocity of all neighboring pixels corresponding to the pixel; For the The velocity distribution diagram of the fog flow field at the sampling time The first The average droplet velocity value of all neighboring pixels corresponding to a pixel; For the The velocity distribution diagram of the fog flow field at the sampling time The first The droplet velocity values ​​of all neighboring pixels corresponding to the pixel are extremely different; is an exponential function with a natural constant as base; is the absolute value symbol.

[0076] Arrange all the pixels in each neighborhood direction of the starting point of droplet diffusion in the order of the corresponding neighborhood direction, and determine the reference pixel sequence in each neighborhood direction including the starting point of droplet diffusion; in the reference pixel sequence, the difference in droplet velocity between each pixel and the next pixel is used as the reference difference of each pixel; the pixels with a reference difference greater than 0 are used as valid pixels of each reference pixel sequence. In each neighborhood direction, the product of the extreme difference of the diffusion starting possibility of all pixels in the corresponding reference pixel sequence and the total number of valid pixels is used as the diffusion direction possibility of each neighborhood direction; the neighborhood direction with the largest corresponding diffusion direction possibility is used as the cluster diffusion direction of each droplet velocity cluster. In a specific implementation of an embodiment of the present invention, all neighborhood directions are set to sixteen neighborhood directions, that is, the angle between adjacent neighborhood directions is 22.5 degrees, which can be adjusted according to the specific implementation environment.

[0077] As the diffusion proceeds, due to the influence of resistance, the droplet velocity value will gradually decrease, and the reference difference value greater than 0 indicates that the corresponding neighborhood direction meets the characteristics of diffusion, and therefore is a valid pixel point; the more valid pixels there are, the more pixels that meet the diffusion characteristics, and the corresponding neighborhood direction as a whole meets the characteristics of diffusion, that is, the more likely it is to be a diffusion direction; further, it is necessary to consider that when the droplets are just sprayed, the force on the droplets is more concentrated, and the corresponding droplet velocities are closer, that is, the probability of diffusion initiation is greater; and as the diffusion proceeds, the influence of air resistance, droplet weight and the adhesion of the droplet drug itself will become greater and greater, the droplet velocity distribution will become more and more chaotic, and as the diffusion continues, the corresponding diffusion initiation probability will become smaller and smaller; therefore, for each neighborhood direction, if the difference in the possibility of diffusion initiation in the corresponding direction is greater, it means that the neighborhood direction as a whole meets the characteristics of diffusion, that is, the more likely it is to be a diffusion direction; therefore, the possibility of diffusion direction is finally determined based on the difference in the possibility of diffusion initiation and the total number of valid pixels; thereby determining the corresponding cluster diffusion direction in each droplet velocity cluster.

[0078] In some possible implementations of the embodiments of the present invention, the process of obtaining the diffusion direction possibility is expressed by the formula: ;in, For the The velocity distribution diagram of the fog flow field at the sampling time The droplet velocity clusters are in the The possibility of diffusion direction in each neighborhood direction; For the The velocity distribution diagram of the fog flow field at the sampling time The droplet velocity clusters are in the The maximum value of the diffusion starting probability of all pixels in the reference pixel sequence corresponding to the neighborhood direction; For the The velocity distribution diagram of the fog flow field at the sampling time The droplet velocity clusters are The minimum value of the diffusion starting probability of all pixels in the reference pixel sequence corresponding to the neighborhood direction; For the The velocity distribution diagram of the fog flow field at the sampling time The droplet velocity clusters are in the The diffusion starting probability of all pixels in the reference pixel sequence corresponding to the neighborhood direction is extremely poor; is the absolute value symbol; For the The velocity distribution diagram of the fog flow field at the sampling time The droplet velocity clusters are in the The number of valid pixels in the reference pixel sequence corresponding to the neighborhood direction.

[0079] Preferably, in some possible implementations of the embodiments of the present invention, the process of acquiring the droplet diffusion direction includes:

[0080] After obtaining the cluster diffusion direction of each droplet velocity cluster, it is further necessary to fuse the diffusion directions of each cluster from the local to the overall perspective, so as to determine the droplet diffusion direction of the corresponding fog flow field velocity distribution diagram. As a carrier with directional attributes, the vector can effectively fuse the attribute of direction, so the unit vector of the cluster diffusion direction of each droplet velocity cluster is further determined with the unit length as the modulus value; in the fog flow field velocity distribution diagram at each sampling moment, the unit vectors of the cluster diffusion directions of all droplet velocity clusters are vector-added to determine the corresponding reference vector; the vector direction corresponding to the reference vector is used as the droplet diffusion direction of the fog flow field velocity distribution diagram at each sampling moment. It should be noted that setting the same unit length can avoid the influence of different modulus lengths, so that the obtained droplet diffusion direction is more accurate. It should be noted that vector addition is a technical means well known to those skilled in the art, and will not be further elaborated here.

[0081] Step S103: Determine the reference abnormality degree of the velocity distribution diagram of the fog flow field at each sampling moment according to the deviation of the standard direction of the droplet diffusion direction and the distribution deviation of the droplet velocity value; and select the initial abnormal velocity distribution diagram according to the reference abnormality degree.

[0082] It is further necessary to consider that the properties of the liquid film sprayed by the nozzle at different sampling moments are usually consistent. Therefore, under normal circumstances, the distribution of droplet velocity values ​​in the fog flow field velocity distribution diagram at different sampling moments should be similar. On this basis, combined with the fact that the more discrete the droplet diffusion direction at a certain sampling moment is relative to the droplet diffusion direction at other sampling moments, the more abnormal the fog flow field velocity distribution diagram at the corresponding sampling moment is. The degree of abnormality is determined by reference based on the deviation of the standard direction of the droplet diffusion direction and the deviation of the droplet velocity value distribution, thereby preliminarily screening out the initial abnormal velocity distribution diagram that may have abnormalities.

[0083] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the reference abnormality degree includes:

[0084] After adjusting the modulus of the reference vector of the fog flow field velocity distribution map at each sampling moment to 1, the corresponding comparison unit vector is determined; after adding the comparison unit vectors of the fog flow field velocity distribution map at all sampling moments, the standard direction vector is determined; the normalized value of the angle between the direction of the comparison unit vector corresponding to the fog flow field velocity distribution map at each sampling moment and the direction of the standard direction vector is used as the corresponding droplet diffusion direction deviation. Similarly, the vector modulus values ​​corresponding to all fog flow field velocity distribution maps are adjusted to 1 to avoid the influence of different modulus lengths, making the analysis of direction deviation more accurate. By adding the comparison unit vectors of the fog flow field velocity distribution map at all sampling moments, the overall diffusion direction characteristics of the fog flow field velocity distribution map at all sampling moments can be determined; then, according to the characteristic that the more discrete the droplet diffusion direction at a certain sampling moment is relative to the droplet diffusion direction at other sampling moments, the more abnormal the fog flow field velocity distribution map at the corresponding sampling moment is, the droplet diffusion direction deviation is calculated, so that when the droplet diffusion direction deviation is larger, the corresponding fog flow field velocity distribution map is more abnormal.

[0085] The mean of the droplet velocity values ​​of all pixels in the fog flow field velocity distribution map at each sampling time is taken as the corresponding droplet velocity characteristic value; the mean of the droplet velocity characteristic values ​​corresponding to all sampling times is taken as the reference velocity characteristic value; the difference between the droplet velocity characteristic value and the reference velocity characteristic value is taken as the droplet velocity distribution deviation of the fog flow field velocity distribution map at each sampling time. Since the properties of the liquid film sprayed by the nozzle at different sampling times are usually consistent, the droplet velocity value distribution of the fog flow field velocity distribution map at different sampling times should be similar under normal circumstances. Therefore, in the dimension of the overall size of the droplet velocity value, the reference velocity characteristic value represents the overall numerical characteristics of the droplet velocity value of the fog flow field velocity distribution map at all sampling times. The greater the deviation of the corresponding droplet velocity characteristic value compared to the reference velocity characteristic value, the more abnormal the overall value of the droplet velocity of the corresponding fog flow field velocity distribution map is, that is, the more abnormal the corresponding fog flow field velocity distribution map is. Since both the droplet diffusion direction deviation and the droplet velocity distribution deviation can characterize the abnormality of the velocity distribution map of the fog flow field, the product of the droplet diffusion direction deviation and the droplet velocity distribution deviation is further normalized to determine the reference abnormality of the velocity distribution map of the fog flow field at each sampling time. Normalization can limit the value range of the reference abnormality to 0 to 1, which is convenient for the subsequent screening of the initial abnormal velocity distribution map.

[0086] In a specific implementation of the embodiment of the present invention, the process of obtaining the reference abnormality degree is expressed by the formula: ;in, For the The reference abnormality degree of the fog flow field velocity distribution diagram at each sampling time; For the The average of the droplet velocity values ​​of all pixels in the fog flow field velocity distribution map at the sampling time, that is, the droplet velocity characteristic value; For the The droplet velocity characteristic value of the fog flow field velocity distribution diagram at each sampling time; is the total number of sampling moments; is the absolute value symbol; is the reference speed characteristic value; For the The droplet velocity distribution deviation corresponding to the fog flow field velocity distribution diagram at each sampling time; For the The normalized value of the angle between the direction of the comparison unit vector and the direction of the standard direction vector corresponding to the fog flow field velocity distribution diagram at each sampling time, that is, the deviation of the droplet diffusion direction; It is a linear normalization function. The implementer can adjust the normalization method according to the specific implementation environment.

[0087] Preferably, in some possible implementations of the embodiments of the present invention, the process of acquiring the initial abnormal velocity distribution map includes:

[0088] The greater the reference abnormality, the more abnormal the corresponding fog flow field velocity distribution map is. Therefore, the fog flow field velocity distribution map corresponding to the reference abnormality greater than the preset abnormality threshold is further used as the initial abnormal velocity distribution map. In a specific implementation of the embodiment of the present invention, the preset abnormality threshold is set to 0.57, which can be adjusted according to the specific implementation environment.

[0089] Step S104: Screen out all the spatial blocking velocity distribution diagrams according to the symmetric changes in the time series of the reference abnormality levels of each of the initial abnormal velocity distribution diagrams that are adjacent in time sequence.

[0090] It is further necessary to consider that the presence of obstacles usually does not affect the disinfection effect, but it will significantly affect the characteristics of the droplet velocity distribution and the droplet diffusion direction in the fog flow field velocity distribution diagram, resulting in the normal fog flow field velocity distribution diagram with obstacles being likely to be divided into an initial abnormal velocity distribution diagram; therefore, it is further necessary to screen out the spatial occlusion velocity distribution diagram according to the characteristics of the fog flow field velocity distribution diagram with obstacles. First, it is necessary to consider that when there are obstacles, the obstacles will continuously affect the fog flow field velocity distribution diagram at multiple sampling moments; secondly, when the obstacles are large, they usually show a characteristic of a gradual increase in the reference abnormality degree in time sequence and a gradual decrease after a period of time; therefore, this application screens out all spatial occlusion velocity distribution diagrams according to the symmetry of the time series changes in the reference abnormality degree of each initial abnormal velocity distribution diagram that is adjacent in time sequence.

[0091] Preferably, in some possible implementations of the embodiments of the present invention, the process of acquiring the spatial occlusion velocity distribution map includes:

[0092] The sampling moment corresponding to the initial abnormal velocity distribution diagram is taken as the initial abnormal moment; the initial abnormal moments that are continuous in time series are merged to determine at least one initial abnormal time period; when there is an obstacle, the obstacle will continuously affect the velocity distribution diagram of the fog flow field at multiple sampling moments, so firstly, the initial abnormal velocity distribution diagram that is continuous in time series is analyzed as a whole.

[0093] In each initial abnormal time period, the difference between the reference abnormality level corresponding to each sampling moment and the reference abnormality level corresponding to the next sampling moment is taken as the abnormality change value at each sampling moment; the sampling moment when the corresponding abnormality change value is greater than 0 is taken as the abnormality rising moment; the sampling moment when the corresponding abnormality change value is less than 0 is taken as the abnormality falling moment;

[0094] The abnormal rise end time and abnormal decline start time of each initial abnormal time period are obtained; wherein, the abnormal change values ​​of all sampling moments between the first sampling moment and the abnormal rise end time of each initial abnormal time period are less than 0, and the abnormal rise end time is not the abnormal degree rise time; the abnormal change values ​​of all sampling moments between the last sampling moment and the abnormal decline start time of each initial abnormal time period are greater than 0, and the abnormal decline start time is not the abnormal degree decrease time; the number of all sampling moments between the first sampling moment and the abnormal rise end time of each initial abnormal time period is taken as the first reference number; the number of all sampling moments between the last sampling moment and the abnormal decline start time of each initial abnormal time period is taken as the second reference number; the difference between the first reference number and the second reference number is negatively correlated and mapped to determine the occlusion time symmetry of each initial abnormal time period.

[0095] Since the atomizer runs at a constant speed along the track and acquires an image once per second, when there is an object blocking the space, the degree of damage to the original liquid film gradually increases from the moment the atomizer passes the object. At this time, the reference abnormality degree will gradually increase. When the atomizer is completely above the blocking object, the liquid film is damaged to the greatest extent, that is, the reference abnormality degree reaches the maximum at this time, and remains until the atomizer begins to move out of the range of the blocking object. At this time, the degree of damage to the liquid film gradually decreases, that is, the reference abnormality degree gradually decreases, until it returns to normal. Therefore, when affected by obstacles, the reference abnormality degree of each sampling moment in the time series will have the characteristics of first increasing, then unchanged, and finally decreasing; and it is easy to understand that since the atomizer moves at a uniform speed, under the influence of obstacles, the time length corresponding to the process of increasing the reference abnormality degree and the time length corresponding to the process of decreasing the abnormality degree should theoretically be consistent. Therefore, firstly, the number of sampling moments when the reference abnormality degree increases under the influence of obstacles, that is, the first reference number, and the number of sampling moments when the reference abnormality degree decreases under the influence of obstacles, that is, the second reference number, are calculated respectively, so that the smaller the difference between the first reference number and the second reference number, the more it conforms to the characteristics of obstacle occlusion; therefore, the greater the symmetry of the occlusion time obtained after negative correlation mapping the difference between the first reference number and the second reference number, the more likely the corresponding initial abnormal time period is to be the time period caused by the influence of obstacles.

[0096] The accumulated value of the absolute value of the abnormal change value at all sampling moments in each initial abnormal time period is taken as the abnormal change degree of each initial abnormal time period. The existence of obstacles will cause the reference abnormal degree at each sampling moment in the time series to increase first, then remain unchanged, and finally decrease. Therefore, the reference abnormal degree in the initial abnormal time period under the influence of obstacles has a strong volatility, and the absolute value of the abnormal change value can characterize the instantaneous change of the reference abnormal degree at the instantaneous moment. Therefore, the greater the abnormal change degree obtained by the accumulated value, the more it conforms to the characteristic of strong volatility of the reference abnormal degree under the influence of obstacles.

[0097] Finally, since both the abnormal change degree and the occlusion time symmetry can characterize the characteristics of obstacle occlusion, the product between the abnormal change degree and the occlusion time symmetry is normalized to determine the spatial occlusion possibility of each initial abnormal time period. In a specific implementation of an embodiment of the present invention, the process of obtaining the spatial occlusion possibility is expressed by the formula: ;in, For the The possibility of spatial occlusion during the initial abnormal period; For the A first reference number of an initial abnormal time period; For the a second reference quantity for an initial abnormal time period; For the The symmetry of the occlusion time of the initial abnormal time period; is an exponential function with a natural constant as the base. Implementers can use other negative correlation mapping methods according to the specific implementation environment, such as inverse, and ,in, is the hyperbolic tangent function, which will not be further described here; For the The first The abnormal change value corresponding to each sampling moment; For the The degree of abnormal change during the initial abnormal time period; It is a linear normalization function. The implementer can adjust the normalization method according to the specific implementation environment.

[0098] The greater the spatial occlusion possibility, the more likely the corresponding initial abnormal time period is to be affected by obstacles. Therefore, the initial abnormal velocity distribution diagram of all sampling moments in the initial abnormal time period corresponding to the spatial occlusion possibility greater than the preset occlusion threshold is further used as the spatial occlusion velocity distribution diagram. In a specific implementation of an embodiment of the present invention, the preset occlusion threshold is set to 0.63, which can be adjusted according to the specific implementation environment.

[0099] Step S105: remove the spatially blocked velocity distribution maps in all initial abnormal velocity distribution maps to determine the real abnormal velocity distribution map; and collect the spraying data of the atomizer according to the real abnormal velocity distribution map.

[0100] The spatial occlusion velocity distribution diagram is a velocity distribution diagram of the fog flow field affected by obstacles, and the existence of obstacles usually does not affect the disinfection effect. Therefore, after further removing the spatial occlusion velocity distribution diagram from all initial abnormal velocity distribution diagrams, a more accurate real abnormal velocity distribution diagram required by this application can be obtained; further, the drug application data is collected according to the real abnormal velocity distribution diagram. Preferably, in some possible implementation methods of the embodiments of the present invention, the process of collecting drug application data of the atomizer according to the real abnormal velocity distribution diagram includes:

[0101] The internal image of the breeding house corresponding to the real abnormal velocity distribution diagram is marked as abnormal, compressed by lossless compression, and then stored in the data acquisition module; other fog flow field velocity distribution diagrams outside the real abnormal velocity distribution diagram are compressed by lossy compression and then stored in the data acquisition module. Since there is an abnormality in the droplet velocity distribution in the real abnormal velocity distribution diagram, it can be considered that its drug application data is abnormal, so it is marked as abnormal, and compressed and stored by lossless compression, so that the real abnormal cause can be clearly explored later, such as damage to the atomizer, abnormal ratio of disinfectant drugs, etc.; and other fog flow field velocity distribution diagrams outside the real abnormal velocity distribution diagram without abnormalities are compressed and stored by lossy compression, which can reduce storage pressure and make the atomizer drug application data collection effect better. It should be noted that in a specific implementation method of an embodiment of the present invention, PNG is selected for lossless compression and JPEG is selected for lossy compression. The implementer can adjust it according to the specific implementation environment, and no further details will be given here.

[0102] In summary, a method for collecting drug application data for an intelligent atomizer first determines the overall droplet diffusion direction of the fog flow field velocity distribution diagram by combining the distribution of droplet velocity values ​​in each droplet velocity cluster in the dimension of local analysis; then, according to the characteristics that the injection direction is usually unchanged and the droplet velocity distribution is usually consistent, an initial abnormal velocity distribution diagram is screened out according to the standard direction deviation of the droplet diffusion direction and the deviation of the droplet velocity value distribution; and on this basis, according to the characteristics that the droplet velocity value under the influence of spatial obstacles will show corresponding characteristic changes, a spatial occlusion velocity distribution diagram that has no effect on the disinfection effect is screened out, so that a more accurate real abnormal velocity distribution diagram is obtained after removing the spatial occlusion velocity distribution diagram, which solves the problem that the prior art is not accurate enough in measuring the abnormal situation of each fog flow field velocity distribution diagram, resulting in low accuracy of the abnormal velocity distribution diagram marked after directly collecting the fog flow field velocity distribution diagram by the particle image velocimetry method, so that the effect of collecting drug application data for the atomizer according to the real abnormal velocity distribution diagram is better.

[0103] This application also provides a drug application data collection system for an intelligent atomizer, please refer to Figure 4 , which shows a structural diagram of a drug application data collection system for an intelligent atomizer provided by an embodiment of the present invention, the system comprising: a fog flow field velocity distribution map acquisition module 401, a droplet diffusion direction determination module 402, a first screening module 403, a second screening module 404 and a fog machine drug application data collection module 405.

[0104] The mist flow field velocity distribution map acquisition module 401 is used to collect the mist flow field velocity distribution map of the internal image of the breeding house at each sampling time when the atomizer moves along the top track of the breeding house and sprays the disinfectant; each pixel point in the mist flow field velocity distribution map corresponds to a droplet velocity value;

[0105] The droplet diffusion direction determination module 402 is used to perform cluster analysis on each pixel point in the fog flow field velocity distribution map in combination with the droplet velocity value, and divide all the pixel points into at least two droplet velocity clusters; determine the cluster diffusion direction of each droplet velocity cluster according to the change trend of the droplet velocity value in each neighborhood direction in each droplet velocity cluster and the overall size of the change rate; determine the droplet diffusion direction of the fog flow field velocity distribution map at each sampling time according to the cluster diffusion direction of each droplet velocity cluster;

[0106] The first screening module 403 is used to determine the reference abnormality degree of the fog flow field velocity distribution diagram at each sampling time according to the standard direction deviation of the droplet diffusion direction and the distribution deviation of the droplet velocity value; and screen out the initial abnormal velocity distribution diagram according to the reference abnormality degree;

[0107] The second screening module 404 is used to screen out all the spatial blocking velocity distribution diagrams according to the symmetry of the time series changes of the reference abnormality degrees of each initial abnormal velocity distribution diagram that is adjacent in time sequence;

[0108] The atomizer spraying data collection module 405 is used to remove the spatially blocked velocity distribution maps in all initial abnormal velocity distribution maps, determine the real abnormal velocity distribution map, and collect the atomizer spraying data according to the real abnormal velocity distribution map.

[0109] It should be noted that the system provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device is divided into different functional modules to complete all or part of the functions described above. In addition, the above embodiment provides a drug application data collection system for an intelligent atomizer and a drug application data collection method for an intelligent atomizer, which belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0110] The present application also provides a computer device. Figure 5, which shows a schematic diagram of the structure of a computer device provided by an embodiment of the present invention, the computer device includes a memory 501, a processor 502, and a computer program 503 stored in the memory 501 and running on the processor 502, wherein when the processor 502 executes the computer program 503, the computer device can execute any one of the aforementioned methods for collecting drug administration data for an intelligent atomizer.

[0111] The embodiment of the present application also provides a computer program product. When the computer program product is run on a computer device, the computer device can execute any one of the aforementioned methods for collecting drug administration data for an intelligent atomizer.

[0112] An embodiment of the present application also provides a computer-readable storage medium, in which a computer program code is stored. When the computer program code is executed on a computer device, the computer device can execute any of the aforementioned methods for collecting drug administration data for an intelligent nebulizer.

[0113] In the embodiments provided in the present application, it should be understood that the provided computer device, computer program product and computer-readable storage medium are all used to execute the corresponding methods provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the methods provided above and will not be repeated here.

[0114] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0115] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for collecting drug application data for an intelligent atomizer, characterized in that: The method comprises: When the atomizer moves along the top track of the breeding house and sprays disinfectants, a fog flow field velocity distribution diagram of the internal image of the breeding house at each sampling time is collected; each pixel point in the fog flow field velocity distribution diagram corresponds to a droplet velocity value; Cluster analysis is performed on each pixel point in the fog flow field velocity distribution map in combination with the droplet velocity value, and all the pixel points are divided into at least two droplet velocity clusters; the cluster diffusion direction of each droplet velocity cluster is determined according to the change trend of the droplet velocity value in each neighborhood direction in each droplet velocity cluster and the overall size of the change rate; the droplet diffusion direction of the fog flow field velocity distribution map at each sampling time is determined according to the cluster diffusion direction of each droplet velocity cluster; Determine the reference abnormality degree of the velocity distribution diagram of the fog flow field at each sampling time according to the deviation of the standard direction of the droplet diffusion direction and the deviation of the droplet velocity value distribution; and select the initial abnormal velocity distribution diagram according to the reference abnormality degree; According to the symmetric situation of the temporal variation of the reference anomaly degree of each initial anomaly velocity distribution diagram that is adjacent in time sequence, all the spatial obstruction velocity distribution diagrams are screened out; Removing the spatially blocked velocity distribution maps from all initial abnormal velocity distribution maps to determine a true abnormal velocity distribution map; and collecting the spraying data of the atomizer according to the true abnormal velocity distribution map; The process of acquiring the spatial occlusion velocity distribution map includes: The sampling time corresponding to the initial abnormal velocity distribution diagram is used as the initial abnormal time; each initial abnormal time in a continuous time series is merged to determine at least one initial abnormal time period; In each initial abnormal time period, the difference between the reference abnormality level corresponding to each sampling moment and the reference abnormality level corresponding to the next sampling moment is taken as the abnormality change value at each sampling moment; the sampling moment when the corresponding abnormality change value is greater than 0 is taken as the abnormality rising moment; the sampling moment when the corresponding abnormality change value is less than 0 is taken as the abnormality falling moment; Obtain the abnormal rise end time and abnormal decline start time of each initial abnormal time period; wherein the abnormal change values ​​of all sampling moments between the first sampling moment of each initial abnormal time period and the abnormal rise end time are less than 0, and the abnormal rise end time is not the abnormal degree rise time; the abnormal change values ​​of all sampling moments between the last sampling moment of each initial abnormal time period and the abnormal decline start time are greater than 0, and the abnormal decline start time is not the abnormal degree decline time; The number of all sampling moments between the first sampling moment of each initial abnormal time period and the abnormal rise end moment is taken as the first reference number; the number of all sampling moments between the last sampling moment of each initial abnormal time period and the abnormal drop start moment is taken as the second reference number; the difference between the first reference number and the second reference number is negatively correlated and mapped to determine the occlusion time symmetry of each initial abnormal time period; The accumulated value of the absolute values ​​of the abnormal change values ​​at all sampling moments in each initial abnormal time period is taken as the abnormal change degree of each initial abnormal time period; The product of the abnormal change degree and the occlusion time symmetry is normalized to determine the spatial occlusion possibility of each initial abnormal time period; the initial abnormal velocity distribution map of all sampling moments in the initial abnormal time period corresponding to the spatial occlusion possibility greater than the preset occlusion threshold is used as the spatial occlusion velocity distribution map.

2. A method for collecting pesticide application data for an intelligent atomizer according to claim 1, characterized in that: The process of obtaining the droplet velocity clustering clusters includes: In the image coordinate system of the fog flow field velocity distribution map, a three-dimensional coordinate system of the fog flow field velocity distribution map is constructed with the droplet velocity value of the pixel point as the z-axis; in the three-dimensional coordinate system, the three-dimensional coordinate points of each pixel point in the fog flow field velocity distribution map are counted; and a cluster analysis is performed using the DBSCAN algorithm based on the three-dimensional coordinate points of all the pixel points to obtain at least two droplet velocity clusters.

3. A method for collecting drug application data for an intelligent atomizer according to claim 1, characterized in that: The process of obtaining the cluster diffusion direction includes: In each droplet velocity cluster, the diffusion starting possibility of each pixel is determined according to the overall size and fluctuation of the droplet velocity in the neighborhood of each pixel; the pixel with the largest diffusion starting possibility in each droplet velocity cluster is taken as the droplet diffusion starting point; All pixel points in each neighborhood direction of the droplet diffusion starting point are arranged in the order of the corresponding neighborhood directions, and a reference pixel point sequence in each neighborhood direction including the droplet diffusion starting point is determined; in the reference pixel point sequence, the difference in droplet velocity value between each pixel point and the next pixel point is used as the reference difference value of each pixel point; and the pixel points with reference difference values ​​greater than 0 are used as valid pixel points of each reference pixel point sequence; In each neighborhood direction, the product of the diffusion starting possibility extreme difference of all pixels in the corresponding reference pixel sequence and the total number of valid pixels is used as the diffusion direction possibility of each neighborhood direction; The neighborhood direction with the greatest probability of corresponding diffusion direction is taken as the cluster diffusion direction of each droplet velocity cluster.

4. A method for collecting drug application data for an intelligent atomizer according to claim 3, characterized in that: The process of obtaining the diffusion starting probability includes: In each droplet velocity cluster, all pixels within a preset neighborhood window of each pixel are taken as corresponding neighborhood pixels; the mean droplet velocity values ​​and the extreme value of the droplet velocity values ​​of all neighborhood pixels are counted; and the product between the negative correlation mapping value of the extreme value of the droplet velocity value and the mean droplet velocity value is taken as the diffusion starting possibility of each pixel in each droplet velocity cluster.

5. The method for collecting drug application data for an intelligent atomizer according to claim 1, characterized in that: The process of obtaining the droplet diffusion direction includes: The unit vector of the cluster diffusion direction of each droplet velocity cluster is determined with unit length as the modulus value; In the fog flow field velocity distribution diagram at each sampling moment, the unit vectors of the cluster diffusion directions of all droplet velocity clusters are vector-added to determine the corresponding reference vector; the vector direction corresponding to the reference vector is used as the droplet diffusion direction of the fog flow field velocity distribution diagram at each sampling moment.

6. A method for collecting drug application data for an intelligent atomizer according to claim 5, characterized in that: The process of obtaining the reference abnormality degree includes: After setting the modulus of the reference vector of the fog flow field velocity distribution diagram at each sampling moment to 1, the corresponding comparison unit vector is determined; After vector addition of the comparison unit vectors of the fog flow field velocity distribution diagram at all sampling moments, a standard direction vector is determined; a normalized value of the angle between the direction of the comparison unit vector corresponding to the fog flow field velocity distribution diagram at each sampling moment and the direction of the standard direction vector is used as the corresponding droplet diffusion direction deviation; The average of the droplet velocity values ​​of all pixels in the fog flow field velocity distribution map at each sampling time is used as the corresponding droplet velocity characteristic value; the average of the droplet velocity characteristic values ​​corresponding to all sampling times is used as the reference velocity characteristic value; the difference between the droplet velocity characteristic value and the reference velocity characteristic value is used as the droplet velocity distribution deviation of the fog flow field velocity distribution map at each sampling time; The product of the droplet diffusion direction deviation and the droplet velocity distribution deviation is normalized to determine the reference abnormality degree of the fog flow field velocity distribution diagram at each sampling moment.

7. The method for collecting drug application data for an intelligent atomizer according to claim 1, characterized in that: The process of obtaining the initial abnormal velocity distribution map includes: The fog flow field velocity distribution map corresponding to the reference abnormality degree greater than the preset abnormality threshold is used as the initial abnormal velocity distribution map.

8. The method for collecting drug application data for an intelligent atomizer according to claim 1, characterized in that: The process of collecting the spraying data of the atomizer according to the real abnormal speed distribution diagram includes: The internal image of the breeding house corresponding to the real abnormal velocity distribution map is marked as abnormal, compressed by lossless compression and stored in the data acquisition module; other fog flow field velocity distribution maps outside the real abnormal velocity distribution map are compressed by lossy compression and stored in the data acquisition module.

9. A drug application data collection system for an intelligent atomizer, characterized in that: The system comprises: The mist flow field velocity distribution map acquisition module is used to collect the mist flow field velocity distribution map of the internal image of the breeding house at each sampling time when the atomizer moves along the top track of the breeding house and sprays disinfectants; each pixel point in the mist flow field velocity distribution map corresponds to a droplet velocity value; A droplet diffusion direction determination module is used to perform cluster analysis on each pixel point in the fog flow field velocity distribution map in combination with the droplet velocity value, and divide all the pixel points into at least two droplet velocity clusters; determine the cluster diffusion direction of each droplet velocity cluster according to the change trend of the droplet velocity value in each neighborhood direction in each droplet velocity cluster and the overall size of the change rate; determine the droplet diffusion direction of the fog flow field velocity distribution map at each sampling time according to the cluster diffusion direction of each droplet velocity cluster; The first screening module is used to determine the reference abnormality degree of the fog flow field velocity distribution diagram at each sampling time according to the standard direction deviation of the droplet diffusion direction and the distribution deviation of the droplet velocity value; and screen out the initial abnormal velocity distribution diagram according to the reference abnormality degree; The second screening module is used to screen out all the spatial blocking velocity distribution diagrams according to the symmetry of the temporal changes of the reference abnormality degrees of each initial abnormal velocity distribution diagram that is adjacent in time sequence; The process of acquiring the spatial occlusion velocity distribution map includes: The sampling time corresponding to the initial abnormal velocity distribution diagram is used as the initial abnormal time; each initial abnormal time in a continuous time series is merged to determine at least one initial abnormal time period; In each initial abnormal time period, the difference between the reference abnormality level corresponding to each sampling moment and the reference abnormality level corresponding to the next sampling moment is taken as the abnormality change value at each sampling moment; the sampling moment when the corresponding abnormality change value is greater than 0 is taken as the abnormality rising moment; the sampling moment when the corresponding abnormality change value is less than 0 is taken as the abnormality falling moment; Obtain the abnormal rise end time and abnormal decline start time of each initial abnormal time period; wherein the abnormal change values ​​of all sampling moments between the first sampling moment of each initial abnormal time period and the abnormal rise end time are less than 0, and the abnormal rise end time is not the abnormal degree rise time; the abnormal change values ​​of all sampling moments between the last sampling moment of each initial abnormal time period and the abnormal decline start time are greater than 0, and the abnormal decline start time is not the abnormal degree decline time; The number of all sampling moments between the first sampling moment of each initial abnormal time period and the abnormal rise end moment is taken as the first reference number; the number of all sampling moments between the last sampling moment of each initial abnormal time period and the abnormal drop start moment is taken as the second reference number; the difference between the first reference number and the second reference number is negatively correlated and mapped to determine the occlusion time symmetry of each initial abnormal time period; The accumulated value of the absolute values ​​of the abnormal change values ​​at all sampling moments in each initial abnormal time period is taken as the abnormal change degree of each initial abnormal time period; Normalizing the product of the abnormal change degree and the occlusion time symmetry to determine the spatial occlusion possibility of each initial abnormal time period; taking the initial abnormal velocity distribution diagram of all sampling moments in the initial abnormal time period corresponding to the spatial occlusion possibility greater than the preset occlusion threshold as the spatial occlusion velocity distribution diagram; The atomizer spraying data acquisition module is used to remove the spatially blocked velocity distribution maps in all initial abnormal velocity distribution maps to determine the real abnormal velocity distribution map; and to collect the atomizer spraying data according to the real abnormal velocity distribution map.

Images