A monitoring data analysis system and method applied to mutton sheep population production evaluation

Abstract

The application discloses a kind of monitoring data analysis system and method applied to mutton sheep population production evaluation, it is related to big data analysis technical field, the present application is by constructing material pile error analysis set, point cloud data are handled in combination with convex hull algorithm, avoid irregular material pile volume calculation error, and introduce temperature, humidity sensitive coefficient, the influence of quantifying environmental factor to natural subsidence, solve the volume misjudgment problem caused by natural subsidence in traditional method.The establishment of feed trough inside and outside material pile density entropy mapping set is associated with the physical density of material pile and the image gray entropy value, and the abnormal points are removed to quickly deduce the density of material pile through image characteristics, which provides an efficient non-contact feed throwing loss monitoring method.Meanwhile, by fusing multi-source data and dynamic correction mechanism, combining the natural subsidence quantification model and the histogram normalization cross-correlation algorithm, the remaining material pile weight is corrected by using the volume change rate, and the effective intake weight is accurately calculated by considering various factors and eliminating interference.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of big data analysis, and particularly relates to a monitoring data analysis system and method applied to production evaluation of a mutton sheep group. BACKGROUND

[0002] In the mutton sheep fattening process, when the feed formula is changed from a high-concentrate daily ration to roughage such as dry grass, a specific period of dry grass adaptability feeding stage is needed to promote the gradual adjustment of rumen microbial flora and digestive function. Due to the significant changes in the physical form, nutritional composition and palatability of the feed, the mutton sheep group is prone to digestive stress and growth performance fluctuations during this stage, and fine monitoring needs to be implemented through data such as feed intake and body weight gain rate. However, during the dry grass feeding period, the sheep group is prone to abnormal behaviors such as food arching and picky eating, which increases the feed spillage rate. Meanwhile, the loose nature of dry grass aggravates the natural settlement of feed in the trough, making it difficult for a visual-based feed residue image analysis system to accurately identify the actual feed intake. The trough weighing method also causes weight monitoring errors due to the mixing of spilled feed and residual feed in the trough, which seriously restricts the implementation effect of the precision feeding management system. SUMMARY

[0003] The application aims to provide a monitoring data analysis system and method applied to production evaluation of a mutton sheep group to solve the problems in the prior art.

[0004] To achieve the above-mentioned purpose, the application provides the following technical scheme: a monitoring data analysis method applied to production evaluation of a mutton sheep group, the monitoring data analysis method comprising the following steps:

[0005] Step S1, acquiring trough material pile image data and environmental data at the moment when the mutton sheep group starts to eat to construct a material pile error analysis set;

[0006] Step S1-1, selecting an arbitrary trough in which the mutton sheep group eats as the research object, and acquiring trough material pile image data through a visual sensor;

[0007] Step S1-2, acquiring environmental data of the mutton sheep group eating through an environmental sensor, the environmental data comprising temperature data and relative humidity data, and constructing a material pile error analysis set according to the acquired material pile image data and environmental data.

[0008] By acquiring the trough material pile image data and environmental data to construct the material pile error analysis set, multi-dimensional basic information at the initial stage of mutton sheep eating can be systematically collected. The image data can visually present the material pile form, and the environmental data can reflect external influencing factors.

[0009] Step S2, calculate the initial volume of the material pile according to the material pile image data in the material pile error analysis set, construct a natural settlement quantitative model according to the environment data of the material pile error analysis set, and calculate the volume change rate of the material pile in the trough through the natural settlement quantitative model;

[0010] The natural settlement quantitative model calculates the volume change rate of the material pile in the trough using the following formula:

[0011]

[0012] In the formula, represents the volume change rate of the material pile per unit time; V0 represents the initial volume of the material pile in the trough; VS represents the volume of the material pile in the trough at the current time; a represents the temperature sensitivity coefficient; △T represents the temperature change value of the meat sheep group feeding per unit time; b represents the relative humidity sensitivity coefficient; and △RH represents the relative humidity change value of the meat sheep group feeding per unit time.

[0013] Step S2-1, according to the material pile error analysis set analysis, extract the three-dimensional point cloud data of the surface of the material pile in the trough, and map the point cloud data to a unified world coordinate system, which is represented as a global reference frame in three-dimensional space, defined as three mutually perpendicular and intersecting coordinate axes;

[0014] Step S2-2, use the convex hull algorithm to extract the boundary of the discrete point set, and construct the smallest convex polyhedron containing all surface points: select the vertices constituting the convex hull boundary through polar angle sorting and cross product judgment, form a closed triangular mesh surface model, decompose the convex polyhedron into several tetrahedral units, and calculate the initial volume of the material pile by geometric integral accumulation of the volume of each unit or using the orthographic projection method to calculate the volume difference of the upper and lower surfaces.

[0015] Decompose the three-dimensional convex hull into multiple tetrahedrons, calculate the volume of all tetrahedrons and sum them up. If the convex hull has n vertices, start from one vertex and form (n-1)(n-2) / 2 tetrahedrons with other vertices.

[0016] For vertices P i (x i , y i , z i ), P j (x j , y j , z j ), P k (x k , y k , z k ) and P u (x u , y u , z u) tetrahedron, whose volume calculation formula is:

[0017]

[0018] Step S2-3, using the percentage of the volume change of the pile in a period of time to the initial volume to obtain the natural sedimentation rate, introducing the temperature data and the relative humidity data to establish a compensation coefficient correction relationship, quantifying the coupling effect of thermal expansion and particle adsorption effect on the sedimentation rate, and constructing a natural sedimentation quantification model;

[0019] Step S2-4, the compensation coefficient correction relationship includes a temperature sensitivity coefficient and a relative humidity sensitivity coefficient; the temperature sensitivity coefficient represents the quantitative influence of unit temperature change on the natural sedimentation rate of the pile; the relative humidity sensitivity coefficient represents the dynamic correction of unit humidity change on particle adsorption and desorption; keeping the relative humidity constant, changing the temperature and recording the volume change, fitting the relationship between the volume change rate and the temperature change to obtain the temperature sensitivity coefficient; keeping the temperature constant, changing the humidity and recording the volume change, fitting the relationship between the volume change rate and the relative humidity change to obtain the relative humidity sensitivity coefficient.

[0020] The convex hull algorithm is used to convert the surface point cloud data of the pile into a three-dimensional geometric model, and the initial volume is calculated by decomposing the tetrahedral unit to avoid the error caused by the irregular shape in the traditional measurement method; secondly, when constructing the natural sedimentation quantification model, the temperature sensitivity coefficient and the relative humidity sensitivity coefficient are introduced to quantify the influence of environmental factors on the pile settlement and correct the volume change deviation caused by thermal expansion and contraction, particle adsorption and other physical effects. This method can comprehensively consider the geometric characteristics of the pile and the environmental factors, and effectively improve the accuracy of the calculation of the volume change rate of the pile.

[0021] Step S3, randomly sampling the pile density in multiple troughs of the sheep group, combining the pile image data to analyze the corresponding gray level histogram entropy value of different pile densities, and constructing a pile density entropy value mapping set in the trough; simultaneously using the construction method of the pile density entropy value mapping set in the trough to analyze the pile spilled outside the trough, and constructing a pile density entropy value mapping set outside the trough;

[0022] Step S3-1, collecting the pile image data in the trough under the conventional feeding environment conditions of the sheep group by using a standardized sampling device, and simultaneously using a pile density tester to randomly sample and detect the density of the sampled pile, and performing gray scale preprocessing on the image;

[0023] Step S3-2, extract the region image data detected by the bulk density tester, calculate the gray level distribution dispersion of the region by histogram statistics, and quantize the entropy value of the gray level histogram according to the information entropy theory; finally, the measured density data of each sampling point is mapped with the corresponding entropy value, and the abnormal points are removed by iterative optimization, which represent the gray level sudden change image data caused by uneven illumination; and a bulk density entropy value mapping set in the feeding trough is constructed according to the mapping relationship;

[0024] Step S3-3, using the processing method of step S3-1 to step S3-2, the bulk outside the feeding trough is processed, the measured density data of each sampling point is mapped with the corresponding entropy value, and the abnormal points are removed by iterative optimization, and a bulk density entropy value mapping set outside the feeding trough is constructed according to the mapping relationship.

[0025] The measured density data of the bulk is obtained by standardized sampling and density tester, combined with the image data after gray level preprocessing, the gray level distribution dispersion of the bulk is quantized by histogram statistics and information entropy theory, the density data is mapped with the entropy value of the gray level histogram, and the connection between the physical properties of the bulk and the image features is effectively established; secondly, when constructing the bulk density entropy value mapping set inside and outside the feeding trough, the abnormal data caused by interference factors such as illumination is removed by iterative optimization, so as to ensure the accuracy and reliability of the mapping relationship. This method can quickly deduce the bulk density by using the image gray level features, provide an intuitive and efficient analysis method for feed throwing loss monitoring, and realize the synchronous monitoring and quantitative evaluation of the feed state inside and outside the feeding trough.

[0026] Step S4, obtaining the image data of the bulk in the feeding trough at the current time of the meat sheep group, combined with the bulk density entropy value mapping set inside the feeding trough, the remaining bulk weight in the feeding trough is calculated and recorded as the remaining bulk weight in the feeding trough; obtaining the image data of the bulk outside the feeding trough at the current time of the meat sheep group, combined with the bulk density entropy value mapping set outside the feeding trough, the bulk weight outside the feeding trough is calculated and recorded as the bulk weight outside the feeding trough;

[0027] Step S4-1, obtaining the image data of the bulk in the feeding trough by the visual sensor, and performing image analysis matching according to the obtained bulk image data and the bulk density entropy value mapping set inside the feeding trough, the image analysis matching is specifically: through the histogram normalization cross correlation algorithm, the entropy value of the real-time collected image is compared with the entropy value in the mapping set, and the density data of the bulk is obtained according to the mapping relationship between the measured density data and the corresponding entropy value;

[0028] The calculation formula of the histogram normalization cross correlation algorithm is as follows:

[0029]

[0030] In the formula, NCC(A, B) represents the entropy value similarity comparison value of A image data and B image data; A(i, j) represents the gray value at pixel coordinate (i, j) in A image data; B(i, j) represents the gray value at pixel coordinate (i, j) in B image data; i represents the row coordinate index of the pixel; j represents the column coordinate index of the pixel; represents the average gray value of A image data; represents the average gray value of B image data;

[0031] Step S4-2, the volume of the feedlot inside the feedlot is calculated by combining the convex hull algorithm, and the remaining feedlot weight inside the feedlot is calculated by multiplying the volume by the density data, which is recorded as the remaining feedlot weight inside the feedlot;

[0032] Step S4-3, the density data is obtained by image analysis and matching of the spilled feedlot image data outside the feedlot using the calculation method of steps S4-1 to S4-2, and the volume of the spilled feedlot outside the feedlot is calculated according to the convex hull algorithm, and the feedlot weight of the spilled feedlot is calculated by multiplying the volume by the density data, which is recorded as the spilled feedlot weight.

[0033] In terms of feedlot inside feedlot weight calculation, image data is obtained by visual sensor, and real-time image entropy value is compared with feedlot inside feedlot density entropy value mapping set by histogram normalization cross-correlation algorithm, and accurate feedlot density data is obtained, and the remaining feedlot weight inside the feedlot is accurately calculated by combining the volume calculated by the convex hull algorithm; for the spilled feedlot outside the feedlot, the same calculation process is used to realize the calculation of the spilled feedlot weight. This method combines image feature analysis, density mapping and geometric volume calculation, breaks through the limitations of traditional weight measurement method, and can realize real-time and accurate monitoring of the weight change of the feedlot inside and outside the feedlot during the feeding process of the sheep.

[0034] Step S5, the data of the feedlot error analysis set, the current collected feedlot image data and the environmental data are input into the natural settlement quantitative model to calculate the volume change rate of the feedlot, the calculated volume change rate is used as the time decay factor to correct the remaining feedlot weight inside the feedlot, and the corrected remaining feedlot weight is obtained, and the effective feeding weight of the sheep group is calculated according to the initial feedlot weight, the corrected remaining feedlot weight and the spilled feedlot weight.

[0035] The data of the feedlot error analysis set, the real-time image and the environmental parameters are input into the natural settlement model, the unit time feedlot volume change rate is calculated as the time decay factor, the volume loss ratio is converted into the weight decay coefficient, the remaining feedlot weight inside the feedlot is multiplied by the decay coefficient to obtain the corrected remaining feedlot weight after eliminating the natural settlement error, and the effective feeding weight of the sheep group is obtained by subtracting the corrected remaining feedlot weight from the initial feedlot weight and the spilled feedlot weight.

[0036] The calculation formula for converting the volume loss ratio into the weight attenuation coefficient is as follows:

[0037]

[0038] In the formula, D decay is expressed as a weight attenuation coefficient; is expressed as a volume change rate of the feed pile per unit time; S pace is expressed as a time length per unit time.

[0039] By introducing the natural sedimentation quantification model and the weight attenuation coefficient, the interference of natural sedimentation on the weight monitoring of the feed pile is effectively corrected, and the effective feeding weight of the sheep group is accurately calculated. The feed pile error analysis set, real-time image and environmental parameters are input into the natural sedimentation quantification model, the volume change rate of the feed pile per unit time is obtained as a time attenuation factor, and a specific formula is used to convert the volume loss ratio into the weight attenuation coefficient. By multiplying the residual feed pile weight in the trough by the coefficient, the weight error caused by natural sedimentation is eliminated, and the corrected residual feed pile weight is obtained. Finally, the initial feeding pile weight and the spilled feed pile weight are combined to accurately calculate the effective feeding weight of the sheep group. Considering the factors such as natural sedimentation and feed spillage, the error caused by natural sedimentation in the traditional feeding weight calculation method is overcome.

[0040] Further, a monitoring data analysis system applied to the production evaluation of a sheep group, the monitoring data analysis system comprising a data acquisition module, a volume modeling module, a density mapping module, a weight calculation module and an effective feeding calculation module;

[0041] The data acquisition module is used to acquire the feeding related data of the sheep group; the volume modeling module is used to construct a feed pile volume related model and calculate the feed pile volume; the density mapping module is used to establish a mapping relationship between the feed pile density and the entropy value; the weight calculation module is used to calculate the weight of the feed pile inside and outside the trough; and the effective feeding calculation module is used to correct the data and calculate the effective feeding weight of the sheep group;

[0042] The output end of the data acquisition module is electrically connected to the input end of the volume modeling module; the output end of the volume modeling module is electrically connected to the input end of the density mapping module; the output end of the density mapping module is electrically connected to the input end of the weight calculation module; and the output end of the weight calculation module is electrically connected to the input end of the effective feeding calculation module;

[0043] The data acquisition module comprises a basic data acquisition unit and an environmental parameter acquisition unit; the basic data acquisition unit is used to acquire the image data of the feed pile in the trough through a visual sensor; and the environmental parameter acquisition unit is used to acquire the temperature data and the relative humidity data of the sheep feeding environment through an environmental sensor;

[0044] The volume modeling module comprises a point cloud processing and modeling unit and a volume calculation and settlement unit; the point cloud processing and modeling unit is used for extracting three-dimensional point cloud data of a material pile surface and mapping to a world coordinate system for modeling; the volume calculation and settlement unit is used for calculating an initial volume of the material pile and constructing a natural settlement quantitative model;

[0045] The density mapping module comprises an in-trough sampling mapping unit and an out-trough sampling mapping unit; the in-trough sampling mapping unit is used for sampling the material pile in the trough and establishing a density and entropy value mapping relationship; the out-trough sampling mapping unit is used for sampling the spilled material pile outside the trough and establishing a density and entropy value mapping relationship;

[0046] The weight measurement module comprises an in-trough weight calculation unit and an out-trough weight calculation unit; the in-trough weight calculation unit is used for calculating the weight of the remaining material pile in the trough in combination with the mapping set and the convex hull algorithm; the out-trough weight calculation unit is used for calculating the weight of the spilled material pile outside the trough in combination with the mapping set and the convex hull algorithm;

[0047] The effective foraging calculation module comprises a data correction unit and a foraging calculation unit; the data correction unit is used for correcting the weight of the remaining material pile by calculating a volume change rate by using the natural settlement quantitative model; and the foraging calculation unit is used for calculating the effective foraging weight of the sheep group according to the initial feeding weight and the corrected weight.

[0048] Compared with the prior art, the present application has the following beneficial effects:

[0049] 1. The present application precisely calculates the initial volume of the material pile and the natural settlement quantitative model by constructing a material pile error analysis set, combining the convex hull algorithm and a multi-factor coupling model, processing point cloud data by using the convex hull algorithm, avoiding the volume calculation error of irregular material piles, introducing temperature and humidity sensitive coefficients, quantifying the influence of environmental factors on settlement, and effectively solving the volume misjudgment problem caused by natural settlement in the traditional method.

[0050] 2. The present application associates the physical density of the material pile with the image gray entropy value by establishing the density entropy value mapping set of the material pile inside and outside the trough. Based on standardized sampling and information entropy theory, abnormal data points are eliminated to realize the rapid back calculation of the density of the material pile through image features, break through the limitation of traditional contact measurement, provide an efficient non-contact analysis means for feed spillage loss monitoring, and improve the real-time performance and accuracy of the feed state monitoring inside and outside the trough.

[0051] 3、The present application fuses multi-source data and dynamic correction mechanism to accurately calculate the effective grazing weight of the mutton sheep group. The natural settlement quantitative model is combined with the histogram normalization cross-correlation algorithm, the remaining material pile weight is corrected by the volume change rate, the initial feeding amount, the correction amount and the spilled weight are comprehensively considered, the natural settlement and the throwing interference are eliminated, the problem of large calculation error of the traditional grazing weight is solved, reliable data support is provided for accurate feeding management, and the efficiency of mutton sheep breeding is improved. BRIEF DESCRIPTION OF DRAWINGS

[0052] Figure 1 A flowchart of a monitoring data analysis method applied to mutton sheep group production evaluation according to the present application is shown.

[0053] Figure 2 A structural diagram of a monitoring data analysis system applied to mutton sheep group production evaluation according to the present application is shown. DETAILED DESCRIPTION

[0054] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0055] Embodiment one: as shown, the present application provides a technical solution, a monitoring data analysis method applied to mutton sheep group production evaluation, the monitoring data analysis method comprising the following steps: Figure 1

[0056] Step S1, acquiring the image data of the feed trough and the environmental data of the mutton sheep group at the start time of grazing to construct a material pile error analysis set;

[0057] Step S1-1, selecting any feed trough of the mutton sheep group as the research object, and acquiring the image data of the feed trough through a visual sensor;

[0058] Step S1-2, acquiring the environmental data of the mutton sheep group through an environmental sensor, the environmental data including temperature data and relative humidity data, and constructing a material pile error analysis set according to the acquired image data and environmental data.

[0059] ​In specific implementation, image data and environmental data of the feed pile in the feeding trough are collected at the moment when the sheep start to eat, aiming to obtain multi-dimensional information of the initial state of the feed feeding. The image data collected by the visual sensor can record the original shape, contour and surface features of the feed pile, and the temperature and humidity data collected by the environmental sensor are used for subsequent analysis of the influence of environmental factors on the state of the feed. The construction of the feed pile error analysis set is to integrate these basic data, which provides the original basis for subsequent calculation of the volume of the feed pile and analysis of natural settlement, so as to ensure that the subsequent analysis is based on accurate initial conditions; the visual sensor needs to be installed reasonably to ensure that the shooting angle can completely cover the feed pile, so as to avoid missing image information due to angle shielding; the environmental sensor should be placed near the feeding trough and in a position that can accurately reflect the environment of the sheep feeding area, so as to prevent deviation of the environmental data; the data collection time needs to be accurately positioned at the moment when the sheep start to eat, so as to avoid the interference of time error with the data in the feeding process and affect the judgment of the initial state.

[0060] Step S2, according to the initial volume of the feed pile calculated from the feed pile image data in the feed pile error analysis set, and according to the environmental data of the feed pile error analysis set, a natural settlement quantification model is constructed, and the volume change rate of the feed pile in the feeding trough is calculated through the natural settlement quantification model;

[0061] Step S2-1, according to the analysis of the three-dimensional point cloud data of the surface of the feed pile in the feeding trough in the feed pile error analysis set, and mapping the point cloud data to a unified world coordinate system, the world coordinate system is represented as a global reference system in three-dimensional space, which is defined as three mutually perpendicular and intersecting coordinate axes;

[0062] Step S2-2, using the convex hull algorithm to extract the boundary of the discrete point set, and constructing the smallest convex polyhedron containing all surface points: selecting the vertices constituting the convex hull boundary through polar angle sorting and cross product judgment, forming a closed triangular mesh surface model, decomposing the convex polyhedron into several tetrahedral units, calculating the initial volume of the feed pile by geometric integral accumulation of the volume of each unit, or using the orthographic projection method to calculate the volume difference of the upper and lower surface projections;

[0063] Step S2-3, using the percentage of the volume change of the feed pile in a period of time to the initial volume to obtain the natural settlement rate, and introducing the temperature data and relative humidity data to establish a compensation coefficient correction relationship, quantifying the coupling effect of thermal expansion and particle adsorption effect on the settlement rate, and constructing a natural settlement quantification model;

[0064] Step S2-4, the compensation coefficient correction relationship includes a temperature sensitivity coefficient and a relative humidity sensitivity coefficient; the temperature sensitivity coefficient represents the quantitative influence of unit temperature change on the natural settling rate of the material pile; the relative humidity sensitivity coefficient represents the dynamic correction of unit humidity change on the adsorption and desorption of particles; keep the relative humidity constant, change the temperature and record the volume change, fit the relationship between the volume change rate and the temperature change to obtain the temperature sensitivity coefficient; keep the temperature constant, change the humidity and record the volume change, fit the relationship between the volume change rate and the relative humidity change to obtain the relative humidity sensitivity coefficient.

[0065] In specific implementation, the shape of the material pile can be converted into quantifiable geometric data by extracting the three-dimensional point cloud data of the surface of the material pile and mapping it to the world coordinate system. The convex hull algorithm constructs the smallest convex polyhedron based on the point cloud data, simulates the external contour of the material pile, decomposes it into tetrahedral units to calculate the volume, and can effectively deal with irregular shapes of the material pile and accurately obtain the initial volume. Combining environmental data to construct a natural settling quantitative model takes into account that temperature and humidity will affect the physical properties of feed particles, such as thermal expansion and contraction, adsorption of moisture leading to changes in weight and volume. By establishing a compensation coefficient correction relationship, the influence of these factors on the settling rate is quantified, and the accurate calculation of the volume change rate of the material pile is realized; the algorithm needs to be reasonably optimized or the data needs to be down-sampled, and when determining the temperature and humidity sensitivity coefficients, the experimental conditions need to be strictly controlled to ensure that the coefficients can truly reflect the influence of the corresponding environmental factors on the settling.

[0066] Step S3, randomly sample the material pile density in multiple troughs of the sheep group, analyze the corresponding gray level histogram entropy value of different material pile densities combined with the material pile image data, and construct a material pile density entropy value mapping set in the trough; simultaneously analyze the material pile spilled outside the trough using the construction method of the material pile density entropy value mapping set in the trough to construct a material pile density entropy value mapping set outside the trough.

[0067] Step S3-1, collect the material pile image data in the trough of the sheep group under the normal feeding environment condition by using the standardized sampling device, and simultaneously use the bulk density tester to randomly sample and detect the density of the sampled material pile, and perform gray scale preprocessing on the image;

[0068] Step S3-2, extract the region image data detected by the bulk density tester of the material pile, calculate the gray scale distribution dispersion of the region by histogram statistics, and quantize the entropy value of the gray level histogram according to the information entropy theory; finally, a mapping relationship between the measured density data of each sampling point and the corresponding entropy value is established, and the abnormal points are removed through iterative optimization, the abnormal points represent the gray scale sudden change image data caused by uneven illumination, and the material pile density entropy value mapping set in the trough is constructed according to the mapping relationship;

[0069] Step S3-3, the processing method of step S3-1 to step S3-2 is used to process the material pile spilled outside the trough, and a mapping relationship is established between the measured density data of each sampling point and the corresponding entropy value, and after iteration and optimization to remove abnormal points, the density entropy mapping set of the material pile outside the trough is constructed according to the mapping relationship.

[0070] In specific implementation, based on the characteristics that there is a correlation between image gray features and material pile density, the material piles inside and outside the trough are sampled, the actual density is measured, and the corresponding image gray histogram entropy value is analyzed to establish a mapping relationship between the two. The gray histogram entropy value can reflect the dispersion degree of the image gray distribution. Different density material piles have different gray features on the image due to differences in surface particle arrangement, voids and other differences, and correspond to different entropy values. After the mapping set is constructed, the material pile density can be quickly calculated by analyzing the image entropy value in the subsequent process, and non-contact monitoring of the feed state is realized; the standardized sampling device needs to ensure the randomness and representativeness of sampling to avoid inaccurate mapping relationship caused by sampling deviation; when the pile density tester measures, the measurement method should be standardized and the operation should be accurate to prevent measurement error.

[0071] Step S4, the image data of the material pile in the trough is obtained at the current time when the mutton sheep group is feeding, and the remaining material pile weight in the trough is calculated by combining the material pile density entropy mapping set in the trough, which is denoted as the remaining material pile weight in the trough; the image data of the spilled material pile outside the trough is obtained at the current time when the mutton sheep group is feeding, and the weight of the material pile spilled in the trough is calculated by combining the material pile density entropy mapping set outside the trough, which is denoted as the spilled material pile weight.

[0072] Step S4-1, the image data of the material pile in the trough is obtained by the vision sensor, and the image analysis matching is performed according to the obtained material pile image data and the material pile density entropy mapping set in the trough. The image analysis matching is specifically: through the histogram normalization cross-correlation algorithm, the entropy value of the real-time collected image is compared with the entropy value in the mapping set for similarity, and the density data of the material pile is obtained according to the mapping relationship between the measured density data and the corresponding entropy value;

[0073] Step S4-2, the volume of the material pile in the trough is calculated by combining the convex hull algorithm, and the remaining material pile weight in the trough is calculated by multiplying the density data by the volume, which is denoted as the remaining material pile weight in the trough.

[0074] Step S4-3, the image analysis matching is performed on the image data of the spilled material pile outside the trough to obtain the density data by using the calculation method of step S4-1 to step S4-2, and the volume of the spilled material pile outside the trough is calculated according to the convex hull algorithm, and the weight of the spilled material pile outside the trough is calculated by multiplying the density data by the volume, which is denoted as the spilled material pile weight.

[0075] In specific implementation, the histogram normalization cross-correlation algorithm is used to compare the entropy value of the real-time collected image of the feed pile in and outside the trough with the constructed mapping set, find the most similar entropy value corresponding relationship, and obtain the pile density. Combined with the pile volume calculated by the convex hull algorithm, the pile weight is obtained by the product of the density and the volume. The conversion from image data to feed weight data is realized, and the remaining and spilled feed in and outside the trough is monitored in real time.

[0076] Step S5, input the data of the pile error analysis set, the current collected image data of the feed pile in the trough and the environmental data into the natural settlement quantitative model to calculate the volume change rate of the feed pile in the trough, take the calculated volume change rate as the time decay factor, correct the remaining feed pile weight in the trough to obtain the corrected remaining feed pile weight, and calculate the effective feeding weight of the sheep group according to the initial feeding weight, the corrected remaining feed pile weight and the spilled feed pile weight.

[0077] Input the data of the pile error analysis set, real-time image and environmental parameters into the natural settlement model, calculate the volume change rate of the pile per unit time as the time decay factor, convert the volume loss proportion into the weight decay coefficient, multiply the remaining feed pile weight in the trough by the decay coefficient to obtain the corrected remaining feed pile weight after eliminating the natural settlement error, subtract the corrected remaining feed pile weight from the initial feeding weight, and obtain the effective feeding weight of the sheep group and the spilled feed pile weight;

[0078] The calculation formula for converting the volume loss proportion into the weight decay coefficient is as follows:

[0079]

[0080] In the formula, D decay is the weight decay coefficient; is the volume change rate of the pile per unit time; S pace is the time length per unit time.

[0081] In specific implementation, input the data of the pile error analysis set, real-time image and environmental parameters into the natural settlement quantitative model to obtain the volume change rate of the feed pile in the trough, which reflects the volume loss of the pile caused by natural settlement and other factors. Take it as the time decay factor to convert it into the weight decay coefficient, correct the remaining feed pile weight in the trough to eliminate the weight misjudgment caused by natural settlement, combine the initial feeding weight and the spilled feed pile weight, and finally obtain the actual effective feeding weight of the sheep group to ensure the accuracy of the feeding data.

[0082] In specific implementation, input the data of the pile error analysis set, real-time image and environmental parameters into the natural settlement quantitative model to obtain the volume change rate of the feed pile in the trough, which reflects the volume loss of the pile caused by natural settlement and other factors. Take it as the time decay factor to convert it into the weight decay coefficient, correct the remaining feed pile weight in the trough to eliminate the weight misjudgment caused by natural settlement, combine the initial feeding weight and the spilled feed pile weight, and finally obtain the actual effective feeding weight of the sheep group to ensure the accuracy of the feeding data. Figure 2As shown, the application provides a monitoring data analysis system applied to production evaluation of a sheep group, which comprises a data acquisition module, a volume modeling module, a density mapping module, a weight calculation module and an effective feeding calculation module.

[0083] The data acquisition module is used to acquire feeding related data of the sheep group; the volume modeling module is used to construct a volume related model of the material pile and calculate the volume of the material pile; the density mapping module is used to establish a mapping relationship between the density of the material pile and the entropy value; the weight calculation module is used to calculate the weight of the material pile in and out of the feeding trough; and the effective feeding calculation module is used to correct data and calculate the effective feeding weight of the sheep group.

[0084] The output end of the data acquisition module is electrically connected to the input end of the volume modeling module; the output end of the volume modeling module is electrically connected to the input end of the density mapping module; the output end of the density mapping module is electrically connected to the input end of the weight calculation module; and the output end of the weight calculation module is electrically connected to the input end of the effective feeding calculation module.

[0085] The data acquisition module comprises a basic data acquisition unit and an environmental parameter acquisition unit; the basic data acquisition unit is used to acquire image data of the material pile in the feeding trough through a visual sensor; and the environmental parameter acquisition unit is used to acquire temperature data and relative humidity data of the feeding environment of the sheep through an environmental sensor.

[0086] The volume modeling module comprises a point cloud processing and modeling unit and a volume calculation and settlement unit; the point cloud processing and modeling unit is used to extract three-dimensional point cloud data of the surface of the material pile and map it to a world coordinate system for modeling; and the volume calculation and settlement unit is used to calculate the initial volume of the material pile and construct a natural settlement quantitative model.

[0087] The density mapping module comprises an in-trough sampling mapping unit and an out-trough sampling mapping unit; the in-trough sampling mapping unit is used to sample the material pile in the feeding trough and establish a mapping relationship between the density and the entropy value; and the out-trough sampling mapping unit is used to sample the spilled material pile out of the feeding trough and establish a mapping relationship between the density and the entropy value.

[0088] The weight calculation module comprises an in-trough weight calculation unit and an out-trough weight calculation unit; the in-trough weight calculation unit is used to calculate the weight of the remaining material pile in the feeding trough in combination with a mapping set and a convex hull algorithm; and the out-trough weight calculation unit is used to calculate the weight of the spilled material pile out of the feeding trough in combination with a mapping set and a convex hull algorithm.

[0089] The effective foraging calculation module comprises a data correction unit and a foraging calculation unit; the data correction unit is used for calculating the volume change rate by using a natural settlement quantification model to correct the residual pile weight; and the foraging calculation unit is used for calculating the effective foraging weight of the sheep group according to the initial feeding weight and the corrected weight.

[0090] It will be obvious to a person skilled in the art that, without departing from the spirit or essential characteristics of the application, the present application can be implemented in other specific forms. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the description given above, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference signs in the claims should not be construed as limiting the claim concerned.

Claims

1. A method for analyzing monitoring data applied to production evaluation of a sheep flock, characterized in that: The monitoring data analysis method comprises the following steps: Step S1, obtaining the image data of the feed pile in the trough and the environmental data of the starting time of the sheep group to construct a feed pile error analysis set; Step S2, according to the image data of the feed pile in the error analysis set, the initial volume of the feed pile is calculated, and the environmental data of the error analysis set is used to construct a natural settlement quantitative model, and the volume change rate of the feed pile in the trough is calculated through the natural settlement quantitative model; Step S3, randomly sampling the density of the feed pile in the trough of the sheep group, combining the image data of the feed pile to analyze the gray histogram entropy value corresponding to different feed pile densities, and constructing a feed pile density entropy value mapping set; the construction method of the feed pile density entropy value mapping set is used to analyze the feed pile outside the trough, and a feed pile density entropy value mapping set outside the trough is constructed; Step S4, obtaining the image data of the feed pile in the trough of the sheep group at the current time, combining the feed pile density entropy value mapping set to calculate the remaining feed pile weight in the trough, which is denoted as the remaining feed pile weight in the trough; obtaining the image data of the spilled feed pile outside the trough of the sheep group at the current time, combining the feed pile density entropy value mapping set to calculate the weight of the feed pile spilled in the trough, which is denoted as the spilled feed pile weight; Step S5, inputting the data of the feed pile error analysis set and the current collected image data of the feed pile in the trough and the environmental data into the natural settlement quantitative model to calculate the volume change rate of the feed pile in the trough, taking the calculated volume change rate as the time decay factor, and correcting the remaining feed pile weight in the trough to obtain the corrected remaining feed pile weight, and calculating the effective feeding weight of the sheep group according to the initial feed pile weight, the corrected remaining feed pile weight and the spilled feed pile weight.

2. The method for analyzing monitoring data applied to the production evaluation of sheep population according to claim 1, characterized in that: The specific steps of step S1 are as follows: Step S1-1, selecting an arbitrary trough of a sheep group as the research object, and obtaining the image data of the feed pile in the trough through a visual sensor; Step S1-2, obtaining the environmental data of the sheep group through an environmental sensor, the environmental data including temperature data and relative humidity data, and constructing a feed pile error analysis set according to the obtained image data and environmental data.

3. The method for analyzing monitoring data applied to the production evaluation of sheep population according to claim 2, characterized in that: The specific steps of step S2 are as follows: Step S2-1, according to the error analysis set, the three-dimensional point cloud data of the surface of the feed pile in the trough is extracted and mapped to a unified world coordinate system, which is represented as a global reference frame in three-dimensional space, defined as three mutually perpendicular and intersecting coordinate axes; Step S2-2, using a convex hull algorithm to extract the boundary of the discrete point set, and constructing a minimum convex polyhedron containing all surface points: selecting the vertices constituting the convex hull boundary through polar angle sorting and cross product judgment, forming a closed triangular mesh surface model, decomposing the convex polyhedron into several tetrahedral units, and calculating the initial volume of the feed pile by geometric integral accumulation of each unit volume or using the normal projection method to calculate the volume difference of the upper and lower surface projections.

4. The method for analyzing monitoring data applied to the production evaluation of sheep population according to claim 3, characterized in that: In step S2, it also includes: Step S2-3, using the volume change of the material pile in a period of time accounts for the percentage of the initial volume to obtain the natural sedimentation rate, and then introducing the temperature data and relative humidity data, a compensation coefficient correction relationship is established to quantify the coupling effect of thermal expansion and particle adsorption effect on the sedimentation rate, and a natural sedimentation quantitative model is constructed; Step S2-4, the compensation coefficient correction relationship includes temperature sensitivity coefficient and relative humidity sensitivity coefficient; the temperature sensitivity coefficient represents the quantitative influence of unit temperature change on the natural sedimentation rate of the material pile; the relative humidity sensitivity coefficient represents the dynamic correction of unit humidity change on particle adsorption and desorption; keeping the relative humidity constant, changing the temperature and recording the volume change, fitting the relationship between the volume change rate and the temperature change to obtain the temperature sensitivity coefficient; keeping the temperature constant, changing the humidity and recording the volume change, fitting the relationship between the volume change rate and the relative humidity change to obtain the relative humidity sensitivity coefficient.

5. The method for analyzing monitoring data applied to the production evaluation of sheep population according to claim 4, characterized in that: The specific steps of step S3 are as follows: Step S3-1, collecting image data of the material pile in the trough under the normal feeding environment condition of the sheep group by using the standardized sampling device, and then randomly sampling and detecting the density of the sampling material pile by using the bulk density tester, and the image is preprocessed by gray scale; Step S3-2, extracting the region image data detected by the bulk density tester, calculating the gray scale distribution dispersion of the region by histogram statistics, and quantifying the entropy value of the gray scale histogram according to the information entropy theory; finally, the measured density data of each sampling point and the corresponding entropy value are mapped, and the abnormal points are removed by iterative optimization, the abnormal points represent the gray scale sudden change image data caused by uneven illumination, and the density entropy mapping set of the material pile in the trough is constructed according to the mapping relationship; Step S3-3, using the processing method of step S3-1 to step S3-2 to process the material pile spilled outside the trough, and the mapping relationship between the measured density data of each sampling point and the corresponding entropy value is established, and the density entropy mapping set of the material pile outside the trough is constructed according to the mapping relationship after iterative optimization of the abnormal points.

6. The method for analyzing monitoring data applied to the production evaluation of sheep population according to claim 5, characterized in that: The specific steps of step S4 are as follows: Step S4-1, obtaining the image data of the material pile in the trough by using the visual sensor, and performing image analysis and matching on the obtained image data of the material pile and the density entropy mapping set of the material pile in the trough, the image analysis and matching specifically includes: comparing the entropy value of the real-time collected image with the entropy value in the mapping set by using the histogram normalization cross-correlation algorithm, and obtaining the density data of the material pile according to the mapping relationship between the measured density data and the corresponding entropy value; Step S4-2, combining the convex hull algorithm to calculate the volume of the material pile in the trough, and calculating the remaining material pile weight in the trough by multiplying the density data by the volume, which is recorded as the remaining material pile weight in the trough; Step S4-3, using the calculation method of step S4-1 to step S4-2 to perform image analysis and matching on the image data of the spilled material pile outside the trough to obtain the density data, and calculating the volume of the spilled material pile outside the trough by using the convex hull algorithm, and calculating the material pile weight of the spilled material pile outside the trough by multiplying the density data by the volume, which is recorded as the spilled material pile weight.

7. The method for analyzing monitoring data applied to the production evaluation of sheep population according to claim 6, characterized in that: In step S5, the data of the heap error analysis set, real-time images and environmental parameters are input into the natural settlement model to calculate the volume change rate of the heap per unit time as a time decay factor, convert the volume loss ratio into a weight decay coefficient, multiply the remaining heap weight in the trough by the decay coefficient to obtain the corrected remaining heap weight after eliminating the natural settlement error, subtract the corrected remaining heap weight from the initial feeding heap weight to obtain the effective feeding weight of the sheep group, and add the spilled heap weight. The calculation formula for converting the volume loss ratio into the weight decay coefficient is as follows: In the formula, Ddecay represents the weight decay coefficient; VSpace represents the volume change rate of the heap per unit time; and Space represents the time length per unit time.

8. A monitoring data analysis system applied to the production evaluation of a meat sheep population, which is applied to the monitoring data analysis method applied to the production evaluation of a meat sheep population according to any one of claims 1-7, characterized in that: The monitoring data analysis system comprises a data acquisition module, a volume modeling module, a density mapping module, a weight calculation module and an effective feeding calculation module. The data acquisition module is used to acquire the feeding-related data of the sheep group; the volume modeling module is used to construct a heap volume-related model and calculate the heap volume; the density mapping module is used to establish a mapping relationship between the heap density and the entropy value; the weight calculation module is used to calculate the weight of the heap inside and outside the trough; and the effective feeding calculation module is used to correct the data and calculate the effective feeding weight of the sheep group. The output end of the data acquisition module is electrically connected to the input end of the volume modeling module; the output end of the volume modeling module is electrically connected to the input end of the density mapping module; the output end of the density mapping module is electrically connected to the input end of the weight calculation module; and the output end of the weight calculation module is electrically connected to the input end of the effective feeding calculation module.

9. The monitoring data analysis system for evaluating the production of a sheep group according to claim 8, wherein: The data acquisition module comprises a basic data acquisition unit and an environmental parameter acquisition unit; the basic data acquisition unit is used to acquire the image data of the heap inside the trough through a visual sensor; and the environmental parameter acquisition unit is used to acquire the temperature data and the relative humidity data of the feeding environment of the sheep through an environmental sensor. The volume modeling module comprises a point cloud processing and modeling unit and a volume calculation and settlement unit; the point cloud processing and modeling unit is used to extract the three-dimensional point cloud data of the heap surface and map it to the world coordinate system for modeling; and the volume calculation and settlement unit is used to calculate the initial volume of the heap and construct a natural settlement quantitative model. The density mapping module comprises an in-trough sampling mapping unit and an out-trough sampling mapping unit; the in-trough sampling mapping unit is used to sample the heap inside the trough and establish a mapping relationship between the density and the entropy value; and the out-trough sampling mapping unit is used to sample the spilled heap outside the trough and establish a mapping relationship between the density and the entropy value.

10. The monitoring data analysis system for evaluating the production of a sheep group according to claim 8, wherein: The weight measurement module comprises an in-feeding trough weight calculation unit and an out-feeding trough weight calculation unit; the in-feeding trough weight calculation unit is configured to calculate the weight of the remaining material pile in the feeding trough by combining the mapping set and the convex hull algorithm; the out-feeding trough weight calculation unit is configured to calculate the weight of the spilled material pile outside the feeding trough by combining the mapping set and the convex hull algorithm; The effective foraging calculation module comprises a data correction unit and a foraging calculation unit; the data correction unit is configured to calculate the volume change rate by using a natural sedimentation quantification model to correct the weight of the remaining material pile; The foraging calculation unit is configured to calculate the effective foraging weight of the sheep group according to the initial feeding weight and the corrected weight.

Images