Monitoring data analysis system and method applied to mutton sheep group production evaluation

Abstract

The invention discloses a monitoring data analysis system and method applied to mutton sheep group production evaluation, and relates to the technical field of big data analysis, and the method comprises the steps: building a material pile error analysis set, processing point cloud data in combination with a convex hull algorithm, avoiding an irregular material pile volume calculation error, and introducing temperature and humidity sensitivity coefficients. The influence of environmental factors on natural settlement is quantified, and the problem of volume misjudgment caused by natural settlement in a traditional method is solved. According to the method, a feed pile density entropy mapping set inside and outside a trough is established, the physical density of the feed pile and the image gray entropy are associated, abnormal points are removed, the feed pile density is quickly and reversely deduced through image features, and an efficient non-contact feed throwing loss monitoring means is provided. Meanwhile, multi-source data and a dynamic correction mechanism are fused, a natural settlement quantitative model and a histogram normalization cross-correlation algorithm are combined, the residual material pile weight is corrected through the volume change rate, factors in multiple aspects are comprehensively considered, interference is eliminated, and the effective ingestion weight is accurately calculated.

Description

Technical Field

[0001] The present invention relates to the technical field of big data analysis, and specifically to a monitoring data analysis system and method for mutton sheep population production evaluation. Background Art

[0002] During the fattening process of mutton sheep, when the feed formula transitions from a high-concentrate diet to roughage, such as feeding hay, etc., a specific cycle of hay adaptation feeding stage is required to promote the gradual adjustment of the rumen microflora and digestive function. During this stage, due to the significant changes in the physical form, nutritional components, and palatability of the feed, it is easy to cause digestive stress and growth performance fluctuations in the mutton sheep population, and refined monitoring needs to be carried out through data such as feed intake and weight gain rate. However, during the hay feeding period, the sheep flock is prone to abnormal behaviors such as arching and picky eating, resulting in an increase in the feed spillage rate. At the same time, the loose characteristics of the hay exacerbate the natural settlement of the feed in the trough, making it difficult for the vision-based feed residue image analysis system to accurately identify the actual feed intake. And the trough weighing method has weight monitoring errors due to the mixing of spilled feed and residual feed in the trough, seriously restricting the implementation effect of the precise feeding management system. Summary of the Invention

[0003] The purpose of the present invention is to provide a monitoring data analysis system and method for mutton sheep population production evaluation to solve the problems raised in the prior art.

[0004] To achieve the above purpose, the present invention provides the following technical solution: A monitoring data analysis method for mutton sheep population production evaluation, the monitoring data analysis method includes the following steps:

[0005] Step S1, obtain the image data of the feed pile in the trough and the environmental data at the moment when the mutton sheep population starts to eat, and construct a feed pile error analysis set;

[0006] Step S1-1, select any trough where the mutton sheep population eats as the research object, and obtain the image data of the feed pile in the trough through a vision sensor;

[0007] Step S1-2, obtain the environmental data of the mutton sheep population eating through an environmental sensor, the environmental data includes temperature data and relative humidity data, and construct a feed pile error analysis set according to the obtained feed pile image data and environmental data.

[0008] By obtaining the image data of the feed pile in the trough and the environmental data to construct a feed pile error analysis set, it is possible to systematically collect multi-dimensional basic information in the initial stage of mutton sheep eating. The image data can visually present the shape of the feed pile, and the environmental data reflects external influencing factors.

[0009] Step S2: Calculate the initial volume of the feed pile based on the analysis of the feed pile image data in the feed pile error analysis set. Construct a natural settlement quantification model according to the environmental data in the feed pile error analysis set, and calculate the volume change rate of the feed pile in the trough through the natural settlement quantification model;

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

[0011]

[0012] In the formula, represents the volume change rate of the feed pile per unit time; V0 represents the initial volume of the feed pile in the trough; VS represents the volume of the feed pile in the trough at the current moment; a represents the temperature sensitivity coefficient; △T represents the temperature change value of the sheep group's foraging per unit time; b represents the relative humidity sensitivity coefficient; △RH represents the relative humidity change value of the sheep group's foraging per unit time.

[0013] Step S2-1: Analyze and extract the three-dimensional point cloud data on the surface of the feed pile in the trough according to the feed pile error analysis set, and map the point cloud data to a unified world coordinate system. The world coordinate system is represented as a global reference frame in three-dimensional space and is defined by 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: Screen out the vertices that form the convex hull boundary through polar angle sorting and cross-product judgment to form a closed triangular mesh surface model. Decompose the convex polyhedron into several tetrahedral units, and calculate the volume of each unit through geometric integration and accumulation, or use the orthographic projection method to calculate the difference in the projection volume of the upper and lower surfaces to calculate the initial volume of the feed pile;

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

[0016] For the vertices P j , u , j , u , i ,

[0016] , k , i , k , k , j , j , i , u , i , u , , k (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)The tetrahedron formed has a volume calculation formula as follows:

[0017]

[0018] Step S2-3: Obtain the natural settlement rate using the percentage of the volume change of the stockpile within a certain period to the initial volume. Then, introduce temperature data and relative humidity data, establish a compensation coefficient correction relationship, quantify the coupled influence of thermal expansion and particle adsorption effects on the settlement rate, and construct a natural settlement quantification model;

[0019] The establishment of the compensation coefficient correction relationship includes a temperature sensitivity coefficient and a relative humidity sensitivity coefficient; the temperature sensitivity coefficient represents the quantified influence of a unit temperature change on the natural settlement rate of the stockpile; the relative humidity sensitivity coefficient represents the dynamic correction of the particle adsorption and desorption effects caused by a unit humidity change. Keep the relative humidity constant, change the temperature and record the volume change, and 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, and fit the relationship between the volume change rate and the relative humidity change to obtain the relative humidity sensitivity coefficient.

[0020] Use the convex hull algorithm to convert the point cloud data on the surface of the stockpile into a three-dimensional geometric model, calculate the initial volume by decomposing tetrahedron units, and avoid the errors caused by irregular shapes in traditional measurement methods; secondly, when constructing the natural settlement quantification model, introduce the temperature sensitivity coefficient and the relative humidity sensitivity coefficient, quantify the influence of environmental factors on the settlement of the stockpile, and correct the volume change deviation caused by physical effects such as thermal expansion and contraction and particle adsorption. This method can comprehensively consider the geometric characteristics of the stockpile and environmental factors, and effectively improve the accuracy of calculating the volume change rate of the stockpile.

[0021] Step S3: Randomly sample the densities of the stockpiles in multiple feed troughs eaten by the meat sheep group, combine the analysis of the gray histogram entropy values corresponding to different stockpile densities with the stockpile image data, and construct a mapping set of the stockpile density entropy values in the feed trough; simultaneously use the construction method of the mapping set of the stockpile density entropy values in the feed trough to analyze the stockpiles spilled outside the feed trough, and construct a mapping set of the stockpile density entropy values outside the feed trough;

[0022] Step S3-1: Collect the stockpile image data in the feed trough under the normal feeding environment conditions of the meat sheep group through a standardized sampling device, and then simultaneously use a stockpile density measuring instrument to randomly sample and detect the density of the sampled stockpile, and perform grayscale preprocessing on the image;

[0023] Step S3-2: Extract the regional image data of the material pile detected by the bulk density meter, calculate the dispersion of the gray distribution of this region through histogram statistics, and quantify the entropy value of the gray histogram according to the information entropy theory; finally, establish a mapping relationship between the measured density data of each sampling point and its corresponding entropy value, and eliminate abnormal points through iterative optimization. The abnormal points are represented as gray mutation image data caused by uneven illumination. Construct a mapping set of the bulk density entropy values in the feeder according to the mapping relationship.

[0024] Step S3-3: Use the processing method of Step S3-1 to Step S3-2 to process the material pile spilled outside the feeder. Establish a mapping relationship between the measured density data of each sampling point and its corresponding entropy value. After iterative optimization to eliminate abnormal points, construct a mapping set of the bulk density entropy values outside the feeder according to the mapping relationship.

[0025] Obtain the measured bulk density data of the material pile through standardized sampling and the density meter. Combine the image data after grayscale preprocessing, use histogram statistics and information entropy theory to quantify the dispersion of the gray distribution of the material pile, and establish a mapping relationship between the density data and the entropy value of the gray histogram, effectively establishing the connection between the physical properties of the material pile and the image features; secondly, when constructing the mapping sets of the bulk density entropy values inside and outside the feeder, eliminate abnormal data caused by interference factors such as illumination through iterative optimization to ensure the accuracy and reliability of the mapping relationship. This method can quickly reverse the bulk density using the gray features of the image, providing an intuitive and efficient analysis method for monitoring feed spillage loss, and realizing synchronous monitoring and quantitative evaluation of the feed status inside and outside the feeder.

[0026] Step S4: Obtain the image data of the material pile inside the feeder that the meat sheep group is eating at the current moment, and analyze and calculate the remaining weight of the material pile in the feeder in combination with the mapping set of the bulk density entropy values inside the feeder, which is recorded as the remaining bulk weight in the feeder; obtain the image data of the spilled material pile outside the feeder that the meat sheep group is eating at the current moment, and analyze and calculate the weight of the spilled material pile in the feeder in combination with the mapping set of the bulk density entropy values outside the feeder, which is recorded as the spilled bulk weight.

[0027] Step S4-1: Obtain the image data of the material pile inside the feeder through the vision sensor, and perform image analysis and matching according to the obtained material pile image data and the mapping set of the bulk density entropy values inside the feeder. The specific image analysis and matching is as follows: Through the histogram normalized cross-correlation algorithm, compare the entropy value of the real-time acquired image with the entropy value in the mapping set, and obtain the density data of the material pile according to the mapping relationship between the measured density data and its corresponding entropy value.

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

[0029]

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

[0031] Step S4-2: Calculate the volume of the feed pile in the trough by combining the convex hull algorithm, and multiply the density data by the volume to calculate the weight of the remaining feed pile in the trough, denoted as the weight of the remaining feed pile in the trough;

[0032] Step S4-3: Adopt the calculation method of Steps S4-1 to S4-2 to perform image analysis and matching on the image data of the spilled feed pile outside the trough to obtain density data, calculate the volume of the spilled feed pile outside the trough according to the convex hull algorithm, and multiply the density data by the volume to calculate the weight of the spilled feed pile outside the trough, denoted as the weight of the spilled feed pile.

[0033] In terms of calculating the weight of the feed pile in the trough, image data is obtained by using a vision sensor, and the real-time image entropy value is compared with the mapping set of the entropy values of the feed pile density in the trough by means of the histogram normalization cross-correlation algorithm to accurately obtain the feed pile density data. Then, combined with the volume calculated by the convex hull algorithm, the weight of the remaining feed pile in the trough is accurately obtained; for the spilled feed pile outside the trough, the same calculation process is adopted to calculate the weight of the spilled feed pile. This method combines image feature analysis, density mapping and geometric volume calculation, breaks through the limitations of traditional weight measurement methods, and can monitor the weight changes of the feed inside and outside the trough in real time and accurately during the feeding process of meat sheep.

[0034] Step S5: Input the data of the feed pile error analysis set, the current collected image data of the feed pile in the trough and the environmental data into the natural settlement quantization model to calculate the volume change rate of the feed pile in the trough, use the calculated volume change rate as the time decay factor to correct the weight of the remaining feed pile in the trough, obtain the corrected remaining feed pile weight, and calculate the effective feeding weight of the meat sheep group according to the initially put feed pile weight, the corrected remaining feed pile weight and the weight of the spilled feed pile.

[0035] Input the data of the feed pile error analysis set, the real-time image and the environmental parameters into the natural settlement model, calculate the volume change rate of the feed pile per unit time as the time decay factor, convert the volume loss ratio into the weight decay coefficient, multiply the weight of the remaining feed pile in the trough by the decay coefficient to obtain the corrected remaining feed pile weight after eliminating the natural settlement error, and use the initially put feed pile weight minus the corrected remaining feed pile weight and the weight of the spilled feed pile to obtain the effective feeding weight of the meat sheep group;

[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 represents the weight attenuation coefficient; represents the volume change rate of the material pile per unit time; S pace represents the duration per unit time.

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

[0040] Furthermore, a monitoring data analysis system applied to the production evaluation of mutton sheep populations, the monitoring data analysis system includes a data acquisition module, a volume modeling module, a density mapping module, a weight measurement module and an effective feeding calculation module;

[0041] The data acquisition module is used to obtain data related to the feeding of mutton sheep populations; the volume modeling module is used to construct a model related to the volume 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 measurement module is used to calculate the weights of the material piles inside and outside the feeding trough; the effective feeding calculation module is used to correct the data and calculate the effective feeding weight of the mutton sheep population;

[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 measurement module; the output end of the weight measurement module is electrically connected to the input end of the effective feeding calculation module;

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

[0044] The volume modeling module includes 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 material pile surface and map it to the world coordinate system for modeling; the volume calculation and settlement unit is used to calculate the initial volume of the material pile and construct a natural settlement quantification model;

[0045] The density mapping module includes an in-feeder sampling and mapping unit and an out-of-feeder sampling and mapping unit; the in-feeder sampling and mapping unit is used to sample the material pile in the feeder and establish a mapping relationship between density and entropy value; the out-of-feeder sampling and mapping unit is used to sample the spilled material pile outside the feeder and establish a mapping relationship between density and entropy value;

[0046] The weight measurement module includes an in-feeder weight calculation unit and an out-of-feeder weight calculation unit; the in-feeder weight calculation unit is used to calculate the weight of the remaining material pile in the feeder by combining the mapping set and the convex hull algorithm; the out-of-feeder weight calculation unit is used to calculate the weight of the spilled material pile outside the feeder by combining the mapping set and the convex hull algorithm;

[0047] The effective forage intake calculation module includes a data correction unit and a forage intake calculation unit; the data correction unit is used to calculate the volume change rate by using the natural settlement quantification model to correct the weight of the remaining material pile; the forage intake calculation unit is used to calculate the effective forage intake weight of the sheep flock according to the initial feeding weight and the corrected weight.

[0048] Compared with the prior art, the beneficial effects of the present invention are:

[0049] 1. By constructing a material pile error analysis set, combining the convex hull algorithm and the multi-factor coupling model, the present invention accurately calculates the initial volume of the material pile and the natural settlement quantification model. The convex hull algorithm is used to process the point cloud data to avoid the volume calculation error of irregular material piles. At the same time, temperature and humidity sensitivity coefficients are introduced to quantify the influence of environmental factors on settlement, effectively solving the problem of volume misjudgment caused by natural settlement in traditional methods.

[0050] 2. By establishing a mapping set of density entropy values of the material piles inside and outside the feeder, the present invention correlates the physical density of the material pile with the image gray entropy value. Based on the standardized sampling and information entropy theory, the interference of abnormal data points is eliminated, and the density of the material pile is quickly deduced through image features, breaking through the limitation of traditional contact measurement, providing an efficient non-contact analysis method for monitoring feed spillage loss, and improving the real-time performance and accuracy of monitoring the feed state inside and outside the feeder.

[0051] 3. By integrating multi-source data with a dynamic correction mechanism, the present invention accurately calculates the effective feeding weight of a flock of meat sheep. By combining a natural sedimentation quantification model with a histogram normalization cross-correlation algorithm, the remaining pile weight is corrected using the volume change rate, and the initial feeding amount, correction margin, and spillage weight are comprehensively considered to eliminate the interference of natural sedimentation and spillage, solve the problem of large errors in traditional feeding weight calculation, provide reliable data support for precise feeding management, and help improve the breeding efficiency of meat sheep. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a schematic flowchart of a monitoring data analysis method for the production evaluation of a flock of meat sheep according to the present invention;

[0053] Figure 2 It is a schematic structural diagram of a monitoring data analysis system for the production evaluation of a flock of meat sheep according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0054] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0055] Embodiment 1: As Figure 1 shown, the present invention provides a technical solution, a monitoring data analysis method for the production evaluation of a flock of meat sheep, and the monitoring data analysis method includes the following steps:

[0056] Step S1. Obtain the image data and environmental data of the feed pile in the feeder at the moment when the flock of meat sheep starts to feed, and construct a feed pile error analysis set;

[0057] Step S1-1. Select any feeder where a flock of meat sheep feeds as the research object, and obtain the image data of the feed pile in the feeder through a vision sensor;

[0058] Step S1-2. Obtain the environmental data of the flock of meat sheep feeding through an environmental sensor, where the environmental data includes temperature data and relative humidity data, and construct a feed pile error analysis set according to the obtained feed pile image data and environmental data.

[0059] In specific implementation, at the moment when the meat sheep starts eating, image data of the feed pile in the feeder and environmental data are collected, aiming to obtain multi-dimensional information on the initial state of feed feeding. The image data collected by the vision sensor can record the original shape, contour and surface features of the feed pile, and the temperature and humidity data obtained by the environmental sensor are used for subsequent analysis of the influence of environmental factors on the feed state. Constructing a feed pile error analysis set is to integrate these basic data, providing the original basis for subsequent calculation of the feed pile volume, analysis of natural settlement, etc., ensuring that subsequent analysis is based on accurate initial conditions; the vision sensor needs to be reasonably installed to ensure that the shooting angle can completely cover the feed pile, avoiding missing image information due to visual occlusion; the environmental sensor should be placed near the feeder and in a position that can accurately reflect the environment of the meat sheep's feeding area to prevent deviation of environmental data; the data collection time needs to be accurately positioned at the moment when the meat sheep starts eating, avoiding data interference during the eating process due to time error and affecting the judgment of the initial state.

[0060] Step S2: Analyze and calculate the initial volume of the feed pile based on the feed pile image data in the feed pile error analysis set, construct a natural settlement quantification model according to the environmental data in the feed pile error analysis set, and calculate the volume change rate of the feed pile in the feeder through the natural settlement quantification model;

[0061] Step S2-1: Analyze and extract the three-dimensional point cloud data on the surface of the feed pile in the feeder according to the feed pile error analysis set, and map the point cloud data to a unified world coordinate system, where the world coordinate system is represented as a global reference frame in three-dimensional space, defined by three mutually perpendicular and intersecting coordinate axes;

[0062] 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: screen out the vertices forming the convex hull boundary through polar angle sorting and cross product judgment to form a closed triangular mesh surface model, decompose the convex polyhedron into several tetrahedral units, accumulate the volumes of each unit through geometric integration, or calculate the difference in projected volumes of the upper and lower surfaces using the orthographic projection method to calculate the initial volume of the feed pile;

[0063] Step S2-3: Obtain the natural settlement rate using the percentage of the volume change of the feed pile over a period of time to the initial volume, then introduce temperature data and relative humidity data, establish a compensation coefficient correction relationship, quantify the coupled influence of thermal expansion and particle adsorption effects on the settlement rate, and construct a natural settlement quantification model;

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

[0065] In specific implementation, by extracting the three-dimensional point cloud data of the material pile surface and mapping it to the world coordinate system, the shape of the material pile can be transformed into quantifiable geometric data. The convex hull algorithm constructs the minimum 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 handle the irregular shape of the material pile to accurately obtain the initial volume. Constructing a natural settlement quantification model in combination with environmental data takes into account that temperature and humidity will affect the physical properties of feed particles, such as thermal expansion and contraction, and the change in weight and volume caused by water absorption. By establishing a compensation coefficient correction relationship, the influence of these factors on the settlement rate is quantified to achieve accurate calculation of the volume change rate of the material pile; the algorithm needs to be reasonably optimized or the data needs to be downsampled. When determining the temperature and humidity sensitivity coefficients, the experimental conditions need to strictly control a single variable to ensure that the coefficients can truly reflect the influence of the corresponding environmental factors on the settlement.

[0066] Step S3: Randomly sample the material pile densities in multiple feed troughs where the meat sheep group feeds, combine the analysis of the gray histogram entropy values corresponding to different material pile densities with the material pile image data, and construct a mapping set of the material pile density entropy values in the feed trough; simultaneously use the construction method of the mapping set of the material pile density entropy values in the feed trough to analyze the material piles spilled outside the feed trough, and construct a mapping set of the material pile density entropy values outside the feed trough.

[0067] Step S3-1: Collect the material pile image data in the feed trough under the conventional feeding environment conditions of the meat sheep group through a standardized sampling device, and then simultaneously use a pile density measuring instrument to randomly sample and detect the density of the sampled material pile, and perform grayscale preprocessing on the image.

[0068] Step S3-2: Extract the regional image data of the material pile detected by the pile density measuring instrument, calculate the dispersion degree of the gray distribution of this region through histogram statistics, and quantify the entropy value of the gray histogram according to the information entropy theory; finally, establish a mapping relationship between the measured density data of each sampling point and its corresponding entropy value, and eliminate the abnormal points through iterative optimization. The abnormal points are represented as the gray mutation image data caused by uneven illumination, and construct a mapping set of the material pile density entropy values in the feed trough according to the mapping relationship.

[0069] Step S3-3: Use the processing method of steps S3-1 to S3-2 to process the material piles spilled outside the feed trough. Establish a mapping relationship between the measured density data at each sampling point and its corresponding entropy value. After iterative optimization to eliminate abnormal points, construct a density-entropy value mapping set of the material pile outside the feed trough according to the mapping relationship.

[0070] In specific implementation, based on the characteristic that there is a correlation between the image gray feature and the material pile density, sample the material piles inside and outside the feed trough, measure the actual density, and analyze the entropy value of the corresponding image gray histogram to establish the mapping relationship between the two. The entropy value of the gray histogram can reflect the discreteness of the gray distribution of the image. Material piles with different densities have different gray features on the image due to differences in surface particle arrangement, voids, etc., and thus correspond to different entropy values. After constructing the mapping set, the material pile density can be quickly calculated by analyzing the image entropy value in the follow-up, realizing non-contact monitoring of the feed state; the standardized sampling device needs to ensure the randomness and representativeness of sampling to avoid inaccurate mapping relationships caused by sampling deviations; when measuring with a bulk density measuring instrument, it is necessary to ensure that the measurement method is standardized and the operation is accurate to prevent measurement errors.

[0071] Step S4: Obtain the image data of the material pile inside the feed trough where the meat sheep group is eating at the current moment, and analyze and calculate the remaining material pile weight in the feed trough in combination with the density-entropy value mapping set of the material pile inside the feed trough, denoted as the remaining material pile weight in the feed trough; obtain the image data of the spilled material pile outside the feed trough where the meat sheep group is eating at the current moment, and analyze and calculate the weight of the spilled material pile in the feed trough in combination with the density-entropy value mapping set of the material pile outside the feed trough, denoted as the spilled material pile weight.

[0072] Step S4-1: Obtain the image data of the material pile inside the feed trough through a vision sensor, and perform image analysis and matching according to the obtained material pile image data and the density-entropy value mapping set of the material pile inside the feed trough. The specific image analysis and matching is as follows: Through the histogram normalization cross-correlation algorithm, compare the entropy value of the real-time collected image with the entropy value in the mapping set, and obtain the density data of the material pile according to the mapping relationship between the measured density data and its corresponding entropy value.

[0073] Step S4-2: Calculate the volume of the material pile inside the feed trough in combination with the convex hull algorithm, and multiply the density data by the volume to calculate the remaining material pile weight in the feed trough, denoted as the remaining material pile weight in the feed trough.

[0074] Step S4-3: Use the calculation method of steps S4-1 to S4-2 to perform image analysis and matching on the image data of the spilled material pile outside the feed trough to obtain density data, calculate the volume of the spilled material pile outside the feed trough according to the convex hull algorithm, and multiply the density data by the volume to calculate the weight of the spilled material pile outside the feed trough, denoted as the spilled material pile weight.

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

[0076] Step S5: Input the data in the feed pile error analysis set, the image data of the feed pile inside the feeding trough currently collected, and the environmental data into the natural settlement quantification model to calculate the volume change rate of the feed pile inside the feeding trough. The calculated volume change rate is used as the time decay factor to correct the remaining weight of the feed pile inside the feeding trough, obtain the corrected remaining weight of the feed pile, and calculate the effective feeding weight of the meat sheep group according to the initially put - in feed pile weight, the corrected remaining weight of the feed pile, and the spilled feed pile weight.

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

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

[0079]

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

[0081] In specific implementation, input the data in the feed pile error analysis set, the real - time image, and the environmental parameters into the natural settlement quantification model to obtain the volume change rate of the feed pile inside the feeding trough. This change rate reflects the volume loss of the feed pile caused by natural settlement and other factors. It is used as the time decay factor to be converted into the weight decay coefficient to correct the remaining weight of the feed pile inside the feeding trough, eliminate the weight misjudgment caused by natural settlement, and combine the initially put - in weight and the spilled feed pile weight to finally obtain the actual effective feeding weight of the meat sheep group, ensuring the accuracy of the feeding data.

[0082] Example two, as Figure 2As shown in the figure, the present invention provides a monitoring data analysis system for the production evaluation of a flock of meat sheep. The monitoring data analysis system includes a data acquisition module, a volume modeling module, a density mapping module, a weight calculation module, and an effective forage intake calculation module;

[0083] The data acquisition module is used to obtain data related to the forage intake of the flock of meat sheep; the volume modeling module is used to construct a model related to the volume of the feed pile and calculate the volume of the feed pile; the density mapping module is used to establish a mapping relationship between the density of the feed pile and the entropy value; the weight calculation module is used to calculate the weights of the feed piles inside and outside the trough; the effective forage intake calculation module is used to correct the data and calculate the effective forage intake weight of the flock of meat sheep;

[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; the output end of the weight calculation module is electrically connected to the input end of the effective forage intake calculation module;

[0085] The data acquisition module includes a basic data acquisition unit and an environmental parameter acquisition unit; the basic data acquisition unit is used to obtain image data of the feed pile inside the trough through a vision sensor; the environmental parameter acquisition unit is used to collect temperature data and relative humidity data of the forage intake environment of the meat sheep through an environmental sensor;

[0086] The volume modeling module includes 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 on the surface of the feed pile and map it to the world coordinate system for modeling; the volume calculation and settlement unit is used to calculate the initial volume of the feed pile and construct a natural settlement quantification model;

[0087] The density mapping module includes an in-trough sampling mapping unit and an out-of-trough sampling mapping unit; the in-trough sampling mapping unit is used to sample the feed pile inside the trough and establish a mapping relationship between density and entropy value; the out-of-trough sampling mapping unit is used to sample the spilled feed pile outside the trough and establish a mapping relationship between density and entropy value;

[0088] The weight calculation module includes an in-trough weight calculation unit and an out-of-trough weight calculation unit; the in-trough weight calculation unit is used to calculate the weight of the remaining feed pile inside the trough by combining the mapping set and the convex hull algorithm; the out-of-trough weight calculation unit is used to calculate the weight of the spilled feed pile outside the trough by combining the mapping set and the convex hull algorithm;

[0089] The effective feeding calculation module includes a data correction unit and a feeding calculation unit; the data correction unit is used to calculate the volume change rate by using the natural sedimentation quantification model to correct the weight of the remaining feed pile; the feeding calculation unit is used to calculate the effective feeding weight of the meat sheep group according to the initial feeding weight and the corrected weight.

[0090] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above-described exemplary embodiments, and without departing from the spirit or basic characteristics of the present invention, the present invention can be implemented in other specific forms. Therefore, in any sense, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present invention. Any reference signs in the claims should not be construed as limiting the claims involved.

Claims

1. A monitoring data analysis method applied to the production assessment of a flock of meat sheep, characterized in that: The monitoring data analysis method includes the following steps: Step S1: Obtain the image data and environmental data of the feed pile in the trough at the moment when the meat sheep group starts eating, and construct a feed pile error analysis set; Step S2: Calculate the initial volume of the feed pile according to the analysis of the feed pile image data in the feed pile error analysis set, construct a natural settlement quantification model based on the environmental data in the feed pile error analysis set, and calculate the volume change rate of the feed pile in the trough through the natural settlement quantification model; Step S3: Randomly sample the feed pile densities in multiple troughs where the meat sheep group eats, combine the analysis of the feed pile image data to analyze the gray histogram entropy values corresponding to different feed pile densities, and construct a feed pile density entropy value mapping set in the trough; simultaneously analyze the feed pile spilled outside the trough using the construction method of the feed pile density entropy value mapping set in the trough, and construct a feed pile density entropy value mapping set outside the trough; Step S4: Obtain the image data of the feed pile in the trough where the meat sheep group eats at the current moment, combine the feed pile density entropy value mapping set in the trough to analyze and calculate the remaining weight of the feed pile in the trough, denoted as the remaining feed pile weight in the trough; obtain the image data of the feed pile spilled outside the trough where the meat sheep group eats at the current moment, combine the feed pile density entropy value mapping set outside the trough to analyze and calculate the weight of the feed pile spilled in the trough, denoted as the spilled feed pile weight; Step S5: Input the data in the feed pile error analysis set, the image data and environmental data of the feed pile in the trough currently collected into the natural settlement quantification model to calculate the volume change rate of the feed pile in the trough, use the calculated volume change rate as the time decay factor to correct the remaining weight of the feed pile in the trough, obtain the corrected remaining feed pile weight, and calculate the effective eating weight of the meat sheep group according to the initially placed feed pile weight, the corrected remaining feed pile weight and the spilled feed pile weight.

2. The monitoring data analysis method for mutton sheep population production evaluation according to claim 1, characterized in that: The specific steps of step S1 are as follows: Step S1-1: Select any trough where a meat sheep group eats as the research object, and obtain the image data of the feed pile in the trough through a visual sensor; Step S1-2: Obtain the environmental data of the meat sheep group eating through an environmental sensor, where the environmental data includes temperature data and relative humidity data, and construct a feed pile error analysis set according to the obtained feed pile image data and environmental data.

3. A monitoring data analysis method applied to the production evaluation of a group of meat sheep according to claim 2, characterized in that: The specific steps of step S2 are as follows: Step S2-1: Analyze and extract the three-dimensional point cloud data of the surface of the feed pile in the trough according to the feed pile error analysis set, and map the point cloud data to a unified world coordinate system, where the world coordinate system is represented as a global reference frame in three-dimensional space, defined by three mutually perpendicular and intersecting coordinate axes; 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: screen out the vertices forming the convex hull boundary through polar angle sorting and cross product judgment to form a closed triangular mesh surface model, decompose the convex polyhedron into several tetrahedral units, and calculate the initial volume of the feed pile by geometric integration to accumulate the volume of each unit or by using the positive projection method to calculate the volume difference of the upper and lower surface projections.

4. A monitoring data analysis method applied to the production assessment of a flock of meat sheep according to claim 3, characterized in that: In step S2, it also includes: Step S2-3: Obtain the natural settlement rate using the percentage of the volume change of the stockpile over a period of time to the initial volume. Then, introduce temperature data and relative humidity data, establish a compensation coefficient correction relationship, quantify the coupled effects of thermal expansion and particle adsorption on the settlement rate, and construct a natural settlement quantification model; Step S2-4: The establishment of the compensation coefficient correction relationship includes a temperature sensitivity coefficient and a relative humidity sensitivity coefficient; the temperature sensitivity coefficient represents the quantified impact of a unit temperature change on the natural settlement rate of the stockpile; the relative humidity sensitivity coefficient represents the dynamic correction of the particle adsorption and desorption effects by a unit humidity change. Keep the relative humidity constant, change the temperature and record the volume change, and 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, and fit the relationship between the volume change rate and the relative humidity change to obtain the relative humidity sensitivity coefficient.

5. A monitoring data analysis method applied to the production evaluation of meat sheep populations according to claim 4, characterized in that: The specific steps of step S3 are as follows: Step S3-1: Collect the image data of the stockpile in the feeding trough under the conventional feeding environment conditions of the meat sheep population through a standardized sampling device. Then, synchronously use a bulk density measuring instrument to conduct random sampling density detection on the sampled stockpile, and perform grayscale preprocessing on the image; Step S3-2: Extract the regional image data of the stockpile detected by the bulk density measuring instrument, calculate the dispersion of the grayscale distribution in this area through histogram statistics, and quantify the entropy value of the grayscale histogram according to the information entropy theory. Finally, establish a mapping relationship between the measured density data of each sampling point and its corresponding entropy value, and eliminate abnormal points through iterative optimization. The abnormal points are represented as the grayscale mutation image data caused by uneven illumination. Construct a density entropy value mapping set of the stockpile in the feeding trough according to the mapping relationship; Step S3-3: Adopt the processing method of steps S3-1 to S3-2 to process the stockpile spilled outside the feeding trough. Establish a mapping relationship between the measured density data of each sampling point and its corresponding entropy value, and construct a density entropy value mapping set of the stockpile outside the feeding trough according to the mapping relationship after eliminating abnormal points through iterative optimization.

6. The monitoring data analysis method for mutton sheep population production evaluation according to claim 5, characterized in that: The specific steps of step S4 are as follows: Step S4-1: Obtain the image data of the stockpile in the feeding trough through a vision sensor, and perform image analysis and matching based on the obtained stockpile image data and the density entropy value mapping set of the stockpile in the feeding trough. The specific image analysis and matching are as follows: Through the histogram normalization cross-correlation algorithm, compare the entropy value of the real-time collected image with the entropy value in the mapping set, and obtain the density data of the stockpile according to the mapping relationship between the measured density data and its corresponding entropy value; Step S4-2: Combine the convex hull algorithm to calculate the volume of the stockpile in the feeding trough, and multiply the density data by the volume to calculate the remaining weight of the stockpile in the feeding trough, which is recorded as the remaining weight of the stockpile in the feeding trough; Step S4-3: Adopt the calculation method of steps S4-1 to S4-2 to perform image analysis and matching on the image data of the spilled stockpile outside the feeding trough to obtain density data, and calculate the volume of the spilled stockpile outside the feeding trough according to the convex hull algorithm. Multiply the density data by the volume to calculate the weight of the spilled stockpile outside the feeding trough, which is recorded as the weight of the spilled stockpile.

7. The monitoring data analysis method applied to the production evaluation of meat sheep groups according to claim 6, characterized in that: In step S5, the data of the stockpile error analysis set, the real-time image, and the environmental parameters are input into the natural settlement model to calculate the volume change rate of the stockpile per unit time as the time decay factor. The volume loss ratio is converted into a weight decay coefficient, and the remaining weight of the stockpile in the trough is multiplied by the decay coefficient to obtain the corrected remaining weight of the stockpile after eliminating the natural settlement error. The effective feeding weight of the meat sheep group is obtained by subtracting the corrected remaining weight of the stockpile from the initially placed weight of the stockpile and adding the weight of the spilled stockpile. The calculation formula for converting the volume loss ratio into a weight decay coefficient is as follows: In the formula, Ddecay represents the weight decay coefficient; VSpace represents the volume change rate of the stockpile per unit time; Space represents the duration per unit time.

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

9. The monitoring data analysis system applied to the production evaluation of a meat sheep group according to claim 8, wherein: The data acquisition module includes a basic data acquisition unit and an environmental parameter acquisition unit; the basic data acquisition unit is used to obtain the image data of the stockpile in the trough through a vision sensor; the environmental parameter acquisition unit is used to collect the temperature data and relative humidity data of the meat sheep feeding environment through an environmental sensor. The volume modeling module includes 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 stockpile surface and map it to the world coordinate system for modeling; the volume calculation and settlement unit is used to calculate the initial volume of the stockpile and construct a natural settlement quantification model. The density mapping module includes an in-trough sampling mapping unit and an out-of-trough sampling mapping unit; the in-trough sampling mapping unit is used to sample the stockpile in the trough and establish a mapping relationship between the density and the entropy value; the out-of-trough sampling mapping unit is used to sample the spilled stockpile outside the trough and establish a mapping relationship between the density and the entropy value.

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

Images