CN111695614B - Dynamic monitoring sensor layout and multi-source information fusion method and system - Google Patents

Abstract

The embodiment of the invention provides a method and a system for fusing sensor layout and multi-source information for dynamic monitoring. The method comprises the following steps: optimizing layout of the environmental sensor and the physiological sensor according to a preset sensor layout model; the sensor layout model is constructed based on a two-dimensional plan view and the relation between the positions of the environmental sensor and the physiological sensor; performing periodic information extraction data on the optimized environmental sensor and physiological sensor, and performing data fusion on the verified data by using a preset multi-source information fusion model to obtain a three-level dynamic evaluation decision; the multi-source information fusion model is obtained based on combination of homogenous sensor data fusion and heterogeneous sensor data fusion. The embodiment of the invention realizes reasonable layout of the sensors, and carries out three-level dynamic evaluation decision of breeding individuals, breeding groups and breeding groups on living animals by utilizing a multisource information fusion model constructed in advance.

Description

A sensor layout and multi-source information fusion method and system for dynamic monitoring

technical field

The invention relates to the technical field of breeding and multi-source information fusion, in particular to a dynamic monitoring sensor layout and multi-source information fusion method and system.

Background technique

At present, when many researchers use multiple sensors to monitor and collect relevant data, they mostly place the relevant sensors roughly based on subjective consciousness. However, randomly placed sensors are not accurate enough for data monitoring and collection. On the one hand, it will cause waste of resources and require more sensors; on the other hand, it will cause inaccuracy of data and redundancy of collected data. Difficulties in subsequent data processing.

There are more or less problems in the existing sensor-related data monitoring and collection methods. Therefore, how to propose a method that can achieve a reasonable layout of sensors and integrate multi-source information to "farming individuals". - Breeding group-breeding population" for three-level dynamic evaluation and decision-making has become an urgent problem to be solved.

Contents of the invention

The purpose of the embodiments of the present invention is to provide a sensor layout and multi-source information fusion method and system for dynamic monitoring that overcome the above problems or at least partially solve the above problems.

In order to solve the above technical problems, on the one hand, an embodiment of the present invention provides a dynamic monitoring sensor layout and multi-source information fusion method, including:

Optimizing the layout of the environmental sensors and the physiological sensors according to a preset sensor layout model; wherein the sensor layout model is constructed based on a two-dimensional plan and the relationship between the positions of the environmental sensors and the physiological sensors;

Periodically extract data from optimized environmental sensors and physiological sensors, and use a preset multi-source information fusion model to perform data fusion on the verified data to obtain a three-level dynamic evaluation decision; wherein, the multi-source information The fusion model is based on the combination of homogeneous sensor data fusion and heterogeneous sensor data fusion.

Further, before optimizing the layout of environmental sensors and physiological sensors according to the preset sensor layout model, it also includes:

Dimensionality reduction is performed on the three-dimensional space living environment and the three-dimensional geometric features of the living animal to obtain the two-dimensional plan view.

Further, the optimized layout of environmental sensors and physiological sensors according to the preset sensor layout model specifically includes:

Preliminarily matching the position of the environmental sensor, the position of the physiological sensor and the two-dimensional plan to obtain a matching set;

Calculate and obtain the maximum monitoring ratio of the environmental sensor and the physiological sensor according to the matching set and the constraint variable;

The optimal layout is determined according to the maximum monitoring ratio of the sensor.

Further, before performing data fusion on the verified data using a preset multi-source information fusion model, it also includes:

Preliminary fusion of homogeneous sensor data according to data spatiotemporal weights to obtain a variety of homogeneous data;

The optimization weight is determined according to the relative time relationship, and heterogeneous sensor data fusion is performed on the various homogeneous data to construct the multi-source information fusion model.

Further, the preliminary fusion of homogeneous sensor data according to data spatiotemporal weights specifically includes:

calculating the data spatiotemporal weight according to the data spatial importance and the data temporal importance, and performing the homogeneous sensor data fusion processing according to the data spatiotemporal weight to obtain the various homogeneous data;

The data space importance is:

∑ l _n is the total number contained in the data set at time t _m , k _{m, n} is the number of occurrences of a certain data value l _{m, n} at time t _m ;

The time importance of the data is:

Y _{μ, v} is the v data contained in the μ time period, n is the number of sensors, k′ _{m, n} is the number of occurrences of the data value l _{m, n} in the μ time period;

The data spatiotemporal weight reconstruction model is:

Further, the determination of the optimization weight according to the relative time relationship to perform heterogeneous sensor data fusion on the various homogeneous data specifically includes:

Set an Euclidean space, there are optimization variables, weight optimizers, and classification optimizers;

converting the relative time relationship required by the optimal path found by the weight optimizer into the optimization weight;

The classification optimizer classifies the optimization variables, and performs heterogeneous sensor data fusion according to the classification results and the optimization weights.

Further, the three-level dynamic evaluation decision-making specifically includes:

Breeding individual health assessment: Perform data fusion on the physiological information data of a single animal according to the multi-source information fusion model to obtain individual data fusion information and individual health assessment decisions;

Breeding group situation assessment: performing data fusion of the individual data fusion information and the physiological information of the current animal group according to the multi-source information fusion model, to obtain group data fusion information and group situation assessment decisions;

Prediction and evaluation of cultured populations: performing data fusion on the population data fusion information and the processed environmental sensor data according to the multi-source information fusion model to obtain population prediction and evaluation decisions.

Further, before the forecast assessment of the breeding population, it also includes:

The group data fusion information is pre-classified, and the data fusion is performed according to the data weight ranking of different groups.

Further, it also includes:

After each level of fusion is completed, the fusion information data will be subjected to data quality assessment and data diagnosis, and if the data quality assessment and data diagnosis pass, it will be used as the next level of input data.

On the other hand, an embodiment of the present invention provides a dynamic monitoring sensor layout and multi-source information fusion system, including:

An optimization module: used to optimize the layout of the environmental sensors and the physiological sensors according to the preset sensor layout model; wherein, the sensor layout model is constructed based on a two-dimensional plan and the relationship between the positions of the environmental sensors and the physiological sensors of;

Evaluation module: used to periodically extract data from the optimized environmental sensor and physiological sensor, and use the preset multi-source information fusion model to perform data fusion on the verified data to obtain a three-level dynamic evaluation decision; among them, The multi-source information fusion model is obtained based on the combination of homogeneous sensor data fusion and heterogeneous sensor data fusion.

The sensor layout and multi-source information fusion method and system for dynamic monitoring provided by the embodiments of the present invention realize the sensor layout by adopting the preset sensor layout model and processing the obtained data through the multi-source information fusion model. The sensors are rationally arranged, and three-level dynamic evaluation and decision-making of "breeding individual-breeding group-breeding population" is realized.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description These are some embodiments of the present invention. Those skilled in the art can also obtain other drawings based on these drawings without creative work.

FIG. 1 is a schematic flow diagram of a dynamic monitoring sensor layout and a multi-source information fusion method provided by an embodiment of the present invention;

FIG. 2 is a schematic diagram of a two-dimensional plan generated by dimensionality reduction provided by an embodiment of the present invention;

FIG. 3 is a schematic flow diagram of a multi-sensor optimization layout process provided by an embodiment of the present invention;

FIG. 4 is a schematic diagram of a three-level dynamic evaluation decision-making process provided by an embodiment of the present invention;

FIG. 5 is a schematic flow diagram of a flow chart of multi-sensor data fusion provided by an embodiment of the present invention;

FIG. 6 is a schematic flow diagram of a dynamic monitoring sensor layout and multi-source information fusion system provided by an embodiment of the present invention;

7 is a schematic flowchart of another dynamic monitoring sensor layout and multi-source information fusion system provided by an embodiment of the present invention;

FIG. 8 is a schematic diagram of a physical structure of an electronic device provided by an embodiment of the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments It is a part of embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

An embodiment of the present invention provides a dynamic monitoring sensor layout and multi-source information fusion method. FIG. 1 is a schematic flowchart of the dynamic monitoring sensor layout and multi-source information fusion method provided by the embodiment of the present invention. As shown in FIG. 1 , the Methods include:

Step S101. Optimizing the layout of environmental sensors and physiological sensors according to a preset sensor layout model; wherein, the sensor layout model is constructed based on a two-dimensional plan and the relationship between the positions of the environmental sensors and physiological sensors;

Specifically, according to the dynamic monitoring sensor layout and multi-source information fusion method of the above-mentioned embodiment, in the above-mentioned step S101, the basic information of the farm such as area, shape, etc. are collected, and the basic information of living animals such as quantity and size are collected. The three-dimensional living environment of animals and the three-dimensional geometric characteristics of living animals are reduced to obtain a two-dimensional plan; use the pre-built sensor layout model to optimize the layout of the farm environment sensors, and use the preset animal density model to calculate the monitoring area based on the farm information. The number of animals, and then use the pre-built sensor layout model to perform physiological sensor layout on the batch of live animals;

The sensor layout model is constructed based on the two-dimensional plan, the relationship between the positions of the environmental sensors and the physiological sensors. The sensor layout model adds certain constraints to the sensors to avoid multiple sensors monitoring the same grid or not all sensors are monitored. The entire two-dimensional plane, but the monitoring ratio is greater than 100%, and considering the impact of the size of the breeding environment, animal size, number of animals, and sensor monitoring mechanism on the layout of the living animal physiological sensor, by combining the environmental sensor with the Preliminarily match the position of the physiological sensor with the two-dimensional plan to obtain a matching set, calculate the maximum monitoring ratio between the environmental sensor and the physiological sensor according to the matching set and constraint variables, and obtain the optimization of the sensor according to the maximum monitoring ratio of the sensor layout.

Step S102, periodically extract data from the optimized environmental sensor and physiological sensor, and use the preset multi-source information fusion model to perform data fusion on the verified data to obtain a three-level dynamic evaluation decision; wherein, the The multi-source information fusion model is based on the combination of homogeneous sensor data fusion and heterogeneous sensor data fusion.

Specifically, according to the dynamic monitoring sensor layout and multi-source information fusion method of the above embodiment, in the above step S102, periodic information extraction is performed on the environmental sensor and physiological sensor data based on the pre-built sensor layout model layout, and the verification The passed data uses the pre-built multi-source information fusion model for data fusion to make a three-level dynamic evaluation decision on "breeding individual-breeding group-breeding population";

Preliminary fusion of homogeneous sensor data according to data spatiotemporal weights to obtain a variety of homogeneous data; then determine the optimization weight according to the relative time relationship to perform heterogeneous sensor data fusion on the various homogeneous data, and construct according to the above method to obtain The multi-source information fusion model;

According to the preset multi-source information fusion model, multi-source information fusion is carried out on the measured environmental sensor and physiological sensor data after the optimized layout, which can in turn verify whether the layout of the environmental sensor and physiological sensor is reasonable. In addition, the multi-source information fusion model processing The data can also be entered by the user.

The sensor layout and multi-source information fusion method for dynamic monitoring provided by the embodiment of the present invention realizes the sensor layout by adopting the preset sensor layout model and processing the obtained data through the multi-source information fusion model. Make a reasonable layout and realize three-level dynamic evaluation and decision-making on "breeding individual-breeding group-breeding population".

Based on any of the above embodiments, further, before optimizing the layout of environmental sensors and physiological sensors according to the preset sensor layout model, it also includes:

Dimensionality reduction is performed on the three-dimensional space living environment and the three-dimensional geometric features of the living animal to obtain the two-dimensional plan view.

Specifically, according to the dynamic monitoring sensor layout and multi-source information fusion method of the above-mentioned embodiment, Fig. 2 is a schematic diagram of a two-dimensional plan generated by dimensionality reduction provided by the embodiment of the present invention. As shown in Fig. 2, the basic information of the farm is collected such as the area , shape, etc., collect the basic information of living animals such as quantity and size, and obtain a two-dimensional grid plan according to the three-dimensional living environment of living animals and the three-dimensional geometric characteristics of living animals. Grid the 3D model of living animals, that is, the 3D geometric characteristics of living animals, and then establish a mapping relationship between the 3D grid and the 2D image, add slits to the 3D grid, and divide it into disc structures for expansion. The three-dimensional coordinates of X, Y, and Z are mapped to the two-dimensional coordinates of U and V of the 2D image, so as to achieve dimensionality reduction processing for the breeding space and the 3D model of the living animal. This dimensionality reduction method can establish a direct mapping relationship between the 3D model and the 2D plane, and the subsequent sensor positions arranged in the 2D plane can be directly projected into the 3D space to represent the real position.

The sensor layout and multi-source information fusion method for dynamic monitoring provided by the embodiment of the present invention realizes the two-dimensional processing of the three-dimensional model by reducing the dimensionality of the three-dimensional space living environment and the three-dimensional geometric characteristics of living animals, reducing the processing difficulty and providing Sensors are properly arranged for preliminary preparations.

Based on any of the above embodiments, further, the optimal layout of the environmental sensors and physiological sensors according to the preset sensor layout model specifically includes:

Preliminarily matching the positions of the environmental sensor and the physiological sensor with the two-dimensional plan view to obtain a matching set;

Calculate and obtain the maximum monitoring ratio of the environmental sensor and the physiological sensor according to the matching set and the constraint variable;

The optimal layout is determined according to the maximum monitoring ratio of the sensor.

Specifically, according to the dynamic monitoring sensor layout and multi-source information fusion method of the above-mentioned embodiment, Fig. 3 is a schematic flow diagram of the multi-sensor optimal layout process flow provided by the embodiment of the present invention. As shown in Fig. 3, the three-dimensional space living environment and living Dimensionality reduction is performed on the three-dimensional geometric features of animals. After obtaining the two-dimensional plan, the two-dimensional plan is divided into several key grids, and the grid set is defined as:

Place the required number of sensors in any position of the two-dimensional plan, then the position set of sensors is:

According to the monitoring mechanism of various sensors and the characteristics of the divided area, the sensor position is initially matched with the key grid, and the matching result is defined as:

f(R,V)＝{(R _(1,1) ,V _(1,1) ),(R _(2,2) ,V _(2,2) ),…(R _(i,j) ,V _(i,j) )}

where (i=1,2...,m), (j=1,2,...,n);

Assuming that the monitoring ranges of k sensors are S={S ₁ , S ₂ ,...,S _k }, the maximum monitoring ratio that can be achieved by k sensors is:

Among them, ∑S represents the maximum monitoring ratio of k sensors, k is the number of sensors measuring the same parameter, S _T is the total monitoring grid area, and ε is a constraint variable;

In particular, in order to avoid over-monitoring in the same type of sensor layout process, that is, multiple sensors monitor the same grid or the sensors do not monitor the entire two-dimensional plane, but the monitoring ratio is greater than 100%. Therefore, certain constraints need to be added, specifically adding a constraint variable to optimize the layout:

Among them, S _c represents the grid area monitored by k sensors of the same type at the same time.

In particular, the calculation results of the model are limited, and the optimal number of sensors and layout positions are obtained, which are judged by the rules in Table 1:

Table 1 shows the judgment of layout level corresponding to different layout rules

layout rules Maximum monitoring ratio ΣV Number of sensors layout level 1 90%-100% optimal number best 2 80%-90% Excellent number excellent 3 60%-80% Suboptimal number suboptimal 4 <60% poor number Difference

For the layout of living animal physiological sensors, factors such as the size of the breeding environment, animal size, number of animals, and sensor monitoring mechanism need to be considered comprehensively. In addition to following the above sensor layout rules, it should be noted that physiological sensors have a specific monitoring mechanism. When monitoring blood sugar in living animals, the sensors are usually placed on the abdomen, when monitoring pulse and heart rate, they are usually placed on the neck, and when monitoring exercise, they are usually placed on the legs. Department etc. The layout of sensors on the animal body needs to make the monitoring range as concentrated as possible to achieve the purpose of accurate data.

Divide the two-dimensional map of living animals obtained by dimensionality reduction into specific functional areas according to the above rules, and directly match the specific sensors that need to be used with the functional areas according to their working principles. The requirement for the layout level is that it is advisable to set the number of wearable sensors that can basically cover a specific functional area, and avoid exceeding the preset functional area range; If there is damage, the number needs to be selected according to the size and health of the animal, and the embedded or implanted sensors that monitor the same indicator are arranged to achieve the maximum monitoring range, and the number should not exceed three.

Set the sampling area as x, the sampling number as m, the number of sampling animals as n, and the breeding range as S _T . Generally, adult sheep occupy an area of about 1.5-2㎡. The length and width of each sample square must be the same, that is, the sampling area must be equal. The total number of living animals in the quadrat is n, and the average area occupied by individual animals can be calculated The activity range of special farmed animals is 2ρ ₀ . Then divide the total breeding area by the average activity area of individual animals to get the animal density that needs to be monitored in this breeding area/> ρ is the animal density that needs to be monitored by installing sensors in a specific monitoring range. Based on this density, the overall information of the animals in the breeding area can be analyzed.

The sensor layout and multi-source information fusion method for dynamic monitoring provided by the embodiment of the present invention realizes a reasonable layout of the sensors by reasonably setting the sensor layout model and various sensor layouts of the sensor layout model.

Based on any of the above embodiments, further, before performing data fusion on the verified data using a preset multi-source information fusion model, it further includes:

Preliminary fusion of homogeneous sensor data according to data spatiotemporal weights to obtain a variety of homogeneous data;

The optimization weight is determined according to the relative time relationship, and heterogeneous sensor data fusion is performed on the various homogeneous data to construct the multi-source information fusion model.

Specifically, according to the dynamic monitoring sensor layout and multi-source information fusion method of the above-mentioned embodiment, the homogeneous sensor data is initially fused according to the data spatiotemporal weight to obtain a variety of homogeneous data; and then the optimal weight pair is determined according to the relative time relationship The multiple homogeneous data are fused with heterogeneous sensor data, and the multi-source information fusion model is constructed according to the above method;

The multi-source information fusion model also has data quality assessment and diagnosis functions, including checking whether the upper-level data is input or missing; fusion data distortion analysis, etc. The specific process includes: the multi-source information fusion model sets the adaptive dynamic data extraction time to extract the sensor data for three-level fusion. After each level of fusion, the fusion data will be transmitted to the data quality assessment and diagnosis center for diagnosis. If the diagnosis is correct, it will be passed as the input data of the next level; if the data is not received in a certain level of decision-making, the sensor data will be extracted again. The level will automatically sleep and wait for valid data to be re-worked; the data quality assessment and diagnosis center will compare the fusion data with the true value of the data, and optimize the weight of the fusion data whose deviation exceeds the preset threshold, and then re-integrate until it meets the requirements. Enter the next level of decision-making model upon request.

The sensor layout and multi-source information fusion method for dynamic monitoring provided by the embodiments of the present invention process data by rationally constructing a multi-source information fusion model, and realize three-level dynamic evaluation and decision-making for "cultured individual-cultured group-cultured population" .

Based on any of the above embodiments, further, the preliminary fusion of homogeneous sensor data according to data spatiotemporal weights specifically includes:

calculating the data spatiotemporal weight according to the data spatial importance and the data temporal importance, and performing the homogeneous sensor data fusion processing according to the data spatiotemporal weight to obtain the various homogeneous data;

The data space importance is:

∑ l _n is the total number contained in the data set at time t _m , k _{m, n} is the number of occurrences of a certain data value l _{m, n} at time t _m ;

The time importance of the data is:

Y _{μ, v} is the v data contained in the μ time period, n is the number of sensors, k′ _{m, n} is the number of occurrences of the data value l _{m, n} in the μ time period;

The data spatiotemporal weight reconstruction model is:

Specifically, according to the dynamic monitoring sensor layout and multi-source information fusion method of the above-mentioned embodiment,

Since the monitoring data basically remains unchanged for a certain period of time, a data threshold φ is given, and whenever the monitoring data exceeds the threshold, it is regarded as an inflection point, and this inflection point is used as the initial value of the next stage.

Therefore, the spatio-temporal weight reconstruction model is:

Step 1. Let y _t =y ₀ ±φ, where y ₀ is the initial value of the data, y _t is the final value of the previous stage (that is, the initial value of the next stage), and φ is the data threshold. Divide each sensor data into a time period set Y={Y _{1, ν} , Y _{2, v} , ..., Y _{μ, ν} }, where v is the number of data contained in a certain time period;

Step 2. Assuming that there are n sensors, the data set monitored at time t _m is l={ _{lm, 1} , _{lm, 2} , ..., lm _{, n} }, so a certain data value lm at time t _{m ,} The number of occurrences of _n k _{m, the ratio of n} to the entire data set as the spatial importance of the data Among them, ∑ l _n is the total number contained in the data set at time t _m . In particular, in the process of spatial importance discrimination, if a certain value appears more times, it means that the measurement results of multiple sensors at the same time are more real.

Step 3. Make the data time important Among them, Y _{μ, ν} represent the v data contained in the μ time period, n is the number of sensors, k′ _{m, n} is the number of occurrences of the data value l _{m, n} in the μ time period. In particular, in the process of judging the importance of time, if a certain value appears more times, it means that the measurement results of multiple sensors do not change significantly within a certain period of time.

Step 4. The final data space-time weight reconstruction model is

Use the weight obtained by this formula to carry out homogeneous sensor data fusion processing: suppose w _QD calculated by this model =(w ₁ ,w ₂ ,...,w _n ), and the monitoring data of homogeneous sensors is a=(a ₁ ,a ₂ ,..., a _n ), then the data fusion result x=w ₁ a ₁ +w ₂ a ₂ +...+w _n a _n .

The sensor layout and multi-source information fusion method for dynamic monitoring provided by the embodiment of the present invention performs fusion processing of homogeneous sensor data through weights, thereby realizing the fusion of homogeneous sensor data.

Based on any of the above-mentioned embodiments, further, determining the optimization weight according to the relative time relationship to perform heterogeneous sensor data fusion on the various homogeneous data, specifically includes:

Set an Euclidean space, there are optimization variables, weight optimizers, and classification optimizers;

converting the relative time relationship required by the optimal path found by the weight optimizer into the optimization weight;

The classification optimizer classifies the optimization variables, and performs heterogeneous sensor data fusion according to the classification results and the optimization weights.

Specifically, according to the dynamic monitoring sensor layout and multi-source information fusion method of the above-mentioned embodiment, the homogeneous sensor data is initially fused by calculating the weight according to the above-mentioned method, and then the heterogeneous sensor data is fused. In the embodiment of the present invention, heterogeneous sensor data fusion is performed after determining the optimization weight based on the relative time relationship. First, set an N×D Euclidean space, the state of variables to be optimized is Y _i =(y _i1 ,y _i2 ,…,y _iD ), the state of weight optimizer i is Xi _i =(x _i1 , x _i2 ,...,x _iD ), the state of classification optimizer j is X _j =(x _j1 ,x _j2 ,...,x _jD ), and an output optimizer k. Where x _id is the position of the nth optimizer in the d (d=1,2,...,D) dimensional variable space to be optimized;

It is stipulated that there is a one-to-one correspondence between the optimizer and the variable to be optimized, and the initial Euclidean distance between each optimizer and the corresponding variable to be optimized is equal. The variables to be optimized are sorted by prior knowledge to obtain the initial weight θ _n .

make Then l(X _i , Y _i ) is the optimal path of the optimizer i, the basic search step size of the optimizer is α, each search takes time t ₀ , and the optimizer i moves from the initial position to any p Move one step in each direction and return to the initial position, choose the direction closest to the target variable value to advance to the next long position, and then repeat the above steps until the variable state position corresponding to the optimizer i is reached, and record the optimal path The required time t.

The relative time relationship required by each optimizer i to find the optimal path is transformed into optimization weight w _n =(w ₁ ,w ₂ ,...,w _D ).

The classification optimizer j corresponds to each variable Y _i to be optimized. The classification optimizer j classifies the data of each variable according to the preset classification rules, and the optimization weight obtained by the weight optimizer through optimization is finally passed to pass formula in output optimizer k Computes the output decision, where /> is the boundary value corresponding to each variable.

The sensor layout and multi-source information fusion method for dynamic monitoring provided by the embodiment of the present invention realizes the fusion of heterogeneous sensor data by determining the optimization weight based on the relative time relationship and then performing heterogeneous sensor data fusion.

Based on any of the above embodiments, further, the three-level dynamic evaluation decision-making specifically includes:

Breeding individual health assessment: Perform data fusion on the physiological information data of a single animal according to the multi-source information fusion model to obtain individual data fusion information and individual health assessment decisions;

Breeding group situation assessment: performing data fusion of the individual data fusion information and the physiological information of the current animal group according to the multi-source information fusion model, to obtain group data fusion information and group situation assessment decisions;

Prediction and evaluation of cultured populations: performing data fusion on the population data fusion information and the processed environmental sensor data according to the multi-source information fusion model to obtain population prediction and evaluation decisions.

Specifically, according to the dynamic monitoring sensor layout and multi-source information fusion method of the above-mentioned embodiment, FIG. 4 is a schematic diagram of the three-level dynamic evaluation decision-making process provided by the embodiment of the present invention. As shown in FIG. 4, the health assessment of breeding individuals: each The information monitored by the sensor is used as the input data, and the above method is used for data fusion decision-making on the data monitored by the physiological information of a single animal. The decision-making and evaluation of the health status can be made based on the animal's body temperature, blood oxygen saturation, etc., and the decision-making and evaluation can be made based on the heart rate, exercise amount, etc. Behavioral state;

Assessment of the situation of the breeding group: the number of animals monitored in the breeding environment obtained through the sensor layout method of this patent, the information after the data fusion of the first-level animal individuals is used as the second-level input data, and the physiological information of the batch of animal groups is analyzed. Data fusion processing, through decision-making, the health status of the animal population in the breeding environment and the behavior information of the population during the monitoring period can be obtained;

Prediction and evaluation of breeding populations: the second-level data fusion information and the processed environmental sensor data information are used as the third-level input data at the same time, and then processed by the pre-set data fusion method to obtain the breeding animals under the environment changes of the farm Prediction of population status.

The sensor layout and multi-source information fusion method for dynamic monitoring provided by the embodiment of the present invention realizes the sensor layout by adopting the preset sensor layout model and processing the obtained data through the multi-source information fusion model. Make a reasonable layout and realize three-level dynamic evaluation and decision-making on "breeding individual-breeding group-breeding population".

Based on any of the above-mentioned embodiments, further, before the prediction and evaluation of the breeding population, it also includes:

The group data fusion information is pre-classified, and the data fusion is performed according to the data weight ranking of different groups.

Specifically, according to the dynamic monitoring sensor layout and multi-source information fusion method of the above-mentioned embodiments, especially when performing dynamic assessment of the breeding population, in different environments of the breeding farm, deviations in the physiological states of living animals of different populations will affect the fusion data. Therefore, the model will pre-classify the data transmitted in the second stage, sort the importance of the data of different populations, and then perform data fusion processing.

The sensor layout and multi-source information fusion method for dynamic monitoring provided by the embodiment of the present invention further increases the accuracy of the three-level dynamic evaluation decision-making of "cultured individual-cultured group-cultured population" by pre-classifying group data fusion information .

Based on any of the above embodiments, it further includes:

After each level of fusion is completed, the fusion information data will be subjected to data quality assessment and data diagnosis, and if the data quality assessment and data diagnosis pass, it will be used as the next level of input data.

Specifically, according to the dynamic monitoring sensor layout and multi-source information fusion method of the above-mentioned embodiment, since the data fusion result of each level of the three-level decision-making model of the present invention is used as the data input of the next level, in order to ensure normal operation, the multi-source information fusion model With data quality assessment and diagnosis functions, including checking whether the upper-level data is input and missing; fusion data distortion analysis, etc. The specific process includes: the multi-source information fusion model sets the adaptive dynamic data extraction time to extract the sensor data for three-level fusion. After each level of fusion, the fusion data will be transmitted to the data quality assessment and diagnosis center for diagnosis. If the diagnosis is correct, it will be passed as the input data of the next level; if the data is not received in a certain level of decision-making, the sensor data will be extracted again. The level will automatically sleep and wait for valid data to be re-worked; the data quality assessment and diagnosis center will compare the fusion data with the true value of the data, and optimize the weight of the fusion data whose deviation exceeds the preset threshold, and then re-integrate until it meets the requirements. Enter the next level of decision-making model upon request.

The sensor layout and multi-source information fusion method for dynamic monitoring provided by the embodiment of the present invention realizes the inspection of each level of data by performing data quality assessment and data diagnosis on the fusion information data, and further increases the "breeding individual-breeding group- The accuracy of the three-level dynamic evaluation decision-making of the "breeding population".

Further, on the basis of the above-mentioned embodiments, FIG. 5 is a schematic flow chart of the multi-sensor data fusion flow chart provided by the embodiment of the present invention. In the embodiment of the present invention, in the three-level breeding risk prediction and evaluation, the ambient temperature, relative humidity, The optimization weight of animal skin temperature, heart rate, etc. is w _n = (0.4, 0.3, 0.2, 0.1), then the range of decision value Y for breeding risk decision-making is safe is

Therefore, when the decision value Y is between 26.2 and 53.3, the output decision of the output optimizer k is safe. Especially in order to avoid the calculated data of the decision value conforming to the safety decision, but a certain evaluation factor has reached the bad condition. It is necessary to add a decision protection mechanism to the classification optimizer. If at least one variable data in a decision is bad, the classification optimizer does not perform subsequent decision calculations, but directly outputs bad decisions, as shown in Table 2.

Table 2 is the output decision rule table

Further, on the basis of the above embodiments, the embodiments of the present invention provide a dynamic monitoring sensor layout and multi-source information fusion system, which is used to implement the dynamic monitoring sensor layout and multi-source information fusion system in the above method embodiments information fusion method. Fig. 6 is a schematic flowchart of a dynamic monitoring sensor layout and multi-source information fusion system provided by an embodiment of the present invention. As shown in Fig. 6, the system includes: an optimization module 601 and an evaluation module 602; wherein,

Optimization module 601: used to optimize the layout of environmental sensors and physiological sensors according to a preset sensor layout model; wherein, the sensor layout model is constructed based on a two-dimensional plan and the relationship between the positions of the environmental sensors and physiological sensors owned;

Specifically, according to the dynamic monitoring sensor layout and multi-source information fusion system of the above-mentioned embodiment, in the above-mentioned optimization module 601, the basic information of the farm such as area, shape, etc. are collected, and the basic information of live animals such as quantity and size are collected, according to The three-dimensional living environment of living animals and the three-dimensional geometric characteristics of living animals are reduced to obtain a two-dimensional plan; the optimization module 601 uses the pre-built sensor layout model to optimize the layout of the farm environment sensors, and uses the preset animal density model based on the farm information Calculate the number of animals that need to be monitored, and then use the pre-built sensor layout model to perform physiological sensor layout on the batch of live animals;

The sensor layout model adds certain constraints to the sensors to avoid multiple sensors monitoring the same grid or sensors not all monitoring the entire two-dimensional plane, but the monitoring ratio is greater than 100%, and comprehensive considerations such as the size of the breeding environment, animal size, The number of animals and the sensor monitoring mechanism affect the layout of physiological sensors of living animals, by initially matching the positions of the environmental sensors and physiological sensors with the two-dimensional plan to obtain a matching set; according to the matching set and constraint variables The maximum monitoring ratio of the environmental sensor and the physiological sensor is obtained through calculation; the sensor layout model is obtained according to the maximum monitoring ratio of the sensor.

Evaluation module 602: used to periodically extract data from the optimized environmental sensor and physiological sensor, and use the preset multi-source information fusion model to perform data fusion on the verified data to obtain a three-level dynamic evaluation decision; , the multi-source information fusion model is obtained based on the combination of homogeneous sensor data fusion and heterogeneous sensor data fusion.

Specifically, according to the dynamic monitoring sensor layout and multi-source information fusion system of the above-mentioned embodiment, in the above-mentioned evaluation module 602, periodic information extraction is performed on the environmental sensor and physiological sensor data based on the pre-built sensor layout model layout, and the evaluation Module 602 uses the pre-built multi-source information fusion model to perform data fusion on the verified data to perform three-level dynamic evaluation and decision-making on "breeding individual-breeding group-breeding population";

Preliminary fusion of homogeneous sensor data according to data spatiotemporal weights to obtain a variety of homogeneous data; then determine the optimization weight according to the relative time relationship to perform heterogeneous sensor data fusion on the various homogeneous data, and construct according to the above method to obtain The multi-source information fusion model;

The evaluation module 602 performs multi-source information fusion on the measured environmental sensors and physiological sensor data after the optimized layout according to the preset multi-source information fusion model, which can in turn verify whether the layout of the environmental sensors and physiological sensors is reasonable. In addition, the multi-source information The data processed by the fusion model can also be input by the user.

It should be noted that the system of the embodiment of the present invention can be used to implement the technical solution of the embodiment of the sensor layout and multi-source information fusion method for dynamic monitoring shown in Fig. repeat.

The sensor layout and multi-source information fusion system for dynamic monitoring provided by the embodiment of the present invention realizes the sensor layout by adopting the preset sensor layout model and processing the obtained data through the multi-source information fusion model. Make a reasonable layout and realize three-level dynamic evaluation and decision-making on "breeding individual-breeding group-breeding population".

Based on any of the above embodiments, it further includes: an interaction module, a knowledge base, and a quality assessment and diagnosis center module.

Interactive module: The user can input data and observe the output decision results through the interactive module, and can also display the explanation of the conclusion and the solution process based on the user's questions.

Knowledge processing module: It is used to provide relevant basic knowledge to the dynamic monitoring sensor layout and multi-source information fusion system. The knowledge is acquired and processed and stored in the knowledge base. Automatic learning function.

Quality assessment and diagnosis center module: used to transfer the relevant basic knowledge to the quality assessment and diagnosis center module after being sent to the knowledge processing module for processing, to conduct qualitative assessment on the acquired expert knowledge, determine the applicability, and Feedback and correction of the situation, and retransmit the corrected data to the knowledge acquisition and processing system, and then the system transmits it to the knowledge inventory reserve;

In the multi-sensor data fusion process, the quality assessment and diagnosis center module dynamically adjusts the data extraction time according to the data (the extraction interval changes adaptively according to the dynamic characteristics of the monitoring data) to extract the multi-sensor data obtained after the layout. Due to the large number of sensors , in order to ensure the normal operation of the system and avoid errors in data extraction by the system, the data must be verified after data input, mainly including whether the data to be fused is successfully input, whether the data is available, etc. If the above errors occur in the data, the center will issue an alarm reminder Staff checks and re-enters the data. If there is still no valid data input in the system after the extraction period expires, the system will send a dormancy application to the user terminal, and the user can choose whether the system is dormant, and check the reason why the data is not input in time and make the system re-extract and process sensor data;

The quality assessment and diagnosis center module is specifically used to include checking whether the upper-level data is input and missing; fusion data distortion analysis, etc. The specific process includes: the multi-source information fusion model sets the adaptive dynamic data extraction time to extract the sensor data for three-level fusion, and after each level of fusion, the fusion data will be transmitted to the data quality assessment and diagnosis center module for diagnosis , if the diagnosis is correct, it will be passed as the input data of the next level; if the data is not received in a certain level of decision-making, the sensor data will be extracted again, if it is still unsuccessful, an alarm will be sent to the user, and the user will solve it. All levels will automatically sleep and wait for valid data to be re-worked; the data quality assessment and diagnosis center will compare the fusion data with the true value of the data, and optimize the weight of the fusion data whose deviation exceeds the preset threshold, and then re-integrate until After meeting the requirements, enter the next level of decision-making model.

The sensor layout and multi-source information fusion method for dynamic monitoring provided by the embodiment of the present invention realizes the sensor layout by adopting the preset sensor layout model and processing the obtained data through the multi-source information fusion model. Make a reasonable layout and realize three-level dynamic evaluation and decision-making on "breeding individual-breeding group-breeding population".

Based on any of the above embodiments, further, Fig. 7 is a schematic flowchart of another dynamic monitoring sensor layout and multi-source information fusion system provided by the embodiment of the present invention. As shown in Fig. 7, the system includes: input and output modules, Comprehensive database, interpreter, knowledge base, knowledge acquisition and processing system, and data quality assessment and diagnosis center; among them,

Input and output module: through this module, users can input data and observe and output decision results.

Comprehensive database: It is specially used to store the original data, intermediate results and final conclusions required in the reasoning process, often as a temporary storage area.

Interpreter: It can explain the conclusion and solution process according to the user's questions.

Knowledge base, knowledge acquisition and processing system: provide relevant basic knowledge to the system, and store the knowledge acquisition and processing in the knowledge base for storage. The knowledge acquisition and processing system can expand and modify the content of the knowledge base, and can also realize the automatic learning function.

Data Quality Assessment and Diagnosis Center: Its functions include,

(1) After the relevant basic knowledge is sent to the knowledge acquisition and processing system for processing, it is transmitted to the data quality assessment and diagnosis center, where the acquired expert knowledge is qualitatively evaluated, applicability is determined, and feedback correction is made according to the actual operation situation , and retransmit the corrected data to the knowledge acquisition and processing system, which is then transferred to the knowledge inventory for storage;

(2) In the multi-sensor data fusion process, the data quality assessment and diagnosis center dynamically adjusts the data extraction time according to the data (the extraction interval changes adaptively according to the dynamic characteristics of the monitoring data) to extract the multi-sensor data obtained after the layout. The quantity is large. In order to ensure the normal operation of the system and avoid errors in data extraction by the system, the data must be verified after data input, mainly including whether the data to be fused is successfully input, whether the data is available, etc. If the above-mentioned errors occur in the data, the center will send Alerts remind workers to verify and re-enter data. If there is still no valid data input in the system after the extraction period has expired, the system will send a dormancy application to the user terminal, and the user can choose whether the system is dormant, and check the reason why the data is not input in time and make the system re-extract and process sensor data.

The sensor layout and multi-source information fusion method for dynamic monitoring provided by the embodiment of the present invention realizes the sensor layout by adopting the preset sensor layout model and processing the obtained data through the multi-source information fusion model. Make a reasonable layout and realize three-level dynamic evaluation and decision-making on "breeding individual-breeding group-breeding population".

An example is as follows:

FIG. 8 is a schematic diagram of the physical structure of an electronic device provided by an embodiment of the present invention. As shown in FIG. 8, the electronic device may include: a processor (processor) 801, a communication interface (Communications Interface) 802, and a memory (memory) 803 and a communication bus 804 , wherein the processor 801 , the communication interface 802 and the memory 803 communicate with each other through the communication bus 804 . The processor 801 can call the logic instructions in the memory 803 to execute the following method: optimize the layout of environmental sensors and physiological sensors according to a preset sensor layout model; wherein, the sensor layout model is based on a two-dimensional plan, the environment The positions of sensors and physiological sensors are constructed; the optimized environmental sensors and physiological sensors are used to periodically extract data, and the data that has passed the verification is fused using a preset multi-source information fusion model to obtain a three-level dynamic Evaluation decision; wherein, the multi-source information fusion model is obtained based on the combination of homogeneous sensor data fusion and heterogeneous sensor data fusion.

In addition, the above logic instructions in the memory 803 may be implemented in the form of software function units and when sold or used as an independent product, may be stored in a computer-readable storage medium. Based on this understanding, the essence of the technical solution of the present invention or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including Several instructions are used to make a computer device (which may be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program codes. .

On the other hand, an embodiment of the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the transmission method provided by the above-mentioned embodiments is implemented, for example, including : optimize the layout of the environmental sensors and physiological sensors according to a preset sensor layout model; wherein, the sensor layout model is constructed based on a two-dimensional plan and the positions of the environmental sensors and physiological sensors; the optimized environmental sensors Periodically extracting data from physiological sensors, and performing data fusion on the verified data using a preset multi-source information fusion model to obtain a three-level dynamic evaluation decision; wherein, the multi-source information fusion model is based on homogeneous It is obtained by combining sensor data fusion and heterogeneous sensor data fusion.

The system embodiments described above are only illustrative, and the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in One place, or it can be distributed to multiple network elements. Part or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. It can be understood and implemented by those skilled in the art without any creative efforts.

Through the above description of the implementations, those skilled in the art can clearly understand that each implementation can be implemented by means of software plus a necessary general-purpose hardware platform, and of course also by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the prior art can be embodied in the form of software products, and the computer software products can be stored in computer-readable storage media, such as ROM/RAM, magnetic discs, optical discs, etc., including several instructions to make a computer device (which may be a personal computer, server, or network device, etc.) execute the methods described in various embodiments or some parts of the embodiments.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it can still be Modifications are made to the technical solutions described in the foregoing embodiments, or equivalent replacements are made to some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims (8)

1. A sensor layout and multi-source information fusion method for dynamic monitoring, characterized in that it comprises: Preliminarily matching the positions of the environmental sensors, the positions of the physiological sensors, and the two-dimensional plan according to the preset sensor layout model to obtain a matching set; calculating the maximum monitoring ratio between the environmental sensor and the physiological sensor according to the matching set and constraint variables; Determine the optimal layout according to the maximum monitoring ratio of the sensors and optimize the layout of the environmental sensors and the physiological sensors; wherein, the sensor layout model is based on a two-dimensional plan, the position of the environmental sensors and the physiological sensors obtained by relationship building; Preliminary fusion of homogeneous sensor data according to data spatiotemporal weights to obtain a variety of homogeneous data; Determining the optimization weight according to the relative time relationship, performing heterogeneous sensor data fusion on the various homogeneous data, and constructing a multi-source information fusion model; Periodically extract data from the optimized environmental sensors and physiological sensors, and use the multi-source information fusion model to perform data fusion on the verified data to obtain a three-level dynamic evaluation decision; wherein, the multi-source information fusion The model is based on the combination of homogeneous sensor data fusion and heterogeneous sensor data fusion. 2. The sensor layout and multi-source information fusion method for dynamic monitoring according to claim 1, characterized in that, according to the preset sensor layout model, the positions of the environmental sensors, the physiological sensors and the two-dimensional plan are initially matched , to obtain a matching set; calculate the maximum monitoring ratio between the environmental sensor and the physiological sensor according to the matching set and the constraint variable; determine the optimal layout according to the maximum monitoring ratio of the sensor and optimize the layout of the environmental sensor and the physiological sensor Before, also include: Dimensionality reduction is performed on the three-dimensional space living environment and the three-dimensional geometric features of the living animal to obtain the two-dimensional plan view. 3. The sensor layout and multi-source information fusion method of dynamic monitoring according to claim 1, wherein said preliminary fusion of homogeneous sensor data according to data spatiotemporal weights specifically includes: calculating the data spatiotemporal weight according to the data spatial importance and the data temporal importance, and performing the homogeneous sensor data fusion processing according to the data spatiotemporal weight to obtain the various homogeneous data; The data space importance is: ∑ l _n is the total number contained in the data set at time t _m , k _{m, n} is the number of occurrences of a certain data value l _{m, n} at time t _m ; The time importance of the data is: Y _{μ, ν} is the ν data contained in the μ time period, n is the number of sensors, k' _{m, n} is the number of occurrences of the data value l _{m, n} in the μ time period; The data spatiotemporal weight reconstruction model is: 4. The sensor layout and multi-source information fusion method of dynamic monitoring according to claim 1, characterized in that, determining the optimization weight according to the relative time relationship to carry out heterogeneous sensor data fusion to the multiple homogeneous data, Specifically include: Set an Euclidean space, there are optimization variables, weight optimizers, and classification optimizers; converting the relative time relationship required by the optimal path found by the weight optimizer into the optimization weight; The classification optimizer classifies the optimization variables, and performs heterogeneous sensor data fusion according to the classification results and the optimization weights. 5. The sensor layout and multi-source information fusion method for dynamic monitoring according to claim 1, wherein the three-level dynamic evaluation decision-making specifically includes: Breeding individual health assessment: Perform data fusion on the physiological information data of a single animal according to the multi-source information fusion model to obtain individual data fusion information and individual health assessment decisions; Breeding group situation assessment: performing data fusion of the individual data fusion information and the physiological information of the current animal group according to the multi-source information fusion model, to obtain group data fusion information and group situation assessment decisions; Prediction and evaluation of cultured populations: performing data fusion on the population data fusion information and the processed environmental sensor data according to the multi-source information fusion model to obtain population prediction and evaluation decisions. 6. The sensor layout and multi-source information fusion method of dynamic monitoring according to claim 5, is characterized in that, before the prediction and evaluation of the cultured population, it also includes: The group data fusion information is pre-classified, and the data fusion is performed according to the data weight ranking of different groups. 7. The sensor layout and multi-source information fusion method of dynamic monitoring according to claim 5, is characterized in that, also comprises: After each level of fusion is completed, the fusion information data will be subjected to data quality assessment and data diagnosis, and if the data quality assessment and data diagnosis pass, it will be used as the next level of input data. 8. A sensor layout and multi-source information fusion system for dynamic monitoring, characterized in that it comprises: Optimization module: used to initially match the position of the environmental sensor, the position of the physiological sensor and the two-dimensional plan according to the preset sensor layout model to obtain a matching set; calculate and obtain the relationship between the environmental sensor and the physiological sensor according to the matching set and constraint variables The maximum monitoring ratio of the sensor; determine the optimal layout according to the maximum monitoring ratio of the sensor and optimize the layout of the environmental sensor and the physiological sensor; wherein, the sensor layout model is based on a two-dimensional plan, the environmental sensor and the physiological sensor The location of the construction is obtained; The fusion module is used to initially fuse the homogeneous sensor data according to the data spatiotemporal weight to obtain a variety of homogeneous data; A building module, used to determine the optimization weight according to the relative time relationship to perform heterogeneous sensor data fusion on the various homogeneous data, and construct a multi-source information fusion model; Evaluation module: used to periodically extract data from the optimized environmental sensor and physiological sensor, and use the multi-source information fusion model to perform data fusion on the verified data to obtain a three-level dynamic evaluation decision; wherein, the The above multi-source information fusion model is based on the combination of homogeneous sensor data fusion and heterogeneous sensor data fusion.