KR20200086579A - Apparatus and method for measuring milk ingredient by specroscopy of spectrum data of raw milk - Google Patents

Abstract

Disclosed are an apparatus and a method for predicting the milk ingredients of raw milk by spectral analysis of spectrum data of raw milk to manage the health and diseases of dairy cattle and raw milk quality. According to the present invention, the prediction apparatus for predicting milk ingredients by spectral analysis of spectrum data sensed in raw milk comprises: an input unit to receive spectrum data obtained by sensing a standard sample of raw milk by a sensor device through a milk ingredient measurement device connected to the sensor device in a short-range wireless manner, and receive milk ingredient data of the standard sample analyzed in a laboratory; a learning unit to analyze the inputted spectrum data to learn and generate a milk ingredient learning model to allow analysis data of each item of milk ingredient data including the number of somatic cells, milk fat, milk protein, and lactose to be within an error with respect to corresponding item data of the inputted milk ingredient of the standard sample; and an output unit to output milk ingredient measurement result information of the milk ingredient data by analysis and prediction of the generated milk ingredient learning model according to the input of the milk ingredient measurement device.

Description

Apparatus and method for measuring milk ingredient by specroscopy of spectrum data of raw milk

The present invention relates to an apparatus and method for predicting the oil component of a crude oil by spectroscopying the spectral data of the crude oil sensed at a milking site as a spectral analysis technology of the crude oil.

The measurement of the quality of the conventional crude oil relies on a sample survey rather than a total survey using samples of crude oil collected from various cows and mixed. In the sample survey, it is determined whether or not the quality of only the milk of the cow to be surveyed corresponds to about 67% of the whole cow. In about 33% of cows excluded from the investigation, if low-quality crude oil is mixed with other crude oils, overall milk quality is deteriorated.

In the conventional crude oil testing service, the Agricultural Cooperative Federation's testing agent visits the ranch once a month in the morning and evening to collect milk samples of cows and send the collected samples to the laboratory to request analysis of milk composition, and report the test results of the analyzed milk ingredients. Was sent to the National Agricultural Cooperative Federation.

Here, since the analysis data relies on a sample of a sample survey once a month, there is a problem of low reliability in quality inspection of crude oil. In addition, since the analysis results are provided after about 20 days, since the crude oil of the sample with poor quality of data cannot be removed immediately, there is a problem in that the quality control is less realistic. In the case of various dairy products processed from crude oil, there is a problem in that the dairy product already contains poor quality crude oil due to results after 20 days.

For reference, the milk component data of crude oil is directly related to the quality of milk, the disease and health of the cow. If the quality of crude milk milked at the milking site is low, it is immediately excluded from the milking site and is required to be discarded. In addition, cows suspected of the disease and health of the cow should be excluded from milking and should be given rest and treatment.

Therefore, for the high-quality crude oil management, disease management and health management of dairy cows, a method of analyzing the crude oil in which the results of the investigation of the milk component is provided in real time is necessary for the full-scale investigation of the milk that is milked in real time at the milking site of a dairy farmer.

Korean Registered Patent 10-1622553 (2016.05.13)

The present invention was created to solve the above-described conventional problems, and in order to determine in real time the quality of the milk, the disease and the health of the cow with the milk component data, the spectral data of the light signal sensed by irradiating the light with the oil. It is an object to provide an apparatus and method for predicting an oil component by spectroscopic analysis.

Prediction device for predicting the oil component by spectral analysis of the spectral data sensed in the crude oil according to an aspect, the sensor device inputs the spectral data sensed by the sample sample of the crude oil through the oil component measuring device wirelessly connected to the sensor device An input unit to receive and input dairy component data of the sample sample analyzed in a laboratory; Analyzing the input spectral data, so that the analysis data for each item of dairy component data including somatic cell number, milk fat, milk protein, and lactose is within error with the corresponding item data of the dairy component data of the input sample sample Learning unit for learning to generate; And an output unit configured to output, based on the input of the dairy component measuring apparatus, result of measuring the dairy component of the dairy component data through analysis and prediction of the generated dairy component learning model.

The sensor device includes a light source that irradiates near-infrared light against crude oil flowing therein; A light receiving lens that receives light that has passed through the crude oil by the irradiation; An optical probe for condensing and outputting an optical signal incident through the light receiving lens by a multiplexing method; A spectroscopic module that converts the outputted optical signal into spectroscopic data; A short-range wireless communication unit transmitting the converted spectral data to the oil component measuring apparatus through short-range wireless communication so that the oil component measurement result information is generated based on the converted spectral data; And a control unit that outputs a control signal of light output to the light source and outputs a communication control signal to the short-range wireless communication unit for transmission of the spectroscopic data during the on operation.

The sensor device is connected between milking pipes, one end of which is worn on the nipple of the cow.

The oil component measuring apparatus, an input unit for receiving a user's key input; A wireless reader unit that reads an ID in which the individual of the cow is identified from the tag of the cow in a short-range wireless method; A wireless communication unit that receives the spectral data sensed by the sensor device in a short-range wireless method; A network communication unit for uploading the ID and the spectral data to the prediction device through a wired or wireless network, requesting measurement of oil component of crude oil, and receiving result of oil component measurement from the prediction device; A display unit that outputs user input information and the result of measuring the dairy component on a screen; And a control unit for controlling short-range communication of the ID and the spectral data and controlling network communication of the milk component data and the cow management information for measuring the milk component.

The learning unit, by using the wavelength band of the analysis section of 900 ~ 1100nm in the available band of the wavelength range of 700 ~ 1100nm of the wavelength graph experimented with the spectrum analysis of the crude oil, each item containing the fat, protein, sugar components of milk The learning model for analyzing the dairy component is generated.

The oil component measuring device receives, in real time, a crude oil test result report obtained by analyzing the crude oil quality from the oil analyzing device as the result of measuring the oil component, and outputs it to the screen, and the oil analysis device receives the oil component data measured for the spectroscopic data. It receives the input from the output unit and analyzes the crude oil quality, and generates the report in which the crude oil quality is analyzed.

The oil component measuring device receives disease and health management information from the disease health management device in real time as the oil component measurement result information and outputs it on a screen, and the disease health management device receives and registers a crude oil test report for each cow. Compare the component ratios of the multiple items in the dairy component data of the old report with the component ratios required for each disease of the cow to manage the disease, and compare the component ratios of the report with the component ratios of the cow's past history information. Determine and manage the health of the cow.

According to another aspect, a method of predicting an oil component by spectroscopically analyzing a spectral data sensed by a prediction device in a crude oil includes a method for measuring the oil component measured by wirelessly connecting the spectroscopic data sensed by a sensor device to a sample sample of the crude oil. Receiving through and receiving dairy component data of the sample sample analyzed in the laboratory; Analyzing the input spectral data, so that the analysis data for each item of milk component data including somatic cell number, milk fat, milk protein, and lactose is within an error with the corresponding item data of the oil component data of the input sample sample Learning to generate; And outputting, based on the input of the dairy component measuring apparatus, result of measuring the dairy component of the dairy component data through analysis and prediction of the generated dairy component learning model.

According to one aspect of the present invention, the time and cost of analysis on the composition and quality of crude oil is reduced by replacing the conventional dairy component analysis processing, which takes more than 20 days in the laboratory, with real-time analysis of a learning model that predictively analyzes spectral data of crude oil. Because of the reduction, a full-scale crude oil quality assessment can be provided in real time to all dairy cows.

In addition, since the sensor device is installed inline in the milking line of a cow, there is no restriction on the installation of the sensor device, and the sensing method through spectral analysis of light according to the non-destructive method of crude oil reduces the sensing time and cost, and ensures the sensing quality. It is possible to introduce a real-time transmission survey of crude oil quality for all cows at an economical cost.

In addition, by using the milk component data and milk quality analyzed in real time to diagnose the disease, health and conception of cows, and provide the diagnosed information in real time during milking at the milking site, to remove the crude oil of abnormal quality and , By excluding the cows in an abnormal state from the milking operation, it is possible to efficiently manage the disease and health of the cows.

The following drawings attached to this specification are intended to illustrate preferred embodiments of the present invention, and serve to further understand the technical idea of the present invention together with the detailed description of the invention described below, and thus the present invention is described in such drawings. It should not be interpreted as being limited to.
1 is a schematic configuration diagram of a system according to an embodiment of the present invention.
FIG. 2 is a schematic internal structural diagram of the sensor device and the oil component measurement device of FIG. 1.
3A and 3B are exemplary views of wavelength graphs of spectral data sensed for crude oil by the sensor device of FIG. 1.
FIG. 4 is a schematic internal structural diagram of the oil component predicting apparatus of FIG. 1 to predict the oil component of crude oil through a learning model.
5 is a schematic illustration of the crude oil analysis device and the disease health management device of FIG. 1.
6 is a schematic flow chart of a milk component measuring device according to an embodiment of the present invention to measure the oil component of crude oil through a sensor device.
7 is a schematic flowchart of a dairy component predicting apparatus predicting an oil component of crude oil through a learning model according to an embodiment of the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. Prior to this, the terms or words used in the present specification and claims should not be construed as being limited to ordinary or lexical meanings, and the inventor appropriately explains the concept of terms to explain his or her invention in the best way. Based on the principle that it can be defined, it should be interpreted as meanings and concepts consistent with the technical spirit of the present invention. Accordingly, the embodiments shown in the embodiments and the drawings described in this specification are only the most preferred embodiments of the present invention and do not represent all of the technical spirit of the present invention, and thus can replace them at the time of application. It should be understood that there may be equivalents and variations.

1 is a schematic configuration diagram of a system 100 according to an embodiment of the present invention.

The system 100 according to an embodiment of the present invention includes a sensor device 110, an oil component measurement device 130, an oil component prediction device 150, a crude oil analysis device 170, and a disease health management device 190. It is composed.

The sensor device 110 irradiates Near Infra-Red light to the milk that is milked from the cow to measure the milk component of the milk, and senses the signal of the received light that has passed through the milk as spectral data. The sensor device 110 transmits the spectral data sensed to the oil component measuring device 130 connected by short-range wireless communication. Examples of short-range wireless communication include RF communication, Bluetooth communication, Wi-Fi communication, and other IoT communication, and various types of short-range network communication may be applied.

The oil component measurement device 130 is a computing device that performs the short-range wireless communication with the sensor device 110 and performs wired or wireless network-based LAN communication with the oil component prediction device 150. In addition, the oil component measuring device 130 receives the oil component data measured in a predictive manner from the oil component predicting device 150 and outputs the screen to measure the oil component of the crude oil in real time.

The dairy component measuring device 130 receives the sensing signal of the spectral data sensed by the sensor device 110 using the short-range wireless communication. The dairy component measurement device 130 uploads the received cow's individual ID (eg, RFID) and spectral data to the dairy component prediction device 150 connected to an external network to request measurement of the dairy component. In addition, the oil component measuring device 130 receives the oil component data measured in a predictive manner from the oil component predicting device 150 and outputs the screen to measure the oil component of the crude oil in real time.

Here, the milk components include milk yield, milk fat, milk protein, milk lactose, milk conductivity, somatic cell count, milk color and Each item such as temperature is included. The value of each item and the component ratio value of a plurality of items are used in the present invention as indicators for determining the quality of milk, the disease and health of the cow.

The oil component prediction device 150 is a computing device that performs LAN communication based on a wired or wireless network. The oil component prediction device 150 receives the spectral data corresponding to the sensing signal from the oil component measurement device 130 and performs spectroscopic analysis of the received spectral data through a learning model. The learning model outputs milk component data of crude oil corresponding to the analysis result. The dairy component prediction device 150 transmits the dairy component data to the crude oil analysis device 170 to request quality analysis of the crude oil.

The crude oil analysis device 170 is a computing device that performs wired or wireless network-based LAN communication. The crude oil analysis device 170 receives the dairy component data from the oil component prediction device 150, analyzes the quality of the crude oil by using each item value of the dairy component data and a component ratio of a plurality of items, and tests the crude oil corresponding to the analysis result Generate grade reports.

The above crude oil test report is a report of milk ingredient and quality data of crude oil managed on the official website of the National Agricultural Cooperative Federation's Cow Improvement Office (www.dcic.co.kr). For reference, the official website of the NACF computerizes and manages individual individual information of cows (eg, breed, weight, etc.) and management history information (eg, dairy ingredient information, childbirth information, disease information, health information, etc.). Of course, the individual individual information and management history information are referred to generate milk component measurement result information for crude oil quality control and cow disease and health management of the present invention.

The disease health management device 190 is a computing device that performs LAN communication based on a wired or wireless network. The disease health management device 19 receives the crude oil test report based on the oil component data from the crude oil analysis device 170, and based on the oil component data and the crude oil quality data of the received report, the reference data, the individual individual information and the management history Compare and analyze using information. Through the processing of the comparison and analysis, the disease health management apparatus 190 determines the disease and health status of the cow, and manages the determined disease information and health information. The managed disease information, health information, and milk component measurement result information of the cow information are provided as inquiry information and notification information (eg, SMS, mail, push message, etc.) to users through the web server of the disease health management device 190. Can.

Here, the oil component prediction device 150, the crude oil analysis device 170, and the disease health management device 190 may be implemented as each independent computing device, and a plurality of devices may be integrated and implemented as one computing device.

FIG. 2 is a schematic illustration of the sensor device 110 and the oil component measuring device 130 of FIG. 1.

The sensor device 110 according to an embodiment of the present invention includes a microprocessor 211, a light source 212, a light receiving lens 214, an optical probe 215, a spectrometer 216, a battery 217, and It comprises a short-range wireless communication device 218.

In addition, the oil component measuring device 130 according to an embodiment of the present invention includes a control device 231, an RF reader 232, a short-range wireless communication device 233, an input device 234, a display device 235 and a LAN It comprises a communication device 236.

The sensor device 110 is connected between the milking pipes 202 and 203 connecting the teat 201 of the milking cow 200 and the milk reservoir 204. The milking pipe is cut into the first milking pipe 202 and the second milking pipe 203, and the sensor device 110 can be connected between the cut first milking pipe 202 and the second milking pipe 203. . Four sensor cells may be configured as one sensor device 110 to simultaneously accommodate four nipples of a cow. Then, crude milk milked from each of the four nipples 201 of the cow flows through the first milking pipe 202, the sensor device 110 and the second milking pipe 203 and is collected in the milk reservoir 204.

When the sensor device 110 is operated from the on signal, the light source 212 receiving the control signal of the microprocessor 211 irradiates the light 213 toward the crude oil flowing therein. Light 213 that has passed through the crude oil is collected in the light receiving lens 214. The optical signal collected by the light receiving lens 214 is incident on the optical probe 215 (optical probe) and condensed. The optical probe 215 is composed of a 4X1 multiplexer and outputs four simultaneous incident light signals to the spectrometer 216 in a time-sharing manner. The spectrometer 216 converts the input optical signal to spectral data of a wavelength value and outputs it. When the power of the battery 217 is output, the sensor device 110 is turned on to output near-infrared light from the light source 212. The short-range wireless device 218 receives the control signal from the MCU 211 and transmits the spectral data of the wavelength value to the oil component measuring device 130 based on the short-range wireless communication.

The dairy component measuring device 130 is a computing device having a wireless communication function, and is a portable measuring device that can be immediately installed at a milking site of a dairy farm.

The control device 231 may wirelessly lower an on signal for the operation of the sensor device 110. The control device 231 controls the process of transmitting the ID and spectral data of the cow 200 to the dairy component prediction device 150 to receive the analysis result and output it to the screen.

The RF reader 232 reads an object identification ID (eg, RFID) from the tag of the cow 200. The short-range wireless communication device 233 receives the spectral data of the sensor device 110 through short-range wireless communication. The input device 234 receives a key input for operation from a user. The display device 235 displays key input information and received dairy component analysis result information. The LAN communication device 236 is connected to a wired or wireless network for data communication.

The measurement result is received in real time since the dairy component measurement device 130 requests the measurement of the oil component by transmitting the object identification ID and the spectral data to the oil component prediction device 150. The dairy component measuring device 130 receives the result of measuring the dairy component including the measured dairy component data, the quality level of the crude oil, disease information, and health status. If there is an abnormality, the dairy component measuring device 130 notifies the abnormality information through the display device 235.

The sensor device 110 and the oil component measuring device 130 of the present invention are independent of the milking machine type of each ranch (eg, a pipeline, a pala system, a robot, etc.), a milking method, and a milking facility to enable sensing of the oil component of crude oil. This is because in-line installation of the sensor device 110 is possible in any path between the nipple 201 and the milk reservoir 204 of the cow 200. In addition, the sensing method through the spectroscopic analysis of light according to the non-destructive method of crude oil reduces the sensing time and cost, and guarantees the sensing quality, so that the dairy farmer can reduce the cost burden and introduce the whole water investigation into the whole cow 200. For reference, laboratory analysis proceeds in a manner that destroys crude oil components. In addition, the milk component measuring device 130 may provide a dairy farmer with real-time a query/notification service of disease information and health information of a cow individual as well as milk component data and quality information at the same time as the milking operation starts.

3A and 3B are exemplary views of wavelength graphs of spectral data sensed by the sensor device 110 of FIG. 1 for crude oil.

Basic optical experiments produce a wavelength graph of the spectral data of the crude oil. In the basic optical experiment, transmission and diffusion of light passing through the crude oil was performed using a spectrometer 216 of a CCD (500 to 1,150 nm), and characteristics of the wavelength band reacting to the crude oil were analyzed. When light of about 10 to 30 mm passes through the crude oil, it has been observed that the spectrum of light maintains an appropriate signal-to-noise ratio (S/N) for the spectrometer 216.

Referring to FIG. 3A, a wavelength band of 700 to 1100 nm was found to be an available band in a wavelength graph of spectrum analysis of crude oil. Referring to FIG. 3B, the wavelength band of 900 to 1100 nm was found to be a significant analysis section in which absorption of fat, protein, and sugar components of milk is generated in a wavelength graph in which an absorption spectrum analysis of crude oil is tested. Accordingly, in the present invention, the oil component is analyzed in the determined wavelength section.

FIG. 4 is a schematic internal structural diagram of the oil component predicting apparatus 150 of FIG. 1 predicting the oil component of crude oil through the learning model 454.

The dairy component measuring device 130 transmits the object identification ID and the spectral data to the dairy component predicting device 150 through the LAN communication device 236 to request an oil component measurement, and the measured oil component result value to the oil component predicting device 150 ).

The oil component prediction apparatus 150 according to an embodiment of the present invention includes an input unit 451, a learning unit 453, a learning model 454, and an output unit 455. After the learning model 454 is completed in the learning mode, the learning model 454 in the prediction mode measures the oil component of the crude oil against the spectral data of the crude oil.

First, the learning mode is described as follows. The input unit 451 receives the spectral data sensed by the oil component measuring device 130 for the crude oil of the sample sample for learning the spectroscopic analysis using the near-infrared light, and the value of the actual laboratory analyzed oil component of the same sample sample Input.

Here, the spectral data of the absorption spectrum input to the input unit 451 is appropriate mathematical preprocessing (math) such as noise removal, smoothing, standard normal variate (SNV), multicaptive scatter correction (MSC), differential, etc. to match the sensing characteristics. Pre-processing may be required.

The learning unit 453 uses the pre-processed spectral data as an input value so that the dairy component data including each item of somatic cell number, milk fat, milk protein, and lactose is within an error and an actual value of the dairy component data of the input sample sample. The learning model 454 is generated through learning.

When the learning model 454 is generated through learning, the physical and chemical properties and components of the crude oil component are statistically predicted through the analysis of the spectrum of light reacting to the raw material of the measurement target based on spectral analysis of near infrared rays. That is, the learning model 454 is a prediction model (Calibration) in which these values are correlated using the values of the spectral data corresponding to the absorption spectrum of the milk sample and the actual values of the milk components (fat, protein, sugar, etc.) to be measured. Model).

The prediction model of the correlation may use PLS Regression (partial least squares regression) as an algorithm of regression analysis. By evaluating the effectiveness of the predictive model for each dairy component, a predictive model with good reliability and reproducibility is established as the learning model 454. The established learning model 454 performs learning to measure values of items of various milk components at the same time using a prediction method from spectral data of the absorption spectrum of a milk sample to be measured. By repeating the learning, the reliability of the learning model 454 is increased by effectively implementing the correction and calibration of the continuous predictive model according to the change of environment such as temperature and various types of crude oil and accumulating data.

Next, the prediction mode will be described as follows. Spectroscopic data of the crude oil to be measured is input to the input unit 451, and pre-processed spectral data is input to the learning model 454. The learning engine of the learning modules 454 outputs values for each item of the oil component through the output unit 455 for the input spectral data.

The oil component prediction device 150 provided in the present invention replaces the analysis processing that takes more than 20 days to analyze the milk component of crude oil in a laboratory with real-time analysis processing. Therefore, since the time and cost of analysis on the composition and quality of the crude oil is reduced, it is possible to evaluate the crude oil quality of the whole cow's milk for all dairy farmers.

FIG. 5 is a schematic illustration of the crude oil analysis device 170 and the disease health management device 190 of FIG. 1.

After the dairy component data is measured in a predictive manner through the learning model 454, the dairy component prediction apparatus 150 outputs the analyzed dairy component data to the crude oil analysis apparatus 170 for quality analysis of the dairy component data.

The crude oil analysis device 170 analyzes the crude oil quality by comparing each item value of the oil component data with a reference value for each corresponding item, and analyzes the crude oil quality by comparing the component blending ratio of a plurality of items with the reference value of the corresponding component blending ratio. do. Then, a crude oil test report corresponding to the analysis result including the item value of the dairy component data and the grade for each item value is generated. The generated crude oil test result report is responded in real time to the dairy component measurement device 130 as the milk component measurement result information, reported in real time on the official site of the NACF server 510, and provided in real time to the disease health management device 190.

For reference, for the analysis processing of crude oil quality, the crude oil analysis device 170 analyzes with reference to oil component standards such as protein, fat, solid not fat (SNF), and milk urea nitrogen (MUN). Compare to the subject's milk component data.

The disease health management apparatus 190 determines the disease based on the ratio of the components of each dairy ingredient data using the crude oil test report provided in real time in the crude oil analysis device 170. When the component ratio of the real-time analyzed dairy component data corresponds to the component ratio mood of each disease, each disease including corresponding hyperacidosis, ketosis, mastitis and SARA (Sub Acute Ruminal Acidosis) is determined. For example, hyperacidosis and ketosis have a high correlation according to the ratio of the components of milk fat and milk protein.

The disease health management device 190 compares the component ratio information with the past history information of the cow 200, and when it exceeds a reference value, judges health conditions such as health abnormalities due to an abnormality in feed formulation and pregnancy status, respectively. do. For the determination of the health status, the disease health management device 190 may refer to various history information corresponding from the official website (www.dcic.co.kr) of the Agricultural Cooperative Federation's central server (510). In addition, the disease health management device 190 may refer to the lineage and ability information of a mother cow in conjunction with the official website of the Korean Breeder Improvement Association (www.aiak.or.kr).

In addition, the disease management device 190 transmits real-time measurement result information of the determined disease information and health information as response data to a request for measuring the milk component of the milk component measurement device 130. The disease management device 190 provides disease and health management information corresponding to a user terminal (eg, a PC, a mobile application) accessing a web server, and provides the result of measuring the milk component. The milk component measurement result information includes milk component data, a crude oil test report, disease information, health information, and herd management information. When the disease information determines that the health abnormality information is determined, the notification information may be notified to the user terminal 530 in real time.

The cow management information is information for managing cows and breeds of cows having excellent quality grade of oil, disease-free, and good health.

Here, the oil component measuring device 130 uploads the spectral data and then receives the oil component measurement result information including the crude oil test report, disease information, and health information in real time. For reference, when milking time per cow is about 7 minutes, the response time of the measurement result within a maximum of 7 minutes is a very long time compared to the actual response time. Actually, the measurement result is received within a few seconds at the same time as milking starts. Crude oil having a quality class below the standard may be provided to the milk component measuring device 130 with guide information that is immediately excluded from the milking site. Guide information that is immediately excluded from the milking target of the cow in which the disease is determined or the health condition is determined may be provided to the milk component measuring device 130. Therefore, in real time, quality control of crude oil, disease and health management of cows are progressed. In addition, the above-mentioned management progress makes the management of friendly forces faster and more reliable.

6 is a schematic flow chart of the oil component measuring device 130 according to an embodiment of the present invention to measure the oil component of the crude oil through the sensor device 110.

While the sensor device 110 is connected to the pipes 202 and 203 of the milking line, the light source 212 of the sensor device 110 irradiates the near-infrared light 213 toward the crude oil by the ON operation (S601). The light passing through the crude oil is received from the light receiving lens 214 (S603). The optical signal is collected from the optical probe 215 through the light receiving lens 214, and the spectrometer 216 converts the collected optical signal into spectral data (S605). The sensor device 110 wirelessly transmits the spectral data as a sensing signal (S607).

The dairy component measuring device 130 receives spectral data as the sensing signal from the sensor device 110 (S611). The dairy component measuring device 130 uploads the individually identified RF ID of the cow 200 and the received spectral data to the dairy component predicting device 150 to request a dairy component measurement (S613).

Thereafter, the dairy component measuring apparatus 130 receives the dairy component data output from the output unit 455 of the dairy component prediction apparatus 150 (S621). The oil component measuring device 130 receives the quality analysis result of the crude oil analyzed in real time based on the oil component data in the oil analysis device 170 (S623). Oil component data and quality analysis results can be included in the crude oil test report. The dairy component measuring device 130 receives disease data and health data from the disease health management device 190 (S625). Since the dairy component measurement device 130 requests the measurement of the dairy component in the step S613, the dairy component data, the crude oil test report, the disease data, and the health data corresponding to the corresponding measurement result are received and displayed on the screen (S627). . Then, by using the milk component measurement result information displayed on the screen, the low-quality crude oil may be excluded from the milking target and discarded, and the cow 200 determined to have a disease or health abnormality may be excluded from the milking operation.

7 is a schematic flow chart of a predicting oil component of crude oil through the learning model 454 by the oil component predicting apparatus 150 according to an embodiment of the present invention.

In the learning mode, a sample sample of crude oil is optically sensed (S701). The light-sensed spectral data is input to the input unit 451 of the oil component prediction apparatus 150 (S703). Spectroscopic data input to the input unit 451 may be pre-processed for learning or prediction. The oil component data of which the sample sample of the same crude oil as the light-sensed sample sample is actually analyzed in the laboratory is input to the input unit 451 (S705).

The learning unit 453 of the oil component prediction apparatus 150 learns the learning model 454 until the input spectral data is output within an allowable range as an actual value of the oil component data (S711).

After learning is completed (S713), the learning model 454 of the oil component prediction device 150 receives spectral data of the crude oil to be measured (S721). The learning model 454 processes predictive analysis of the dairy component data with respect to the input spectral data (S723). The learning model 454 outputs and provides the measurement result of the processed dairy component data (S725).

Although the present invention has been described by means of limited embodiments and drawings, the present invention is not limited by this, and the technical idea of the present invention and claims to be described below by those skilled in the art to which the present invention pertains Of course, various modifications and variations are possible within the scope of the scope.

100: system 110: sensor
130: oil component measurement device 150: oil component prediction device
170: crude oil analysis device 190: disease health management device
451: input unit 453: learning unit
454: learning model 455: output unit

Claims (14)

In the prediction device for predicting the oil component by spectral analysis of the spectral data sensed in the crude oil,
An input unit for receiving the spectral data sensing a sample sample of crude oil from a sensor device through a dairy component measuring device wirelessly connected to the sensor device, and receiving milk component data of the sample sample analyzed in a laboratory;
Analyzing the input spectral data, so that the analysis data for each item of dairy component data including somatic cell number, milk fat, milk protein, and lactose is within error with the corresponding item data of the dairy component data of the input sample sample Learning unit for learning to generate; And
According to the input of the dairy component measuring device, an output unit for outputting the result of measuring the dairy component of the dairy component data through analysis and prediction of the generated dairy component learning model
Prediction device comprising a. According to claim 1,
The sensor device,
A light source irradiating near-infrared light against the crude oil flowing therein;
A light receiving lens that receives light that has passed through the crude oil by the irradiation;
An optical probe for condensing and outputting an optical signal incident through the light receiving lens by a multiplexing method;
A spectroscopic module that converts the outputted optical signal into spectroscopic data;
A short-range wireless communication unit transmitting the converted spectral data to the oil component measuring apparatus through short-range wireless communication so that the oil component measurement result information is generated based on the converted spectral data; And
Control unit for outputting a control signal of the light output to the light source, and outputting a communication control signal to the short-range wireless communication unit for the transmission of the spectral data in the ON operation
Prediction device comprising a. According to claim 1,
The sensor device,
Apparatus characterized in that it is connected between the milking pipe is one end is worn on the nipple of the cow. According to claim 1,
The oil component measuring device,
An input unit that receives a user's key input;
A wireless reader unit that reads an ID in which the individual of the cow is identified from the tag of the cow in a short-range wireless method;
A wireless communication unit that receives the spectral data sensed by the sensor device in a short-range wireless method;
A network communication unit for uploading the ID and the spectral data to the prediction device through a wired or wireless network, requesting measurement of oil component of crude oil, and receiving result of measurement of oil component from the prediction device;
A display unit that outputs user input information and the result of measuring the dairy component on a screen; And
Control unit for controlling short-range communication of the ID and the spectral data and controlling network communication of the milk component data and the cow management information for the measurement of the milk component
Prediction device comprising a. According to claim 1,
The learning unit,
In the available band of the 700~1100nm wavelength band of the spectrum graph where the spectrum analysis of crude oil was experimented, the milk component of each item including fat, protein, and sugar components of milk is analyzed using the wavelength band of the analysis section of 900~1100nm Prediction device characterized in that for generating the learning model. According to claim 1,
The oil component measuring device receives, in real time, a crude oil test result report obtained by analyzing the oil quality from the oil analyzing device as the result of measuring the oil component, and outputs the result to the screen.
The crude oil analysis device receives the oil component data measured for the spectroscopic data from the output unit, analyzes the crude oil quality, and generates a report in which the crude oil quality is analyzed. According to claim 1,
The oil component measurement device receives disease and health management information from the disease health management device as real-time measurement result information and outputs it on a screen,
The disease health management apparatus receives a crude oil test score report for each cow, compares the component ratios of a plurality of items of the dairy component data of the registered report with the component ratios required for each disease of the cow, and manages diseases to be matched. Prediction device characterized in that by comparing the ratio of the components of the report with the ratio of the components of the past history of the cow's cows, the health status of each cow is determined and managed. In the method for predicting the oil component by spectral analysis of the spectral data sensed in the crude oil prediction device,
Receiving the spectroscopic data from which a sample sample of crude oil is sensed by a sensor device through an oil component measuring device wirelessly connected to the sensor device, and receiving milk component data of the sample sample analyzed in a laboratory;
Analyzing the input spectral data, so that the analysis data for each item of dairy component data including somatic cell number, milk fat, milk protein, and lactose is within error with the corresponding item data of the dairy component data of the input sample sample Learning to generate; And
Outputting, based on the input of the dairy component measuring apparatus, the result of measuring the dairy component of the dairy component data through analysis and prediction of the generated dairy component learning model.
How to include. The method of claim 8,
The sensor device,
Irradiating near-infrared light against the crude oil through which the light source flows;
Receiving a light passing through the crude oil by the light receiving lens;
Condensing and outputting an optical signal incident through the light receiving lens through a multiplexing method;
A spectroscopic module converting the outputted optical signal into spectroscopic data; And
The short-range wireless communication unit transmits the spectral data to the oil-based component measuring device through short-range wireless communication so that the oil component measurement result information is generated based on the converted spectral data.
Method characterized in that the execution. The method of claim 8,
The sensor device,
A method characterized by being connected between milking pipes, one end of which is worn on the nipple of the cow. The method of claim 8,
The oil component measuring device,
Reading an ID in which the individual of the cow is identified from the tag of the cow in a short-range wireless method;
Receiving the spectral data from the sensor device;
Uploading the ID and the spectral data to the prediction device through a wired or wireless network to request measurement of oil component of crude oil; And
Receiving the oil component measurement result information from the prediction device and outputting the screen
Method characterized in that the execution. The method of claim 8,
The step of learning and generating,
In the available band of the 700~1100nm wavelength band of the spectrum graph where the spectrum analysis of crude oil was experimented, the milk component of each item including fat, protein, and sugar components of milk is analyzed using the wavelength band of the analysis section of 900~1100nm And generating the learning model. The method of claim 8,
The oil component measuring device receives, in real time, a crude oil test result report obtained by analyzing the oil quality from the oil analyzing device as the result of measuring the oil component, and outputs the result to the screen.
The method for analyzing the crude oil by receiving the oil component data measured for the spectral data from the output unit and analyzing the crude oil quality and generating the report analyzed by the crude oil quality. The method of claim 8,
The oil component measurement device receives disease and health management information from the disease health management device as real-time measurement result information and outputs it on a screen,
The disease health management apparatus receives a crude oil test score report for each cow, compares the component ratios of a plurality of items of dairy component data in the registered report with the component ratios required for each disease of the cow, and manages the disease that matches. And comparing the component ratio of the report with the component ratio of the past history information of the cow to determine and manage the health status of each cow.

Images