A device and method for collecting and storing infrared spectrum data of raw milk

Abstract

The present application belongs to the technical field of raw milk detection, and provides a device and method for infrared spectrum data collection and storage of raw milk, which comprises: taking a raw milk sample, homogenizing the raw milk sample by using a homogenizing device, evaluating the homogenization by using ultrasonic detection, determining whether to end the homogenization, if yes, obtaining the homogenized sample to be detected, deploying an infrared spectrum detection light path system, a multi-modal sensor group and a communication networking for infrared spectrum collection of the sample to be detected, collecting the infrared spectrum of the sample to be detected, marking the actual collection period and obtaining the actual sample infrared spectrum of the sample to be detected, collecting a multi-modal data group by using the multi-modal sensor group, analyzing the multi-modal data group, determining whether the sample to be detected has a high abnormal risk, if not, triggering multi-modal data compensation, and compensating the actual sample infrared spectrum.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of fresh milk detection, in particular to a device and method for infrared spectrum data collection and storage in fresh milk. BACKGROUND

[0002] Quality detection of fresh milk is crucial for food safety supervision, dairy processing and market circulation. Due to the advantages of rapidity, non-destructiveness and simultaneous analysis of multiple components, infrared spectrum technology is widely used in the detection of milk components. However, the existing methods for collecting and storing infrared spectrum data of fresh milk still have many technical problems in practical application, resulting in insufficient accuracy and poor reliability of spectrum data, which is difficult to meet the demand of high-precision detection.

[0003] In the prior art, the homogenization of fresh milk relies on mechanical operation with fixed parameters, and there is a lack of real-time evaluation mechanism for homogenization effect. Fat globules in fresh milk are prone to aggregation or floating due to insufficient homogenization, resulting in uneven distribution of sample components, and thus insufficient representation of components in the infrared spectrum collection area, causing characteristic absorption peak shift or intensity distortion. In addition, traditional homogenization processing is prone to data deviation due to single homogenization failure for samples with high viscosity or abnormal fat content.

[0004] The deployment of existing infrared spectrum collection hardware focuses on the construction of a single optical path system, ignoring the influence of multi-modal environmental parameters on the spectrum. On the one hand, the coaxial adjustment of infrared light source, sample cell and detector relies on manual experience, which is prone to mechanical errors, resulting in inconsistent optical path and poor repeatability of multiple spectrum collection results for the same sample. On the other hand, multi-modal sensors such as temperature, pH and conductivity are not integrated, which cannot capture the environmental changes of the sample during the collection process. These factors result in spectrum data reflecting only the instantaneous state, lacking correction basis for environmental interference, and reducing the long-term comparability of data.

[0005] The existing method lacks quality monitoring of the infrared spectrum collection process and does not establish an effective recheck triggering mechanism. Due to the colloidal properties of fresh milk, bubbles or impurities are easily attached to the sample cell, resulting in sharp noise peaks or baseline drift in the spectrum. Traditional methods directly store the original spectrum without identifying and removing low-quality data. In addition, the processing of spectrum data and multi-modal data is independent of each other, and there is no spectrum compensation model based on multi-modal parameters. At the same time, data storage only records the spectrum curve without associating homogenization parameters, environmental parameters and abnormal markers. When the detection result is controversial, it is difficult to trace the source of deviation, reducing the traceability and credibility of data.

[0006] To solve the above problems, the application provides a device and method for infrared spectrum data collection and storage in fresh milk. SUMMARY

[0007] To make up for the deficiencies of the prior art, at least one technical problem raised in the background art is solved.

[0008] The technical solution adopted by the present application to solve the technical problems is: a method for collecting and storing infrared spectrum data of fresh milk, comprising:

[0009] Take a fresh milk sample, use a homogenizing device to homogenize the fresh milk sample, and use ultrasonic detection to evaluate the homogenization process to determine whether to end the homogenization process. If so, a homogenized sample for detection is obtained.

[0010] Deploy a collection hardware for the infrared spectrum collection of the sample for detection. The collection hardware includes an infrared spectrum detection optical system, a multi-modal sensor group, and a communication network.

[0011] Use the infrared spectrum detection optical system to collect the infrared spectrum of the sample for detection, mark the actual collection period, and obtain the actual sample infrared spectrum of the sample for detection. Use the multi-modal sensor group to collect a multi-modal data set.

[0012] Analyze the multi-modal data set to determine whether the sample for detection has a high abnormal risk. If not, trigger multi-modal data compensation and compensate the actual sample infrared spectrum.

[0013] The determination method of whether to end the homogenization process is:

[0014] Set a homogenization value for each sample for detection and assign an initial value of 0. Obtain the variation coefficient of ultrasonic propagation speed and the variation coefficient of attenuation coefficient.

[0015] If the variation coefficient of ultrasonic propagation speed and the variation coefficient of attenuation coefficient are both less than a preset variation coefficient threshold, or the homogenization value of the sample for detection reaches a preset flag threshold, it is determined that the homogenization of the sample for detection is completed. Otherwise, it is determined that the homogenization of the sample for detection is not completed. The sample for detection is homogenized again, and the homogenization value is incremented by one.

[0016] The acquisition method of the variation coefficient of ultrasonic propagation speed and the variation coefficient of attenuation coefficient is:

[0017] Obtain the sample for detection, use an ultrasonic detector to detect the sample for detection, uniformly select several sampling points on the surface of the sample for detection, perform ultrasonic detection multiple times for each sampling point, record the propagation distance and propagation time of the ultrasonic signal in the sample for detection, and perform data processing to obtain the ultrasonic propagation speed.

[0018] For each ultrasonic detection, record the ultrasonic signal incident ultrasonic intensity and the ultrasonic intensity after penetrating the sample to be detected, and calculate the attenuation coefficient;

[0019] For each sampling point, the ultrasonic propagation speed and the attenuation coefficient of each ultrasonic detection are respectively processed, and the variation coefficient of the ultrasonic propagation speed and the variation coefficient of the attenuation coefficient are calculated.

[0020] The acquisition method of the sample to be detected is:

[0021] A homogenizing device is provided for the fresh milk to be collected for infrared spectrum, and the homogenizing device is used to homogenize the fresh milk to be collected for infrared spectrum, and the homogenizing device includes a pressure adjusting device and a temperature adjusting device;

[0022] An adaptive adjusting module is added to the pressure adjusting device and the temperature adjusting device, which collects homogenizing pressure data and homogenizing temperature data in real time, and respectively calculates the deviation with the set homogenizing pressure and homogenizing temperature in real time. If the deviation is out of standard, the adaptive control algorithm is used to set the homogenizing pressure or homogenizing temperature as the target to adjust the homogenizing pressure data, and the homogenized fresh milk sample is introduced into the detection container to obtain the sample to be detected;

[0023] The acquisition method of the actual sample infrared spectrum is:

[0024] The period of collecting infrared spectrum for a sample to be detected is marked as a collection period, and an infrared detection value is set for each sample to be detected, and the initial value is 0;

[0025] The signal-to-noise ratio of the sample infrared spectrum is obtained, if it is less than the preset signal-to-noise ratio threshold, the recheck is triggered to collect infrared spectrum for the sample to be detected again, and the infrared detection value is added once, otherwise, or the infrared detection value reaches the preset threshold, the current collection period is marked as the actual collection period, and the sample infrared spectrum in the actual collection period is marked as the actual sample infrared spectrum;

[0026] The acquisition method of the signal-to-noise ratio of the sample infrared spectrum is:

[0027] The sample infrared spectrum is obtained, the average signal power of the characteristic absorption peak region and the average noise power of the noise region in the sample infrared spectrum are selected for data processing, and the signal-to-noise ratio of the sample infrared spectrum is calculated;

[0028] The acquisition method of the sample infrared spectrum is:

[0029] The infrared spectrum detection light path system is used for collecting and smoothing the infrared spectrum of the sample to be detected to obtain an original sample infrared spectrum, and a blank correction method is used to set distilled water without any component to be detected as a blank sample, and the infrared spectrum detection light path system is used to collect the infrared spectrum of the blank sample to obtain a baseline infrared spectrum by fitting;

[0030] The baseline infrared spectrum is used for baseline calibration of the collected original sample infrared spectrum to obtain a sample infrared spectrum;

[0031] The determination of whether the sample to be detected has a high abnormal risk is as follows:

[0032] The actual sample infrared spectrum and the multi-modal data set of the sample to be detected within the actual collection period are obtained, a normal threshold range of the multi-modal data is set, if the multi-modal data set is not within the normal threshold range of the multi-modal data, it is determined that the sample to be detected has a high abnormal risk, otherwise, it is determined that the sample to be detected does not have a high abnormal risk;

[0033] The multi-modal data compensation includes:

[0034] A standard value of the multi-modal data set is set, the actual sample infrared spectrum and the multi-modal data set of the sample to be detected within the actual collection period are combined, a compensation coefficient of each data in the multi-modal data set is obtained by fitting experimental data through a multi-linear regression model, and the actual sample infrared spectrum of the sample to be detected is compensated by data processing using the compensation coefficient to obtain the compensated actual sample infrared spectrum.

[0035] A fresh milk infrared spectrum data collection and storage device includes the following modules:

[0036] The sample processing module takes a fresh milk sample, performs homogenization treatment on the fresh milk sample using a homogenization device, and evaluates the homogenization treatment using ultrasonic detection to determine whether to end the homogenization treatment, and if so, obtains a homogenized sample to be detected.

[0037] The hardware arrangement module deploys collection hardware for infrared spectrum collection of the sample to be detected, and the collection hardware includes an infrared spectrum detection light path system, a multi-modal sensor group, and a communication group network.

[0038] The data collection module collects the infrared spectrum of the sample to be detected using the infrared spectrum detection light path system, marks the actual collection period, and obtains the actual sample infrared spectrum of the sample to be detected, and collects the multi-modal data set using the multi-modal sensor group.

[0039] The spectrum compensation module analyzes the multi-modal data set to determine whether the sample to be detected has a high abnormal risk, and if not, triggers multi-modal data compensation to compensate the actual sample infrared spectrum.

[0040] The beneficial effects of the present application are as follows:

[0041] 1、The present application ensures uniformity of fresh milk samples through homogenization processing combined with ultrasonic detection evaluation, greatly improves the accuracy of subsequent infrared spectrum acquisition, at the same time, the specially deployed acquisition hardware covers infrared spectrum detection light path system, multi-modal sensor group and communication networking, can synchronously acquire actual sample infrared spectrum and multi-modal data group, realizes comprehensive data acquisition, provides rich and reliable data basis for subsequent accurate analysis, and effectively improves the detection quality.

[0042] 2、The present application analyzes the multi-modal data group, judges whether the to-be-detected sample has high abnormal risk in advance, avoids invalid or wrong data processing, saves time and cost, and when the multi-modal data compensation is triggered, the actual sample infrared spectrum can be compensated, the possible error can be effectively corrected, and the accuracy and reliability of the data are further improved. This kind of comprehensive data processing mechanism is helpful to more accurately grasp the quality of fresh milk, and guarantees the product quality and safety. BRIEF DESCRIPTION OF DRAWINGS

[0043] The present application will be further described below in conjunction with the drawings.

[0044] Figure 1 It is a step flow chart of a fresh milk infrared spectrum data acquisition and storage method according to an embodiment of the present application;

[0045] Figure 2 It is a module architecture diagram of a fresh milk infrared spectrum data acquisition and storage device according to an embodiment of the present application. DETAILED DESCRIPTION

[0046] In order to make the technical means, creative features, purposes and effects realized by the present application easy to understand, the present application will be further described below in conjunction with specific embodiments.

[0047] Embodiment 1

[0048] Please refer to Figure 1 The fresh milk infrared spectrum data acquisition and storage method according to an embodiment of the present application includes the following steps:

[0049] S1: Take a fresh milk sample, use a homogenization device to homogenize the fresh milk sample, and use ultrasonic detection to evaluate the homogenization, and determine whether to end the homogenization, if yes, get the homogenized to-be-detected sample;

[0050] Set a homogenization device for the fresh milk to be collected infrared spectrum, and use the homogenization device to homogenize the fresh milk to be collected infrared spectrum;

[0051] Specifically, the homogenizing device is selected as a high-pressure homogenizer, and the homogenizing device mainly comprises a feeding port, a homogenizing cavity, a pressure adjusting device, a temperature adjusting device and a discharging port. The homogenizing cavity is internally provided with a homogenizing valve. The particles in the raw fresh milk are broken and uniformly dispersed through high-pressure action, and the large fat globule structure in the raw fresh milk is destroyed to prevent the floating of milk fat from causing uneven composition in the infrared spectrum measurement area;

[0052] The key parameters of the homogenizing device are set, including the homogenizing pressure, the homogenizing temperature and the homogenizing time. After the key parameters of the homogenizing device are set, the raw fresh milk to be collected for infrared spectrum is taken from the raw fresh milk storage container using a sterilized sampler, poured into the feeding port of the homogenizing device, and the homogenizing device is started. The period of time during which the homogenizing device homogenizes a raw fresh milk sample is marked as a homogenizing period.

[0053] During the homogenizing period, the pressure adjusting device of the homogenizing device collects homogenizing pressure data in real time through an internal pressure sensor, and the temperature adjusting device of the homogenizing device collects homogenizing temperature data in real time through an internal temperature sensor.

[0054] An adaptive adjustment module is added to the pressure adjusting device and the temperature adjusting device. The adaptive adjustment module obtains the collected homogenizing pressure data and homogenizing temperature data in real time, and respectively calculates the deviation from the set homogenizing pressure and homogenizing temperature in real time to obtain real-time pressure deviation and real-time temperature deviation. If the real-time pressure deviation exceeds the preset pressure deviation standard, the adaptive control algorithm is used to adjust the homogenizing pressure data with the set homogenizing pressure as the target. Similarly, if the real-time temperature deviation exceeds the preset temperature deviation standard, the adaptive control algorithm is used to adjust the homogenizing temperature data with the set homogenizing temperature as the target.

[0055] The homogenized raw fresh milk sample is introduced into a detection container and left to stand until the raw fresh milk sample in the detection container stops flowing to obtain a raw fresh milk sample to be detected.

[0056] The ultrasonic detector is used to detect the sample to be detected. A plurality of sampling points are uniformly selected on the surface of the sample to be detected to ensure that the positions of the sampling points are representative. The probe of the ultrasonic detector is vertically placed on the surface of the sample to be detected at each sampling point. The ultrasonic detector is used to detect the sample at each sampling point multiple times. For each ultrasonic detection, the propagation distance d of the ultrasonic signal in the sample to be detected and the propagation time are recorded. The propagation distance d is determined by the fixed size of the detection container, and the propagation time is obtained by a timing module of the ultrasonic detector. The propagation distance and the propagation time are processed by ratio to obtain the ultrasonic propagation speed v of the ultrasonic signal collected at each sampling point.

[0057] For each ultrasonic detection, the incident ultrasonic intensity of the ultrasonic signal is recorded and the intensity I of the ultrasonic wave after passing through the sample to be detected, the attenuation coefficient is calculated , the calculation formula is:

[0058] ;

[0059] The average value and the standard deviation of the ultrasonic propagation speed and the attenuation coefficient are calculated respectively for each sampling point and each ultrasonic detection, and the average value of the ultrasonic propagation speed , the standard deviation of the ultrasonic propagation speed , the average value of the attenuation coefficient , the standard deviation of the attenuation coefficient ;

[0060] The coefficient of variation of the ultrasonic propagation speed and the coefficient of variation of the attenuation coefficient ;

[0061] ;

[0062] ;

[0063] It should be noted that the ultrasonic propagation speed and the attenuation coefficient reflect the uniformity of the internal homogeneity of the sample to be detected. The more uniform the homogeneity, the more consistent the particle distribution, and the smaller the coefficient of variation of the ultrasonic propagation speed and the attenuation coefficient in the sample;

[0064] Set a homogeneity processing value for each sample to be detected and assign an initial value of 0 to the homogeneity processing value. The coefficient of variation of the ultrasonic propagation speed and the coefficient of variation of the attenuation coefficient are compared with the preset coefficient of variation threshold respectively;

[0065] If the coefficient of variation of the ultrasonic propagation speed and the coefficient of variation of the attenuation coefficient are both less than the preset coefficient of variation threshold, or the homogeneity processing value of the sample to be detected reaches the preset flag threshold, it is determined that the homogeneity processing of the sample to be detected is completed, otherwise, it is determined that the homogeneity processing of the sample to be detected is not completed. The sample to be detected is subjected to homogeneity processing again, and the homogeneity processing value is subjected to a self-adding operation;

[0066] It should be noted that the purpose of this step is to destroy the large fat ball structure in fresh milk by high-pressure homogenization, prevent the floating of milk fat to cause uneven composition in the spectral measurement area, evaluate the homogenization effect by ultrasonic detection, stabilize the homogenization pressure and temperature through the adaptive adjustment module, ensure the sample uniformity, eliminate the interference of uneven composition distribution on infrared spectrum measurement, improve the stability of spectrum data, introduce the adaptive adjustment module, realize real-time dynamic correction of homogenization pressure and temperature, solve the problem of easy drift of traditional homogenization parameters, ensure the stability of homogenization parameters, and control the ultrasonic evaluation and homogenization value in a cycle to avoid insufficient or excessive homogenization, and improve the sample homogeneity.

[0067] S2: Deploying acquisition hardware for the infrared spectrum of the sample to be detected, the acquisition hardware including an infrared spectrum detection optical path system, a multi-modal sensor group, and a communication network;

[0068] Specifically, the components of the infrared spectrum detection optical path system include a light source, a sample cell, an absorption detector, and an optical adjustment rack. The light source is selected to be a mid-infrared light source with a wavelength range of 2.5-25 μm, which covers the characteristic absorption peaks of the main components such as protein, fat, and lactose in the sample to be detected, providing a basis for subsequent accurate analysis of the composition of the milk. The sample cell is made of quartz material, and the optical path is set to ensure that the sample to be detected absorbs the infrared light moderately. The absorption detector is selected to be a mercury-cadmium-telluride detector, which has high detection sensitivity in the mid-infrared waveband and can accurately capture the changes in the infrared light after being absorbed by the sample to be detected. The light source, sample cell, and absorption detector are mechanically fixed according to the optical path sequence. The positions of the components are adjusted through the optical adjustment rack to ensure that the optical path is coaxial, so that the infrared light is emitted by the light source, passes through the sample to be detected in the sample cell, and is then received by the absorption detector, forming a complete infrared spectrum detection optical path system.

[0069] The multi-modal sensor group includes a temperature sensor, a pH sensor, and a conductivity sensor, which are used to collect multi-modal data of the sample to be detected. The multi-modal data includes the temperature, pH value, and conductivity of the sample to be detected. The temperature sensor is selected to be a platinum resistance temperature sensor, which is fixed on the inner wall of the sample cell through a bracket. The pH sensor is a glass electrode pH sensor, which is inserted into the sample cell at an angle and avoids contact with the wall of the sample cell. The conductivity sensor is selected to be an electrode type conductivity sensor, which is installed vertically at the central position of the bottom of the sample cell.

[0070] The communication network adopts Ethernet communication protocol to connect the homogenization device, the infrared spectrum detection optical path system, the multi-modal sensor group, and the data storage module. IP addresses and port numbers are set to ensure stable data transmission.

[0071] It should be noted that the purpose of this step is to build a hardware system for infrared spectrum acquisition, multi-modal parameter monitoring, and data transmission, which provides a foundation for subsequent spectrum acquisition and data processing. The system integrates mid-infrared light path and multi-modal sensor group, realizes the collaborative acquisition of spectrum and environmental parameters, breaks through the limitations of traditional single spectrum acquisition, optimizes the sensor layout according to the characteristics of the sample cell, reduces measurement interference, and improves the accuracy of parameter acquisition.

[0072] S3: Using the infrared spectrum detection optical path system to collect the infrared spectrum of the sample to be detected, marking the actual acquisition period and obtaining the actual sample infrared spectrum of the sample to be detected, and using the multi-modal sensor group to collect the multi-modal data set.

[0073] The deployed acquisition hardware is used to collect infrared spectrum and multi-modal data of the to-be-tested sample;

[0074] Specifically, for infrared spectrum acquisition, a time period for one infrared spectrum acquisition of a to-be-tested sample is marked as an acquisition time period, the to-be-tested sample is transferred to a sample cell of an infrared spectrum detection optical path system through a sterile pipette, the infrared spectrum detection optical path system is used to collect infrared spectrum of the to-be-tested sample and perform smoothing processing, and an original sample infrared spectrum is obtained , wherein represents wavelength;

[0075] The blank correction method is used, distilled water without any to-be-tested component is set as a blank sample, the blank sample is transferred to the sample cell of the infrared spectrum detection optical path system through the sterile pipette, the infrared spectrum detection optical path system is used to collect infrared spectrum of the blank sample, and a baseline infrared spectrum is obtained through fitting ;

[0076] For the collected original sample infrared spectrum , baseline calibration is performed, a sample infrared spectrum is obtained through baseline calibration, and a calculation formula of the sample infrared spectrum is as follows:

[0077] ;

[0078] It should be noted that the role of baseline calibration is to eliminate the interference of the infrared spectrum detection optical path system itself and environmental factors on infrared spectrum measurement, so as to ensure that the calculated fresh milk sample infrared spectrum more truly reflects the component information of the to-be-tested sample;

[0079] The signal-to-noise ratio SNR of the sample infrared spectrum is calculated, the average signal power of a characteristic absorption peak region in the sample infrared spectrum is selected , and the average noise power of a noise region is selected , and a calculation formula is as follows:

[0080] ;

[0081] An infrared detection value is set for each to-be-tested sample, and an initial value of the infrared detection value is 0; the signal-to-noise ratio SNR of the sample infrared spectrum is compared with a preset signal-to-noise ratio threshold;

[0082] If the signal-to-noise ratio SNR of the sample infrared spectrum is less than the preset signal-to-noise ratio threshold, it is judged that there is a serious error in infrared spectrum acquisition, retesting is triggered to perform infrared spectrum acquisition again on the to-be-tested sample, and the infrared detection value is subjected to a self-adding operation once;

[0083] If the signal-to-noise ratio SNR of the infrared spectrum of the sample is greater than or equal to a preset signal-to-noise ratio threshold, or the infrared detection value reaches a preset mark threshold, it is judged that the infrared spectrum acquisition of the to-be-detected sample is completed, the current acquisition time period is marked as an actual acquisition time period, and the sample infrared spectrum of the to-be-detected sample in the actual acquisition time period is marked as an actual sample infrared spectrum;

[0084] For multi-modal data acquisition, at the beginning of each acquisition time period, a multi-modal sensor group is used to acquire multi-modal data of the to-be-detected sample, the multi-modal data including temperature, PH value and conductivity of the to-be-detected sample, and a multi-modal data group of the to-be-detected sample in the acquisition time period is obtained by integration;

[0085] It should be noted that the purpose of this step is to acquire the infrared spectrum of the homogenized sample, to improve the spectrum quality through baseline calibration and signal-to-noise ratio screening, to synchronously acquire multi-modal data, to provide original data for subsequent analysis, to eliminate optical path and environmental interference through baseline calibration, to make the spectrum more truly reflect the sample composition, to remove low-quality spectra through signal-to-noise ratio cyclic detection, to ensure data reliability, and to synchronously acquire multi-modal data with the spectrum to provide time-matched environmental parameters for subsequent compensation;

[0086] S4: analyzing the multi-modal data group to determine whether the to-be-detected sample has a high abnormal risk, and if not, triggering multi-modal data compensation and compensating the actual sample infrared spectrum;

[0087] The actual sample infrared spectrum and the multi-modal data group of the to-be-detected sample in the actual acquisition time period are acquired, the multi-modal data group is analyzed, and it is determined whether the to-be-detected sample has a high abnormal risk;

[0088] A normal threshold range of the multi-modal data is set, including a temperature normal range, a PH normal range and a conductivity normal range, if the temperature of the to-be-detected sample in the multi-modal data group exceeds the temperature normal range, there may be a risk of deterioration or component denaturation of the fresh milk of the to-be-detected sample, if the PH value of the to-be-detected sample in the multi-modal data group exceeds the PH normal range, there may be a risk of microbial contamination of the fresh milk of the to-be-detected sample, and if the conductivity of the to-be-detected sample in the multi-modal data group exceeds the conductivity normal range, there may be a risk of adulteration or abnormality of the fresh milk of the to-be-detected sample;

[0089] The multi-modal data group is compared with the normal threshold range of the multi-modal data, if the multi-modal data group is not within the normal threshold range of the multi-modal data, it is determined that the to-be-detected sample has a high abnormal risk, the to-be-detected sample is marked as a high-risk sample, and the actual sample infrared spectrum and the multi-modal data group of the high-risk sample are sent to the data storage module for storage through communication networking;

[0090] If the multi-modal data set is within the normal threshold range of multi-modal data, it is determined that the sample to be detected does not have a high abnormal risk, and multi-modal data compensation is triggered;

[0091] If the multi-modal data compensation is triggered, the standard values of the multi-modal data set are set, including temperature standard values, PH standard values and conductivity standard values, and the actual sample infrared spectrum of the sample to be detected is compensated in combination with the actual sample infrared spectrum of the sample to be detected in the actual collection period and the multi-modal data set, to obtain the compensated actual sample infrared spectrum ;

[0092] ;

[0093] wherein, represents the compensation coefficient of the kth data in the multi-modal data set, which is obtained by fitting experimental data through a multiple linear regression model, and reflects the influence weight of each data on the infrared spectrum absorbance, represents the absolute difference between the kth data in the multi-modal data set and the standard value of the corresponding multi-modal data set;

[0094] The sample to be detected is marked as a normal sample, and the compensated actual sample infrared spectrum of the normal sample is sent to the data storage module for storage through communication networking;

[0095] It should be noted that the role of this step is to judge the abnormal risk of the sample through multi-modal data, to optimize the spectrum of the normal sample through multi-modal compensation, and to finally store the data in a hierarchical manner, so as to provide a high-quality data set for subsequent application, to quickly identify deteriorated, contaminated or adulterated samples through multi-modal abnormal risk assessment, to ensure data effectiveness, to eliminate the interference of environmental parameters on the spectrum based on the compensation model of multiple linear regression, to improve the accuracy of the spectrum, to facilitate data tracing and subsequent analysis through hierarchical storage, to replace the traditional single parameter compensation by using a multi-parameter compensation coefficient to quantify the influence of temperature, PH and conductivity on the spectrum, and to improve the accuracy of spectrum correction;

[0096] The technical scheme of the embodiment of the application is as follows: fresh milk samples are taken, a homogenization device is used to homogenize the fresh milk samples, and ultrasonic detection is used to evaluate the homogenization process to determine whether the homogenization process is completed, if so, a homogenized sample to be detected is obtained, collection hardware is deployed for infrared spectrum collection of the sample to be detected, the collection hardware includes an infrared spectrum detection optical system, a multi-modal sensor group and communication networking, the infrared spectrum detection optical system is used to collect the infrared spectrum of the sample to be detected, an actual collection period is marked and the actual sample infrared spectrum of the sample to be detected is obtained, the multi-modal sensor group is used to collect a multi-modal data set, the multi-modal data set is analyzed, it is determined whether the sample to be detected has a high abnormal risk, if not, multi-modal data compensation is triggered, and the actual sample infrared spectrum is compensated.

[0097] Embodiment 2

[0098] As Figure 2 shown, the fresh milk infrared spectrum data acquisition and storage device provided by the embodiment of the application comprises the following modules:

[0099] The sample processing module takes the fresh milk sample, uses a homogenization device to perform homogenization treatment on the fresh milk sample, and uses ultrasonic detection to evaluate the homogenization treatment and determine whether to end the homogenization treatment, if yes, a homogenization-treated sample to be detected is obtained.

[0100] The hardware arrangement module deploys acquisition hardware for infrared spectrum acquisition of the sample to be detected, and the acquisition hardware comprises an infrared spectrum detection light path system, a multi-modal sensor group and a communication network.

[0101] The data acquisition module uses the infrared spectrum detection light path system to perform infrared spectrum acquisition on the sample to be detected, marks an actual acquisition time period and obtains an actual sample infrared spectrum of the sample to be detected, and uses the multi-modal sensor group to acquire a multi-modal data group.

[0102] The spectrum compensation module analyzes the multi-modal data group, determines whether the sample to be detected has a high abnormal risk, if not, triggers multi-modal data compensation, and compensates the actual sample infrared spectrum.

[0103] The above shows and describes the basic principles, main features and advantages of the application. It should be understood by those skilled in the art that the application is not limited by the above embodiments, and the above embodiments and descriptions in the specification are only to illustrate the principles of the application, and various changes and improvements can be made without departing from the spirit and scope of the application, and these changes and improvements all fall within the scope of the application. The scope of protection of the application is defined by the appended claims and their equivalents.

Claims

1. A method for collecting and storing infrared spectrum data of raw milk, characterized in that: The method comprises the following steps: Take a fresh milk sample, homogenize the fresh milk sample by using a homogenizing device, and evaluate the homogenization by using ultrasonic detection to determine whether the homogenization is completed. If yes, a homogenized sample for detection is obtained. Deploy a collection hardware for infrared spectrum collection of the sample for detection. The collection hardware comprises an infrared spectrum detection light path system, a multi-modal sensor group, and a communication network. Collect the infrared spectrum of the sample for detection by using the infrared spectrum detection light path system, mark the actual collection period, and obtain the actual sample infrared spectrum of the sample for detection. Collect the multi-modal data set by using the multi-modal sensor group. Analyze the multi-modal data set to determine whether the sample for detection has a high abnormal risk. If not, trigger multi-modal data compensation, and compensate the actual sample infrared spectrum.

2. The method for collecting and storing the infrared spectrum data of fresh milk according to claim 1, characterized in that: The determination method of whether the homogenization is completed is as follows: Set a homogenization value for each sample for detection, and assign an initial value of 0. Obtain the variation coefficient of ultrasonic propagation speed and the variation coefficient of attenuation coefficient. If the variation coefficient of ultrasonic propagation speed and the variation coefficient of attenuation coefficient are both less than a preset variation coefficient threshold, or the homogenization value of the sample for detection reaches a preset flag threshold, it is determined that the homogenization of the sample for detection is completed. Otherwise, it is determined that the homogenization of the sample for detection is not completed. The sample for detection is homogenized again, and the homogenization value is increased by one.

3. The method for collecting and storing the mid-infrared spectrum data of fresh milk according to claim 2, characterized in that: The obtaining method of the variation coefficient of ultrasonic propagation speed and the variation coefficient of attenuation coefficient is as follows: Obtain the sample for detection, detect the sample for detection by using an ultrasonic detector, uniformly select a plurality of sampling points on the surface of the sample for detection, perform ultrasonic detection on each sampling point multiple times, record the propagation distance and propagation time of the ultrasonic signal in the sample for detection, and perform data processing to obtain the ultrasonic propagation speed. For each ultrasonic detection, record the incident ultrasonic intensity of the ultrasonic signal and the ultrasonic intensity after penetrating the sample for detection, and calculate the attenuation coefficient. Perform data processing on the ultrasonic propagation speed and the attenuation coefficient of each ultrasonic detection of each sampling point, and calculate the variation coefficient of ultrasonic propagation speed and the variation coefficient of attenuation coefficient.

4. The method for collecting and storing the mid-infrared spectrum data of fresh milk according to claim 3, characterized in that: The obtaining method of the sample for detection is as follows: Set a homogenizing device for the fresh milk whose infrared spectrum is to be collected. Homogenize the fresh milk whose infrared spectrum is to be collected by using the homogenizing device. The homogenizing device comprises a pressure adjusting device and a temperature adjusting device. Add an adaptive adjusting module to the pressure adjusting device and the temperature adjusting device. The adaptive adjusting module collects homogenization pressure data and homogenization temperature data in real time, and performs deviation calculation with the set homogenization pressure and homogenization temperature in real time, respectively. If the deviation exceeds the standard, use the adaptive control algorithm to adjust the homogenization pressure data with the set homogenization pressure or homogenization temperature as the target. The homogenized fresh milk sample is introduced into a detection container to obtain a sample for detection.

5. The method for collecting and storing the mid-infrared spectrum data of fresh milk according to claim 1, characterized in that: The obtaining method of the actual sample infrared spectrum is as follows: Mark a time period for one infrared spectrum acquisition of one sample to be detected as an acquisition time period, set an infrared detection value for each sample to be detected and assign an initial value of 0 to it; Obtain the signal-to-noise ratio of the sample infrared spectrum, if less than a preset signal-to-noise ratio threshold, trigger re-inspection to perform infrared spectrum acquisition again on the sample to be detected, and make the infrared detection value perform a self-adding operation once, otherwise, or if the infrared detection value reaches a preset threshold, mark the current acquisition time period as an actual acquisition time period, and mark the sample infrared spectrum in the actual acquisition time period as an actual sample infrared spectrum.

6. The method for collecting and storing the mid-infrared spectrum data of fresh milk according to claim 5, characterized in that: The signal-to-noise ratio of the sample infrared spectrum is obtained in the following manner: Obtain the sample infrared spectrum, select the average signal power of the characteristic absorption peak region and the average noise power of the noise region in the sample infrared spectrum for data processing, and calculate the signal-to-noise ratio of the sample infrared spectrum.

7. The method for collecting and storing the mid-infrared spectrum data of fresh milk according to claim 6, characterized in that: The sample infrared spectrum is obtained in the following manner: Use the infrared spectrum detection optical path system to collect the infrared spectrum of the sample to be detected and perform smoothing processing to obtain the original sample infrared spectrum, use the blank correction method to set distilled water containing no measured component as a blank sample, use the infrared spectrum detection optical path system to collect the infrared spectrum of the blank sample, and fit to obtain the baseline infrared spectrum; Based on the baseline infrared spectrum, baseline calibration is performed on the collected original sample infrared spectrum to obtain the sample infrared spectrum.

8. The method for collecting and storing the mid-infrared spectrum data of fresh milk according to claim 1, characterized in that: The judgment method for whether the sample to be detected has a high abnormal risk is as follows: Obtain the actual sample infrared spectrum and the multi-modal data set of the sample to be detected in the actual acquisition time period, set a normal threshold range of the multi-modal data, if the multi-modal data set is not within the normal threshold range of the multi-modal data, judge that the sample to be detected has a high abnormal risk, otherwise, judge that the sample to be detected has no high abnormal risk.

9. The method for collecting and storing the mid-infrared spectrum data of fresh milk according to claim 1, characterized in that: The multi-modal data compensation includes: Set a standard value of the multi-modal data set, combine the actual sample infrared spectrum and the multi-modal data set of the sample to be detected in the actual acquisition time period, fit the experimental data through a multi-linear regression model to obtain a compensation coefficient of each data in the multi-modal data set, and perform data processing compensation on the actual sample infrared spectrum of the sample to be detected using the compensation coefficient to obtain the compensated actual sample infrared spectrum.

10. A raw milk mid-infrared spectroscopy data acquisition storage device for implementing the acquisition storage method according to any one of claims 1 to 9, characterized in that it comprises: It includes the following modules: Sample processing module: take fresh milk samples, use a homogenizing device to homogenize the fresh milk samples, and use ultrasonic detection to evaluate the homogenization process to determine whether to end the homogenization process, if so, obtain the homogenized sample to be detected; Hardware arrangement module: deploy acquisition hardware for infrared spectrum acquisition of the sample to be detected, the acquisition hardware includes an infrared spectrum detection optical path system, a multi-modal sensor group, and a communication group network; Data acquisition module: use the infrared spectrum detection optical path system to collect the infrared spectrum of the sample to be detected, mark the actual acquisition time period and obtain the actual sample infrared spectrum of the sample to be detected, and use the multi-modal sensor group to collect the multi-modal data set; Spectrum compensation module: analyze the multi-modal data set to determine whether the sample to be detected has a high abnormal risk, if not, trigger multi-modal data compensation to compensate the actual sample infrared spectrum.

Images