CN116012383A

CN116012383A - Data processing method for chondroitin sulfate production monitoring

Info

Publication number: CN116012383A
Application number: CN202310308330.6A
Authority: CN
Inventors: 姜开岭
Original assignee: Shandong Xinsheng Bio Tech Co ltd
Current assignee: Shandong Xinsheng Bio Tech Co ltd
Priority date: 2023-03-28
Filing date: 2023-03-28
Publication date: 2023-04-25
Anticipated expiration: 2043-03-28
Also published as: CN116012383B

Abstract

The invention relates to the technical field of chondroitin sulfate preparation, in particular to a data processing method for chondroitin sulfate production monitoring, which comprises the steps of acquiring monitoring data and solution state monitoring images of a precipitation inverse solution corresponding to each set moment in the past in the precipitation generation process of chondroitin sulfate production; determining each suspected abnormal data point according to the monitoring data corresponding to each set time; according to the pixel value of each pixel point in each solution state monitoring image and the sedimentation connected domain in the image, determining a sedimentation fluctuation value corresponding to each set moment, and further determining monitoring data corresponding to a real abnormal data point in the suspected abnormal data points. According to the scheme, the problem that abnormal monitoring data in the existing chondroitin sulfate production and preparation process is not accurate enough in determination can be effectively solved in the chondroitin sulfate production and preparation process, and therefore accurate monitoring of the abnormal data can be achieved.

Description

Data processing method for chondroitin sulfate production monitoring

Technical Field

The invention relates to the technical field of chondroitin sulfate preparation, in particular to a data processing method for chondroitin sulfate production monitoring.

Background

The chondroitin sulfate has a great deal of biological functions and has wide application in clinical medical treatment, and in the existing preparation process of the chondroitin sulfate, cartilage tissues such as pig laryngeal bones, nasal bone, trachea and the like are generally used as raw materials, and the chondroitin sulfate is prepared by the reaction extraction processes such as alkaline hydrolysis, salt hydrolysis, enzymolysis and the like, and precipitation. Taking pig cartilage as a raw material to extract chondroitin sulfate as an example, the specific reaction extraction process mainly comprises a leaching stage, an enzymolysis stage, an adsorption stage and a precipitation stage. Wherein, in the leaching stage, naOH is used for stirring and dissolving the filtered pig cartilage residues, hydrochloric acid is added into the filtrate for neutralization and alkalinity so as to adjust the pH value to be weak alkalinity; in the enzymolysis stage, pancreatin is added for hydrolysis, the temperature is controlled, the hydrolysis time is usually about 7 hours, and NaOH is added at any time to adjust the pH value; in the adsorption stage, hydrochloric acid is added after the hydrolysis is finished to make the pH value neutral, tao Baitu and active carbon are added for adsorption, the pH value is stirred and adjusted, and the pH value is adjusted to be weak acidity and precipitation and filtration are carried out after the adsorption is finished; and in the precipitation stage, adding NaOH into the filtrate to adjust the pH value, adding sodium chloride solution to dissolve, filtering to clarify, adding ethanol, stirring, and standing for more than 8 hours to obtain the chondroitin sulfate precipitation product.

In the process of reaction extraction of chondroitin sulfate, conventionally, the reaction condition is usually judged empirically to determine the addition amount of raw materials, or excessive materials are added, which results in waste of raw materials to some extent. Along with standardization of the preparation process, in the process of reactive extraction of chondroitin sulfate, a sensor is used to monitor the temperature value and the pH value in the process of reactive extraction, and whether the current reaction is completed or not is judged according to the monitored temperature value and the monitored pH value which change with time, so that whether raw materials are continuously added or not and the specific amount of the raw materials are continuously added are determined. Because the temperature and the pH value need to be limited in the reaction extraction process of the chondroitin sulfate, the chondroitin sulfate belongs to the factors of heat absorption, heat release and reversibility of organic matters when the reaction occurs, and the temperature value and the pH value obtained by monitoring can be caused to generate abnormal data when the pH value and the temperature are fixed. For example, in the precipitation generation stage in the process of extracting chondroitin sulfate, that is, from the beginning of precipitation to the end of precipitation, more off-cluster data appears at the temperature value and the pH value detected by the sensor, and it should be noted that the off-cluster data is data exceeding the temperature or pH limit compared with the data acquired at the normal limit temperature. For example, the pH is limited to 8.8-9.0 during hydrolysis, and the temperature is kept at 53-54 ℃; or ph=6.0 need to be adjusted when adding alcohol to precipitate out the finished product, and the difference from the defined interval is identified as the off-cluster data. Specifically, outliers obtained by extracting the collected data by using an LOF outlier factor algorithm are the outlier data; the cluster data is partly abnormal data and partly error data, and the reasons for generating the error data include uneven distribution of solution materials around the sensor, deposition coverage of the sensor and the like, so that the abnormal data in the cluster data needs to be screened. However, when monitoring data such as a temperature value and a pH value in the reaction extraction process of chondroitin sulfate are processed, the cluster-separated data is generally directly used as abnormal noise monitoring data, which leads to inaccurate finally determined abnormal monitoring data, further influences the judgment of the subsequent extraction reaction result of chondroitin sulfate, and finally influences the extraction process of chondroitin sulfate.

Disclosure of Invention

The invention aims to provide a data processing method for chondroitin sulfate production monitoring, which is used for solving the problem that abnormal monitoring data in the existing chondroitin sulfate production process is not accurate enough to determine.

In order to solve the technical problems, the invention provides a data processing method for monitoring production of chondroitin sulfate, which comprises the following steps:

in the precipitation generation process of chondroitin sulfate production, monitoring data and solution state monitoring images corresponding to each set time in the past of a reaction solution are obtained, wherein the monitoring data comprise at least one of a temperature value and a pH value;

constructing data points corresponding to all the set moments according to the monitoring data corresponding to all the set moments, and determining all suspected abnormal data points in all the data points;

determining a sedimentation intensity value corresponding to each solution state monitoring image according to the pixel value of each pixel point in each solution state monitoring image;

determining each precipitation connected domain in each solution state monitoring image, determining the precipitation equivalent distance corresponding to each solution state monitoring image according to the position of the precipitation connected domain, and determining the precipitation equivalent radius corresponding to each solution state monitoring image according to the distribution of the precipitation connected domains;

Determining a precipitation degree value corresponding to each solution state monitoring image according to the precipitation intensity value, the precipitation equivalent distance and the precipitation equivalent radius corresponding to each solution state monitoring image;

determining a sediment fluctuation value corresponding to each set time according to the sediment degree value of the solution state monitoring image at each set time and the adjacent set time before and after the set time;

and determining real abnormal data points in the suspected abnormal data points according to the precipitation fluctuation values at the set time corresponding to the suspected abnormal data points, and determining monitoring data corresponding to the real abnormal data points as abnormal data.

Further, determining a sedimentation intensity value corresponding to each solution state monitoring image includes:

according to the pixel value of each pixel point in each solution state monitoring image, the pixel value comprises an R channel value, a G channel value and a B channel value, and the minimum value in the pixel value of each pixel point is determined as the dark channel value of the pixel point;

and calculating the average value of the dark channel values of all pixel points in each solution state monitoring image, and determining the average value of the dark channel values as a sedimentation intensity value corresponding to each solution state monitoring image.

Further, determining a precipitation equivalent distance corresponding to each solution state monitoring image includes:

Determining the mass centers of all the sedimentation connected domains in each solution state monitoring image, thereby obtaining a mass center set;

selecting one centroid as an initial target centroid in the centroid set at will, determining Euclidean distance between the initial target centroid and each centroid in the centroid set, determining the minimum value in all Euclidean distances, and determining the minimum value as the minimum Euclidean distance corresponding to the initial target centroid;

deleting the initial target centroid from the centroid set to obtain an updated centroid set, taking another centroid corresponding to the minimum Euclidean distance corresponding to the initial target centroid in the updated centroid set as a new target centroid, determining the Euclidean distance between the new target centroid and each other centroid in the updated centroid set to determine the minimum Euclidean distance corresponding to the new target centroid, and repeating the steps to continuously determine the new target centroid and determine the minimum Euclidean distance corresponding to the new target centroid until the updated centroid set is an empty set, thereby obtaining each minimum Euclidean distance;

and calculating the average value of each minimum Euclidean distance, and determining the average value of each minimum Euclidean distance as the precipitation equivalent distance corresponding to each solution state monitoring image.

Further, determining a precipitation equivalent radius corresponding to each solution state monitoring image includes:

determining the perimeter and the area of each sedimentation communicating domain in each solution state monitoring image according to the distribution of each sedimentation communicating domain in each solution state monitoring image;

determining a first radius of each sedimentation communicating domain in each solution state monitoring image according to a circumference of each sedimentation communicating domain in each solution state monitoring image and a circular circumference calculation formula;

determining a second radius of each sedimentation communicating domain in each solution state monitoring image according to the area of each sedimentation communicating domain in each solution state monitoring image and a circular area calculation formula;

determining the weight corresponding to the first radius and the second radius of each sedimentation communicating domain in each solution state monitoring image according to the perimeter and the area of each sedimentation communicating domain in each solution state monitoring image;

according to the weight corresponding to the first radius and the second radius of each sedimentation communicating domain in each solution state monitoring image, carrying out weighted summation on the first radius and the second radius of each sedimentation communicating domain, so as to obtain the comprehensive radius corresponding to each sedimentation communicating domain in each solution state monitoring image;

And calculating the average value of the comprehensive radiuses corresponding to all the sedimentation connected domains in each solution state monitoring image, and determining the average value of the comprehensive radiuses as the sedimentation equivalent radius corresponding to each solution state monitoring image.

Further, a calculation formula corresponding to the weight corresponding to each of the first radius and the second radius of each sedimentation connected domain in each solution state monitoring image is determined as follows:

wherein ,

the weight corresponding to the first radius of the jth precipitate connected domain in the ith solution state monitoring image,

the weight corresponding to the second radius of the jth precipitate connected domain in the ith solution state monitoring image,

the area of the jth precipitate connected domain in the image is monitored for the ith solution state,

and (3) monitoring the perimeter of the jth sedimentation connected domain in the image for the ith solution state, wherein I is the absolute value sign, and pi is the perimeter rate.

Further, determining a precipitation degree value corresponding to each solution state monitoring image includes:

calculating a product value of a sedimentation intensity value, a sedimentation equivalent distance and a sedimentation equivalent radius corresponding to each solution state monitoring image, carrying out normalization processing on the product value, and determining a normalization processing result as a sedimentation degree value corresponding to each solution state monitoring image.

Further, constructing data points corresponding to each set time includes:

the monitoring data comprises a temperature value and a pH value;

comparing the pH value corresponding to each set time with the prior standard pH value corresponding to each set time, thereby obtaining a pH residual value corresponding to each set time;

comparing the temperature value corresponding to each set time with the prior standard temperature value corresponding to each set time, thereby obtaining a temperature residual value corresponding to each set time;

and taking each set time and the corresponding pH residual value and temperature residual value as three coordinate values of a three-dimensional coordinate system, determining a three-dimensional coordinate point by the three coordinate values, and determining the three-dimensional coordinate point as a data point corresponding to each set time, thereby obtaining the data point corresponding to each set time.

Further, determining suspected outlier data points of all data points includes:

calculating local outlier factors of all data points by using a local outlier detection algorithm according to three coordinate values of all data points;

and comparing the local outlier factors of all the data points with the set outlier factor threshold, and determining the data points with the local outlier factors larger than the set outlier factor threshold as suspected abnormal data points.

Further, determining a sediment fluctuation value corresponding to each set time includes:

and calculating the absolute value of the difference between the sedimentation degree values of the solution state monitoring images at each set time and each adjacent set time according to the sedimentation degree values of the solution state monitoring images at each set time and the front and rear adjacent set times, and determining the average value of the absolute values of the differences between each set time and all the adjacent set times as the sedimentation fluctuation value corresponding to each set time.

Further, determining a true outlier data point of the suspected outlier data points includes:

and comparing the precipitation fluctuation value of each suspected abnormal data point at the set moment with a set precipitation fluctuation threshold value, and determining the suspected abnormal data point corresponding to the precipitation fluctuation value smaller than the set precipitation fluctuation threshold value as a real abnormal data point.

The invention has the following beneficial effects: in the preparation process of chondroitin sulfate, the abnormal data in the monitoring data can be accurately determined by acquiring the solution state monitoring image corresponding to the monitoring data. Specifically, in the precipitation generation process of chondroitin sulfate production, in order to know the precipitation generation condition of the reaction solution in real time, monitoring data corresponding to each set time of the reaction solution in the past are obtained, and because abnormal data may be mixed in the monitoring data, in order to facilitate subsequent screening of the abnormal monitoring data, solution state monitoring images corresponding to each set time of the reaction solution in the past are also required to be obtained. And constructing data points according to the monitoring data corresponding to each set time, and primarily screening the data points to screen out each suspected abnormal data point which is possibly corresponding to the abnormal monitoring data. Considering that the transparency of the solution changes along with the generation of the sediment in the solution, the density of the sediment becomes larger and larger, and the sediment caking degree becomes stronger, calculating the sediment intensity value corresponding to each solution state monitoring image according to the pixel value of each pixel point in the solution state monitoring image, and accurately representing the current sediment degree condition of the solution by using the sediment intensity value. As the precipitation generation time in the solution advances, the size, shape, loosening degree and the like of the precipitation in the solution are changed, in order to accurately extract the precipitation characteristics of the reaction solution, each precipitation connected domain in each solution state monitoring image is determined, each precipitation connected domain corresponds to one precipitation block, and then the precipitation equivalent distance and the precipitation equivalent radius corresponding to each solution state monitoring image are determined according to the precipitation connected domains. As the precipitation particles in the solution are dispersed and evenly distributed in the initial stage of precipitation formation, the distance between the precipitation particles is smaller, the size of the precipitation particles is smaller, the precipitation particles are mutually aggregated along with the progress of precipitation reaction, the number of the precipitation particles is reduced, the size and the mass of the precipitation particles are continuously increased, the precipitation particles are clustered, the distance between the precipitation is gradually increased, and the precipitation size is also gradually increased, so that the size of the precipitation size is represented by determining the equivalent precipitation distance and the equivalent precipitation radius and utilizing the equivalent precipitation distance to represent the size of the precipitation, and the precipitation distribution condition of the current solution is facilitated to be determined. And comprehensively considering the precipitation intensity value, the precipitation equivalent distance and the precipitation equivalent radius, determining a precipitation degree value, and accurately measuring the comprehensive characteristics of precipitation in the current solution by using the precipitation degree value. Considering that the sediment is continuously formed in the sediment generation stage, the characteristics of the sediment are continuously changed, and therefore the sediment degree value of the solution state monitoring image at each set time and the adjacent set time before and after the set time is compared, the sediment fluctuation value corresponding to each set time can be determined, and the change condition of the current sediment distribution characteristics is measured by utilizing the sediment fluctuation value. And finally, based on the sediment fluctuation value corresponding to each set time, taking the monitoring data corresponding to each suspected abnormal data point into consideration, and finally, accurately determining the monitoring data corresponding to the real abnormal data point in each suspected abnormal data point, wherein the monitoring data is the real abnormal data. According to the method, the solution state monitoring images corresponding to the monitoring data are synchronously acquired, the solution state monitoring images are subjected to image processing, the precipitation fluctuation value corresponding to each set moment is obtained, the abnormal data summarized by the monitoring data are accurately determined by utilizing the precipitation fluctuation value, and the problem that the abnormal monitoring data in the existing chondroitin sulfate production and preparation process are not accurately determined is effectively solved.

Drawings

In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.

FIG. 1 is a flow chart of a data processing method for chondroitin sulfate production monitoring according to an embodiment of the present invention;

FIG. 2 is a schematic structural view of a reaction vessel according to an embodiment of the present invention;

FIG. 3 is a solution status monitoring image according to an embodiment of the present invention;

FIG. 4 is a graph showing the change of the pH and temperature values with time of the monitored portion according to the embodiment of the present invention;

FIG. 5 is a graph showing the relationship between the alkali concentration (pH) and the yield in the example of the present invention;

wherein: 1 is a stirring rod, 2 is a motor, 3 is a feeding window, 4 is a discharging window, 5 is an impurity outlet, and 6 is a window.

Detailed Description

In order to further describe the technical means and effects adopted by the present invention to achieve the preset purpose, the following detailed description is given below of the specific implementation, structure, features and effects of the technical solution according to the present invention with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. In addition, all parameters or indices in the formulas referred to herein are values after normalization that eliminate the dimensional effects.

The embodiment provides a data processing method for monitoring production of chondroitin sulfate, and a corresponding flow chart is shown in fig. 1, and the method comprises the following steps:

step S1: in the precipitation generation process of chondroitin sulfate production, monitoring data and solution state monitoring images of the reaction solution corresponding to all set moments in the past are obtained, wherein the monitoring data comprise at least one of a temperature value and a pH value.

In the production process of chondroitin sulfate, the whole reaction extraction process of the chondroitin sulfate is finished in an industrial reaction kettle, as shown in fig. 2, a stirring rod 1 is arranged in the reaction kettle, the stirring rod 1 is driven by a motor 2 to work, a feeding window 3 and a discharging window 4 are arranged on the side wall of the reaction kettle, and an impurity outlet 5 is arranged at the bottom of the reaction kettle. The inner wall of the reaction kettle is also provided with a monitoring sensor (the monitoring sensor is not shown in fig. 2) for monitoring the pH value and the temperature value in the reaction extraction process of the acid soft ossein. In addition, a window 6 is opened at the top of the reaction kettle, and a solution state monitoring photographing device (the solution state monitoring photographing device is not shown in fig. 2) is provided at the window 6 to monitor the solution state during the reaction extraction process of the acid soft bone element, thereby obtaining a solution state monitoring image.

In the process of the reaction extraction of the chondroitin sulfate, the extraction reaction solution can change in properties by adjusting the pH value, the temperature value and the like. In order to monitor the temperature value and the pH value of the sediment generation stage, namely the adsorption stage, the pH value and the temperature value are collected once at each set time interval by a temperature value and pH value monitoring sensor arranged in the reaction kettle at the beginning of sediment formation, and each collection time is taken as a set time, so that the pH value and the temperature value of each set time can be obtained, namely the monitoring data of each set time can be obtained. In the present embodiment, the set time interval is set to 5s, that is, the pH value and the temperature value are collected every 5s, and of course, the 5s herein is only given as a specific example, and the appropriate time interval may be set according to the actual length of the precipitate generating stage. Meanwhile, in order to facilitate the subsequent analysis of the monitoring data such as the temperature value and the pH value to determine abnormal data in the monitoring data, the upper surface of the solution in the reaction kettle is photographed at each set time by a solution state monitoring photographing device set at the top of the reaction kettle, so as to obtain a solution state monitoring image at each set time, wherein the solution state monitoring image is an RGB image, as shown in FIG. 3.

FIG. 3 is a microscopic image of free precipitate; it is to be noted that the content of the precipitated block is different between the solutions in different states, the components in the turbid solution are uniformly distributed, the solution with the precipitate forms a glue-like solution under the stirring of the stirring rod, and the existing research shows that the excessive temperature can deepen the color of the finished product of the generated cartilage cream, and the content of the produced finished product is reduced, namely the precipitation is reduced, due to the fact that the pH is not in accordance with the requirement. Precipitation of the solution can cause turbidity of the solution, i.e., the transparency of the solution changes, and the strength of the precipitate and the transparency of the precipitate, the changes in transparency and color can reflect the changes in the solution based on time sequence. Suspended sediment can appear in the solution through different reactions, the quantity, the size and the like of the sediment can reflect the change of the reactions, along with the time and the degree of the reactions, the sediment gradually forms larger to obtain suspended matters from smaller, and finally the suspended matters are deposited to the bottom of the solution, and the size, the shape, the loosening degree of the sediment, the shape of flocs and the like of free sediment of the solution can be changed in the process.

Step S2: and constructing data points corresponding to all the set moments according to the monitoring data corresponding to all the set moments, and determining all suspected abnormal data points in all the data points.

In order to facilitate the subsequent screening of abnormal data in the monitored monitoring data, according to the monitoring data corresponding to each set time, constructing data points corresponding to each set time, the implementation steps include:

the monitoring data comprises a temperature value and a pH value;

Specifically, during the production of chondroitin sulfate, the pH and temperature values at each set point are monitored to change in real time, but when the reaction solutions are in the same steady state, the pH and temperature values are often relatively close. Fig. 4 shows a three-dimensional schematic of a solution from the start of precipitation to the end of precipitation, in which two data clusters in a comparative set of data, each corresponding to a steady state of the solution, are present together. During the transition of a solution from one stable state to another, there are also a plurality of discrete outlier points, which may be monitoring data of normal reactions or abnormal data, so that they need to be identified first for further screening of abnormal data therein. Since the purpose of fig. 4 is to illustrate the general distribution of data, specific coordinate values are not given.

The local outlier detection algorithm (Local Outlier Factor, LOF) is considered to accurately identify outlier data points, so that the LOF algorithm can be utilized to screen the monitored pH and temperature values at each set time, thereby obtaining the local outlier pH and temperature values. However, since the monitored pH and temperature values have fluctuations on a certain trend, and the fluctuation intervals of the pH and temperature values in different time periods are different, the detection of outlier data points by directly using the LOF algorithm may result in inaccurate outlier data points that are finally identified.

In order to improve the accuracy of the outlier data points determined later, the embodiment also needs to obtain the prior standard pH value and the prior standard temperature value corresponding to each set time in advance, where the prior standard pH value and the prior standard temperature value refer to the normal standard pH value and the normal standard temperature value corresponding to each set time under normal conditions, and these normal standard pH value and temperature value can be determined by multiple measurement and averaging to determine, etc., and the specific determination mode does not belong to the focus of attention of the scheme and is not repeated here. And comparing the monitored pH value corresponding to each set time with the prior standard pH value corresponding to each set time, and calculating residual data to obtain the pH residual value corresponding to each set time. And meanwhile, comparing the monitored temperature value corresponding to each set time with the prior standard temperature value corresponding to each set time, and calculating residual data so as to obtain the temperature residual value corresponding to each set time. Since the specific calculation process of the residual data belongs to the prior art, the description is omitted here.

After each set time and the corresponding pH residual value and temperature residual value are obtained, for any set time, the set time is taken as an x-axis coordinate value, the pH value corresponding to the set time is taken as a y-axis coordinate value, and the temperature value corresponding to the set time is taken as a z-axis coordinate value, so that one three-dimensional coordinate point of the set time under a three-dimensional coordinate system can be determined. In this way, three-dimensional coordinate points corresponding to the respective set times, that is, data points corresponding to the respective set times, can be determined.

After the data points corresponding to each set time are obtained, local outlier detection is carried out on the data points by using an LOF algorithm, so that each suspected outlier data point in the data points can be obtained, and the implementation steps comprise:

Specifically, local outlier detection is performed on the data points by using an LOF algorithm according to the data points corresponding to each set time, so that local outlier factors corresponding to each data point can be obtained. When the local outlier is less than 1, the higher the density of the corresponding data point is, the more likely the corresponding data point is a dense point, and when the local outlier is greater than 1, the lower the density of the corresponding data point is, the more likely the corresponding data point is an outlier. Therefore, a set outlier factor threshold is set, the value of the set outlier factor threshold is set to be 1, local outlier factors of all data points are respectively compared with the set outlier factor threshold 1, and when the local outlier factors are larger than the set outlier factor threshold 1, the corresponding data points are determined to be suspected abnormal data points. In this way, each suspected abnormal data point in all data points formed by the monitored pH value and the temperature value at each set time can be determined, and then the monitored solution state monitoring image is analyzed, so that the real abnormal data point in each suspected abnormal data point can be identified.

Step S3: and determining a sedimentation intensity value corresponding to each solution state monitoring image according to the pixel value of each pixel point in each solution state monitoring image.

In the production process of chondroitin sulfate, when precipitation occurs in a reaction extraction solution, the precipitation can cause turbidity of the solution, the transparency of the solution is reduced, the solution is available according to the prior principle of dark channels, when the solution is more opaque, the gray value of the dark channel value of a pixel point in an image corresponding to the solution is larger, the corresponding precipitation strength is larger, the density of precipitation caking is higher, the precipitation caking is tighter, and the density of the corresponding precipitation caking is higher. Therefore, the pixel value of each pixel point in each solution state monitoring image is analyzed, the dark channel value of each pixel point is determined, and then the sedimentation intensity value corresponding to each solution state monitoring image is determined, and the implementation steps comprise:

Specifically, for each solution state monitoring image, the solution state monitoring image is an RGB image, as shown in fig. 3, according to the pixel value of each pixel point in the solution state monitoring image, that is, the R channel value, the G channel value, and the B channel value of each pixel point, the minimum value of the three channel values of each pixel point is determined, and the minimum value is the dark channel value of the corresponding pixel point. And then calculating the average value of the dark channel values of all pixel points in the solution state monitoring image, and determining the average value as a sedimentation intensity value corresponding to the solution state monitoring image. In this way, a corresponding sedimentation intensity value for each of the acquired solution state monitoring images can be determined. When the precipitation intensity value is larger, the solution at the acquisition time of the corresponding solution state monitoring image is more turbid, the corresponding solution state monitoring image cannot see the color of the original solution clearly, and at the moment, the higher the precipitation intensity of the solution is, the higher the precipitation degree of the solution is, and the higher the precipitation density is. The precipitation intensity value corresponding to each solution state monitoring image is extracted, and the precipitation condition of the solution during the acquisition of the solution state monitoring images is accurately represented by the precipitation intensity value, so that the distribution characteristics of precipitation in the solution are accurately extracted conveniently, and the real abnormal data points in each suspected abnormal data point are screened.

Step S4: determining each precipitation connected domain in each solution state monitoring image, determining the precipitation equivalent distance corresponding to each solution state monitoring image according to the position of the precipitation connected domain, and determining the precipitation equivalent radius corresponding to each solution state monitoring image according to the distribution of the precipitation connected domains.

In the precipitation generation process of chondroitin sulfate production, suspended precipitation occurs through different reactions in the solution, the quantity, the size and the like of the precipitation can reflect the change of the reactions, smaller precipitation gradually forms larger suspended matters along with the promotion of the reaction time, and finally the suspended matters are precipitated to the bottom of the solution, and the size, the shape, the loosening degree of the precipitation, the shape of flocs and the like of free precipitation of the solution are changed in the whole process. In order to accurately extract the precipitation characteristics of the reaction solution, for each solution state monitoring image, edge detection is performed on the solution state monitoring image by using a Canny edge segmentation operator, so as to obtain each precipitation connected domain, each precipitation connected domain corresponds to one suspended precipitate, and the precipitation equivalent distance corresponding to each solution state monitoring image can be determined by analyzing the position of each precipitation connected domain, and the implementation steps comprise:

Specifically, in the process of generating the precipitate in the production of chondroitin sulfate, the size and shape of the corresponding precipitate at different times are different, and the distribution of the precipitate blocks in the solution is also different due to the different amounts of the precipitate precipitated from the solution. As the precipitation reaction proceeds, suspended precipitated nodules coagulate into clusters, the size of the nodules becomes larger and the shape becomes more regular, the density of the precipitates becomes larger and the number of suspended precipitates in the solution becomes smaller. In order to measure the distribution condition of the sediment in the solution, for each solution state monitoring image, the mass center of each sediment connected domain in the solution state monitoring image is determined, the mass center is used as the position of the corresponding sediment connected domain, and all mass centers are utilized to form a mass center set.

For each solution state monitoring image, determining a precipitation equivalent distance corresponding to the solution state monitoring image based on a centroid set formed by centroids of all precipitation connected domains in the solution state monitoring image, wherein the specific determination process comprises the following steps: for the ith solution state monitoring image, a centroid set formed by centroids of all precipitate connected domains in the ith solution state monitoring image

Optionally selecting a centroid as the initial target centroid

Calculating the initial target centroid

And centroid set

The Euclidean distance between other centroids in the model is taken as the initial target centroid, and the Euclidean distance corresponding to the minimum value in all Euclidean distances

Corresponding minimum euclidean distance. Assuming an initial target centroid

Is combined with centroid

Other centroids in (1)

The Euclidean distance between the two is minimum, and the other barycenter is then

As a new target centroid, while the initial target centroid is to be taken

From a centroid set

To delete, thereby realizing centroid set

Updates of (i) i.e. updated centroid sets

Not including the target centroid

. Then calculate a new target centroid

With updated centroid sets

The Euclidean distance between other centroids in the model is taken as the new target centroid, wherein the Euclidean distance corresponding to the minimum value in all Euclidean distances

Corresponding minimum euclidean distance. Assuming a new target centroid

Is combined with the updated centroid set

Other centroids in (1)

As a new target centroid, and simultaneously, the new target centroid

From updated centroid sets

To delete, thereby realizing centroid set

Further updates of (a), i.e. further updated centroid sets

Not including the target centroid

And

. Calculating a new target centroid

And further updated centroid sets

Repeating the steps until the updated centroid set is an empty set, and obtaining each minimum Euclidean distance. After each minimum Euclidean distance corresponding to the ith solution state monitoring image is obtained, calculating an average value among all the minimum Euclidean distances, and determining the average value as a precipitation equivalent distance corresponding to the ith solution state monitoring image.

Through the steps, the sedimentation equivalent distance corresponding to each solution state monitoring image can be determined, and in the process of determining the sedimentation equivalent distance, the Euclidean distance between each target centroid and the nearest centroid is determined, and the average value of the Euclidean distances is calculated to obtain the sedimentation equivalent distance. In the initial stage of precipitate formation, the precipitate particles in the solution are distributed discretely and uniformly, and the value of the equivalent distance of the precipitate is relatively small. As the precipitation reaction proceeds, the precipitation particles aggregate with each other, the number of precipitation particles decreases, the size and mass of the precipitation particles are continuously increased, the precipitation particles aggregate into clusters, and the value of the precipitation equivalent distance becomes gradually larger at this time. Therefore, according to the precipitation equivalent distance corresponding to each solution state monitoring image, the precipitation distribution condition of the current solution can be determined, so that the distribution characteristics of precipitation in the current solution are determined, and the real abnormal data points in each suspected abnormal data point are screened out.

Meanwhile, in order to further extract the characteristics of the sediment in the reaction solution accurately, for each solution state monitoring image, the distribution condition of each sediment connected domain in the solution state monitoring image is analyzed, and the equivalent radius of the sediment corresponding to each solution state monitoring image can be determined, and the implementation steps comprise:

Specifically, for each solution state monitoring image, according to the perimeter of each sedimentation connected domain in the solution state monitoring image, determining a first radius of each sedimentation connected domain in each solution state monitoring image according to a circular perimeter calculation formula, and simultaneously, according to the area of each sedimentation connected domain in the solution state monitoring image, determining a second radius of each sedimentation connected domain in each solution state monitoring image according to a circular area calculation formula, wherein the corresponding calculation formula is as follows:

wherein ,

the first radius of the jth precipitate connected domain in the image is monitored for the ith solution state,

a second radius of the jth precipitate connected domain in the ith solution state monitoring image,

the perimeter of the jth precipitate connected domain in the image is monitored for the ith solution state,

the area of the jth sedimentation connected domain in the image is monitored for the ith solution state, and pi is the circumference ratio.

In order to comprehensively consider the first radius and the second radius of each sedimentation connected domain in each solution state monitoring image, thereby determining the comprehensive radius of each sedimentation connected domain in each solution state monitoring image, the respective corresponding weights of the first radius and the second radius of each sedimentation connected domain in each solution state monitoring image are acquired according to the perimeter and the area of each sedimentation connected domain in each solution state monitoring image, and the corresponding calculation formula is as follows:

wherein ,

According to the calculation formula of the weights corresponding to the first radius and the second radius of each sedimentation connected domain in each solution state monitoring image, when the shape of the sedimentation connected domain is more similar to a regular circle, the weight corresponding to the first radius

The smaller the value of (2), the corresponding weight of the corresponding second radius

The more the value of (2) approaches 1, then the more focused the second radius determined by the area of the precipitate connected domain is subsequently in determining the integrated radius of the precipitate connected domain; when the shape of the sedimentation connected domain is more irregular, the radiation radius is larger, the perimeter value is larger, and the weight corresponding to the first radius is given

The larger the value of (2), the corresponding weight of the corresponding second radius

The smaller the value of (c) then the later in determining the overall radius of the sedimentation connected domain, the more focused is on the first radius determined by the perimeter of the sedimentation connected domain.

After determining the first radius and the second radius of each sedimentation connected domain in each solution state monitoring image and the weights corresponding to the first radius and the second radius, respectively, carrying out weighted summation on the first radius and the second radius of each sedimentation connected domain and the weights corresponding to the first radius and the second radius, respectively, so as to obtain the comprehensive radius of each sedimentation connected domain in each solution state monitoring image, wherein the corresponding calculation formula is as follows:

wherein ,

the comprehensive radius of the jth precipitate connected domain in the image is monitored for the ith solution state,

and monitoring the second radius of the jth sedimentation connected domain in the image for the ith solution state.

By the method, the comprehensive radius of each sedimentation connected domain in each solution state monitoring image is determined, the comprehensive radius can accurately characterize the size of each sedimentation connected domain, and the sedimentation size condition in the solution state monitoring image can be accurately reflected. And then, the comprehensive radius of all the sedimentation connected domains in each solution state monitoring image is averaged, and the average value is determined as the sedimentation equivalent radius corresponding to each solution state monitoring image. The precipitation equivalent radius corresponding to each solution state monitoring image is extracted, the size of the solution precipitation during the acquisition of the solution state monitoring images is accurately characterized by utilizing the precipitation equivalent radius, and the subsequent accurate extraction of the precipitation distribution characteristics in the solution is facilitated, so that the real abnormal data points in each suspected abnormal data point are screened out.

Step S5: and determining a precipitation degree value corresponding to each solution state monitoring image according to the precipitation intensity value, the precipitation equivalent distance and the precipitation equivalent radius corresponding to each solution state monitoring image.

Through the steps, the sedimentation intensity value, the sedimentation equivalent distance and the sedimentation equivalent radius corresponding to each solution state monitoring image can be determined, the sedimentation degree value corresponding to each solution state monitoring image is calculated based on three indexes of the sedimentation intensity value, the sedimentation equivalent distance and the sedimentation equivalent radius, so that the distribution characteristics of sediments in the solution corresponding to the acquisition time of each solution state monitoring image are accurately represented, and the specific implementation process is as follows:

Specifically, for each solution state monitoring image, calculating a precipitation degree value corresponding to the solution state monitoring image according to a precipitation intensity value, a precipitation equivalent distance and a precipitation equivalent radius corresponding to the solution state monitoring image, wherein a corresponding calculation formula is as follows:

wherein ,

the precipitation degree value corresponding to the image is monitored for the ith solution state,

the corresponding sedimentation intensity value of the image is monitored for the ith solution state,

the precipitation equivalent distance corresponding to the image is monitored for the ith solution state,

the equivalent radius of the sediment corresponding to the ith solution state monitoring image,

as a normalization function for

The value of (2) is normalized to be in the range of 0-1.

In the precipitation generation stage in the reaction extraction process of chondroitin sulfate, as the precipitation reaction occurs, the number of precipitates precipitated from the solution is reduced to increased, the precipitate communicating domains become larger gradually, and the distance between the precipitate communicating domains becomes larger gradually. Along with the progress of the precipitation process, the number of precipitation particles in the solution is reduced, and the amount of precipitated precipitation is increased, so that the precipitation intensity value of the solution is increased; the longer the sedimentation agglomeration time, the larger the sedimentation equivalent radius of the sediment, the larger the sedimentation density, i.e. after sedimentation agglomeration, the less discrete and small number of sedimentation connected domains, and only larger sedimentation connected domains remain, at this time, the distance between the sedimentation blocks becomes larger, i.e. the sedimentation equivalent distance becomes larger, the overall sedimentation degree of the solution becomes stronger, and the sedimentation degree value becomes larger.

Step S6: and determining a sediment fluctuation value corresponding to each set time according to the sediment degree value of the solution state monitoring image at each set time and the adjacent set time before and after the set time.

Since there is fluctuation in the sedimentation degree value of the solution state monitoring image at each set time and the adjacent set time, the fluctuation in the sedimentation degree value is expressed as the change of the sedimentation process, so that the change of the sedimentation process can be analyzed by the sedimentation degree value of the solution state monitoring image at each set time and the adjacent set time before and after, the implementation steps include:

Specifically, for each set time, calculating the absolute value of the difference between the set time and the precipitation degree value of the solution state monitoring image corresponding to the two set times, and determining the average value of the two absolute values of the difference as the precipitation fluctuation value corresponding to the set time, wherein the corresponding calculation formula is as follows:

wherein ,

setting a sediment fluctuation value corresponding to the ith set time,

the precipitation degree value of the image is monitored for the solution state at the ith set time,

the precipitation degree value of the image is monitored for the solution state at the i-1 th set time,

and (3) monitoring the precipitation degree value of the image for the solution state at the (i+1) th set time, wherein I is an absolute value sign.

In addition, for the first set time, since there is no previous set time, the absolute value of the difference between the precipitation degree values of the solution state monitoring image corresponding to the first set time and the next set time is directly determined as the precipitation fluctuation value corresponding to the first set time. Meanwhile, for the last set time, since the last set time does not have the next set time, the absolute value of the difference value between the precipitation degree values of the solution state monitoring image corresponding to the last set time and the previous set time is directly determined as the precipitation fluctuation value corresponding to the last set time.

According to the method, the precipitation fluctuation value corresponding to each set time can be determined by calculating the difference condition of the precipitation degree values between the solution state monitoring images corresponding to each set time and the front and rear adjacent set time, when the precipitation fluctuation value is larger, the precipitation distribution characteristics between the current set time and the front and rear adjacent set time are indicated to be more dissimilar, the precipitation reaction at the current set time is reflected, and the precipitation characteristic change is obvious; when the fluctuation value of the precipitation is smaller, the characteristic of the precipitation distribution between the current setting time and the adjacent setting time is similar, and the fact that the precipitation reaction is not started or is finished at the current setting time is reflected.

Step S7: according to the precipitation fluctuation value at the set time corresponding to each suspected abnormal data point, determining the real abnormal data point in each suspected abnormal data point, and determining the pH value and the temperature value corresponding to the real abnormal data point as abnormal data.

After determining the sedimentation fluctuation value corresponding to each set time through the above step S6, since each suspected abnormal data point corresponds to one set time, the sedimentation fluctuation value of the set time corresponding to each suspected abnormal data point can be determined. Since each suspected abnormal data point is a local outlier data point, the precipitation fluctuation value of the set time corresponding to each suspected abnormal data point should be relatively large, at this time, the precipitation reaction should occur at the set time corresponding to each suspected abnormal data point, and when the actual value of the precipitation fluctuation value of the set time corresponding to some suspected abnormal data points is small, it is indicated that the precipitation reaction does not actually occur, and then the corresponding suspected abnormal data point is a true abnormal data point.

In this embodiment, compared with the normal precipitation reaction, the difference value is obtained by residual error with the prior data, and the precipitation fluctuation value is obtained by difference value analysis, wherein the specific reaction of the precipitation fluctuation value is to obtain the finished product amount of chondroitin sulfate, and the yield of the chondroitin sulfate is calculated by the finished product amount:

Yield = chondroitin sulfate finished product amount/(cartilage mass (dry weight)) ×100%

Specifically, the amount of the precipitated finished product depends on the alkali concentration, the temperature and the precipitation time, the precipitation time is usually more than 8 hours in the production and preparation process, the alkali concentration is the pH value in the actual monitoring process, the chondroitin sulfate is degraded due to the excessive alkali concentration, so that the content of the finished product is reduced, and the existing research shows that the excessive temperature can cause the deepening of the color of the extract instead of white precipitation, and the relation between the specific alkali concentration (pH) and the yield is shown as shown in fig. 5.

In this example, the factors and the response value data were optimally analyzed by Design-Expert software according to the relationship between the alkali concentration (pH) and the yield, and the yield predictive value of chondroitin sulfate was 59.31% when the alkali concentration (NaOH) was 0.49 mol/L. That is, the chondroitin sulfate data affects the precipitation yield when the alkali concentration is not 0.49mol/L, and the corresponding data may be abnormal data points when compared with the alkali concentration of 0.49 mol/L. Based on the above analysis, a set sedimentation fluctuation threshold value, which can be determined from experimental tests, is set to a value of 0.4 in the present embodiment. And comparing the precipitation fluctuation value of each suspected abnormal data point at the set moment with a set precipitation fluctuation threshold value, and determining the suspected abnormal data point corresponding to the precipitation fluctuation value smaller than the set precipitation fluctuation threshold value as a real abnormal data point. That is, comparing the precipitation fluctuation value of the set time corresponding to each suspected abnormal data point with the set precipitation fluctuation threshold, when the precipitation fluctuation value is smaller than the set precipitation fluctuation threshold, indicating that the precipitation degree of the set time corresponding to the precipitation fluctuation value is not larger than the fluctuation of the precipitation degree of the set time adjacent to the set time, indicating that the monitored pH value and the temperature value of the set time corresponding to the precipitation fluctuation value are not consistent with the actual situation, and determining the suspected abnormal data point corresponding to the set time corresponding to the precipitation fluctuation value as a real abnormal data point; when the precipitation fluctuation value is greater than or equal to the set precipitation fluctuation threshold value, the fact that the precipitation degree of the set moment corresponding to the precipitation fluctuation value is larger than the fluctuation of the precipitation degree of the set moment adjacent to the set moment is indicated, the fact that the pH value and the temperature value monitored at the set moment corresponding to the precipitation fluctuation value are consistent with the actual situation is indicated, and the suspected abnormal data point corresponding to the set moment corresponding to the precipitation fluctuation value is determined to be a normal data point.

In the above manner, the true abnormal data point in each suspected abnormal data point can be determined, and after the true abnormal data point in each suspected abnormal data point is determined, the pH value and the temperature value corresponding to the true abnormal data point are determined as abnormal data. The abnormal data are removed from all pH values and temperature values which are monitored currently, the pH values and the temperature values which are obtained after the removal operation are reliable monitoring data which are obtained currently, the current precipitation reaction condition can be accurately determined according to the reliable monitoring data of the pH values and the temperature values, so that the addition condition of raw materials is determined, whether the current pH values and the temperature values meet the production requirements can also be determined, and the existing research shows that the excessive high temperature values can deepen the color of the finished product of the cartilage cream, and the pH is inconsistent with the requirements and can cause the reduction of the content of the produced finished product. Because the key point of the scheme is that in the precipitation generation stage in the reaction extraction process of the chondroitin sulfate, the abnormal data in the pH value and the temperature value which are monitored currently are accurately analyzed and extracted, and then the subsequent work such as the current precipitation reaction condition is determined according to the normal monitoring data such as the pH value and the temperature value from which the abnormal data are removed, the key point of the scheme is not included, and the technical personnel can determine the abnormal data according to own experience, and the abnormal data are not repeated here.

In the precipitation generation stage in the reaction extraction process of chondroitin sulfate, the invention acquires the monitoring data of the pH value and the temperature value of the reaction solution, and acquires the solution state monitoring image corresponding to the monitoring data. The monitoring data, pH and temperature, are analyzed to determine outlier detection data therein. And analyzing and screening the outlier data in the monitoring data according to the actual solution state monitoring image corresponding to the monitoring data, so that the real abnormal monitoring data can be determined, and the determination accuracy of the abnormal monitoring data is effectively improved.

It should be noted that: the above embodiments are only for illustrating the technical solution of the present application, and are not limiting; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present application, and are intended to be included in the scope of the present application.

Claims

1. A data processing method for chondroitin sulfate production monitoring, comprising the steps of:

2. A data processing method for chondroitin sulfate production monitoring according to claim 1, wherein determining a corresponding sedimentation intensity value for each solution state monitoring image comprises:

3. A data processing method for chondroitin sulfate production monitoring according to claim 1, wherein determining a precipitation equivalent distance corresponding to each solution state monitoring image comprises:

4. A data processing method for chondroitin sulfate production monitoring according to claim 1, wherein determining the corresponding precipitation equivalent radius of each solution state monitoring image comprises:

5. The data processing method for monitoring chondroitin sulfate production according to claim 4, wherein the calculation formula corresponding to the weight corresponding to the first radius and the second radius of each precipitate communicating domain in each solution state monitoring image is determined as follows:

wherein ,

weight corresponding to the first radius of the jth precipitate connected domain in the ith solution state monitoring image, +.>

Weight corresponding to the second radius of the jth precipitate connected domain in the ith solution state monitoring image, < ->

Monitoring the area of the jth sedimentation connected domain in the image for the ith solution state, +.>

6. A data processing method for chondroitin sulfate production monitoring according to claim 1, wherein determining a corresponding precipitation degree value for each solution state monitoring image comprises:

7. The data processing method for monitoring chondroitin sulfate production according to claim 1, wherein constructing data points corresponding to respective set moments comprises:

the monitoring data comprises a temperature value and a pH value;

8. The method of claim 7, wherein determining suspected abnormal data points among all data points comprises:

9. A data processing method for chondroitin sulfate production monitoring according to claim 1, wherein determining a precipitation fluctuation value corresponding to each set time comprises:

10. A data processing method for chondroitin sulfate production monitoring according to claim 1, wherein determining a true abnormal data point among the suspected abnormal data points comprises: