CN117763477A

CN117763477A - Cement plant factory inlet waste classification management method and system

Info

Publication number: CN117763477A
Application number: CN202311772557.2A
Authority: CN
Inventors: 韩韶辉; 户宁; 贾利军; 胡超龙; 韩屹勍; 于朋辉; 曹军科; 魏虹凤
Original assignee: Hebei Jinyu Dingxin Cement Co ltd
Current assignee: Hebei Jinyu Dingxin Cement Co ltd
Priority date: 2023-12-21
Filing date: 2023-12-21
Publication date: 2024-03-26
Anticipated expiration: 2043-12-21
Also published as: CN117763477B

Abstract

According to the method and the system for classifying and managing the factory entering waste of the cement factory, whether the factory entering waste information range which is required to be detected and corresponds to the factory entering waste information range feature is abnormal or not is determined according to the feature difference value between the factory entering waste information range feature which is required to be detected and the corresponding example factory entering waste information range feature, the factory entering waste information detection is carried out by fully utilizing the non-abnormal factory entering waste information feature covered by the example factory entering waste information, and the accuracy of factory entering waste information detection is improved; on the other hand, through detecting each factory entering waste information range that needs to detect, the degree of factory entering waste information detection has been thinned, factory entering waste information detection effect can be improved to can pinpoint the factory entering waste information range that exists the abnormality in the factory entering waste information that needs to detect, further improve the detection reliability and the accuracy of detection of unusual factory entering waste information.

Description

Cement plant factory inlet waste classification management method and system

Technical Field

The application relates to the technical field of information management, in particular to a method and a system for classifying and managing factory-entering wastes of a cement factory.

Background

A large amount of unused waste exists in cement factories, the waste is reasonably piled up, and the clean treatment of the waste can be guaranteed when the subsequent waste is treated.

However, in the actual operation process, there is a problem that the waste is confusing, so that the cleanest treatment mode of the waste cannot be realized when the waste is treated, and thus, the environment is damaged, and the concentrated treatment of the waste is required due to the increasing awareness of the state on the environment, but in the actual operation process, the treatment modes of the waste are different, so that the analysis of the waste is required, but the classification of the waste in a cement factory is difficult, and therefore, a technical scheme is needed to improve the technical problem.

Disclosure of Invention

In order to improve the technical problems existing in the related art, the application provides a method and a system for classifying and managing factory-entering wastes of a cement factory.

In a first aspect, there is provided a method of classification management of factory-entering waste of a cement plant, the method comprising:

acquiring factory entering waste information to be detected and example factory entering waste information corresponding to the factory entering waste information to be detected;

Extracting the factory entering waste information range to be detected and the factory entering waste information characteristic of the example factory entering waste information range in the example factory entering waste information respectively through a target factory entering waste information detection thread to obtain the factory entering waste information range characteristic to be detected and the example factory entering waste information range characteristic; the target in-plant waste information detection thread is obtained through detection configuration of abnormal ranges of two in-plant waste information examples and consistent configuration of key contents of in-plant waste information characteristics of the two in-plant waste information examples in non-abnormal ranges;

and determining an abnormal range of the incoming waste information to be detected by combining the characteristic differences between the range characteristics of the incoming waste information to be detected and the range characteristics of the example incoming waste information through the target incoming waste information detection thread, classifying according to the abnormal range, and determining a classification processing result.

Further, the method further comprises:

obtaining an original in-plant waste information detection thread, wherein the original in-plant waste information detection thread comprises a first main unit, a second main unit and a dividing unit, the first main unit and the second main unit are the same, the first main unit is used for extracting the characteristics of one in-plant waste information example of the two in-plant waste information examples, the second main unit is used for extracting the characteristics of the other in-plant waste information example of the two in-plant waste information examples, and the dividing unit is used for determining the abnormal range of the two in-plant waste information examples by combining the extracted characteristics of the first main unit and the second main unit;

Performing the key content consistent configuration of the factory entrance waste information characteristics of the two factory entrance waste information examples in a non-abnormal range through the extracted characteristics of the first main unit and the second main unit so as to optimize the weight of the original factory entrance waste information detection thread;

performing detection configuration of abnormal ranges of the two factory-entering waste information examples through the extracted features of the first main unit and the second main unit so as to optimize the weight of the original factory-entering waste information detection thread;

and determining the original factory entrance waste information detection thread with the configuration completed as the target factory entrance waste information detection thread.

Further, the feature extracted by the first main unit and the second main unit performs a key content consistent configuration of the two factory entrance waste information examples in a non-abnormal range of factory entrance waste information features to optimize the weight of the original factory entrance waste information detection thread, including:

acquiring a plurality of target integration benchmarks corresponding to a plurality of factory entrance waste information range features based on the plurality of factory entrance waste information range features in a first factory entrance waste information example of the target item;

Obtaining, by the first main unit, a number of in-plant waste information scope characteristics in a second in-plant waste information example of the target item;

obtaining the association condition of a plurality of factory entering waste information range characteristics in a non-abnormal range and a plurality of target integration references in the second factory entering waste information example;

obtaining, by the second main unit, a number of in-plant waste information range characteristics in a third in-plant waste information example of the target item, the second in-plant waste information example and the third in-plant waste information example belonging to the two in-plant waste information examples; obtaining the association condition of a plurality of factory entering waste information range characteristics in a non-abnormal range and a plurality of target integration references in the third factory entering waste information example;

determining the association situation of the plurality of in-plant waste information range features in the non-abnormal range and the plurality of target integration benchmarks in the second in-plant waste information example as potential records of the association situation of the plurality of in-plant waste information range features in the non-abnormal range and the plurality of target integration benchmarks in the third in-plant waste information example, and carrying out key content consistent configuration on the original in-plant waste information detection thread so as to optimize the weight of the original in-plant waste information detection thread;

And determining the association situation of the plurality of factory entrance waste information range features in the non-abnormal range and the plurality of target integration references in the third factory entrance waste information example as potential records of the association situation of the plurality of factory entrance waste information range features in the non-abnormal range and the plurality of target integration references in the second factory entrance waste information example, and carrying out key content consistent configuration on the original factory entrance waste information detection thread so as to optimize the weight of the original factory entrance waste information detection thread.

Further, the first factory-entering waste information example includes first factory-entering waste information and second factory-entering waste information; the method for obtaining the target integration benchmarks corresponding to the plurality of factory entering waste information range features in the first factory entering waste information example based on the target items comprises the following steps:

obtaining a plurality of first factory entrance waste information range characteristics corresponding to the first factory entrance waste information through the first main unit;

obtaining a plurality of second factory entrance waste information range characteristics corresponding to the second factory entrance waste information through the second main unit;

And integrating the first factory entering waste information range features and the second factory entering waste information range features to obtain a plurality of target integration references.

Further, the integrating processing is performed on the first plurality of factory entering waste information range features and the second plurality of factory entering waste information range features to obtain a plurality of target integration benchmarks, including:

optimizing a plurality of original integration references by combining the plurality of first factory entering waste information range characteristics to obtain a plurality of pending integration references;

and optimizing the plurality of pending integration benchmarks by combining the plurality of second factory entering waste information range characteristics to obtain a plurality of target integration benchmarks.

Further, the optimizing the plurality of original integrated references by combining the plurality of first factory entering waste information range features to obtain a plurality of pending integrated references includes:

calculating a first distinction between the first in-plant waste information scope feature and an integration benchmark that the first in-plant waste information scope feature is associated with in real-time, and calculating a second distinction between the first in-plant waste information scope feature and an integration benchmark that the first in-plant waste information scope feature is not associated with in real-time;

Calculating a relative integration quality assessment result in combination with the first distinction and the second distinction;

and optimizing the integration standard based on the relative integration quality evaluation result until the relative integration quality evaluation result reaches the specified requirement, so as to obtain a plurality of undetermined integration standards.

Further, before optimizing the integration benchmark based on the relative integration quality assessment results, the method further comprises: calculating an absolute integrated quality assessment result in combination with a first distinction between the first in-plant waste information range feature and an integration benchmark associated with the first in-plant waste information range feature in real time;

optimizing the integration benchmark based on the relative integration quality assessment results until the relative integration quality assessment results meet specified requirements, comprising: and optimizing the integration benchmark by combining the relative integration quality assessment result and the relative integration quality assessment result until the relative integration quality assessment result and the relative integration quality assessment result reach specified requirements.

Further, optimizing the integration benchmark based on the relative integration quality assessment results comprises: depolarizing the first factory waste information range characteristic to obtain a depolarization value; and calculating an optimized integration benchmark by combining the depolarization value, the designated coefficient and the real-time integration benchmark so as to optimize the integration benchmark.

Further, the obtaining the association between the plurality of factory entering waste information range features in the non-abnormal range and the plurality of target integration benchmarks in the second factory entering waste information example includes:

calculating a distinction between the non-abnormal range of the in-process waste information range features and the respective target integration benchmarks in the second in-process waste information example, and determining the target integration benchmarks corresponding to the minimum distinction as the target integration benchmarks associated with the non-abnormal range of the in-process waste information range features in the second in-process waste information example;

and determining target integration benchmarks associated with the plurality of factory entering waste information range features in the non-abnormal range in the second factory entering waste information example as the association situation of the plurality of factory entering waste information range features in the non-abnormal range and the plurality of target integration benchmarks in the second factory entering waste information example.

Further, the obtaining the association between the plurality of factory entering waste information range features in the non-abnormal range and the plurality of target integration benchmarks in the second factory entering waste information example includes: and determining the probability distribution of the plurality of factory entering waste information range characteristics in the non-abnormal range in the second factory entering waste information example, which belong to the plurality of target integration benchmarks, as the association condition of the plurality of factory entering waste information range characteristics in the non-abnormal range in the second factory entering waste information example and the plurality of target integration benchmarks.

Further, the method further comprises:

calculating the distinction between the non-abnormal range of the in-process waste information range features and the target integration references in the second in-process waste information example;

and determining a value obtained by dividing the difference between the non-abnormal range of the in-process waste information range features in the second in-process waste information example and the target integration references by the sum of the non-abnormal range of the in-process waste information range features in the second in-process waste information example and the target integration references as a probability distribution that the non-abnormal range of the in-process waste information range features in the second in-process waste information example belongs to the target integration references.

Further, optimizing the weight of the original mill-waste information detection thread includes:

calculating a first thread quality assessment by combining key content regression analysis results of the characteristic of the factory entering waste information of the two factory entering waste information examples in a non-abnormal range, which are output in the key content consistency configuration process, and potential records corresponding to the two factory entering waste information examples;

calculating a second thread quality assessment by combining the abnormal range regression analysis results for the two factory entering waste information examples output in the abnormal range detection process with the specified labels corresponding to the two factory entering waste information examples; optimizing the weight of the raw mill waste information detection thread in combination with the first thread quality assessment and the second thread quality assessment.

In a second aspect, there is provided a cement plant waste classification management system comprising a processor and a memory in communication with each other, the processor being adapted to read a computer program from the memory and execute the computer program to implement the method as described above.

According to the method and the system for classifying and managing the factory entering waste of the cement factory, the factory entering waste information range characteristics and the factory entering waste information range characteristics of the example are obtained by extracting the factory entering waste information range needing to be detected and the factory entering waste information characteristics of the example in the factory entering waste information; determining whether the factory entering waste information range to be detected corresponding to each factory entering waste information range feature to be detected is abnormal or not according to the feature difference value between the factory entering waste information range feature to be detected and the corresponding example factory entering waste information range feature, fully utilizing the factory entering waste information feature without abnormality covered by the example factory entering waste information to detect the factory entering waste information, and improving the accuracy of the factory entering waste information detection; on the other hand, through detecting each factory entering waste information range that needs to detect, the degree of factory entering waste information detection has been thinned, factory entering waste information detection effect can be improved to can pinpoint the factory entering waste information range that exists the abnormality in the factory entering waste information that needs to detect, further improve the detection reliability and the accuracy of detection of unusual factory entering waste information.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the embodiments will be briefly described below, it being understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered limiting the scope, and that other related drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Fig. 1 is a flow chart of a method for classifying and managing factory-entering wastes of a cement factory according to an embodiment of the present application.

Detailed Description

In order to better understand the technical solutions described above, the following detailed description of the technical solutions of the present application is provided through the accompanying drawings and specific embodiments, and it should be understood that the specific features of the embodiments and embodiments of the present application are detailed descriptions of the technical solutions of the present application, and not limit the technical solutions of the present application, and the technical features of the embodiments and embodiments of the present application may be combined with each other without conflict.

Referring to fig. 1, a method for classifying and managing waste from a cement plant is shown, which may include the following steps 210-230.

Step 210, obtaining the factory entrance waste information to be detected and the example factory entrance waste information corresponding to the factory entrance waste information to be detected.

For example, the incoming waste information to be detected can be obtained by various photographing devices and component detecting devices; the example factory entering waste information can be understood as waste information standards set in advance, such as: image information and composition properties of the carbon residue.

For example, the incoming waste information to be detected is the incoming waste information which is not determined in real time whether there is an abnormality or not and is to be detected, the example incoming waste information is the incoming waste information which is not abnormal, the incoming waste information to be detected has a corresponding relation with the example incoming waste information, and the incoming waste information items covered by the incoming waste information to be detected and the example incoming waste information thereof should be the same.

In one possible implementation, the example in-process waste information may be normal in-process waste information pre-stored by the in-process waste information detection device. On some preconditions, the in-plant waste information detecting means may determine normal in-plant waste information determined in the in-plant waste information detecting process before in-plant waste information that needs to be detected as the example in-plant waste information. For example, for a certain point on the cement production line, when the in-plant waste information to be detected is detected at the real-time instant, the in-plant waste information detecting device may obtain the example in-plant waste information corresponding to the in-plant waste information to be detected, which is detected before the real-time instant and for which the in-plant waste information detection result determines that the cement in-plant waste information of the point is normal, is determined as the in-plant waste information to be detected at the real-time instant. The in-plant waste information detecting means may obtain the normal in-plant waste information determined to distinguish the latest real-time as the example in-plant waste information, or may periodically optimize the example in-plant waste information based on the normal in-plant waste information determined by the in-plant waste information detection.

Step 220, extracting the factory entering waste information range to be detected and the factory entering waste information range to be detected in the example factory entering waste information respectively through a target factory entering waste information detection thread to obtain the factory entering waste information range characteristics to be detected and the factory entering waste information range characteristics to be detected; the target in-plant waste information detection thread is obtained through detection configuration of abnormal ranges of two in-plant waste information examples and consistent configuration of key contents of in-plant waste information characteristics of the two in-plant waste information examples in non-abnormal ranges.

For example, the target mill waste information detection thread may perform feature extraction on mill waste information and detect an abnormal range of mill waste information according to the extracted features. In order to improve the reliability of the detection of the abnormal range, in this embodiment, the detection configuration of the abnormal range of the two incoming waste information examples is combined with the consistent configuration of the key content of the incoming waste information characteristics of the two incoming waste information examples in the non-abnormal range, so as to obtain the target incoming waste information detection thread. The detection configuration of the abnormal range of the factory entering waste information example can enable the factory entering waste information detection thread to accurately identify the abnormal range in the factory entering waste information, and the key content consistency configuration can enable the factory entering waste information detection thread to accurately identify the key content category to which the factory entering waste information range belongs. The key content category is a category generated by classifying the incoming waste information from the key content information.

And if the thread with incomplete configuration is the original factory entrance waste information detection thread, the abnormal range detection configuration is to detect the abnormal range according to the factory entrance waste information range characteristics of the factory entrance waste information examples extracted by the original factory entrance waste information detection thread, and the key content consistent configuration is to identify the key content category of the factory entrance waste information range according to the factory entrance waste information range characteristics of the factory entrance waste information examples extracted by the original factory entrance waste information detection thread. It can be seen that the abnormal range detection configuration and the key content consistent configuration are processed according to the characteristics of the range of the incoming waste information example extracted by the original incoming waste information detection thread, so that the characteristics of the range of the incoming waste information example extracted by the original incoming waste information detection thread are required to cover the information required by the abnormal detection and the information required by the key content type identification, the target incoming waste information detection thread which completes the configuration can extract the characteristics of the incoming waste information by combining the key content information and the information required by the abnormal detection, the abnormal range of the incoming waste information which needs to be detected can be determined by combining the key content information, the classification processing is performed according to the abnormal range, the classification processing result is determined, and the reliability and the accuracy of the identification of the abnormal range are improved.

When the feature extraction of the entering waste information is performed, the entering waste information is firstly divided into a plurality of entering waste information ranges, namely, the entering waste information to be detected is divided into a plurality of entering waste information ranges to be detected, the example entering waste information is divided into a plurality of example entering waste information ranges, and then the feature extraction of the entering waste information of each entering waste information range is performed to obtain the corresponding feature of the entering waste information range. In the embodiment of the present application, the minimum scale of the incoming waste information range is a characteristic point. The division of the waste-in-process information range to be detected is the same as the division of the sample waste-in-process information range to ensure that the waste-in-process information range to be detected corresponds one-to-one to the sample waste-in-process information range. Extracting the waste information characteristics of each incoming waste information range to be detected to obtain corresponding waste information range characteristics of the incoming waste to be detected, and extracting the waste information characteristics of each incoming waste information range to be detected to obtain corresponding waste information range characteristics of the incoming waste.

In one possible implementation, the target mill waste information detection thread may be implemented by a convolutional network including a machine learning unit, a deep learning unit, or the like. In the embodiment of the application, the unit for performing the extraction of the characteristics of the factory entering waste information in the target factory entering waste information detection thread is called a main unit, the target factory entering waste information detection thread extracts the characteristics of the factory entering waste information in the range of the factory entering waste information to be detected in the factory entering waste information to be detected through a first main unit, extracts the characteristics of the factory entering waste information in the range of the factory entering waste information in the example factory entering waste information through a second main unit, and the first main unit and the second main unit are different units, but the first main unit and the second main unit are twin units, namely the unit structures of the first main unit and the second main unit and related parameters (such as unit weights) are the same. The two units which are twin units are used for respectively extracting the characteristics of the factory entering waste information to be detected and the example factory entering waste information, so that the characteristics of the example factory entering waste information can be fully utilized, the distinction between the factory entering waste information to be detected and the example factory entering waste information can be judged based on the extracted factory entering waste information characteristics, whether the abnormality exists in the factory entering waste information to be detected or not is judged, and the accuracy of detecting the abnormal factory entering waste information is improved.

Step 230, determining an abnormal range of the in-plant waste information to be detected by combining the feature differences between the in-plant waste information range features to be detected and the plurality of example in-plant waste information range features through the target in-plant waste information detection thread, and performing classification processing according to the abnormal range to determine a classification processing result.

Illustratively, after obtaining the in-plant waste information range feature to be detected and the example in-plant waste information range feature, calculating a feature difference between the in-plant waste information range feature to be detected and the example in-plant waste information range feature corresponding thereto, and recording the feature difference between the in-plant waste information range feature to be detected and the example in-plant waste information range feature corresponding thereto as the in-plant waste information range feature difference corresponding to the in-plant waste information range feature to be detected.

In one possible implementation, the feature difference between the in-process waste information range feature that is to be detected and its corresponding example in-process waste information range feature may be represented by a distinction between the in-process waste information range feature that is to be detected and its corresponding example in-process waste information range feature, which may be a European distinction, a cosine distinction, or the like.

The difference between the factory entering waste information range which needs to be detected and the example factory entering waste information range which corresponds to the factory entering waste information range which needs to be detected is reflected by the poor range characteristics which correspond to the factory entering waste information range characteristics which need to be detected. Since the example in-plant waste information is normal in-plant waste information, the smaller the difference between in-plant waste information that needs to be detected and the example in-plant waste information, the more similar the in-plant waste information that needs to be detected and the example in-plant waste information, and thus the less likely that the in-plant waste information that needs to be detected is abnormal. Taking the range characteristic difference as an example for distinguishing between the waste entering information range characteristic to be detected and the corresponding example waste entering information range characteristic, when the range characteristic difference is smaller than the specified distinguishing target value, the corresponding waste entering information range to be detected and the example waste entering information range have higher similarity, and the waste entering information range to be detected and the example waste entering information range to be detected can be considered to be consistent, namely the waste entering information range to be detected is not abnormal. When the range characteristic difference is larger than the specified discrimination target value, a larger difference is indicated between the corresponding incoming waste information range to be detected and the example incoming waste information range, and at this time, the incoming waste information range to be detected and the example incoming waste information range are not consistent, that is, the incoming waste information range to be detected is abnormal. It is understood that the case where the range characteristic difference is equal to the specified discrimination target value may be classified into the case where the range characteristic difference is greater than the specified discrimination target value or the case where the range characteristic difference is less than the specified discrimination target value according to the actual demand.

In one possible implementation embodiment, the apparatus for detecting the incoming waste information can directly determine whether the range of the incoming waste information to be detected is abnormal as the detection result of the incoming waste information to be detected, and when the scale of the range of the incoming waste information to be detected is the smallest, the detection result of the incoming waste information with the characteristic point level can be obtained, and the range of the incoming waste information with the abnormality in the incoming waste information to be detected can be accurately positioned, so that the detection reliability and the detection accuracy of the abnormal incoming waste information are improved.

In one possible embodiment, the in-plant waste information detecting means may determine the detection result of in-plant waste information to be detected based on the number of in-plant waste information ranges to be detected for which abnormality exists in each in-plant waste information range to be detected. For example, when the number of the in-plant waste information ranges for which the detection is required is greater than the target value, or the number of the in-plant waste information ranges for which the detection is required is greater than the specified value, the in-plant waste information for which the detection is required is considered to be abnormal; otherwise, the incoming waste information that needs to be detected is considered to be free of anomalies.

In the technical scheme provided by the embodiment of the application, the target factory entering waste information detection thread is used for respectively extracting the factory entering waste information range needing to be detected and the factory entering waste information characteristics of the plurality of example factory entering waste information ranges in the example factory entering waste information to obtain the factory entering waste information range characteristics needing to be detected and the plurality of example factory entering waste information range characteristics; according to the characteristic difference between the characteristic of the factory entering waste information range to be detected and the characteristic of the corresponding example factory entering waste information range, the abnormal range of the factory entering waste information to be detected is determined, on one hand, because the target factory entering waste information detection thread is obtained through the detection configuration of the abnormal ranges of the two factory entering waste information examples and the consistent configuration of the key contents of the factory entering waste information characteristics of the two factory entering waste information examples in the non-abnormal ranges, the target factory entering waste information detection thread has strong key content sensing capability, can combine the key content information and the abnormal information of the factory entering waste information to perform abnormal detection, and further improves the accuracy of the abnormal detection of the factory entering waste information. On the other hand, the non-abnormal factory entrance waste information features covered by the example factory entrance waste information are fully utilized to detect the factory entrance waste information, the accuracy of the factory entrance waste information detection is improved, the degree of the factory entrance waste information detection is refined by detecting each factory entrance waste information range needing to be detected, the factory entrance waste information detection effect can be improved, the abnormal factory entrance waste information range in the factory entrance waste information needing to be detected can be accurately positioned, and the detection reliability and the detection accuracy of the abnormal factory entrance waste information are further improved.

The method for classifying and managing the factory-entering waste of the cement factory provided by the embodiment of the application comprises the following steps 310 to 370, wherein the method comprises the following steps: step 310, obtaining an original factory entering waste information detection thread, wherein the original factory entering waste information detection thread comprises a first main unit, a second main unit and a dividing unit, the first main unit and the second main unit are the same, the first main unit is used for extracting the characteristics of one factory entering waste information example in two factory entering waste information examples, the second main unit is used for extracting the characteristics of the other factory entering waste information example in the two factory entering waste information examples, and the dividing unit is used for determining the abnormal range of the two factory entering waste information examples according to the extracted characteristics of the first main unit and the second main unit.

Specifically, the original factory entering waste information detection thread comprises a first main unit, a second main unit and a dividing unit, wherein the first main unit and the second main unit are twin units, namely the first main unit and the second main unit are identical. The first main unit and the second main unit are used for extracting the characteristics of the entering factory waste information, and the two main units can be used for extracting the characteristics of the entering factory waste information of different entering factory waste information at the same time. The dividing unit is used for identifying an abnormal range of the in-plant waste information, and the abnormal ranges of the two in-plant waste information examples are determined according to the difference between the characteristics extracted by the first main unit and the second main unit.

Step 320, performing the key content consistent configuration of the factory entrance waste information features of the two factory entrance waste information examples in the non-abnormal range through the features extracted by the first main unit and the second main unit, so as to optimize the weight of the original factory entrance waste information detection thread.

Illustratively, the critical content reconciliation configuration is a critical content category that identifies a scope of in-plant waste information based on in-plant waste information scope characteristics of an example of in-plant waste information extracted by an original in-plant waste information detection thread, which process may also be referred to as a critical content identification portion of in-plant waste information. The critical content category identification portion of the in-plant waste information resembles a multi-category process with a large number of tags that would increase the collection pressure of the in-plant waste information examples if the critical content category tags were added to all in-plant waste information examples.

In the key content consistent configuration process, the first main unit extracts the advance factory waste information characteristics of each advance factory waste information range of one advance factory waste information example in the two advance factory waste information examples, and obtains the corresponding advance factory waste information range characteristics; meanwhile, the first main unit extracts the incoming factory waste information characteristics for each incoming factory waste information example range of the other incoming factory waste information example of the two incoming factory waste information examples, and obtains the corresponding incoming factory waste information example range characteristics. Then, a potential record corresponding to the extent of the in-process waste information instance in another in-process waste information instance is generated based on the in-process waste information instance extent feature of one in-process waste information instance, the potential record identifying a key content category to which the corresponding in-process waste information instance extent in the other in-process waste information instance belongs. Finally, the potential records can be used for carrying out 'supervised' configuration, namely, the potential records are used for determining regression analysis targets for identifying the key contents, and thread quality assessment is calculated according to regression analysis values for identifying the key contents in the configuration process and the corresponding regression analysis targets (potential records), wherein the thread quality assessment is used for optimizing the weight of the original factory-entering waste information detection thread.

Step 330, performing detection configuration of the abnormal range of the two factory entrance waste information examples through the features extracted by the first main unit and the second main unit, so as to optimize the weight of the original factory entrance waste information detection thread.

Illustratively, the characteristics of the incoming waste information according to the abnormal range detection configuration are the same as those according to the key content consistent configuration of the previous steps, that is, the first main unit extracts the incoming waste information characteristics of each incoming waste information range of one of the two incoming waste information examples, and obtains the corresponding incoming waste information range characteristics; meanwhile, the first main unit extracts the incoming factory waste information characteristics for each incoming factory waste information example range of the other incoming factory waste information example of the two incoming factory waste information examples, and obtains the corresponding incoming factory waste information example range characteristics. Differences between the range characteristics of the factory entering waste information examples of the two factory entering waste information examples are then identified through the dividing unit, and the abnormal range in the factory entering waste information examples is determined.

When the detection configuration of the abnormal range is performed using the in-plant waste information example, the in-plant waste information example has a tag for identifying whether or not there is an abnormality in the in-plant waste information range in the in-plant waste information example. Illustratively, when there is an abnormal in-plant waste information range in the in-plant waste information example, the tag of the in-plant waste information range is set to 1; otherwise, the tag of the normal factory entering waste information range is set to 0. Therefore, the detection configuration process of the abnormal range can adopt a supervision configuration mode. The abnormality detection part of the factory entering waste information is equivalent to two classification of the factory entering waste information through the dividing unit, the number of labels is small, and the accuracy of abnormality detection of the threads can be improved through supervised configuration of the labeled data.

In one possible implementation, the thread weight optimization process specifically includes: calculating a first thread quality assessment according to key content regression analysis results of the characteristic of the factory entering waste information of the two factory entering waste information examples in a non-abnormal range and potential records corresponding to the two factory entering waste information examples, which are output in the key content consistency configuration process; calculating a second thread quality assessment according to the abnormal range regression analysis results for the two factory entering waste information examples and the specified labels corresponding to the two factory entering waste information examples, which are output in the abnormal range detection process; the weights of the raw mill waste information detection threads are optimized according to the first thread quality assessment and the second thread quality assessment.

Specifically, a first thread quality assessment is obtained based on a key content consistency configuration process, a second thread quality assessment is obtained based on an abnormal range detection configuration process, and finally the first thread quality assessment and the second thread quality assessment are fused to optimize the weight of an original factory-entering waste information detection thread, for example, a back propagation method and a gradient descent method can be adopted to realize thread parameter optimization. The first thread quality assessment shows the learning of the thread on key content information, the second thread quality assessment shows the learning of the thread on the abnormality information of the incoming waste information, the thread parameters are optimized by combining the two quality assessments, and the key content category perception capability is injected into the abnormality detection process of the incoming waste information of the thread, so that the abnormality detection accuracy of the thread is improved.

And 340, determining the original factory entrance waste information detection thread with the configuration completed as a target factory entrance waste information detection thread.

Illustratively, the original mill-waste information detection thread that is ultimately configured has both mill-waste information anomaly detection capability and mill-waste information key content category identification capability. In this embodiment, when the target factory entrance waste information detection thread is used to detect the factory entrance waste information to be detected, the factory entrance waste information key content category identification capability of the thread is abandoned, but the factory entrance waste information abnormality detection capability of the thread is not affected, and the target factory entrance waste information detection thread has higher practical value.

Step 350, obtaining the factory entrance waste information to be detected and the example factory entrance waste information corresponding to the factory entrance waste information to be detected.

Step 360, extracting the factory entering waste information range to be detected and the factory entering waste information range to be detected in the example factory entering waste information respectively through a target factory entering waste information detection thread to obtain the factory entering waste information range characteristics to be detected and the factory entering waste information range characteristics to be detected; the target in-plant waste information detection thread is obtained through detection configuration of abnormal ranges of two in-plant waste information examples and consistent configuration of key contents of in-plant waste information characteristics of the two in-plant waste information examples in non-abnormal ranges.

And 370, determining an abnormal range of the in-process waste information to be detected according to the characteristic difference between the in-process waste information range characteristics to be detected and the example in-process waste information range characteristics by the target in-process waste information detection thread, and performing classification processing according to the abnormal range to determine a classification processing result.

The specific implementation process of steps 350-370 is the same as that of steps 210-230 in the previous embodiment, and will not be repeated here.

In one possible implementation, the key content consistent configuration process includes steps 510 through 570, as follows: step 510, obtaining a plurality of target integration benchmarks corresponding to the plurality of factory entrance waste information range features based on the plurality of factory entrance waste information range features in the first factory entrance waste information example of the target item.

For example, the target items refer to items covered in the factory entrance waste information examples, and configuration should be performed based on the factory entrance waste information examples of the same target item, that is, in a configuration process, the factory entrance waste information examples corresponding to the same target item should be used by the key content consistent configuration and the abnormal range detection configuration, for example, items covered by cloth at different positions on the cloth production line are different, and in a configuration process, the factory entrance waste information examples of cloth at the same position should be used. The incoming waste information examples may also be divided into several batches, the incoming waste information examples of different batches may correspond to different target matters, and the incoming waste information examples of the same batch correspond to the same target matters.

The integrating process is performed on the plurality of factory entering waste information range features in the first factory entering waste information example, that is, the classifying process of the key content categories is performed on the plurality of factory entering waste information range features, and the obtained target integrating reference represents the corresponding key content category.

In one possible implementation, the first factory-entering waste information example includes first factory-entering waste information and second factory-entering waste information, and the integrating process specifically includes: obtaining a plurality of first factory entrance waste information range characteristics corresponding to the first factory entrance waste information through the first main unit; obtaining a plurality of second factory entrance waste information range characteristics corresponding to the second factory entrance waste information through the second main unit; and integrating the first factory entering waste information range features and the second factory entering waste information range features to obtain a plurality of target integration references.

In order to realize the detection of the in-plant waste information of the feature point level and the identification of the key content category, when the feature extraction of the in-plant waste information is carried out, the first in-plant waste information and the second in-plant waste information are divided, the first in-plant waste information is divided into a plurality of first in-plant waste information ranges, and the second in-plant waste information is divided into a plurality of second in-plant waste information ranges. Then, extracting features of a plurality of first factory entering waste information ranges in the first factory entering waste information through the first main unit to obtain a plurality of first factory entering waste information range features; and extracting the characteristics of a plurality of second factory entrance waste information ranges in the second factory entrance waste information through the second main unit to obtain the characteristics of the plurality of second factory entrance waste information ranges.

In one possible implementation, the first and second in-process waste information range features are integrated by sharing the same set of integration criteria, that is, the first and second in-process waste information range features are integrated (twice integrated), but the integration criteria in the integration process are shared, so that the first and second in-process waste information range features correspond to the same set of target integration criteria.

Specifically, the integration process includes: optimizing a plurality of original integration references according to the range characteristics of the first factory entering waste information to obtain a plurality of undetermined integration references; and optimizing the plurality of undetermined integration benchmarks according to the plurality of second factory entering waste information range characteristics to obtain a plurality of target integration benchmarks.

First, a set of primitive integration criteria is obtained, and then, the range features of the incoming waste information corresponding to one of the two incoming waste information are selected for integration. The integration standard obtained by performing unsupervised integration on the plurality of first factory-entering waste information range features is taken as a pending integration standard, the integration process starts from the original integration standard, and the final pending integration standard is obtained by continuously and iteratively optimizing the integration standard, wherein the original integration standard is preset, and under some preconditions, the original integration standard can be randomly determined in the plurality of first factory-entering waste information range features. After the integration of the first in-plant waste information range feature is completed, the second in-plant waste information range feature is continuously integrated on the basis of the integration result of the first in-plant waste information range feature, that is, in the integration process of the second in-plant waste information range, the original integration standard is a pending integration standard obtained by integrating the first in-plant waste information range feature. Finally, the integration benchmark obtained after the integration of the second factory-entering waste information range characteristics is completed is taken as a target integration benchmark.

In one possible implementation, the integration process involves optimization of an integration benchmark, the integration process of the first in-plant waste information scope feature is the same as the integration process of the second in-plant waste information scope feature, the optimization step of the integration benchmark is the same, and the integration benchmark optimization process at the time of integration is described below by taking the first in-plant waste information scope feature as an example. In optimizing the integration benchmarks, a first distinction between the first in-process waste information range feature and the integration benchmarks associated with the first in-process waste information range feature in real-time is first calculated, and a second distinction between the first in-process waste information range feature and the integration benchmarks not associated with the first in-process waste information range feature in real-time is calculated. In general, the integration references associated with the first in-plant waste information scope feature are one, and there are a plurality of integration references (marked as unassociated integration references) unassociated with the first in-plant waste information scope feature in real time, so the second distinction refers to the sum of the distinction between the first in-plant waste information scope feature and all unassociated integration references. Then, a relative integrated quality assessment result is calculated from the first and second distinctions, which can be calculated by means of a comparative quality assessment. The first distinction and the second distinction embody the relative distinction between the first in-plant waste information range feature and the corresponding integration reference, so that the relative integration quality assessment result at this time is constrained by the relative distinction between the first in-plant waste information range feature and the respective integration reference in real time. And finally, optimizing the integration standard based on the relative integration quality evaluation result until the relative integration quality evaluation result reaches the specified requirement, and completing iterative optimization of the integration standard to obtain the undetermined integration standard.

And after the relative integration quality evaluation result is calculated, optimizing the integration standard so as to carry out integration again based on the optimized integration standard until the relative integration quality evaluation result reaches the specified requirement. For example, the relative integration quality evaluation result and the designated relative target value are determined, when the relative integration quality evaluation result is greater than the designated relative target value, the integration standard is optimized, and integration is performed again according to the optimized integration standard until the relative integration quality evaluation result is less than or equal to the designated relative target value. Alternatively, the iteration number may be set, and when the iteration number does not reach the set number, the integration criterion is optimized to be integrated again.

In one possible implementation, in addition to employing a relative integration quality assessment to constrain the relative distinction between the in-plant waste information scope characteristics and the respective integration benchmarks, the embodiments of the present application incorporate in performing integration benchmarks optimization, absolute distinction between in-plant waste information scope characteristics and the respective integration benchmarks to determine optimization requirements, including in particular: calculating an absolute integration quality assessment result according to a first distinction between the in-plant waste information range feature and an integration benchmark associated with the in-plant waste information range feature in real time; and optimizing the integration standard according to the relative integration quality evaluation result and the relative integration quality evaluation result until the relative integration quality evaluation result and the relative integration quality evaluation result reach the specified requirement.

Step 520, obtaining, by the first main unit, a number of in-plant waste information scope characteristics in the second in-plant waste information example of the target item.

For example, the second in-plant waste information example may be the same as the first in-plant waste information example, or may be different from the first in-plant waste information example, but both cover the same target matter. The process of obtaining the factory entrance waste information range feature of the second factory entrance waste information example is the same as the process of extracting the factory entrance waste information feature in the foregoing step, and will not be described here again.

Step 530, obtaining the association condition of the plurality of factory entrance waste information range features in the non-abnormal range and the plurality of target integration benchmarks in the second factory entrance waste information example.

Specifically, since the key content consistent configuration is configured based on the non-abnormal-range in-process waste information range feature in the in-process waste information example, the association condition obtained in this step is only required for the non-abnormal-range in-process waste information range feature.

In one possible implementation, the association of the in-plant waste information scope feature with the plurality of target integration benchmarks is represented by the target integration benchmarks to which the in-plant waste information scope feature belongs, the steps specifically including: calculating a distinction between the in-process waste information range feature in the non-abnormal range and each of the target integration benchmarks in the second in-process waste information example, and determining a target integration benchmark corresponding to the minimum distinction as a target integration benchmark associated with the in-process waste information range feature in the non-abnormal range in the second in-process waste information example; a target integration benchmark associated with the plurality of in-plant waste information range features in the non-abnormal range in the second in-plant waste information example is determined as an association of the plurality of in-plant waste information range features in the non-abnormal range with the plurality of target integration benchmarks in the second in-plant waste information example.

In one possible implementation, the association of the in-plant waste information scope feature with the plurality of target integration benchmarks may be further represented by a probability distribution of the in-plant waste information scope feature over the plurality of target integration benchmarks, the steps specifically comprising: the probability distribution of the plurality of in-plant waste information range features in the non-abnormal range in the second in-plant waste information example, each belonging to the plurality of target integration benchmarks, is determined as the association situation of the plurality of in-plant waste information range features in the non-abnormal range in the second in-plant waste information example with the plurality of target integration benchmarks.

Specifically, the probability distribution represents the association condition between the range features of the factory waste information and the integration standards of the targets, and the calculation process of the probability distribution comprises the following steps: firstly, calculating the distinction between the characteristic of the factory entering waste information range in a non-abnormal range and the integration standard of each target in the second factory entering waste information example; and then dividing the difference between the non-abnormal range of the second in-process waste information example and the target integration standard by the sum of the non-abnormal range of the second in-process waste information example and the target integration standard to determine the probability distribution of the non-abnormal range of the second in-process waste information example.

Step 540, obtaining, by the second main unit, a number of in-plant waste information range characteristics in a third in-plant waste information example of the target item, the second in-plant waste information example and the third in-plant waste information example belonging to two in-plant waste information examples.

Step 550, obtaining a correlation of the plurality of factory entering waste information range features in the non-abnormal range and the plurality of target integration benchmarks in the third factory entering waste information example.

The process of obtaining the factory entering waste information range feature of the third factory entering waste information example and the process of calculating the association situation of the factory entering waste information range feature and the target integration references are the same as the process of extracting the factory entering waste information feature and calculating the association situation of the second factory entering waste information example in the foregoing steps, and will not be described herein. The third incoming waste information example may be the same as or different from the first incoming waste information example, but the third incoming waste information example and the second incoming waste information example do not belong to the same incoming waste information.

Step 560, determining the association situation of the plurality of in-process waste information range features in the non-abnormal range and the plurality of target integration benchmarks in the second in-process waste information example as the potential record of the association situation of the plurality of in-process waste information range features in the non-abnormal range and the plurality of target integration benchmarks in the third in-process waste information example, and performing the key content consistent configuration on the original in-process waste information detection thread so as to optimize the weight of the original in-process waste information detection thread.

Step 570, determining the association situation of the plurality of in-process waste information range features in the non-abnormal range and the plurality of target integration benchmarks in the third in-process waste information example as the potential record of the association situation of the plurality of in-process waste information range features in the non-abnormal range and the plurality of target integration benchmarks in the second in-process waste information example, and performing the key content consistent configuration on the original in-process waste information detection thread to optimize the weight of the original in-process waste information detection thread.

Based on the rule, when the second factory entrance waste information example is identified in the key content category, determining the association condition of a plurality of factory entrance waste information range features in a non-abnormal range and a plurality of target integration references in the third factory entrance waste information example as corresponding potential records; in the critical content category identification of the third factory entering waste information example, the association condition of the plurality of factory entering waste information range features in the non-abnormal range and the plurality of target integration benchmarks in the second factory entering waste information example needs to be determined to be corresponding potential records.

In one possible implementation, finally, when calculating the thread quality assessment for optimizing the thread weight, each branch calculates the quality assessment according to the regression analysis result and the corresponding potential record, and then fuses the quality assessments of the branches to obtain the thread quality assessment when the key content is configured consistently.

In a possible implementation embodiment, the apparatus further comprises: an original thread obtaining module, configured to obtain an original incoming waste information detection thread, where the original incoming waste information detection thread includes a first main unit, a second main unit, and a dividing unit, where the first main unit and the second main unit are the same, the first main unit is configured to extract a feature of one incoming waste information example of the two incoming waste information examples, the second main unit is configured to extract a feature of another incoming waste information example of the two incoming waste information examples, and the dividing unit is configured to determine an abnormal range of the two incoming waste information examples in combination with the features extracted by the first main unit and the second main unit; the key content consistency configuration module is used for carrying out key content consistency configuration of the factory-entering waste information characteristics of the two factory-entering waste information examples in a non-abnormal range through the characteristics extracted by the first main unit and the second main unit so as to optimize the weight of the original factory-entering waste information detection thread; the abnormality detection configuration module is used for carrying out detection configuration of the abnormality ranges of the two factory entering waste information examples through the characteristics extracted by the first main unit and the second main unit so as to optimize the weight of the original factory entering waste information detection thread; and the target thread generation module is used for determining the original factory entering waste information detection thread which completes configuration as the target factory entering waste information detection thread.

In one possible implementation, the key content agreement configuration module includes: an integrating unit, configured to obtain a plurality of target integration benchmarks corresponding to a plurality of factory entrance waste information range features in a first factory entrance waste information example of a target item based on the plurality of factory entrance waste information range features; a first feature obtaining unit configured to obtain, by the first main unit, a plurality of in-plant waste information range features in a second in-plant waste information example of the target item; a first association situation obtaining unit, configured to obtain association situations of a plurality of factory entering waste information range features in a non-abnormal range and the plurality of target integration benchmarks in the second factory entering waste information example; a second feature obtaining unit configured to obtain, by the second main unit, a plurality of in-plant waste information range features in a third in-plant waste information example of the target item, the second in-plant waste information example and the third in-plant waste information example belonging to the two in-plant waste information examples; a second association situation obtaining unit, configured to obtain association situations of a plurality of factory entering waste information range features in a non-abnormal range and the plurality of target integration benchmarks in the third factory entering waste information example; a first configuration unit, configured to determine, as a potential record of a correlation situation between the plurality of in-plant waste information range features in the non-abnormal range and the plurality of target integration benchmarks in the third in-plant waste information example, a correlation situation between the plurality of in-plant waste information range features in the non-abnormal range and the plurality of target integration benchmarks in the second in-plant waste information example, and perform a critical content consistent configuration on the original in-plant waste information detection thread so as to optimize a weight of the original in-plant waste information detection thread; and the second configuration unit is used for determining the association situation of the plurality of factory entrance waste information range features in the non-abnormal range and the plurality of target integration references in the third factory entrance waste information example as the potential record of the association situation of the plurality of factory entrance waste information range features in the non-abnormal range and the plurality of target integration references in the second factory entrance waste information example, and performing key content consistent configuration on the original factory entrance waste information detection thread so as to optimize the weight of the original factory entrance waste information detection thread.

In one possible implementation, the first in-plant waste information example includes first in-plant waste information and second in-plant waste information; the integration unit includes: the first local feature extraction unit is used for obtaining a plurality of first factory entrance waste information range features corresponding to the first factory entrance waste information through the first main unit; the second local feature extraction unit is used for obtaining a plurality of second factory entrance waste information range features corresponding to the second factory entrance waste information through the second main unit; and the integration subunit is used for integrating the plurality of first factory entering waste information range characteristics and the plurality of second factory entering waste information range characteristics to obtain a plurality of target integration references.

On the basis of the above, there is provided a cement plant inlet waste classification management device, the device comprising:

the information acquisition module is used for acquiring the factory entering waste information to be detected and the example factory entering waste information corresponding to the factory entering waste information to be detected;

the characteristic determining module is used for respectively extracting the factory entering waste information range needing to be detected and the factory entering waste information characteristic of the plurality of the example factory entering waste information ranges in the example factory entering waste information through a target factory entering waste information detection thread to obtain the factory entering waste information range characteristic needing to be detected and the example factory entering waste information range characteristic; the target in-plant waste information detection thread is obtained through detection configuration of abnormal ranges of two in-plant waste information examples and consistent configuration of key contents of in-plant waste information characteristics of the two in-plant waste information examples in non-abnormal ranges;

And the result classification module is used for determining the abnormal range of the incoming waste information to be detected by combining the characteristic difference value between the range characteristics of the incoming waste information to be detected and the range characteristics of the example incoming waste information through the target incoming waste information detection thread, classifying according to the abnormal range, and determining the classification result.

On the above, there is shown a cement plant in-plant waste sort management system comprising a processor and a memory in communication with each other, the processor being adapted to read a computer program from the memory and execute it to carry out the method described above.

On the basis of the above, there is also provided a computer readable storage medium on which a computer program stored which, when run, implements the above method.

In summary, based on the above-mentioned scheme, by extracting the factory entering waste information range to be detected and the factory entering waste information feature of the example factory entering waste information range from the factory entering waste information to be detected, respectively, the factory entering waste information range feature to be detected and the example factory entering waste information range feature are obtained; determining whether the factory entering waste information range to be detected corresponding to each factory entering waste information range feature to be detected is abnormal or not according to the feature difference value between the factory entering waste information range feature to be detected and the corresponding example factory entering waste information range feature, fully utilizing the factory entering waste information feature without abnormality covered by the example factory entering waste information to detect the factory entering waste information, and improving the accuracy of the factory entering waste information detection; on the other hand, through detecting each factory entering waste information range that needs to detect, the degree of factory entering waste information detection has been thinned, factory entering waste information detection effect can be improved to can pinpoint the factory entering waste information range that exists the abnormality in the factory entering waste information that needs to detect, further improve the detection reliability and the accuracy of detection of unusual factory entering waste information.

It should be appreciated that the systems and modules thereof shown above may be implemented in a variety of ways. For example, in some embodiments, the system and its modules may be implemented in hardware, software, or a combination of software and hardware. Wherein the hardware portion may be implemented using dedicated logic; the software portions may then be stored in a memory and executed by a suitable instruction execution system, such as a microprocessor or special purpose design hardware. Those skilled in the art will appreciate that the methods and systems described above may be implemented using computer executable instructions and/or embodied in processor control code, such as provided on a carrier medium such as a magnetic disk, CD or DVD-ROM, a programmable memory such as read only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The system and its modules of the present application may be implemented not only with hardware circuitry, such as very large scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., but also with software, such as executed by various types of processors, and with a combination of the above hardware circuitry and software (e.g., firmware).

It should be noted that, the advantages that may be generated by different embodiments may be different, and in different embodiments, the advantages that may be generated may be any one or a combination of several of the above, or any other possible advantages that may be obtained.

Claims

1. A method for classifying and managing factory-entering wastes of a cement factory, the method comprising:

2. The method of claim 1, further comprising:

3. The cement plant incoming waste classification management method of claim 2, wherein said feature extracted by said first main unit and said second main unit, performs a critical content consistent configuration of said two incoming waste information instances in non-anomaly range of incoming waste information features to optimize the weight of said original incoming waste information detection thread, comprises:

4. The cement plant waste sort management method of claim 3, wherein the first mill waste information example includes first mill waste information and second mill waste information; the method for obtaining the target integration benchmarks corresponding to the plurality of factory entering waste information range features in the first factory entering waste information example based on the target items comprises the following steps:

5. The method of claim 4, wherein the integrating the first plurality of factory entering waste information range features and the second plurality of factory entering waste information range features to obtain the target integration references comprises:

6. The method for classifying and managing industrial waste in a cement plant according to claim 5, wherein said optimizing a plurality of primitive integration benchmarks by combining said plurality of first industrial waste information range features to obtain a plurality of pending integration benchmarks comprises:

7. The method of claim 6, further comprising, prior to optimizing the integration benchmark based on the relative integration quality assessment results: calculating an absolute integrated quality assessment result in combination with a first distinction between the first in-plant waste information range feature and an integration benchmark associated with the first in-plant waste information range feature in real time;

8. The method of claim 6, wherein optimizing the integration benchmark based on the relative integration quality assessment results comprises: depolarizing the first factory waste information range characteristic to obtain a depolarization value; and calculating an optimized integration benchmark by combining the depolarization value, the designated coefficient and the real-time integration benchmark so as to optimize the integration benchmark.

9. The method for classifying and managing waste in a cement plant according to claim 3, wherein said obtaining the association of the plurality of waste in-process information range features in the non-abnormal range with the plurality of target integration benchmarks in the second waste in-process information example includes:

10. The method for classifying and managing waste in a cement plant according to claim 3, wherein said obtaining the association of the plurality of waste in-process information range features in the non-abnormal range with the plurality of target integration benchmarks in the second waste in-process information example includes: determining the probability distribution of the plurality of in-plant waste information range features in the second in-plant waste information example in the non-abnormal range to each of the plurality of target integration benchmarks as the association condition of the plurality of in-plant waste information range features in the second in-plant waste information example in the non-abnormal range and the plurality of target integration benchmarks;

Wherein the method further comprises:

determining a value obtained by dividing the difference between the non-abnormal range of the in-process waste information range features in the second in-process waste information example and the target integration references by the sum of the non-abnormal range of the in-process waste information range features in the second in-process waste information example and the target integration references as a probability distribution that the non-abnormal range of the in-process waste information range features in the second in-process waste information example belongs to the target integration references;

wherein optimizing the weight of the raw mill waste information detection thread comprises:

11. A cement plant waste classification management system comprising a processor and a memory in communication with each other, the processor being adapted to read a computer program from the memory and execute the computer program to implement the method of any one of claims 1-10.