An environmental supervision method and system for a clean factory

Abstract

The present invention relates to an environmental supervision method and system for a clean factory, belonging to the technical field of clean factory operation and maintenance. The method includes obtaining denoised particle concentration information by processing according to the sequential dust particle concentration information by introducing the maximum allowable deviation, obtaining the factory cleanliness and cleanliness warning information by processing and analyzing the denoised particle concentration information, obtaining environmental anomaly information according to the sequential environmental parameter information and the environmental parameter threshold, obtaining the environmental anomaly probability value and the environmental anomaly early warning information by regression analysis of the sequential environmental parameter information according to the environmental anomaly information, extracting the real-time image information of the staff according to the access control monitoring image, obtaining the wearing risk score by processing according to the real-time image information of the staff, generating wearing detection information according to the wearing risk score, and the wearing detection information includes wearing self-check warning information and passable prompt information, realizing the remote supervision of the clean factory and improving the real-time performance and efficiency of the environmental management of the clean factory.

Description

Technical Field

[0001] The present invention belongs to the technical field of operation and maintenance of clean factories, and particularly relates to an environmental supervision method and system for clean factories. Background Art

[0002] The development history of clean factories can be traced back to the American aviation industry in the early 20th century. With the rapid development of the electronics industry, clean technology has been increasingly widely used. In the early days, due to low production requirements, the structural form of electronic factories was simple. However, with the continuous progress of electronic science and technology, the performance of electronic products has become better and better, and the requirements for the production environment have also become higher and higher. This has prompted the continuous improvement of the cleanliness requirements of clean factories, and the structural form has also tended to be more complex.

[0003] Although clean factories play an important role in modern industrial production, there are still some problems and challenges. For example, traditional clean factory designs usually adopt high-energy-consuming air conditioning systems, which will increase production costs and waste energy. In addition, the environmental control of traditional clean factories requires staff to enter and exit the production environment regularly, which not only increases labor costs, but also has low control efficiency and cannot monitor the environmental information of clean factories in real time. As a result, when the cleanliness of clean factories fluctuates, it cannot be adjusted immediately, further affecting the production quality of clean factories. Summary of the Invention

[0004] To solve the above problems existing in the prior art, the present invention provides an environmental supervision method and system for clean factories, which acquires the sequential environmental parameter information, sequential dust particle concentration information, and access control monitoring images of preset sampling points, processes the sequential dust particle concentration information by introducing the maximum allowable deviation to obtain the denoised particle concentration information, analyzes and processes the denoised particle concentration information to obtain the factory cleanliness, generates a cleanliness warning information according to the factory cleanliness, obtains the environmental parameter difference according to the sequential environmental parameter information and the environmental parameter threshold, generates an environmental anomaly information according to the environmental parameter difference, obtains the environmental anomaly probability value by performing regression analysis on the sequential environmental parameter information according to the environmental anomaly information, generates an environmental anomaly warning information according to the environmental anomaly probability value, extracts the real-time image information of staff from the access control monitoring images, processes the staff image information to obtain a wearing risk score, generates a wearing detection information according to the wearing risk score, the wearing detection information includes a wearing self-check warning information and a passable prompt information, generates a work log according to the cleanliness warning information, environmental anomaly warning information, and wearing detection information, and uploads it to a display terminal, realizing the remote supervision of clean factories and improving the real-time performance and efficiency of clean factory environmental management.

[0005] The object of the present invention can be achieved by the following technical solutions:

[0006] An environmental supervision method for a clean factory, specifically including the following steps:

[0007] S1: Obtain the sequential environmental parameter information, sequential dust particle concentration information, and access control monitoring images of preset sampling points;

[0008] S2: Process the sequential dust particle concentration information by introducing the maximum allowable deviation to obtain denoised particle concentration information, perform sliding filtering on the denoised particle concentration information to obtain the first particle concentration information, obtain the factory cleanliness based on the first particle concentration information, generate cleanliness warning information based on the factory cleanliness, and generate a fresh air control instruction based on the cleanliness warning information;

[0009] S3: Obtain the environmental parameter threshold, obtain the environmental parameter difference based on the sequential environmental parameter information and the environmental parameter threshold, generate environmental anomaly information based on the environmental parameter difference, obtain the environmental anomaly probability value by performing regression analysis on the sequential environmental parameter information based on the environmental anomaly information, generate environmental anomaly warning information based on the environmental anomaly probability value, and generate an environmental regulation instruction based on the environmental anomaly warning information;

[0010] S4: Extract the real-time image information of the staff from the access control monitoring images, obtain the staff wearing characteristics by feature extraction based on the real-time image information of the staff, obtain the wearing risk score by processing the staff wearing characteristics through a wearing risk model, and generate wearing detection information based on the wearing risk score. The wearing detection information includes wearing self-check warning information and passable prompt information;

[0011] S5: Generate a work log based on the cleanliness warning information, the environmental anomaly warning information, and the wearing detection information, and upload it to the display terminal.

[0012] Preferably, the step S2 specifically includes the following steps:

[0013] S201: Calculate the denoised particle concentration information according to the sequential dust particle concentration information through the formula where x i represents the i-th denoised particle concentration information, and d represents the maximum allowable deviation;

[0014] S202: Calculate the second particle concentration information according to the denoised particle concentration information through the formula where P n represents n second particle concentration information, d represents the number of information after removing the initial value, maxx n,d represents the maximum value of the denoised particle concentration information, minxn,d represents the minimum value of the denoised particle concentration information, c represents the number of samplings, and the second particle concentration information is windowed to obtain the first particle concentration information, which is expressed as: where P1 n represents n pieces of the first particle concentration information, E represents the data length of the sliding window, and d represents the number of information after removing the initial value;

[0015] S203: Obtain the particle concentration mean value according to the first particle concentration information, and calculate the particle concentration mean error through the formula where WE represents the particle concentration mean error, P1 n represents n pieces of the first particle concentration information, r represents the particle concentration mean value, and a represents the total number of plant sampling points;

[0016] S204: Calculate the plant cleanliness through the formula M = r + g * WE according to the particle concentration mean value and the particle concentration mean error, where M represents the plant cleanliness, r represents the particle concentration mean value, WE represents the particle concentration mean error, and g represents the confidence distribution coefficient.

[0017] Preferably, step S3 specifically includes the following steps:

[0018] S301: Calculate the environmental parameter difference through the formula ΔP = P1 - P0 according to the time-series environmental parameter information and the environmental parameter threshold, where ΔP represents the environmental parameter difference, P1 represents the time-series environmental parameter information, P0 represents the environmental parameter threshold. When ΔP > 0, generate the environmental anomaly information; when ΔP ≤ 0, generate the environmental non-anomaly information;

[0019] S302: Obtain N groups of the environmental parameter information and calculate the environmental parameter mean value array, and the calculation formula is: Average[i] represents the environmental parameter mean value array, and num[i] represents the i-th group of the time-series environmental parameter information;

[0020] S303: Obtain the environmental parameter predicted value through regression analysis according to the environmental parameter mean value array, and calculate the environmental anomaly probability value through the formula where P represents the environmental anomaly probability value, L represents the environmental parameter threshold, Ki represents the environmental parameter predicted value, ∑(all) represents the time-series environmental parameter information. When the environmental anomaly probability value is greater than or equal to the preset threshold, generate the environmental anomaly warning information.

[0021] Preferably, the step S4 specifically includes the following steps:

[0022] Obtain standard wearing image information, respectively obtain the staff wearing features and standard wearing features through feature extraction based on the real-time image information of the staff and the standard wearing image information, and obtain the wearing feature distance by calculating the Euclidean distance between the staff wearing features and the standard wearing features. The calculation formula is: where OP represents the wearing feature distance, A1 and A2 represent feature weight coefficients, Di represents the staff wearing features, and Ei represents the standard wearing features;

[0023] Obtain the wearing risk score threshold, process the wearing feature distance through the wearing risk model to obtain the wearing risk score. When the wearing risk score is greater than or equal to the wearing risk score threshold, generate the wearing self-check warning information. When the wearing risk score is less than the wearing risk score threshold, generate the passable prompt information.

[0024] Preferably, the wearing risk model is expressed as: JK = Con(OP * OW), where JK represents the wearing risk score, OP represents the wearing feature distance, OW represents the weight coefficient, Con(·) represents the fusion scoring function, and * represents the convolution operator.

[0025] An environmental supervision system for a clean factory includes the following modules:

[0026] A data acquisition module for obtaining the sequential environmental parameter information, sequential dust particle concentration information, and access control monitoring images of preset sampling points;

[0027] A cleanliness calculation module for processing the sequential dust particle concentration information by introducing the maximum allowable deviation to obtain the denoised particle concentration information, obtaining the first particle concentration information through sliding filtering based on the denoised particle concentration information, obtaining the factory cleanliness based on the first particle concentration information, generating a cleanliness warning information based on the factory cleanliness, and generating a fresh air control instruction based on the cleanliness warning information;

[0028] An environmental monitoring module for obtaining environmental parameter thresholds, obtaining the environmental parameter difference based on the sequential environmental parameter information and the environmental parameter thresholds, generating environmental anomaly information based on the environmental parameter difference, obtaining the environmental anomaly probability value by performing regression analysis on the sequential environmental parameter information based on the environmental anomaly information, generating an environmental anomaly early warning information based on the environmental anomaly probability value, and generating an environmental regulation instruction based on the environmental anomaly early warning information;

[0029] The wearing detection module is used to extract the real-time image information of the staff according to the access control monitoring image, obtain the wearing characteristics of the staff through feature extraction according to the real-time image information of the staff, obtain the wearing risk score through processing by the wearing risk model according to the wearing characteristics of the staff, and generate wearing detection information according to the wearing risk score. The wearing detection information includes wearing self-check warning information and passable prompt information;

[0030] The supervision log module is used to generate a work log according to the cleanliness warning information, the environmental anomaly early warning information and the wearing detection information, and upload it to the display terminal.

[0031] The beneficial effects of the present invention are as follows:

[0032] 1. The denoised particle concentration information is obtained by processing the sequential dust particle concentration information by introducing the maximum allowable deviation. The first particle concentration information is obtained by sliding filtering processing according to the denoised particle concentration information. The factory cleanliness is obtained according to the first particle concentration information. The cleanliness warning information is generated according to the factory cleanliness. The fresh air control instruction is generated according to the cleanliness warning information. The factory cleanliness is calculated according to the sequential dust particle concentration information. The impurity signal is eliminated by sliding filtering processing according to the collected sequential dust particle concentration information, further improving the accuracy of the sensor transmission value;

[0033] 2. The real-time image information of the staff is extracted according to the access control monitoring image. The wearing characteristics of the staff are obtained through feature extraction according to the staff image information. The wearing risk score is obtained through processing by the wearing risk model according to the wearing characteristics of the staff. The wearing detection information is generated according to the wearing risk score. The wearing detection information includes wearing self-check warning information and passable prompt information. The wearing self-check warning information is generated according to the wearing risk score to remind the staff to self-check whether the wearing is compliant, ensuring that the wearing of the staff entering the clean area of the clean factory meets the standards and ensuring the cleanliness of the clean area;

[0034] 3. By monitoring the sequential environmental parameter information, the sequential dust particle concentration information and the access control monitoring image, and obtaining the cleanliness warning information, the environmental anomaly early warning information and the wearing detection information through processing, the remote supervision of the clean factory is realized, and the real-time performance and efficiency of the environmental management of the clean factory are improved. Description of the Drawings

[0035] For the convenience of those skilled in the art to understand, the present invention will be further described below with reference to the drawings.

[0036] Figure 1 It is a schematic flow chart of the environmental supervision method for a clean factory of the present invention. Detailed Embodiments

[0037] To further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following will, in conjunction with the accompanying drawings and preferred embodiments, elaborate in detail on the specific implementation manners, structures, features, and effects according to the present invention as follows.

[0038] Please refer to Figure 1 , an environmental supervision method for a clean factory, specifically including the following steps:

[0039] S1: Obtain the time-series environmental parameter information, time-series dust particle concentration information, and access control monitoring images of preset sampling points;

[0040] S2: Process the time-series dust particle concentration information by introducing the maximum allowable deviation to obtain the denoised particle concentration information, perform sliding filtering on the denoised particle concentration information to obtain the first particle concentration information, obtain the factory cleanliness based on the first particle concentration information, generate cleanliness warning information based on the factory cleanliness, and generate a fresh air control instruction based on the cleanliness warning information;

[0041] S3: Obtain the environmental parameter threshold, obtain the environmental parameter difference based on the time-series environmental parameter information and the environmental parameter threshold, generate environmental anomaly information based on the environmental parameter difference, obtain the environmental anomaly probability value by performing regression analysis on the time-series environmental parameter information based on the environmental anomaly information, generate environmental anomaly warning information based on the environmental anomaly probability value, and generate an environmental regulation instruction based on the environmental anomaly warning information;

[0042] S4: Extract the real-time image information of the staff from the access control monitoring images, obtain the wearing characteristics of the staff through feature extraction based on the real-time image information of the staff, obtain the wearing risk score through processing by the wearing risk model based on the wearing characteristics of the staff, generate wearing detection information based on the wearing risk score, and the wearing detection information includes wearing self-check warning information and passable prompt information;

[0043] S5: Generate a work log based on the cleanliness warning information, the environmental anomaly warning information, and the wearing detection information, and upload it to the display terminal.

[0044] Sample the environmental parameters at preset sampling points in the clean factory, set up access control for the clean area of the clean factory, and identify whether the wearing conditions of the staff meet the standards.

[0045] Specifically, the step S2 specifically includes the following steps:

[0046] S201: Calculate the denoised particle concentration information according to the time-series dust particle concentration information through the formula where xi represents the i-th denoised particle concentration information, and d represents the maximum allowable deviation;

[0047] S202: According to the denoised particle concentration information, through the formula calculate to obtain the second particle concentration information, where P n represents n pieces of the second particle concentration information, d represents the number of information after removing the initial value, maxx n,d represents the maximum value of the denoised particle concentration information, minx n,d represents the minimum value of the denoised particle concentration information, c represents the number of sampling times, perform windowing processing on the second particle concentration information to obtain the first particle concentration information, and the first particle concentration information is expressed as: where P1 n represents n pieces of the first particle concentration information, E represents the data length of the sliding window, and d represents the number of information after removing the initial value;

[0048] S203: Obtain the particle concentration mean according to the first particle concentration information, and calculate the particle concentration mean error through the formula where WE represents the particle concentration mean error, P1 n represents n pieces of the first particle concentration information, r represents the particle concentration mean, and a represents the total number of sampling points in the factory building;

[0049] S204: Calculate the factory cleanliness according to the particle concentration mean and the particle concentration mean error through the formula M = r + g * WE, where M represents the factory cleanliness, r represents the particle concentration mean, WE represents the particle concentration mean error, and g represents the confidence distribution coefficient.

[0050] Calculate the factory cleanliness according to the time-series dust particle concentration information. Eliminate the impurity signal through sliding filtering processing of the collected time-series dust particle concentration information to further improve the accuracy of the sensor transmission value. It should be further noted that when |a i - a i-1 | is less than or equal to the maximum allowable deviation, x i takes the i-th time-series particle concentration information. When |a i - a i-1 | is greater than the maximum allowable deviation, x i takes the (i - 1)-th time-series particle concentration information.

[0051] Specifically, step S3 specifically includes the following steps:

[0052] S301: Calculate the environmental parameter difference according to the timing environmental parameter information and the environmental parameter threshold through the formula ΔP = P1 - P0, where ΔP represents the environmental parameter difference, P1 represents the timing environmental parameter information, and P0 represents the environmental parameter threshold. When ΔP > 0, generate the environmental anomaly information; when ΔP ≤ 0, generate the environmental non - anomaly information;

[0053] S302: Obtain N groups of the environmental parameter information and calculate the environmental parameter mean array. The calculation formula is: Average[i] represents the environmental parameter mean array, and num[i] represents the i - th group of the timing environmental parameter information;

[0054] S303: Obtain the environmental parameter predicted value through regression analysis according to the environmental parameter mean array. Calculate the environmental anomaly probability value according to the environmental parameter threshold and the environmental parameter predicted value through the formula where P represents the environmental anomaly probability value, L represents the environmental parameter threshold, Ki represents the environmental parameter predicted value, and ∑(all) represents the timing environmental parameter information. When the environmental anomaly probability value is greater than or equal to the preset threshold, generate the environmental anomaly warning information.

[0055] Specifically, step S4 specifically includes the following steps:

[0056] Obtain the standard wearing image information. Respectively extract the staff wearing features and the standard wearing features from the staff real - time image information and the standard wearing image information. Obtain the wearing feature distance by calculating the Euclidean distance between the staff wearing features and the standard wearing features. The calculation formula is: where OP represents the wearing feature distance, A1 and A2 represent the feature weight coefficients, Di represents the staff wearing features, and Ei represents the standard wearing features;

[0057] Obtain the wearing risk score threshold. Process the wearing feature distance through the wearing risk model to obtain the wearing risk score. When the wearing risk score is greater than or equal to the wearing risk score threshold, generate the wearing self - inspection warning information; when the wearing risk score is less than the wearing risk score threshold, generate the passable prompt information.

[0058] Specifically, the wearing risk model is expressed as: JK = Con(OP * OW), where JK represents the wearing risk score, OP represents the wearing feature distance, OW represents the weight coefficient, Con(·) represents the fusion scoring function, and * represents the convolution operator.

[0059] Generate a self-check warning message for wearing based on the wearing risk score to remind the staff to check whether their wearing is compliant, ensuring that the wearing of the staff entering the clean area of the clean factory meets the standards and guaranteeing the cleanliness of the clean area.

[0060] Furthermore, the present application provides an environmental monitoring system for a clean factory, including the following modules:

[0061] A data acquisition module, configured to obtain the sequential environmental parameter information, sequential dust particle concentration information, and access control monitoring images of preset sampling points;

[0062] A cleanliness calculation module, configured to process the sequential dust particle concentration information by introducing a maximum allowable deviation to obtain denoised particle concentration information, obtain first particle concentration information through sliding filtering based on the denoised particle concentration information, obtain the factory cleanliness based on the first particle concentration information, generate a cleanliness warning message based on the factory cleanliness, and generate a fresh air control instruction based on the cleanliness warning message;

[0063] An environmental monitoring module, configured to obtain environmental parameter thresholds, obtain an environmental parameter difference based on the sequential environmental parameter information and the environmental parameter thresholds, generate an environmental anomaly message based on the environmental parameter difference, obtain an environmental anomaly probability value by performing regression analysis on the sequential environmental parameter information based on the environmental anomaly message, generate an environmental anomaly warning message based on the environmental anomaly probability value, and generate an environmental regulation instruction based on the environmental anomaly warning message;

[0064] A wearing detection module, configured to extract real-time image information of the staff based on the access control monitoring images, obtain the wearing characteristics of the staff through feature extraction based on the real-time image information of the staff, obtain a wearing risk score through processing by a wearing risk model based on the wearing characteristics of the staff, and generate wearing detection information based on the wearing risk score, where the wearing detection information includes a self-check warning message for wearing and a passable prompt message;

[0065] A supervision log module, configured to generate a work log based on the cleanliness warning message, the environmental anomaly warning message, and the wearing detection information, and upload it to a display terminal.

[0066] The working principle and usage process of the present invention:

[0067] Obtain the timing environmental parameter information, timing dust particle concentration information, and access control monitoring images of preset sampling points. Process the timing dust particle concentration information by introducing the maximum allowable deviation to obtain the denoised particle concentration information. Analyze and process the denoised particle concentration information to obtain the factory cleanliness. Generate cleanliness warning information based on the factory cleanliness. Obtain the environmental parameter difference based on the timing environmental parameter information and the environmental parameter threshold. Generate environmental anomaly information based on the environmental parameter difference. Use regression analysis on the timing environmental parameter information according to the environmental anomaly information to obtain the environmental anomaly probability value. Generate environmental anomaly warning information based on the environmental anomaly probability value. Extract the real-time image information of the staff from the access control monitoring images. Process the staff image information to obtain the wearing risk score. Generate wearing detection information based on the wearing risk score. The wearing detection information includes wearing self-check warning information and passable prompt information. Generate a work log based on the cleanliness warning information, environmental anomaly warning information, and wearing detection information, and upload it to the display terminal.

[0068] The program code included in the method in the embodiments of the present invention can be transmitted by any suitable medium, including but not limited to wireless, wire, optical cable, etc., or any suitable combination of the above. The computer program code for performing the operations of the present invention can be written in one or more programming languages or combinations thereof. The programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, executed as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, by using an Internet service provider to connect through the Internet).

[0069] As mentioned above, it is only a preferred embodiment of the present invention and does not impose any form of limitation on the present invention. Although the present invention has been disclosed above with a preferred embodiment, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to the above-disclosed technical content within the scope of the technical solution of the present invention to obtain equivalent embodiments with equivalent changes. However, as long as it does not depart from the content of the technical solution of the present invention, any brief modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. An environmental monitoring method for a clean factory, characterized in that: The specific steps include: S1: Obtaining the time-series environmental parameter information, time-series dust particle concentration information and access control monitoring images of the preset sampling points; S2: According to the time series dust particle concentration information, a maximum allowable deviation is introduced to process to obtain denoised particle concentration information, according to the denoised particle concentration information, a first particle concentration information is obtained by sliding filtering, according to the first particle concentration information, a factory cleanliness is obtained, according to the factory cleanliness, a cleanliness warning information is generated, and a fresh air control instruction is generated according to the cleanliness warning information; S3: Acquire an environmental parameter threshold, obtain an environmental parameter difference according to the time series environmental parameter information and the environmental parameter threshold, generate environmental anomaly information according to the environmental parameter difference, obtain an environmental anomaly probability value by performing regression analysis on the time series environmental parameter information according to the environmental anomaly information, generate environmental anomaly warning information according to the environmental anomaly probability value, and generate an environmental control instruction according to the environmental anomaly warning information; S4: extracting real-time image information of the staff according to the access control monitoring image, obtaining wearing features of the staff by feature extraction according to the real-time image information of the staff, obtaining wearing risk scores according to the wearing features of the staff by processing through a wearing risk model, and generating wearing detection information according to the wearing risk scores, wherein the wearing detection information includes wearing self-check warning information and passable prompt information; S5: Generate a work log according to the cleanliness warning information, the abnormal environment warning information and the wearing detection information, and upload it to the display terminal.

2. The environmental monitoring method for a clean factory according to claim 1, characterized in that: The step S2 specifically includes the following steps: S201: According to the time series dust particle concentration information, the formula The denoised particle concentration information is calculated, where x i represents the denoised particle concentration information of the i-th particle, and d represents the maximum allowable deviation; S202: According to the denoised particle concentration information, the formula The second particle concentration information is calculated, where P n represents the concentration information of n second particles, d represents the number of information after removing the initial value, maxx n,d Indicates the maximum value of the denoised particle concentration information, minx n,d represents the minimum value of the denoised particle concentration information, c represents the number of sampling times, and the second particle concentration information is subjected to windowing processing to obtain the first particle concentration information, which is expressed as: Among them, P1 n represents the concentration information of n first particles, E represents the data length of the sliding window, and d represents the amount of information after removing the initial value; S203: Obtain a particle concentration mean value according to the first particle concentration information, and calculate the particle concentration mean value according to the formula The mean error of particle concentration is calculated, where WE represents the mean error of particle concentration, P1 n represents n pieces of the first particle concentration information, r represents the mean value of the particle concentration, and a represents the total number of sampling points in the plant; S204: The factory cleanliness is calculated according to the mean particle concentration and the mean error of the particle concentration by using the formula M=r+g*WE, wherein M represents the factory cleanliness, r represents the mean particle concentration, WE represents the mean error of the particle concentration, and g represents the confidence distribution coefficient.

3. The environmental monitoring method for a clean factory according to claim 1, characterized in that: The step S3 specifically comprises the following steps: S301: Calculate the environmental parameter difference according to the time series environmental parameter information and the environmental parameter threshold value by using the formula ΔP=P1-P0, wherein ΔP represents the environmental parameter difference, P1 represents the time series environmental parameter information, and P0 represents the environmental parameter threshold value. When ΔP>0, generate the environmental abnormality information, and when ΔP≤0, generate the environmental normal abnormality information; S302: Obtain N groups of environmental parameter information and calculate the mean array of environmental parameters. The calculation formula is: Average[i] represents the mean value array of the environmental parameters, and num[i] represents the time series environmental parameter information of the i-th group; S303: Obtain an environmental parameter prediction value through regression analysis according to the environmental parameter mean value array, and obtain an environmental parameter prediction value through a formula according to the environmental parameter threshold and the environmental parameter prediction value. The environmental abnormality probability value is calculated, wherein P represents the environmental abnormality probability value, L represents the environmental parameter threshold, Ki represents the environmental parameter prediction value, ∑(all) represents the time series environmental parameter information, and when the environmental abnormality probability value is greater than or equal to the preset threshold, the environmental abnormality warning information is generated.

4. The environmental monitoring method for a clean factory according to claim 1, characterized in that: The step S4 specifically comprises the following steps: The standard wearing image information is obtained, and the staff wearing features and the standard wearing features are obtained by feature extraction according to the staff real-time image information and the standard wearing image information, and the wearing feature distance is obtained by calculating the Euclidean distance between the staff wearing features and the standard wearing features. The calculation formula is: Among them, OP represents the wearing feature distance, A1 and A2 represent the feature weight coefficients, Di represents the staff wearing feature, and Ei represents the standard wearing feature; Obtain a wearing risk score threshold, obtain the wearing risk score through the wearing risk model according to the wearing feature distance, generate the wearing self-check warning information when the wearing risk score is greater than or equal to the wearing risk score threshold, and generate the passable prompt information when the wearing risk score is less than the wearing risk score threshold.

5. The environmental monitoring method for a clean factory according to claim 4, characterized in that: The wearing risk model is expressed as: JK=Con(OP*OW), wherein JK represents the wearing risk score, OP represents the wearing feature distance, OW represents the weight coefficient, Con(·) represents the fusion scoring function, and * represents the convolution operator.

6. An environmental monitoring system for a clean factory, the environmental monitoring system for a clean factory adopts the environmental monitoring method for a clean factory as claimed in claim 1, characterized in that: Includes the following modules: A data acquisition module is used to obtain the time-series environmental parameter information, time-series dust particle concentration information and access control monitoring images of preset sampling points; a cleanliness calculation module, configured to obtain denoised particle concentration information by introducing a maximum allowable deviation according to the time series dust particle concentration information, obtain first particle concentration information by sliding filtering according to the denoised particle concentration information, obtain factory cleanliness according to the first particle concentration information, generate cleanliness warning information according to the factory cleanliness, and generate a fresh air control instruction according to the cleanliness warning information; An environmental monitoring module is used to obtain an environmental parameter threshold, obtain an environmental parameter difference according to the time series environmental parameter information and the environmental parameter threshold, generate environmental anomaly information according to the environmental parameter difference, obtain an environmental anomaly probability value by performing regression analysis on the time series environmental parameter information according to the environmental anomaly information, generate environmental anomaly warning information according to the environmental anomaly probability value, and generate an environmental control instruction according to the environmental anomaly warning information; A wearing detection module, used to extract real-time image information of staff according to the access control monitoring image, obtain wearing features of staff by feature extraction according to the real-time image information of staff, obtain wearing risk scores according to the wearing features of staff by processing through a wearing risk model, and generate wearing detection information according to the wearing risk scores, wherein the wearing detection information includes wearing self-check warning information and passable prompt information; The supervision log module is used to generate a work log according to the cleanliness warning information, the environmental abnormality warning information and the wearing detection information, and upload it to the display terminal.

Images