Monitoring system and method applied to factory production safety

Abstract

The invention discloses a monitoring system and method applied to factory production safety, relates to the technical field of production safety, and solves the technical problems that the management of factory safety production at present mainly focuses on the aspects of equipment monitoring, production environment monitoring and the like and neglects the behavior monitoring of factory operators, resulting in safety accidents. The method comprises the steps of collecting face images of operators on a factory production line; extracting a face feature image of the operator from the face image; acquiring a plurality of analysis coordinates of the facial feature image based on the facial feature image; analyzing behavior parameters of the operator based on the analysis coordinates; based on the behavior parameters, the behavior state of the operator is analyzed, and safety early warning is conducted on the operator; the monitoring of the behaviors of factory operators is realized, the production safety is ensured, the production efficiency is improved, and the product quality is ensured, so that the safety production level of a factory is comprehensively improved.

Description

Technical Field

[0001] The present invention belongs to the field of production safety, and in particular relates to a monitoring system and method for factory production safety. Background Art

[0002] Currently, factory safety management primarily focuses on equipment monitoring and production environment monitoring, which undoubtedly plays a vital role in preventing equipment failures, ensuring smooth production processes, and maintaining a good working environment. However, an often overlooked yet crucial aspect is monitoring the behavior of factory operators. Operators are the primary force in factory production, and their behavior directly impacts production safety. Irregular operations, negligence, and fatigue can all lead to safety accidents. Therefore, the present invention provides a monitoring system and method for factory production safety. Summary of the Invention

[0003] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a monitoring system and method for factory production safety, which is used to solve the technical problem that the current management of factory production safety mainly focuses on equipment monitoring and production environment monitoring, while ignoring the monitoring of factory operator behavior, leading to safety accidents.

[0004] To achieve the above-mentioned object, a first aspect of the present invention provides a monitoring system for factory production safety, comprising an analysis module, an image acquisition module and an early warning module connected thereto;

[0005] Image acquisition module: used to collect facial images of operators on the factory production line;

[0006] Analysis module: used for extracting facial feature images of the operator from the facial image; wherein the facial feature images include lip feature images and eye feature images;

[0007] Based on the facial feature image, several analysis coordinates of the facial feature image are obtained; based on the analysis coordinates, the operator's behavior parameters are analyzed; based on the behavior parameters, the operator's behavior state is analyzed; wherein, the analysis coordinates include lip analysis coordinates and eye analysis coordinates, the behavior parameters include lip behavior parameters and eye behavior parameters, and the behavior state includes normal behavior state and abnormal behavior state.

[0008] Preferably, the step of extracting the operator's facial feature image from the facial image includes:

[0009] Input the face image into the feature recognition model and output the face feature label; the face feature label is matched with the face feature image one by one;

[0010] The construction process of the feature recognition model is as follows:

[0011] extracting a plurality of facial images and a plurality of facial feature images from a factory database;

[0012] Integrate several facial images and several facial feature labels into several groups of training data and test data; use the training data to train an artificial intelligence model; use the test data to test the trained artificial intelligence model, and adjust the artificial intelligence model according to the test results; finally obtain a feature recognition model with facial images as input and facial feature labels as output; wherein the artificial intelligence model is a BP neural network model or an RBF neural network model.

[0013] The present invention automatically receives input facial images and quickly outputs corresponding facial feature labels without manual intervention, greatly improving processing efficiency and accuracy. Both BP neural networks and RBF neural networks have the ability to self-learn and adjust, and can continuously optimize model performance based on new data to adapt to more extensive and complex facial feature recognition needs.

[0014] Preferably, the step of obtaining several analysis coordinates of the facial feature image includes:

[0015] Establishing a coordinate system of the facial feature image with the center position of the facial feature image as the origin of the coordinate system and a number of pixels as one coordinate unit; and extracting the contour lines of the lip area and the eye area;

[0016] Based on the coordinate system of the facial feature image and the contour lines of the lip area and the eye area, a number of analysis coordinates are determined; wherein the analysis coordinates include lip analysis coordinates and eye analysis coordinates.

[0017] This invention establishes a coordinate system with the center of the facial feature image as the origin, standardizing and unifying the positions of facial features. Using a number of pixels as a coordinate unit, the accuracy of the coordinate system is ensured. This high-precision coordinate system accurately reflects subtle changes in facial feature points, providing a reliable foundation for subsequent contour extraction and analysis. By using image processing techniques to precisely extract the contours of the lip and eye regions, key coordinate points in these areas can be accurately determined, laying an important foundation for subsequent behavioral analysis of operators.

[0018] Preferably, calculating the lip behavior parameter based on the lip feature image includes:

[0019] Acquire a number of lip analysis coordinates based on the lip feature image; wherein the lip analysis coordinates include the coordinates of the corner of the mouth, the center coordinates of the upper lip, and the center coordinates of the lower lip;

[0020] The lip opening and closing angle θ is calculated using the following formula:

[0021]

[0022] Among them, the lip opening and closing angle value is the lip behavior parameter, (X1, Y1) is the coordinate of the mouth corner, (X2, Y2) is the center coordinate of the upper lip, (X3, Y3) is the center coordinate of the lower lip, γ is the opening and closing angle correction coefficient, γ∈(0,1];

[0023] Determine whether the opening and closing angle value is greater than a preset angle threshold; if yes, mark the opening and closing angle value as an abnormal angle; if not, mark the opening and closing angle as a normal angle.

[0024] This method accurately extracts the coordinates of the mouth corners, upper lip center, and lower lip center from lip feature images and constructs a formula to calculate the lip opening and closing angle. This method avoids the subjectivity and inconsistency of manual judgment. Furthermore, a preset angle threshold is used to determine whether the opening and closing angle is abnormal, achieving standardized lip behavior analysis. By introducing the opening and closing angle correction coefficient γ, the method can be adjusted based on the lip characteristics of different individuals or situations, enhancing its adaptability and flexibility.

[0025] Preferably, calculating the lip behavior parameter based on the lip feature image further includes:

[0026] Obtain the circumscribed rectangle of the lip, extract the coordinates of the four intersection points of the circumscribed rectangle, and calculate the area of the circumscribed rectangle of the lip using the rectangle area formula; determine whether the area of the circumscribed rectangle is greater than the preset area threshold; if so, mark the opening and closing angle value as an abnormal angle; otherwise, mark the opening and closing angle value as a normal angle.

[0027] The present invention uses image processing technology (such as edge detection, morphological operations, contour detection, etc.) to identify the lip area, calculates its circumscribed rectangle based on the identified lip area, extracts the coordinates of the four intersection points of the circumscribed rectangle, calculates the area of the lip circumscribed rectangle, and intuitively represents the lip area through the circumscribed rectangle, further reflecting the opening and closing of the lips.

[0028] Preferably, calculating the eye behavior parameter based on the eye feature image includes:

[0029] Acquire a number of eye analysis coordinates based on the eye feature image; wherein the eye analysis coordinates include the center coordinates of the upper eyelid, the center coordinates of the lower eyelid, and the center coordinates of the pupil;

[0030] Based on the center coordinate of the upper eyelid and the center coordinate of the lower eyelid, the longitudinal distance between the upper eyelid and the lower eyelid is calculated and marked as the eye distance value; determining whether the eye distance value is less than a preset distance threshold; if so, marking the eye distance value as abnormal eye distance;

[0031] If not, the center coordinates of the pupil will be extracted to determine whether the center coordinates of the pupil are the same as the center coordinates of the eye; if yes, the pupil position will be marked as normal; if not, the pupil position will be marked as abnormal.

[0032] The present invention can accurately quantify eye features by extracting the center coordinates of the upper eyelid, lower eyelid, and pupil. Based on a preset distance threshold and the eye center coordinates, it can determine whether the interocular distance and pupil position are abnormal, and monitor the operator's eye status in real time, avoiding errors caused by human observation and improving the objectivity of the analysis.

[0033] Preferably, analyzing the operator's behavior status based on the behavior parameters includes:

[0034] Set the video analysis cycle and divide the video analysis cycle into several processing frames;

[0035] Determine whether the number of occurrences of abnormal angles or abnormal eye distances within a video analysis period exceeds a corresponding preset number threshold, and whether the duration of the occurrence interval is less than a corresponding preset interval duration threshold; if so, mark the lip state as abnormal lip state or mark the eye state as abnormal eye state; if not, mark the lip state as normal lip state or mark the eye state as normal eye state;

[0036] Determine whether pupil position abnormality occurs continuously for a preset time within the video analysis period; if so, mark the eye state as abnormal; if not, mark the eye state as normal;

[0037] When the operator has any of abnormal lip state, abnormal eye state or abnormal pupil position, the operator's behavior state is marked as abnormal behavior state; otherwise, the operator's behavior state is marked as normal behavior state.

[0038] The present invention monitors in real time the number of occurrences of key indicators such as lip angle, eye distance, and pupil position within a video analysis cycle, as well as parameters such as interval duration or continuous appearance duration, to accurately judge abnormal conditions of the lips and eyes, avoiding the problem of misjudgment based on short-term observations alone. When an abnormal condition is detected within the analysis cycle, the system can quickly issue an early warning to remind operators or managers to take corresponding measures to prevent potential problems from occurring.

[0039] Preferably, the preset angle threshold and the preset distance threshold are the lip opening and closing angle value and the eye distance value of each person in a normal behavior state.

[0040] The present invention sets specific thresholds for each person based on the lip opening and closing angle values and eye distance values of each person in a normal behavior state, which can more accurately determine whether the lip opening and closing angle and eye distance are abnormal, thereby improving the accuracy of the analysis. This method fully takes individual differences into account, can more accurately reflect the actual status of each person, and avoids one-size-fits-all misjudgments.

[0041] Preferably, the early warning module is used to warn the operator of abnormal behavior; wherein the operator wears a communication device, and when an abnormal behavior occurs, the communication device sends a warning signal to remind the operator.

[0042] Preferably, a monitoring method for factory production safety includes the following steps:

[0043] Collect facial images of operators on a factory production line; extract facial feature images of operators from the facial images;

[0044] Based on the facial feature image, a number of analysis coordinates of the facial feature image are obtained; based on the analysis coordinates, behavioral parameters of the operator are analyzed;

[0045] Based on behavioral parameters, the operator's behavioral status is analyzed and safety warnings are issued to the operator.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] To protect the safety of factory operators, the present invention uses a camera to capture the operator's facial image in real time, extract facial feature images from them, and obtain several analysis coordinates of the facial feature images. Using these analysis coordinates, the operator's lip behavior parameters and eye behavior parameters are analyzed, and the analyzed behavior parameters are compared with the corresponding set thresholds. This allows for a more accurate assessment of the operator's behavioral state, timely warning of abnormal behavior states, and effective prevention of accidents, ensuring the operator's personal safety and the safe and stable operation of the production line. In summary, the present invention not only monitors the state of the lips and eyes, but also comprehensively considers changes in pupil position, achieving comprehensive coverage and comprehensive assessment of the operator's behavioral state, helping to more accurately reflect the operator's actual state and improve the accuracy and comprehensiveness of behavioral state judgment. In addition, the system has a high level of automation and intelligence, capable of automatic collection, analysis, and warning, greatly reducing the burden of manual monitoring and improving work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0049] Figure 1 Schematic diagram of the system structure of the present invention;

[0050] Figure 2 This is a schematic diagram of the lip and eye status evaluation and analysis of the present invention;

[0051] Figure 3 Schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0052] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the embodiments described are only a part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0053] See also Figure 1 , a first aspect of the present invention provides a monitoring system for factory production safety, comprising an analysis module, and an image acquisition module and an early warning module connected thereto;

[0054] The image acquisition module uses a camera to capture facial images of operators on the factory production line;

[0055] The analysis module extracts the operator's facial feature image from the facial image; wherein the facial feature image includes a lip feature image and an eye feature image;

[0056] Specifically, the face image is input into the feature recognition model, and the face feature labels are output; the face feature labels are matched one-to-one with the face feature images;

[0057] The construction process of the feature recognition model is as follows:

[0058] extracting a number of facial images and facial feature images from a factory database;

[0059] Integrate several facial images and several facial feature labels into several groups of training data and test data; use the training data to train an artificial intelligence model; use the test data to test the trained artificial intelligence model, and adjust the artificial intelligence model according to the test results; finally obtain a feature recognition model with facial images as input and facial feature labels as output; wherein the artificial intelligence model is a BP neural network model or an RBF neural network model.

[0060] For example, facial images of several people are obtained from the factory database, and natural number labels are set for the facial feature images, that is, the lip feature image is labeled as 0 and the eye feature image is labeled as 1; the facial images and facial feature labels are respectively combined into several groups of training data and test data to train the artificial intelligence model.

[0061] Based on the facial feature image, obtaining a number of analysis coordinates of the facial feature image;

[0062] Specifically, the SUSAN algorithm is used to detect the corner points of the lip area and the eye area in the facial feature image respectively, and then the contour lines of the lip area and the eye area are obtained by NURBS cubic curve fitting; wherein, the coordinate system of the facial feature image is established with the center position of the facial feature image as the origin of the coordinate system and a number of pixel points as a coordinate unit.

[0063] Analyze operator behavior parameters based on analysis coordinates;

[0064] The analysis method of lip behavior parameters is as follows:

[0065] Example 1: Acquire a number of lip analysis coordinates based on a lip feature image; wherein the lip analysis coordinates include the coordinates of the mouth corners, the center coordinates of the upper lip, and the center coordinates of the lower lip;

[0066] The lip opening and closing angle θ is calculated using the following formula:

[0067]

[0068] Among them, the lip opening and closing angle value is the lip behavior parameter, (X1, Y1) is the coordinate of the mouth corner, (X2, Y2) is the center coordinate of the upper lip, (X3, Y3) is the center coordinate of the lower lip, γ is the opening and closing angle correction coefficient, γ∈(0,1];

[0069] It should be noted that the mouth corner coordinates can be selected from any side of the mouth corner, and the mouth corner, the center point of the upper lip and the center point of the lower lip form an isosceles triangle.

[0070] Determine whether the opening and closing angle value is greater than a preset angle threshold; if yes, mark the opening and closing angle value as an abnormal angle; if not, mark the opening and closing angle value as a normal angle.

[0071] See also Figure 2 For example, assume that the corner coordinates are (X1, Y1) = (50, 80), the upper lip center coordinates are (X2, Y2) = (60, 70), the lower lip center coordinates are (X3, Y3) = (60, 90), and the opening and closing angle correction coefficient γ is set to 0.8;

[0072] Substitute the above lip coordinates into the calculation formula of the opening and closing angle, |Y2-Y3|=|70-90|=20, We get arcsin(0.707)≈0.785 radians (approximately 45 0 ), and finally the solution is θ≈72 0 .

[0073] Example 2: Obtain the circumscribed rectangle of the lips, extract the coordinates of the four intersection points of the circumscribed rectangle, and calculate the area of the circumscribed rectangle of the lips using the area formula of the rectangle; determine whether the area of the circumscribed rectangle is greater than a preset area threshold; if so, mark the opening and closing angle value as an abnormal angle; if not, mark the opening and closing angle value as a normal angle.

[0074] For example: Upper left corner: (Xlt, Ylt) = (50, 70) (Xlt, Ylt) = (50, 70), Upper right corner: (Xrt, Yrt) = (90, 70) (Xrt, Yrt) = (90, 70), Lower left corner: (Xlb, Ylb) = (50, 100) (Xlb, Ylb) = (50, 100), Lower right corner: (Xrb, Yrb) = (90, 100) (Xrb, Yrb) = (90, 100);

[0075] The width of the rectangle (horizontally) is: Width = |Xrt - Xlt |= |90 - 50 |= 40;

[0076] The height of the rectangle (vertical direction) is: Height = |Ylb - Ylt |= |100 - 70 |= 30;

[0077] The area of the rectangle A = 40 × 30 = 1200.

[0078] The analysis method of eye behavior parameters is as follows:

[0079] Acquire a number of eye analysis coordinates based on the eye feature image; wherein the eye analysis coordinates include the center coordinates of the upper eyelid, the center coordinates of the lower eyelid, and the center coordinates of the pupil;

[0080] Based on the center coordinate of the upper eyelid and the center coordinate of the lower eyelid, the longitudinal distance between the upper eyelid and the lower eyelid is calculated and marked as the eye distance value; determining whether the eye distance value is less than a preset distance threshold; if so, marking the eye distance value as abnormal eye distance;

[0081] If not, it is determined whether the center coordinates of the pupil are the same as the center coordinates of the eye; if yes, the pupil position is marked as normal; if not, the pupil position is marked as abnormal.

[0082] See also Figure 2 ,Based on the behavioral parameters, the operator's behavioral status is analyzed. The analysis process is as follows:

[0083] Set the video analysis cycle, divide the video analysis cycle into several processing frames evenly, and analyze the behavioral parameters of the processing frames;

[0084] Count the number of abnormal angles or abnormal eye distances that occur during the video analysis period, and determine whether the number exceeds a corresponding preset number threshold, and whether the duration of the interval between occurrences is less than a corresponding preset interval duration threshold; if so, mark the lip state as abnormal lip state or the eye state as abnormal eye state; if not, mark the lip state as normal lip state or the eye state as normal eye state;

[0085] Determine whether pupil position abnormality occurs continuously for a preset time within the video analysis period; if so, mark the eye state as abnormal; if not, mark the eye state as normal;

[0086] It should be noted that the larger the lip opening and closing angle value, the more likely the operator's lips are moving, and they may be talking or yawning; the smaller the eye distance, the more likely the operator is fatigued; and abnormal pupil position, i.e., left / right shift of the pupil, indicates that the operator may be distracted.

[0087] Therefore, when the operator has any of the abnormal lip state, abnormal eye state or abnormal pupil position, the operator's behavior state is marked as an abnormal behavior state; otherwise, the operator's behavior state is marked as a normal behavior state.

[0088] The preset angle threshold and the preset distance threshold are the lip opening and closing angle value and the eye distance value of each person in a normal behavior state.

[0089] For example: assuming that the preset angle threshold and the preset distance threshold of operator A are analyzed, several facial images of operator A in a normal behavior state are extracted from the video database, and their facial feature images are extracted, and the lip opening and closing angle value and the eye distance value in each facial feature image are calculated. The calculated several opening and closing angle values and several eye distance values are respectively averaged to obtain the preset angle threshold and the preset distance threshold.

[0090] The early warning module is used to warn the operator of abnormal behavior; the operator wears a communication device, and when an abnormal behavior occurs, the communication device sends a warning signal to remind the operator.

[0091] For example, the early warning system sends the following warning signal through operator A's communication device (such as a smart bracelet or headphones):

[0092] Vibration reminder: The smart bracelet vibrates slightly to remind Xiao Li to pay attention to his status.

[0093] Voice warning: The headset plays a prompt tone: "Please note that you may be feeling tired. Please take a break."

[0094] At the same time, the early warning system records the abnormal status and sends it to the supervisor so that further measures can be taken (such as arranging operator A to rest).

[0095] See also Figure 3 The second aspect of the present invention provides a monitoring method for factory production safety, comprising the following steps:

[0096] Step 1: Collect facial images of operators on the factory production line; extract facial feature images of operators from the facial images;

[0097] Step 2: Based on the facial feature image, obtain a number of analysis coordinates of the facial feature image; based on the analysis coordinates, analyze the operator's behavior parameters;

[0098] Step 3: Based on the behavioral parameters, analyze the operator's behavioral status and issue a safety warning to the operator.

[0099] Some of the data in the above formula are calculated by removing the dimensions and taking their numerical values. The formula is a formula that is closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0100] The above embodiments are only used to illustrate the technical method of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A monitoring system for factory production safety, characterized in that: Including analysis module and image acquisition module; Image acquisition module: used to collect facial images of operators on the factory production line; Analysis module: used for extracting facial feature images of the operator from the facial image; wherein the facial feature images include lip feature images and eye feature images; Based on the facial feature image, several analysis coordinates of the facial feature image are obtained; based on the analysis coordinates, the operator's behavior parameters are analyzed; based on the behavior parameters, the operator's behavior state is analyzed; wherein, the analysis coordinates include lip analysis coordinates and eye analysis coordinates, the behavior parameters include lip behavior parameters and eye behavior parameters, and the behavior state includes normal behavior state and abnormal behavior state.

2. A monitoring system for factory production safety according to claim 1, characterized in that: The step of extracting the operator's facial feature image from the facial image includes: Input the face image into the feature recognition model and output the face feature label; the face feature label is matched with the face feature image one by one; The construction process of the feature recognition model is as follows: extracting a plurality of facial images and a plurality of facial feature images from a factory database; Integrate several facial images and several facial feature labels into several groups of training data and test data; use the training data to train an artificial intelligence model; use the test data to test the trained artificial intelligence model, and adjust the artificial intelligence model according to the test results; finally obtain a feature recognition model with facial images as input and facial feature labels as output; wherein the artificial intelligence model is a BP neural network model or an RBF neural network model.

3. A monitoring system for factory production safety according to claim 2, characterized in that: The step of obtaining the analysis coordinates of the facial feature image includes: Establishing a coordinate system of the facial feature image with the center position of the facial feature image as the origin of the coordinate system and a number of pixels as one coordinate unit; and extracting the contour lines of the lip area and the eye area; Based on the coordinate system of the facial feature image and the contour lines of the lip area and the eye area, a number of analysis coordinates are determined; wherein the analysis coordinates include lip analysis coordinates and eye analysis coordinates.

4. A monitoring system for factory production safety according to claim 3, characterized in that: Calculating the lip behavior parameters based on the lip feature image includes: Acquire a number of lip analysis coordinates based on the lip feature image; wherein the lip analysis coordinates include the coordinates of the corner of the mouth, the center coordinates of the upper lip, and the center coordinates of the lower lip; The lip opening and closing angle θ is calculated using the following formula: Among them, the lip opening and closing angle value is the lip behavior parameter, (X1, Y1) is the coordinate of the mouth corner, (X2, Y2) is the center coordinate of the upper lip, (X3, Y3) is the center coordinate of the lower lip, γ is the opening and closing angle correction coefficient, γ∈(0,1]; Determine whether the opening and closing angle value is greater than a preset angle threshold; if yes, mark the opening and closing angle value as an abnormal angle; if not, mark the opening and closing angle as a normal angle.

5. The monitoring system for factory production safety according to claim 3, characterized in that: Calculating the lip behavior parameter based on the lip feature image further includes: Obtain the circumscribed rectangle of the lip, extract the coordinates of the four intersection points of the circumscribed rectangle, and calculate the area of the circumscribed rectangle of the lip using the rectangle area formula; determine whether the area of the circumscribed rectangle is greater than the preset area threshold; if so, mark the opening and closing angle value as an abnormal angle; otherwise, mark the opening and closing angle value as a normal angle.

6. A monitoring system for factory production safety according to claim 4 or 5, characterized in that: Calculating the eye behavior parameter based on the eye feature image includes: Acquire a number of eye analysis coordinates based on the eye feature image; wherein the eye analysis coordinates include the center coordinates of the upper eyelid, the center coordinates of the lower eyelid, and the center coordinates of the pupil; Based on the center coordinate of the upper eyelid and the center coordinate of the lower eyelid, the longitudinal distance between the upper eyelid and the lower eyelid is calculated and marked as the eye distance value; determining whether the eye distance value is less than a preset distance threshold; if so, marking the eye distance value as abnormal eye distance; If not, the center coordinates of the pupil will be extracted to determine whether the center coordinates of the pupil are the same as the center coordinates of the eye; if yes, the pupil position will be marked as normal; if not, the pupil position will be marked as abnormal.

7. A monitoring system for factory production safety according to claim 6, characterized in that: The analysis of the operator's behavior status based on the behavior parameters includes: Set the video analysis cycle and divide the video analysis cycle into several processing frames; Determine whether the number of occurrences of abnormal angles or abnormal eye distances within a video analysis period exceeds a corresponding preset number threshold, and whether the duration of the occurrence interval is less than a corresponding preset interval duration threshold; if so, mark the lip state as abnormal lip state or mark the eye state as abnormal eye state; if not, mark the lip state as normal lip state or mark the eye state as normal eye state; Determine whether pupil position abnormality occurs continuously for a preset time within the video analysis period; if so, mark the eye state as abnormal; if not, mark the eye state as normal; When the operator has any of abnormal lip state, abnormal eye state or abnormal pupil position, the operator's behavior state is marked as abnormal behavior state; otherwise, the operator's behavior state is marked as normal behavior state.

8. The monitoring system for factory production safety according to claim 7, characterized in that: The preset angle threshold and the preset distance threshold are the lip opening and closing angle value and the eye distance value of each person in a normal behavior state.

9. The monitoring system for factory production safety according to claim 7, characterized in that: Also includes: Early warning module: used to warn the operator of abnormal behavior; the operator wears a communication device, and when abnormal behavior occurs, the communication device sends a warning signal to remind the operator.

10. A monitoring method for factory production safety, based on the monitoring system for factory production safety according to any one of claims 1 to 9, characterized in that: The following steps are involved: Collect facial images of operators on a factory production line; extract facial feature images of operators from the facial images; Based on the facial feature image, a number of analysis coordinates of the facial feature image are obtained; based on the analysis coordinates, behavioral parameters of the operator are analyzed; Based on behavioral parameters, the operator's behavioral status is analyzed and safety warnings are issued to the operator.

Images