Intelligent monitoring operation and maintenance system for road traffic tourism signboard

Abstract

The invention discloses an intelligent monitoring operation and maintenance system for a road traffic tourism signboard, relates to the technical field of intelligent monitoring, and solves the problem that the traffic tourism signboard is difficult to monitor and maintain according to video image data monitored and collected in real time. Analyzing a head posture change rate, a sight line direction change rate, a surface dust coverage degree value and a fading change rate by utilizing a face recognition technology, an eyeball tracking algorithm and a target detection model; environmental influence coefficients are difficult to analyze according to environmental parameters; a residual service life prediction model is difficult to construct to analyze the residual service life ratio; the performance index is difficult to comprehensively analyze according to the guiding effect of the signboard, the dust covering degree value of the surface, the color fading change rate, the environmental influence and the remaining service life, and operation and maintenance are difficult to carry out according to the performance index. According to the method, the guiding effect, the surface dust covering degree value, the color fading change rate, the environment influence coefficient and the remaining service life ratio of the signboard are monitored in real time, and operation and maintenance are conducted by analyzing the performance index.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent monitoring, and specifically relates to an intelligent monitoring and operation and maintenance system for road traffic tourism signs. Background Art

[0002] With the growth of road traffic and tourism travel demands, the traditional manual inspection method has low efficiency and strong subjectivity, making it difficult to meet the demands. Against this background, an intelligent monitoring and operation and maintenance system for road traffic tourism signs that integrates advanced technologies such as the Internet of Things, big data, and artificial intelligence has emerged. By deploying various sensors to collect various data in real time, it analyzes the guiding effect of the effective attention area of the signs, the degree of surface dust coverage, the fading change situation, etc. By comprehensively evaluating the performance of the signs, it provides a scientific basis for operation and maintenance decisions. It can achieve real-time monitoring, accurate analysis, and efficient operation and maintenance of the signs, improving management efficiency and service quality.

[0003] The following problems exist in the prior art: It is difficult to analyze the head pose change rate, line-of-sight direction change rate, surface dust coverage value, and fading change rate using face recognition technology, eye tracking algorithms, and object detection models based on the video image data collected by real-time monitoring; it is difficult to analyze the environmental impact coefficient based on environmental parameters; it is difficult to construct a remaining service life prediction model to analyze the remaining service life ratio; it is difficult to comprehensively analyze the performance index based on the guiding effect of the signs, the surface dust coverage value and fading change rate, environmental impact, and remaining service life and perform operation and maintenance accordingly. Summary of the Invention

[0004] To solve the problems existing in the above prior art, the first aspect of the present invention provides an intelligent monitoring and operation and maintenance system for road traffic tourism signs, including the following modules:

[0005] Data collection module: Monitor and collect the physical parameters, environmental parameters, and video image data of road traffic tourism signs in real time; using Internet of Things technology, transmit the collected physical parameters, environmental parameters, and video image data to the data processing module in real time through wireless communication;

[0006] Data processing module: Responsible for receiving the transmitted physical parameters, environmental parameters, and video image data, and performing data cleaning, denoising, and standardization on the physical parameters and environmental parameters, and performing filtering denoising and image enhancement on the video image data;

[0007] Data analysis module: According to the preprocessed video image data, using face recognition technology and eye tracking algorithms, analyze and calculate the head pose change rate and the line-of-sight direction change rate; use the object detection model to automatically identify the categories and positions of road traffic tourism signs and analyze and calculate the guiding effect of the effective attention area of road traffic tourism signs; analyze the dust coverage value and fading change rate on the surface of road traffic tourism signs according to the automatic recognition results of the object detection model; analyze and calculate the environmental impact coefficient of road traffic tourism signs according to environmental parameters; use machine learning to construct a remaining service life prediction model and calculate the remaining service life ratio of road traffic tourism signs;

[0008] Operation and maintenance management module: Comprehensively evaluate the performance of road traffic tourism signs according to the guiding effect of the effective attention area, the dust coverage value, the fading change rate, the environmental impact coefficient, and the remaining service life ratio.

[0009] Preferably, the physical parameters, environmental parameters, and video image data of road traffic tourism signs are monitored and collected in real time, including the following steps: The physical parameters include: tilt angle, wind loads on the sign and the column, vibration frequency; The environmental parameters include: environmental temperature, environmental humidity, wind force, rainfall, snowfall; The video image data is road monitoring video images.

[0010] Preferably, according to the preprocessed video image data, using face recognition technology and eye tracking algorithms, analyze and calculate the head pose change rate and the line-of-sight direction change rate, including the following steps:

[0011] Use a multi-task cascaded convolutional neural network as the deep learning model for face detection. By collecting a historical video image dataset containing faces, use bounding boxes to label the positions of faces in the video image data, and input the labeled historical video image dataset into the multi-task cascaded convolutional neural network for training;

[0012] By monitoring road traffic tourism signs in real time, input the collected and preprocessed video image data into the trained multi-task cascaded convolutional neural network to identify the face position and output the face bounding box coordinates (x, y, w, h); where, (x, y) represents the coordinate position of the upper left corner of the face bounding box; w represents the width of the face bounding box; h represents the height of the face bounding box;

[0013] According to the face bounding box output by the multi-task cascaded convolutional neural network, the 68-point face landmark model in the Dlib library for facial landmark detection is used to obtain the facial landmarks including eyebrows, eyes, nose, and mouth; based on the obtained landmarks and combined with the iris localization algorithm, the iris region is located within the face bounding box; the iris is detected by the method of ellipse fitting to obtain the iris center coordinates and iris radius, which are used as the reference points for the line-of-sight direction;

[0014] By calculating the Euler angles of the head, the following formula for calculating the head pose change rate is obtained:

[0015] where, H i represents the head pose change rate of the i-th road traffic tourism sign, and α t , β t and γ t are respectively the Euler angles of the head in consecutive frames, t is the frame number, and α 均 , β 均 and γ 均 represent the average values of the Euler angles of the head in consecutive frames;

[0016] By connecting the iris center coordinates with the eyeball center coordinates, a direction vector of the eyeball in the head coordinate system is obtained, and the eyeball direction vector is transformed from the head coordinate system to the world coordinate system. Combining the eyeball position and head pose, the line-of-sight vector is calculated through the eyeball-head linkage model; the line-of-sight direction change rate is obtained by calculating the average value of the modulus change of the line-of-sight vector in consecutive frames of the video image.

[0017] Preferably, the target detection model is used to automatically identify the category and position of the road traffic tourism sign and analyze and calculate the guiding effect of the effective attention area of the road traffic tourism sign, including the following steps:

[0018] Label the category and position of the road traffic tourism signs in the historical video image dataset. Among them, the categories of road traffic tourism signs include: prohibition signs, warning signs, indication signs, and guide signs; the position of the road traffic tourism signs is labeled by using bounding boxes; the labeled historical video image dataset is input into the target detection model for training; the video image data collected in real time and preprocessed is input into the trained target detection model to automatically identify the category and position of the road traffic tourism signs;

[0019] According to the identified position of the road traffic tourism sign, within the range of ±10% of the central area of the bounding box of the road traffic tourism sign, it is defined as the effective attention area of the road traffic tourism sign;

[0020] According to the four corner point coordinates of the \(i\)th road traffic tourism signboard in the video image, fit the plane where the signboard is located; calculate the normal direction of the \(i\)th road traffic tourism signboard through the plane equation;

[0021] The formula for calculating the angle between the line of sight vector and the normal vector of the \(i\)th road traffic tourism signboard is:

[0022] where \(g\) i · \(f\) i represents the dot product of the line of sight vector and the normal vector of the \(i\)th road traffic tourism signboard, and \(\vert\vert g\vert\vert\) i and \(\vert\vert f\vert\vert\) i represent the magnitudes of the line of sight vector and the normal vector respectively;

[0023] When the angle is lower than the threshold of 15 degrees, it is determined that the extension line of the line of sight passes through the effective attention area of the road traffic tourism signboard; the residence time of the line of sight direction in the \(i\)th road traffic tourism signboard in the continuous frames is obtained by dividing the number of frames in which the effective attention area of the \(i\)th road traffic tourism signboard is gazed by the frame rate of the video image;

[0024] The formula for evaluating the guiding effect of the effective attention area of the \(i\)th road traffic tourism signboard is:

[0025]

[0026] where \(\theta\) i and \(\theta_0\) represent the angle and the angle threshold respectively, \(T\) i represents the residence time of the effective attention area of the \(i\)th road traffic tourism signboard, \(H\) i represents the head pose change rate of the effective attention area of the \(i\)th road traffic tourism signboard, \(H\) max represents the maximum value of the head pose change rate, \(V\) i is the line of sight direction change rate, and \(w_1\), \(w_2\), \(w_3\) and \(w_4\) represent the adjustment coefficients, which are 0.4, 0.3, 0.2 and 0.1 respectively.

[0027] Preferably, according to the automatic recognition result of the target detection model, analyze the dust coverage value and the fading change rate on the surface of the road traffic tourism signboard, including the following steps:

[0028] Automatically identify the position of the i-th road traffic tourism sign in the video image according to the object detection model, crop the surface image of the i-th road traffic tourism sign and grayscale it; by dividing the surface image of the i-th road traffic tourism sign into n small regions, calculate the texture feature values of the local binary pattern for each small region to obtain the texture feature vector of each small region; compare the texture feature vector of each region with the texture feature vector of the clean region to determine whether there is dust coverage on the surface image of the i-th road traffic tourism sign;

[0029] When there is dust coverage, by calculating the average values of the local binary pattern feature vectors of all pixels in the dust-covered area and the non-dust-covered area respectively, obtain the texture feature value of the dust-covered area and the texture feature value of the non-dust-covered area; through a dust particle counter, real-time monitor and collect the air dust concentration;

[0030] Formula for calculating the dust coverage degree value of the i-th road traffic tourism sign:

[0031] Obtain the dust coverage degree value D of the i-th road traffic tourism sign i ; where A i represents the total area of the i-th road traffic tourism sign, A d,i is the texture feature value of the dust-covered area of the i-th road traffic tourism sign, A c,i is the texture feature value of the non-dust-covered area of the i-th road traffic tourism sign; C represents the air dust concentration;

[0032] Using image processing algorithms to extract features from the video image of the road traffic tourism sign includes: grayscale histogram, hue value, saturation and lightness. Calculate the formula for the fading change rate of the i-th road traffic tourism sign:

[0033]

[0034] Obtain the fading change rate S i ; where C h and O h are the average hue values of the current and initial video images respectively, C s and O s are the average saturations of the current and initial video images respectively, C h and O h are the average lightnesses of the current and initial video images respectively; C k and C o are the values of the grayscale histogram of the current and initial video images at the gray level k respectively.

[0035] Preferably, the environmental impact coefficient of the road traffic tourism sign is analyzed and calculated according to environmental parameters, including the following steps: according to the collected and preprocessed environmental parameters including environmental temperature, environmental humidity, wind force, rainfall, and snowfall, the environmental impact coefficient of the road traffic tourism sign is calculated by weighted average.

[0036] Preferably, a remaining service life prediction model is constructed using machine learning and the remaining service life ratio of the road traffic tourism sign is calculated, including the following steps:

[0037] According to the physical parameters and environmental impact coefficient of the road traffic tourism sign, a remaining service life prediction model of the road traffic tourism sign is constructed using a regression model in machine learning;

[0038] Collect the physical parameters and environmental impact coefficients of historical road traffic tourism signs as the training data set of the remaining service life prediction model; input the training data set into the remaining service life prediction model for training;

[0039] By collecting the current physical parameters and environmental impact coefficients of historical road traffic tourism signs in real time and inputting them into the trained remaining service life prediction model; according to the input physical parameters and environmental impact coefficients, the remaining service life duration of the i-th road traffic tourism sign is output in real time; according to the predicted remaining service life duration of the i-th road traffic tourism sign divided by the standard service life duration, the remaining service life ratio is obtained.

[0040] Preferably, the performance of the road traffic tourism sign is comprehensively evaluated according to the guiding effect of the effective attention area, the dust coverage degree value, the fading change rate, the environmental impact coefficient, and the remaining service life ratio, including the following steps:

[0041] Calculate the performance index formula of the i-th road traffic tourism sign:

[0042]

[0043] Obtain the performance index P of the i-th road traffic tourism sign i ; where E i 、D i 、S i 、M i and H i respectively represent the guiding effect of the effective attention area, the dust coverage degree value, the fading change rate, the remaining service life ratio, and the environmental impact coefficient of the i-th road traffic tourism sign; ε i is the material strength of the i-th road traffic tourism sign;

[0044] When the performance index of the i-th road traffic tourism sign is lower than the threshold of 0.5, perform operation and maintenance processing on the i-th road traffic tourism sign; otherwise, continue to monitor the road traffic tourism sign in real time.

[0045] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0046] Through the data collection module, the present invention uses the Internet of Things technology to monitor and collect the physical parameters, environmental parameters, and video image data of road traffic tourism signs in real time, achieving a comprehensive and real-time grasp of the sign status and environment.

[0047] Through the data analysis module, the present invention uses technologies such as face recognition, eye tracking, object detection, and machine learning to achieve a comprehensive analysis of the guiding effect, surface condition, environmental impact, and remaining service life of road traffic tourism signs, providing a multi-dimensional basis for the performance evaluation of signs.

[0048] Through the operation and maintenance management module, the present invention comprehensively considers various evaluation indicators, scientifically quantifies the performance of road traffic tourism signs, and formulates reasonable operation and maintenance strategies according to the evaluation results, achieving refined management of signs. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0050] Figure 1 It is a system module diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0051] The following will clearly and completely describe the technical solutions of the present invention in conjunction with the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0052] Please refer to Figure 1 , an embodiment of the first aspect of the present invention provides an intelligent monitoring and operation and maintenance system for road traffic tourism signs, including the following modules:

[0053] Data collection module: Monitor and collect the physical parameters, environmental parameters, and video image data of road traffic tourism signs in real time; Using Internet of Things technology, transmit the collected physical parameters, environmental parameters, and video image data to the data processing module in real time through wireless communication;

[0054] Data processing module: Responsible for receiving the transmitted physical parameters, environmental parameters, and video image data, and performing data cleaning, denoising, and standardization on the physical parameters and environmental parameters, and performing filtering denoising and image enhancement on the video image data;

[0055] Data analysis module: According to the preprocessed video image data, use face recognition technology and eye tracking algorithm to analyze and calculate the head pose change rate and gaze direction change rate; Use the object detection model to automatically identify the category and location of road traffic tourism signs and analyze and calculate the guiding effect of the effective attention area of road traffic tourism signs; Analyze the dust coverage value and fading change rate of the surface of road traffic tourism signs according to the automatic recognition results of the object detection model; Analyze and calculate the environmental impact coefficient of road traffic tourism signs according to environmental parameters; Use machine learning to build a remaining service life prediction model and calculate the remaining service life ratio of road traffic tourism signs;

[0056] Operation and maintenance management module: Comprehensively evaluate the performance of road traffic tourism signs according to the guiding effect of the effective attention area, dust coverage value, fading change rate, environmental impact coefficient, and remaining service life ratio.

[0057] Specifically, install sensors on the road traffic tourism signboards, such as temperature and humidity sensors, vibration sensors, etc., to monitor the physical parameters and environmental parameters of the signboards in real time. Install cameras near the signboards to ensure that the images of the signboards can be clearly captured to obtain the best road monitoring video image data. Configure wireless communication devices, such as 4G / 5G modules, Wi-Fi, etc., to ensure that the collected data can be transmitted to the data processing module in real time. Use sensors to monitor the physical parameters and environmental parameters of the signboards in real time. Utilize the Internet of Things technology to transmit the collected physical parameters, environmental parameters, and video image data to the data processing module in real time through wireless communication. The data processing module receives the transmitted physical parameters and environmental parameters and performs data cleaning to remove outliers and incorrect data; at the same time, perform denoising processing to improve the accuracy and reliability of the data. Perform filtering denoising and image enhancement processing on the video image data to improve the clarity and quality of the images for better subsequent object detection and analysis. Use face recognition technology and eye tracking algorithms to analyze the people in the video images and calculate the head pose change rate and the line-of-sight direction change rate. Apply an object detection model to automatically identify the categories and positions of the road traffic tourism signboards in the video images, providing a basis for subsequent guidance effect analysis. According to the categories and positions of the signboards, combined with the head pose and the line-of-sight direction change rate, analyze and calculate the guidance effect of the effective attention area of the signboards to determine whether the signboards can effectively guide the traffic flow and tourism activities. Through the analysis of the video images, calculate the dust coverage value and the fading change rate on the surface of the signboards to evaluate the appearance condition and maintenance requirements of the signboards. According to the environmental parameters, analyze and calculate the environmental impact coefficient of the signboards to evaluate the impact of environmental factors on the performance and service life of the signboards. Use machine learning algorithms, combined with the physical parameters and environmental parameters of the signboards, to construct a remaining service life prediction model and calculate the remaining service life ratio of the signboards, providing a scientific basis for the maintenance and replacement of the signboards. According to indicators such as the guidance effect of the effective attention area, the dust coverage value, the fading change rate, the environmental impact coefficient, and the remaining service life ratio, comprehensively evaluate the performance of the road traffic tourism signboards to obtain a performance evaluation result. According to the performance evaluation result, realize the real-time monitoring and operation and maintenance of the road traffic tourism signboards. The operation and maintenance decisions include: regularly cleaning the surface of the signboards, timely repairing the faded parts, adjusting the position or angle of the signboards, etc., to ensure the normal operation of the signboards and good guidance effects.

[0058] In this embodiment, the steps for real-time monitoring and collecting the physical parameters, environmental parameters, and video image data of the road traffic tourism signboards are as follows: The physical parameters include: tilt angle, wind loads on the signboard and the column, vibration frequency; The environmental parameters include: environmental temperature, environmental humidity, wind force, rainfall, snowfall; The video image data is the road monitoring video image.

[0059] Specifically, select a high-precision inclination sensor, such as an electronic level or an accelerometer-type inclination sensor, and install it on the column of the signboard or the back of the signboard to ensure that the inclination angle of the signboard can be accurately measured. When installing, pay attention to the horizontality and firm fixation of the sensor to avoid measurement errors caused by improper installation. Use a wind speed and direction sensor to indirectly measure the wind load on the signboard and the column. Install the sensor in an open and unobstructed area near the signboard to ensure that the wind speed and direction can be accurately measured, and then calculate the wind load based on the wind speed and the windward area of the signboard. The installation height is generally about 2 meters above the ground, and the sensor should be kept horizontal. Select a vibration acceleration sensor and install it on the column of the signboard or the back of the signboard to measure the vibration frequency. The installation position should be selected at a more sensitive part on the vibration transmission path, such as the middle of the column or the four corners of the signboard. Use special installation accessories to firmly fix the sensor to ensure good contact between the sensor and the object to be measured. Select a temperature and humidity integrated sensor, such as a capacitive temperature and humidity sensor, and install it near the signboard to ensure that the ambient temperature and humidity can be accurately measured. Similar to the wind speed sensor in the wind load sensor, it is installed in an open area near the signboard, with a height of about 2 meters above the ground, for directly measuring the wind force. The rain sensor can use a tipping bucket rain gauge, and the snow depth sensor can select an ultrasonic snow depth sensor. The rain gauge is installed on the ground near the signboard to ensure that the rain gauge is in a horizontal state; the ultrasonic snow depth sensor is installed at a certain height above the ground at a position near the signboard that is not easily covered by snow, and measures the snow depth through the principle of ultrasonic reflection. Select a high-definition road monitoring camera with a resolution of 1080P or higher, and with functions such as low light and wide dynamic range to meet the video acquisition requirements under different lighting conditions. The camera is installed at a suitable position near the signboard to ensure that the monitoring video images of the signboard and the surrounding road can be clearly captured. Its installation height is generally about 2.5 to 3 meters above the ground, and the angle can be adjusted according to the actual position of the signboard and the road conditions to make the signboard centered and completely displayed in the picture.

[0060] In this embodiment, according to the preprocessed video image data, using face recognition technology and eye tracking algorithms, analyze and calculate the head pose change rate and the gaze direction change rate, including the following steps:

[0061] Use a multi-task cascaded convolutional neural network as the deep learning model for face detection. By collecting a historical video image data set containing faces, use bounding boxes to label the positions of faces in the video image data, and input the labeled historical video image data set into the multi-task cascaded convolutional neural network for training;

[0062] By real-time monitoring of road traffic tourism signboards, the collected and preprocessed video image data is input into the trained multi-task cascaded convolutional neural network to identify the face position and output the face bounding box coordinates (x, y, w, h); where (x, y) represents the coordinate position of the upper left corner of the face bounding box; w represents the width of the face bounding box; h represents the height of the face bounding box.

[0063] According to the face bounding box output by the multi-task cascaded convolutional neural network, use the 68-point face landmark model in the Dlib library for facial landmark detection to obtain the face landmarks including eyebrows, eyes, nose, and mouth; based on the obtained landmarks and combined with the iris localization algorithm, locate the iris area in the face bounding box; use the method of ellipse fitting to detect the iris to obtain the iris center coordinates and iris radius, which are used as the reference points for the line-of-sight direction.

[0064] By calculating the Euler angles of the head, the head pose change rate formula is calculated as follows:

[0065] where H i represents the head pose change rate of the i-th road traffic tourism signboard, and α t , β t and γ t are the Euler angles of the head in consecutive frames respectively, t is the frame number, and α 均 , β 均 and γ 均 represent the average values of the Euler angles of the head in consecutive frames.

[0066] By connecting the iris center coordinates with the eyeball center coordinates, a direction vector of the eyeball in the head coordinate system is obtained, and the eyeball direction vector is transformed from the head coordinate system to the world coordinate system. Combining the eyeball position and head pose, the line-of-sight vector is calculated through the eyeball-head linkage model; the line-of-sight direction change rate is obtained by calculating the average value of the modulus change of the line-of-sight vector in consecutive frames of the video image.

[0067] Specifically, collect video image data of various human faces under different scenarios, lighting conditions, and angles to ensure the diversity and representativeness of the dataset. The dataset should include human face images of different genders, ages, and ethnicities, as well as human face images with different expressions, postures, and occlusion situations. Use annotation tools such as LabelImg to annotate the boundaries of human faces in the collected video image data, mark the positions of human faces in the images, generate corresponding annotation files, and the annotation files contain the coordinate information of the human face bounding boxes (such as the upper left corner coordinates (x, y), width w, and height h). Input the annotated historical video image dataset into a multi-task cascaded convolutional neural network for training. During the training process, adjust the hyperparameters of the network such as the learning rate and batch size, and use data augmentation techniques such as rotation, flipping, and brightness adjustment to improve the generalization ability and robustness of the model until the model achieves satisfactory human face detection results on the validation set. Through the cameras near the road traffic tourism signs, collect and preprocess video image data in real time. The preprocessing includes operations such as filtering and denoising the images and image enhancement to improve the image quality for subsequent human face detection. Input the preprocessed video image data into the trained model, and the model will automatically identify the positions of human faces in the images and output the human face bounding box coordinates (x, y, w, h). Use the 68-point face landmark model in the Dlib library to detect 68 facial landmark points in the human face area according to the human face bounding box output by MTCNN, including the landmark points of parts such as eyebrows, eyes, nose, and mouth. Through these landmark points, the various feature parts of the human face can be accurately located, providing a basis for subsequent iris localization and head pose calculation. Based on the facial landmark point detection results, focus on locating the landmark points in the eye area, and combine with the iris localization algorithm to further locate the iris area in the human face bounding box. Methods such as those based on gray-scale features, edge detection, or template matching can be used to determine the approximate position of the iris. Use the method of ellipse fitting to detect the iris area, and obtain the center coordinates and radius of the iris. The ellipse fitting can be achieved through the least squares method or other optimization algorithms to fit the ellipse closest to the iris boundary, and its center is the iris center coordinates, and the radius is used to determine the size and shape of the iris. The iris center coordinates are used as the reference points for the line of sight direction. According to the coordinate information of the facial landmark points, calculate the Euler angles of the head, including the pitch angle, yaw angle, and roll angle. The Euler angles reflect the rotation state of the head in three-dimensional space and can be calculated through the geometric relationship and trigonometric functions between the facial landmark points. For example, calculate the pitch angle by calculating the angle between the line connecting the tip of the nose and the center point of the face and the horizontal plane, and calculate the yaw angle by calculating the angle between the center line of the face and the symmetry axis of the front face, etc. For each sign, obtain the head Euler angle data of consecutive frames in its corresponding video image sequence. Calculate the mean value of each Euler angle, including the pitch angle, yaw angle, and roll angle, in the consecutive frames.According to the given formula for the head pose change rate, subtract the Euler angles of consecutive frames from their mean value and substitute them to calculate the head pose change rate. Connect the iris center coordinates and the eyeball center coordinates to obtain a direction vector of the eyeball in the head coordinate system. Then, through coordinate transformation, transform this eyeball direction vector from the head coordinate system to the world coordinate system. Coordinate transformation needs to consider parameters such as the head pose and position, as well as the internal and external parameters of the camera, and a projection matrix or a rotation and translation matrix can be used for transformation. Combining the eyeball position and the head pose, using the eyeball-head linkage model, calculate the line-of-sight vector according to the transformed eyeball direction vector. In consecutive frames of the video image, calculate the line-of-sight vector of each frame, and then calculate the norm of the change in the line-of-sight vector between adjacent frames. Average these norms to obtain the line-of-sight direction change rate.

[0068] In this embodiment, the target detection model is used to automatically identify the category and position of the road traffic tourism sign and analyze and calculate the guiding effect of the effective attention area of the road traffic tourism sign, including the following steps:

[0069] Annotate the category and position of the road traffic tourism signs in the historical video image dataset. Among them, the categories of road traffic tourism signs include: prohibition signs, warning signs, indication signs, and guide signs; the position of the road traffic tourism signs is annotated using bounding boxes; input the annotated historical video image dataset into the target detection model for training; input the video image data collected in real time and preprocessed into the trained target detection model to automatically identify the category and position of the road traffic tourism signs;

[0070] According to the identified position of the road traffic tourism sign, within the range of ±10% of the center area of the bounding box of the road traffic tourism sign, it is defined as the effective attention area of the road traffic tourism sign;

[0071] According to the four corner coordinates of the i-th road traffic tourism sign in the video image, fit the plane where the sign is located; calculate the normal direction of the i-th road traffic tourism sign through the plane equation;

[0072] The formula for calculating the angle between the line-of-sight vector of the i-th road traffic tourism sign and the normal vector of the road traffic tourism sign:

[0073] Among them, g i ·f i represents the dot product of the line-of-sight vector and the normal vector of the i-th road traffic tourism sign, ||g i || and ||f i || represent the norms of the line-of-sight vector and the normal vector respectively;

[0074] When the included angle is lower than the threshold value of 15 degrees, it is determined that the extended line of sight passes through the effective attention area of the road traffic tourism sign; the residence time of the line of sight in the effective attention area of the i-th road traffic tourism sign in consecutive frames is obtained by calculating the number of frames in which the effective attention area of the i-th road traffic tourism sign is gazed at divided by the frame rate of the video image;

[0075] The guiding effect evaluation formula for the effective attention area of the i-th road traffic tourism sign is calculated as follows:

[0076]

[0077] where θ i and θ0 represent the included angle and the included angle threshold respectively, T i represents the residence time of the effective attention area of the i-th road traffic tourism sign, H i represents the head pose change rate of the effective attention area of the i-th road traffic tourism sign, H max represents the maximum value of the head pose change rate, V i is the line of sight direction change rate, and w1, w2, w3, and w4 represent the adjustment coefficients, which are 0.4, 0.3, 0.2, and 0.1 respectively.

[0078] Specifically, collect a historical video image dataset containing various road traffic and tourism signboards, ensuring that the categories of signboards in the dataset include, but are not limited to: prohibition signs, warning signs, indication signs, and guide signs. Use annotation tools to perform bounding box annotation on each signboard in the dataset, recording its category and location information. When annotating, precisely select the edges of the signboard to ensure the accuracy of the location information. Take Faster R-CNN as an example in the target detection model, configure its framework, and organize the annotated dataset in the format required by the model. Divide the training dataset into a training set and a validation set, usually in a ratio of 8:2. Use the training set to train the Faster R-CNN model, set appropriate hyperparameters such as the learning rate and the number of iterations, and use the validation set for model validation and adjustment during the training process until the model achieves a high detection accuracy and recall rate on the validation set. Real-time collect video images of the road traffic scene through a camera and perform preprocessing operations such as image enhancement and filtering denoising to improve the image quality and enhance the visibility of the signboards. Input the preprocessed video image frames into the trained Faster R-CNN model, and the model will automatically identify the categories and locations of the road traffic and tourism signboards in the image, outputting the bounding box coordinates and corresponding category labels of each signboard. For each identified road traffic and tourism signboard, define an effective attention area within a range of ±10% of the central area of its bounding box. Usually, the four corner coordinates of the identified road traffic and tourism signboard in the video image can be further accurately extracted through the bounding box of the signboard, or some corner detection algorithms such as Harris corner detection can be used, and the plane equation where the signboard is located can be fitted using the least squares method or other plane fitting algorithms. The general form of the plane equation is Ax + By + Cz + D = 0, where (A, B, C) is the normal direction vector of the plane. Normalize the obtained normal direction vector to make its modulus 1 for convenient subsequent angle calculation. Combine the line-of-sight vector information obtained by the eye tracking algorithm, and this line-of-sight vector is represented in the world coordinate system. Calculate the angle between the line-of-sight vector and the signboard normal vector according to the formula; if the calculated angle is lower than the threshold of 15 degrees, it is determined that the current line of sight extension passes through the effective attention area of the road traffic and tourism signboard; the threshold is dynamically adjusted according to the actual situation and historical data. In consecutive frames of the video image, count the number of frames in which the line of sight direction falls within the effective attention area of the i-th road traffic and tourism signboard, and then divide the number of frames by the frame rate of the video image to obtain the residence time of the line of sight in the effective attention area of the signboard in consecutive frames. Obtain parameters such as the angle and its threshold, residence time, head pose change rate, the maximum value of the head pose change rate, and line-of-sight direction change rate, and substitute them into the formula to calculate the guiding effect evaluation value of the effective attention area of the i-th road traffic and tourism signboard.

[0079] In this embodiment, analyzing the dust coverage value and fading change rate of the surface of road traffic tourism signs according to the automatic recognition results of the target detection model includes the following steps:

[0080] According to the position of the i-th road traffic tourism sign automatically recognized by the target detection model in the video image, the surface image of the i-th road traffic tourism sign is intercepted and grayscaled; by dividing the surface image of the i-th road traffic tourism sign into n small regions, the texture feature values of the local binary pattern are calculated for each small region to obtain the texture feature vector of each small region; the texture feature vector of each region is compared with the texture feature vector of the clean region to determine whether there is dust coverage in the surface image of the i-th road traffic tourism sign.

[0081] When there is dust coverage, by calculating the average values of the local binary pattern feature vectors of all pixels in the dust-covered area and the non-dust-covered area respectively, the texture feature value of the dust-covered area and the texture feature value of the non-dust-covered area are obtained; the air dust concentration is monitored and collected in real time by a dust particle counter.

[0082] The formula for calculating the dust coverage value of the i-th road traffic tourism sign:

[0083]

[0084] Obtain the dust coverage value D of the i-th road traffic tourism sign i ; where A i represents the total area of the i-th road traffic tourism sign, A d,i is the texture feature value of the dust-covered area of the i-th road traffic tourism sign, A c,i is the texture feature value of the non-dust-covered area of the i-th road traffic tourism sign; C represents the air dust concentration;

[0085] Using image processing algorithms to extract features from the video image of the road traffic tourism sign includes: grayscale histogram, hue value, saturation, and lightness. The formula for calculating the fading change rate of the i-th road traffic tourism sign:

[0086]

[0087] Obtain the fading change rate S i ; where C h and O h are the average hue values of the current and initial video images respectively, C s and O s are the average saturations of the current and initial video images respectively, C h and O hare the average brightness of the current and initial video images, respectively; C k and C o are the values of the gray - level histograms of the current and initial video images at gray - level k, respectively.

[0088] Specifically, according to the position information of the i - th road traffic tourism signboard recognized by the target detection model, the surface image of the signboard is intercepted from the video image. The intercepted surface image of the signboard is grayscale - processed to convert the color image into a grayscale image, so as to reduce the data volume and computational complexity while retaining the main features of the image. The grayscale - processed surface image of the signboard is divided into n small regions. For example, it can be divided in a grid form, and the size of each small region can be determined according to the size of the signboard and the image resolution. The texture feature value of the local binary pattern (LBP) is calculated for each small region to obtain the texture feature vector of each small region. This vector contains statistical information such as the LBP histogram of the region. The texture feature vector of each region is compared with the texture feature vector of the clean region. The texture feature vector of the clean region can be obtained by performing the same texture feature extraction process on the clean surface image of the signboard. If the difference between the texture feature vector of a certain region and the texture feature vector of the clean region exceeds a certain threshold, it is determined that there is dust coverage in this region. The difference can be measured by calculating the distance between the feature vectors, such as the Euclidean distance, Manhattan distance, etc. The threshold can be taken as the mean plus the standard deviation of the difference statistical analysis as a preliminary threshold and adjusted according to the actual situation and historical data. When there is dust coverage, the average values of the local binary pattern feature vectors of all pixels in the dust - covered region and the non - dust - covered region are calculated respectively to obtain the texture feature value of the dust - covered region and the texture feature value of the non - dust - covered region. The dust concentration in the air is monitored in real - time through a dust particle counter. The dust particle counter can be installed near the road traffic tourism signboard to obtain the dust concentration data related to the environment where the signboard is located. The calculated dust coverage degree value is used as an index to evaluate the cleanliness of the signboard surface. The larger the value, the more serious the dust coverage. Image - processing algorithms are used to extract features from the video image of the road traffic tourism signboard, including the gray - level histogram, hue value, saturation, and brightness. The gray - level histogram reflects the distribution of pixels with different gray - levels in the image and can be obtained by counting the frequency of each gray - level. The hue value, saturation, and brightness are the color - space features of the image, and these features can be obtained by converting the image from the RGB color space to the HSV color space. These feature values are substituted into the fading change rate formula to calculate the fading change rate.

[0089] In this embodiment, calculating the environmental impact coefficient of a road traffic tourism sign according to environmental parameters includes the following steps: According to the collected and preprocessed environmental parameters including: environmental temperature, environmental humidity, wind force, rainfall, and snowfall, the environmental impact coefficient of the road traffic tourism sign is calculated by weighted average.

[0090] Specifically, environmental parameters such as environmental temperature, environmental humidity, wind force, rainfall, and snowfall are collected in real time through sensors deployed near the road traffic tourism sign. These sensors can include temperature sensors, humidity sensors, wind speed and direction sensors, rain gauges, and snow depth sensors, etc. The collected environmental parameter data is preprocessed, including data cleaning, denoising, and standardization. Data cleaning can remove obviously incorrect or abnormal data points; denoising can use filtering algorithms such as moving average filtering, Kalman filtering, etc. to reduce noise interference in the data; standardization can convert data with different dimensions and units to the same numerical range, such as between 0 and 1, for subsequent weighted average calculation. According to the degree of influence of environmental parameters on the road traffic tourism sign, the weight of each parameter is determined. Environmental temperature may have a greater impact on the thermal expansion, contraction, and aging of the sign, and a relatively high weight of 0.6 can be assigned; environmental humidity may have a greater impact on the corrosion and visibility of the sign, and a relatively high weight can also be assigned; wind force, rainfall, and snowfall have an impact on the structural stability and surface cleanliness of the sign, and the weights can be relatively low. The weights can be determined through expert evaluation or historical data analysis; in this implementation, the weight of environmental temperature is 0.3, the weight of environmental humidity is 0.25, the weight of wind force is 0.15, the weight of rainfall is 0.15, and the weight of snowfall is 0.15. The environmental impact coefficient is calculated by adding the products of the collected and preprocessed environmental parameters multiplied by their corresponding weights.

[0091] In this embodiment, using machine learning to construct a remaining service life prediction model and calculate the remaining service life ratio of a road traffic tourism sign includes the following steps:

[0092] According to the physical parameters and environmental impact coefficient of the road traffic tourism sign, using a regression model in machine learning to construct a remaining service life prediction model for the road traffic tourism sign;

[0093] Collect the physical parameters and environmental impact coefficients of historical road traffic tourism signs as the training data set for the remaining service life prediction model; input the training data set into the remaining service life prediction model for training;

[0094] By collecting the current physical parameters and environmental impact coefficients of historical road traffic tourism signs in real time and inputting them into the trained remaining service life prediction model; according to the input physical parameters and environmental impact coefficients, the remaining service life duration of the i-th road traffic tourism sign is output in real time; according to the predicted remaining service life duration of the i-th road traffic tourism sign divided by the standard service life duration, the remaining service life ratio is obtained.

[0095] Specifically, collect the physical parameters and environmental impact coefficient data of historical road traffic tourism signboards as the training data set for the remaining service life prediction model. The physical parameters include the tilt angle, wind loads on the signboard and the column, vibration frequency, etc.; the environmental impact coefficient is obtained by weighted average calculation based on environmental parameters. At the same time, collect the actual service life data of these signboards, that is, the time length from installation to the need for replacement due to damage, aging, etc., as a reference for the target variable of the training model, namely the remaining service life duration. Clean the collected data, remove missing values, outliers and noise data, and perform standardization or normalization processing on the data to convert data with different dimensions and magnitudes to the same numerical range, improving the training effect of the model. Select a suitable machine learning regression model, such as linear regression, decision tree regression, random forest regression, support vector regression (SVR), or neural network, etc. Taking random forest regression as an example, it is an ensemble learning method that can handle complex relationships between multiple variables and has high prediction accuracy and robustness. Divide the preprocessed training data set into a training set and a test set, usually in a ratio of 8:2 or 7:3. Use the training set to train the random forest regression model, and optimize the model performance by adjusting the hyperparameters of the model, such as the number of trees, the depth of the trees, etc. During the training process, use the physical parameters and environmental impact coefficient as input features, and the actual service life of the signboard as the target variable, allowing the model to learn the mapping relationship between the input features and the remaining service life. Use the test set to evaluate the trained model, and calculate the prediction error of the model, such as mean square error, root mean square error, mean absolute error, etc. According to the evaluation results, if the prediction error of the model is large, the hyperparameters of the model can be further adjusted, or other types of regression models can be tried to improve the prediction accuracy of the model. Through a real-time monitoring system, collect the physical parameters of the current road traffic tourism signboards, such as the tilt angle, wind load, vibration frequency, etc., and the environmental impact coefficient is calculated based on the real-time environmental parameters. Perform the same preprocessing operations on the real-time collected data as the training data, and then input them into the trained remaining service life prediction model. The model outputs the predicted value of the remaining service life duration of the i-th road traffic tourism signboard according to the input real-time physical parameters and environmental impact coefficient. Calculate the remaining service life ratio based on the predicted remaining service life duration and the standard service life duration. The standard service life duration refers to the service life designed for the signboard under normal environmental conditions, which is determined according to the actual situation.

[0096] In this embodiment, comprehensively evaluate the performance of the road traffic tourism signboard according to the guiding effect of the effective attention area, the dust coverage value, the fading change rate, the environmental impact coefficient and the remaining service life ratio, including the following steps:

[0097] The formula for calculating the performance index of the i-th road traffic tourism signboard:

[0098]

[0099] Obtain the performance index P of the i-th road traffic tourism sign i ; where E i , D i , S i , M i and H i respectively represent the guiding effect of the effective attention area, the dust coverage value, the fading change rate, the remaining service life ratio, and the environmental impact coefficient of the i-th road traffic tourism sign; ε i is the material strength of the i-th road traffic tourism sign;

[0100] When the performance index of the i-th road traffic tourism sign is lower than the threshold value of 0.5, perform operation and maintenance processing on the i-th road traffic tourism sign, otherwise continue to monitor the road traffic tourism sign in real time.

[0101] Specifically, according to the real-time monitoring of the i-th road traffic tourism sign, comprehensively evaluate the performance of the road traffic tourism sign by analyzing and calculating the guiding effect of the effective attention area, the dust coverage value, the fading change rate, the environmental impact coefficient, and the remaining service life ratio, and obtain the performance index of the i-th road traffic tourism sign by substituting into the formula. Among them, the material strength of the i-th road traffic tourism sign is obtained according to the regulations of the national or regional transportation department, the material specification provided by the manufacturer, or laboratory tests. According to the actual situation and experience, formulate the evaluation criteria for the performance index. A performance index lower than 0.5 indicates poor performance and requires timely operation and maintenance processing. For each road traffic tourism sign, judge according to the calculated performance index according to the evaluation criteria. If the performance index is lower than the threshold value of 0.5, trigger the operation and maintenance processing flow, and perform corresponding operation and maintenance operations on the sign, including: cleaning the surface of the sign, repairing the faded part, strengthening the structure, replacing damaged components, etc. If the performance index is higher than or equal to 0.5, continue to monitor the sign in real time, regularly collect data and recalculate the performance index to timely grasp the performance change of the sign.

[0102] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An intelligent monitoring and operation and maintenance system for road traffic tourism signboards, characterized in that, It includes the following modules: Data collection module: Real-time monitor and collect the physical parameters, environmental parameters and video image data of road traffic tourism signs; Using Internet of Things technology, transmit the collected physical parameters, environmental parameters and video image data to the data processing module in real time through wireless communication; Data processing module: Responsible for receiving the transmitted physical parameters, environmental parameters and video image data, and performing data cleaning, denoising and standardization on the physical parameters and environmental parameters, and performing filtering denoising and image enhancement on the video image data; Data analysis module: According to the preprocessed video image data, use face recognition technology and eye tracking algorithm to analyze and calculate the head pose change rate and line-of-sight direction change rate; Automatically identify the category and location of road traffic tourism signs using a target detection model and analyze and calculate the guiding effect of the effective attention area of road traffic tourism signs; Analyze the dust coverage value and fading change rate on the surface of road traffic tourism signs according to the automatic recognition results of the target detection model; Analyze and calculate the environmental impact coefficient of road traffic tourism signs according to environmental parameters; Use machine learning to construct a remaining service life prediction model and calculate the remaining service life ratio of road traffic tourism signs; Operation and maintenance management module: Comprehensively evaluate the performance of road traffic tourism signs according to the guiding effect of the effective attention area, dust coverage value, fading change rate, environmental impact coefficient and remaining service life ratio.

2. The intelligent monitoring, operation and maintenance system for a road traffic tourism signboard according to claim 1, wherein, Real-time monitor and collect the physical parameters, environmental parameters and video image data of road traffic tourism signs, including the following steps: The physical parameters include: tilt angle, wind load of sign and column, vibration frequency; The environmental parameters include: environmental temperature, environmental humidity, wind force, rainfall, snowfall; The video image data is road monitoring video images.

3. The intelligent monitoring, operation and maintenance system for road traffic tourism signboards according to claim 1, characterized in that, According to the preprocessed video image data, use face recognition technology and eye tracking algorithm to analyze and calculate the head pose change rate and line-of-sight direction change rate, including the following steps: Use a multi-task cascaded convolutional neural network as the deep learning model for face detection. By collecting a historical video image data set containing faces, use bounding boxes to label the positions of faces in the video image data, and input the labeled historical video image data set into the multi-task cascaded convolutional neural network for training; By real-time monitoring road traffic tourism signs, input the collected and preprocessed video image data into the trained multi-task cascaded convolutional neural network to identify the face position and output the face bounding box coordinates (x, y, w, h); Among them, (x, y) represents the coordinate position of the upper left corner of the face bounding box; w represents the width of the face bounding box; h represents the height of the face bounding box; According to the face bounding box output by the multi-task cascaded convolutional neural network, use the 68-point face landmark model in the Dlib library for facial landmark detection to obtain the landmarks of the face, including eyebrows, eyes, nose, and mouth; based on the obtained landmarks and combined with the iris localization algorithm, locate the iris region in the face bounding box; detect the iris using the method of ellipse fitting to obtain the iris center coordinates and iris radius, which are used as the reference points for the line of sight direction; By calculating the Euler angles of the head, the head pose change rate formula is calculated as follows: Among them, H i represents the head attitude change rate of the i-th road traffic tourism sign, and α t , β t and γ t are the Euler angles of the head in consecutive frames respectively, t is the frame number, and α 均 , β 均 and γ 均 represent the mean values of the Euler angles of the head in consecutive frames; Connect the iris center coordinates with the eyeball center coordinates to obtain a direction vector of the eyeball in the head coordinate system, and transform the eyeball direction vector from the head coordinate system to the world coordinate system. Combine the eyeball position and head pose, and calculate the line of sight vector through the eyeball-head linkage model; calculate the average value of the modulus of the change in the line of sight vector in consecutive frames of the video image to obtain the line of sight direction change rate.

4. The intelligent monitoring, operation and maintenance system for road traffic tourism signboards according to claim 3, wherein, Use the object detection model to automatically identify the category and position of road traffic tourism signs and analyze and calculate the guiding effect of the effective attention area of road traffic tourism signs, including the following steps: Label the category and position of road traffic tourism signs in the historical video image dataset. Among them, the categories of road traffic tourism signs include: prohibition signs, warning signs, indication signs, and guide signs; the position of road traffic tourism signs is labeled using bounding boxes; input the labeled historical video image dataset into the object detection model for training; input the preprocessed video image data collected in real time into the trained object detection model to automatically identify the category and position of road traffic tourism signs; According to the identified position of the road traffic tourism sign, within the range of ±10% of the central area of the bounding box of the road traffic tourism sign, it is defined as the effective attention area of the road traffic tourism sign; According to the four corner coordinates of the i-th road traffic tourism sign in the video image, fit the plane where the sign is located; calculate the normal direction of the i-th road traffic tourism sign through the plane equation; Formula for calculating the angle between the line-of-sight vector and the normal vector of the i-th road traffic tourism sign: Among them, g i ·f i represents the dot product of the line-of-sight vector and the normal vector of the i-th road traffic tourism sign, ||g i || and ‖f i ‖ respectively represent the magnitudes of the line-of-sight vector and the normal vector; When the included angle is lower than the threshold of 15 degrees, it is determined that the extension line of the line of sight passes through the effective attention area of the road traffic tourism sign; calculate the residence time of the line of sight in the i-th road traffic tourism sign in consecutive frames by dividing the number of frames in which the effective attention area of the i-th road traffic tourism sign is gazed at by the frame rate of the video image; Calculate the guiding effect evaluation formula for the effective attention area of the i-th road traffic tourism sign: Among them, θ i and θ0 represent the included angle and the included angle threshold respectively, T i represents the residence time of the effective attention area of the i-th road traffic tourism sign, H i represents the head pose change rate of the effective attention area of the i-th road traffic tourism sign, H max represents the maximum value of the head pose change rate, V i is the line-of-sight direction change rate, and w1, w2, w3, and w4 represent the adjustment coefficients, which are 0.4, 0.3, 0.2, and 0.1 respectively.

5. The intelligent monitoring, operation and maintenance system for road traffic tourism signboards according to claim 1, characterized in that Analyze the dust coverage value and fading change rate of the surface of road traffic tourism signs according to the automatic recognition results of the object detection model, including the following steps: Automatically identify the position of the i-th road traffic tourism sign in the video image according to the target detection model, crop the surface image of the i-th road traffic tourism sign and grayscale it; by dividing the surface image of the i-th road traffic tourism sign into n small regions, calculate the texture feature values of the local binary pattern for each small region to obtain the texture feature vector of each small region; compare the texture feature vector of each region with the texture feature vector of the clean region to determine whether there is dust coverage in the surface image of the i-th road traffic tourism sign; When there is dust coverage, calculate the average value of the local binary pattern feature vectors of all pixels in the dust-covered area and the non-dust-covered area respectively to obtain the texture feature value of the dust-covered area and the texture feature value of the non-dust-covered area; use a dust particle counter to monitor the air dust concentration in real time; Formula for calculating the dust coverage value of the i-th road traffic tourism signboard: Obtain the dust coverage value D of the i-th road traffic tourism sign i ; where A i represents the total area of the i-th road traffic tourism sign, and A d,i is the texture feature value of the dust-covered area of the i-th road traffic tourism sign, and A c,i is the texture feature value of the non-dust-covered area of the i-th road traffic tourism sign; C represents the air dust concentration; Using image processing algorithms to extract features from the video image of the road traffic tourism sign includes: grayscale histogram, hue value, saturation and lightness, and calculate the fading change rate formula of the i-th road traffic tourism sign: Obtain the fading change rate S i ; where, C h and O h are the average hue values of the current and initial video images respectively, C s and O s are the average saturations of the current and initial video images respectively, C h and O h are the average lightness of the current and initial video images respectively; C k and C o are the values of the gray histograms of the current and initial video images at the gray level k.

6. The intelligent monitoring, operation and maintenance system for a road traffic tourism signboard according to claim 1, characterized in that, Analyze and calculate the environmental impact coefficient of the road traffic tourism sign according to the environmental parameters, including the following steps: According to the collected and preprocessed environmental parameters including: environmental temperature, environmental humidity, wind force, rainfall, snowfall, calculate the environmental impact coefficient of the road traffic tourism sign by weighted average.

7. The intelligent monitoring, operation and maintenance system for a road traffic tourism signboard according to claim 1, characterized in that Use machine learning to construct a remaining service life prediction model and calculate the remaining service life ratio of the road traffic tourism sign, including the following steps: According to the physical parameters and environmental impact coefficient of the road traffic tourism sign, use the regression model in machine learning to construct a remaining service life prediction model for the road traffic tourism sign; Collect the physical parameters and environmental impact coefficients of historical road traffic tourism signs as the training data set of the remaining service life prediction model; input the training data set into the remaining service life prediction model for training; By collecting the current physical parameters and environmental impact coefficients of historical road traffic tourism signs in real time and inputting them into the trained remaining service life prediction model; according to the input physical parameters and environmental impact coefficients, output the remaining service life duration of the i-th road traffic tourism sign in real time; divide the predicted remaining service life duration of the i-th road traffic tourism sign by the standard service life duration to obtain the remaining service life ratio.

8. The intelligent monitoring, operation and maintenance system for a road traffic tourism signboard according to claim 1, characterized in that, Comprehensively evaluate the performance of the road traffic tourism sign according to the guiding effect of the effective attention area, the dust coverage degree value, the fading change rate, the environmental impact coefficient and the remaining service life ratio, including the following steps: Calculate the performance index formula of the i-th road traffic tourism sign: Obtain the performance index P of the i-th road traffic tourism sign i ; where, E i 、D i 、S i 、M i and H i respectively represent the guiding effect of the effective attention area, the dust coverage value, the fading change rate, the remaining service life ratio and the environmental impact coefficient of the i-th road traffic tourism sign; ε i is the material strength of the i-th road traffic tourism sign When the performance index of the i-th road traffic tourism sign is lower than the threshold of 0.5, perform operation and maintenance processing on the i-th road traffic tourism sign, otherwise continue to monitor the road traffic tourism sign in real time.

Images