Highway high slope displacement integrated monitoring method based on image processing

Abstract

The invention relates to an alarm for responding to disaster events, and discloses a highway high slope displacement integrated monitoring method based on image processing, which comprises installation of high slope monitoring piles, an image processing module, a laser ranging module, a slope stability prediction module and an alarm module. According to the invention, the monitoring pile with the identification circle is installed on the high slope; the camera is used for shooting a side slope picture and carrying out image processing to obtain a center point coordinate, the center point coordinate is compared with historical data monitored during starting, and whether alarm information needs to be sent or not is determined according to whether the position measured by the laser range finder changes or not. When the slope displacement is insufficient to cause an alarm, the slope stability is pre-warned and pre-warning information is sent through slope stability prediction, and the method has the advantages of being timely in response, saving manpower and being high in monitoring breadth.

Description

Technical Field

[0001] The present invention relates to the technical field of alarms for responding to disaster events, and specifically to an integrated monitoring method for highway high slope displacement based on image processing. Background Technique

[0002] High slopes are important projects related to the national economy and are structures that are more or less used in projects such as civil construction, municipal engineering, water conservancy and hydropower, roads, railway transportation, and mining. Most of these high slopes are built on the edges of mountains in mountainous towns. The stability of high slopes is extremely important for the construction and use of construction projects. Serious landslide accidents not only cause heavy economic losses but may also claim precious lives, with extremely great harm. With the rapid development of the national economy, high slopes are increasingly used in engineering, with increasing height and scale, and high slopes can be seen everywhere. In addition, in recent years, there have been more and more abnormal climates globally, and abnormal climates pose a huge threat to both engineering slopes and natural slopes.

[0003] Slope disasters are disasters that seriously threaten human survival and production. Slope disasters mainly refer to slope landslides. The prediction and prevention of slope disasters are the common wish of all mankind. Landslides are often encountered in the operation of water conservancy and hydropower projects, mountain roads, highway construction, and geological disasters, and it is also one of the important factors affecting road safety and traffic safety.

[0004] Currently, in the field of slope safety detection, the main detection methods are mainly divided into two categories: stress-based detection and displacement-based detection. Stress-based detection methods, such as the sensor method, have disadvantages such as high professionalism and high professional requirements; displacement-based detection methods, such as the GPS detection method and the manual on-site detection method, all have disadvantages such as high professionalism, high cost, and danger to varying degrees.

[0005] Chinese Patent with Publication No. CN107843204A discloses a three-dimensional deformation monitoring method and system for slopes based on surveillance cameras; specifically disclosed: including controlling multiple surveillance cameras to synchronously collect images; calibrating the internal and external parameters of multiple surveillance cameras through a total station or the Global Positioning System (GPS); performing aberration correction on the images collected by multiple surveillance cameras; using a network cloud platform to analyze the images in real time to obtain the three-dimensional deformation of the slope, and transmitting the results for display. This method can transmit the results of the three-dimensional deformation of the slope collected from the images of multiple surveillance cameras for processing, improve the measurement efficiency, improve the accuracy of detection, and effectively reduce the occurrence of disasters.

[0006] However, this existing technology relies on three-dimensional surveying and mapping technology, which has relatively high requirements for the professional capabilities of surveying personnel, the optical environment at the surveying site, and the geological stability. It also has relatively high requirements for computing power during subsequent data processing, as well as for the hardware of arithmetic units and memories. The equipment cost is relatively high, making it difficult to promote on a large scale.

[0007] Chinese Patent No. CN112697050A discloses a nighttime slope displacement monitoring system based on luminous bodies; specifically, it discloses: including three subsystems: 1. The luminous body marker matrix system, where the luminous body markers are numbered in a nine-square grid manner and fixed on the slope. The fixing plate is prefabricated as a black square plate, forming an obvious color difference contrast with the slope and preventing the influence caused by the reflection of the fixing plate under nighttime lights; 2. The image acquisition system, with two cameras set on both sides in front of the monitored slope. The cameras are used to capture images of the slope and the luminous body marker matrix; 3. The data processing system, which is connected to the image acquisition system. The data processing system identifies the coordinates and changes of the luminous body markers through algorithms to monitor the slope displacement situation and achieve fully automatic slope displacement monitoring, which can meet the slope monitoring work under bad weather conditions.

[0008] However, this existing technology relies on visible light visual surveying and mapping technology and lacks effective processing of the obtained optical pixel data, especially the lack of processing technology based on visible light filtering. The original data is greatly affected by light and shadow and color light reflection, and the recognition accuracy of slope hazards is not high.

[0009] Chinese Patent No. CN113267128A discloses a binocular vision automated slope displacement monitoring method; specifically, it discloses: including the following steps: conducting real-time monitoring of the slope surface of the monitored object; using binocular vision technology and machine vision means to calculate the real displacement value of the entire slope in real time; evaluating its real-time displacement value through a pre-set warning value. If it exceeds the pre-set warning value, an alarm will be triggered, and it will also be displayed on the screen to inform the management personnel. By combining binocular vision technology and machine vision technology, through the automated extraction of information in the captured images, a safety status early warning system for slope surface displacement in slope engineering is constructed, with simple operation, improved efficiency and timeliness of safety monitoring, and greatly reduced probability of safety accidents.

[0010] However, this existing technology also has the technical problems of lack of data processing and filtering when relying on visible light for slope hazard identification, large influence of the original data by light and shadow or color light reflection, and low recognition accuracy of slope hazards. In practical applications, the risk of false alarms or missed alarms is relatively high. Summary of the Invention

[0011] To solve the problems of high specialization, high risk, and high labor cost in the existing high-slope monitoring technology, the present invention provides the following technical solution: An integrated monitoring method for highway high-slope displacement based on image processing, comprising the following steps:

[0012] Red circular monitoring piles are installed on each level of the high slope platform;

[0013] Monitoring cameras are installed and calibrated;

[0014] When the camera is started, the images of the high slope are processed to obtain the data of the marked circles of the monitoring piles and stored in the storage module as historical data;

[0015] When the camera is monitoring, the images of the high slope are processed to compare the marked data of the monitoring piles with the historical data;

[0016] If the horizontal or vertical displacement of the center data of the marked circle of the monitoring pile detected and the historical data exceeds Q1, the laser rangefinder is started for verification;

[0017] If the horizontal or vertical displacement of the center data of the marked circle of the monitoring pile detected and the historical data does not exceed Q1, slope stability prediction is performed, and warning information is sent according to the result;

[0018] If the data measured by the laser rangefinder exceeds the threshold, an alarm message is sent;

[0019] The received alarm information and warning information are processed.

[0020] Preferably, the Zhang-Zhengyou calibration method is used for calibrating the camera, specifically as follows:

[0021] (1) Prepare a Zhang-Zhengyou calibration standard checkerboard with a known size;

[0022] (2) Take pictures of the standard checkerboard from different angles;

[0023] (3) Detect the feature points in the image, such as the corner points of the standard checkerboard, to obtain the pixel coordinate values of the corner points of the standard checkerboard. According to the known size of the checkerboard and the origin of the world coordinate system, obtain the physical coordinate values of the corner points of the standard checkerboard;

[0024] (4) Obtain the internal and external parameters of the camera;

[0025] (5) Obtain the distortion coefficients;

[0026] (6) The calibration is completed.

[0027] Preferably, the image processing includes: image preprocessing, edge detection, and center point detection;

[0028] Preferably, it includes image preprocessing: performing grayscale processing on the captured image by the maximum value method, and the formula is as follows:

[0029] [f(i,j)] = [R′(i,j),G′(i,j),B′(i,j)]

[0030] R′(i,j) = G′(i,j) = B′(i,j) = max(R(i,j),G(i,j),B(i,j))

[0031] In the formula, i and j respectively represent the horizontal and vertical coordinates of each pixel point in the picture, [f(i,j)] represents the RGB value of the pixel point at coordinate (i,j) after grayscale processing by the maximum value method, R(i,j) and R'(i,j) respectively represent the red base component of the pixel point, G(i,j) and G'(i,j) respectively represent the green base component of the pixel point, and B(i,j) and B'(i,j) respectively represent the blue base component of the pixel point;

[0032] Image filtering. Median filtering is a non - linear filtering algorithm based on the pixel gray - level sorting within the image region. The gray - level value of the filtering pixel is determined according to the pixel gray - level sorting, and the median value of the gray levels within the pixel neighborhood is used as the new gray - level value of the filtering pixel;

[0033] Image enhancement. Perform histogram equalization processing on the captured high - slope image to remove and weaken useless information and highlight useful information such as color, edges, and shapes; Histogram equalization is to transform the gray - level histogram of the original gray - level image into a uniformly distributed gray - level histogram, and correct and calculate the gray - level of the pixel points in the image according to the equalized histogram, so that the image after equalization has information volume. By recalculating the gray - level value of the image points, the number of pixels within a certain gray - level range is approximately the same;

[0034] Image binarization. For image binarization, use the following formula to divide the grayscale image of the captured image into two parts: background and target according to the gray - level characteristics of the image; The formula:

[0035]

[0036] In the formula, image(i,j) represents the average value of the RGB values of the pixel point at coordinate (i,j) of the unprocessed picture, output(i,j) represents the RGB value of the pixel point at coordinate (i,j) of the processed picture, and t represents the threshold.

[0037] The larger the between - class variance between the background and the target, the greater the difference between the two parts that make up the image.

[0038] Preferably, use the Roberts operator for edge detection.

[0039] Preferably, perform center point detection on the image after edge detection:

[0040] Start detecting from the upper left corner of the image after edge detection; scan line by line. If all the gray values of a certain line are 0, then move to the next line and continue scanning; when scanning to a point where the gray value is not 0, record the coordinates of this point, and starting from the coordinates of this point, detect its 8 neighborhoods in sequence. If the pixel of a certain point is detected as 255, record the coordinates of this point, and repeat this operation; after the scanning is completed, calculate the average value of the coordinates of all marked points to obtain the center point coordinates of the identified circle as historical data.

[0041] Preferably, when the camera is started, the detected center point coordinates are saved as historical data, and the center point coordinates obtained by monitoring are compared with the historical data. If it exceeds Q1, start the laser rangefinder for distance detection.

[0042] Preferably, when the camera is monitoring, perform image processing on the captured image, extract the shape features and color features of the red circular mark, identify the image with the circular mark through the trained neural network, and perform edge detection and center point detection on the identified circular mark image; compare the obtained center point coordinates with the historical data;

[0043] Among them, for training the neural network, collect multiple pictures containing the monitoring piles with red circular marks and pictures without monitoring piles to form a data set, and divide the data set into a training set and a test set;

[0044] Normalize the data of the training set and input it into the neural network model. Through continuous forward propagation and error backpropagation, the result tends to be accurate. When the number of training times reaches the learning number T, stop the training process and save the training result to obtain the trained neural network model;

[0045] For the trained neural network model, input the pictures of the normalized test set, compare the obtained recognition results with the pictures of the test set. If the accuracy reaches u0, the training is completed; otherwise, increase the maximum number of training times and retrain until the accuracy reaches u0 to obtain the trained neural network model.

[0046] Preferably, start the laser rangefinder for distance detection verification, specifically:

[0047]

[0048] In the formula, c represents the speed of light, represents the phase delay generated by the light traveling back and forth once, and ω represents the modulation light angular frequency; if the detected distance from the calibration point is greater than Q2, then give an alarm.

[0049] Preferably, for the slope displacement that does not exceed Q1, conduct slope stability prediction:

[0050] After preprocessing and normalizing the warning of the monitored slope displacement data, input it into the slope prediction model to obtain the displacement situation of the slope in the future for a period of time. From the beginning of slope deformation to the final instability failure, the slope cumulative displacement-time curve slope is constantly changing. Therefore, the slope deformation evolution stage can be judged by the change of the slope of the cumulative displacement-time curve, that is, the tangent angle. Since the stretching of the horizontal and vertical coordinates will cause the change of the tangent angle, transformation processing is required to keep the dimension of the coordinates consistent. The transformation method is to divide the cumulative displacement by the rate so that the vertical coordinate and the horizontal coordinate have the same time dimension. The formula is as follows:

[0051]

[0052] In the formula, T i is the vertical coordinate value at the i-th moment after dimension unification, S i is the displacement within a monitoring period, and v represents the displacement rate; further obtain the tangent angle as follows:

[0053]

[0054] In the formula, a represents the tangent angle, t i represents the monitoring moment, t i-1 is the moment before the monitoring moment, and T i-1 is the vertical coordinate value at the (i - 1)-th moment after dimension unification;

[0055] If a is less than or equal to 45°, the slope is in the initial deformation stage and no warning is required;

[0056] If a is greater than 45° and less than 85°, the slope is in the accelerated deformation stage and a secondary warning is sent;

[0057] If a is greater than or equal to 85° and less than 90°, the slope is in the high-accelerated deformation stage and a primary warning is sent.

[0058] The secondary warning indicates the probability of a landslide risk in the next week.

[0059] The primary warning indicates the probability of a landslide risk within the next few hours, such as within 24 hours.

[0060] Preferably, process the alarm and warning information:

[0061] Set a siren alarm device at the site near the highway slope. Use the siren alarm to remind and block the road section in time for the alarm information, send a reminder message to the monitoring equipment of the site management personnel, and take corresponding measures to process the slope warning information;

[0062] If a secondary warning is detected, it is necessary to reinforce the monitored slope to prevent landslides;

[0063] If a first-level warning is detected, the road section needs to be urgently blocked to avoid vehicle accidents.

[0064] Beneficial effects

[0065] The beneficial effects of the present invention are as follows: The present invention captures the identification circle of the monitoring pile installed on the high slope through a camera, and obtains the center coordinates of the identification circle through steps such as image segmentation, edge detection, and center detection. The obtained center coordinates of the identification circle are compared with historical data. When the comparison difference is greater than the threshold Q1, the laser rangefinder is activated to measure the distance to the high slope calibration point and compare it with historical data. When the over-determined threshold Q2 is exceeded, a warning is immediately issued and corresponding safety treatment measures are taken.

[0066] Further, the center point detection starts from the upper left corner of the image after edge detection; scan line by line. If all the gray values of a certain line are 0, then transfer to the next line and continue scanning; when scanning to a point where the gray value is not 0, record the coordinates of this point, and starting from the coordinates of this point, detect its 8 neighborhoods in sequence. If the pixel of a certain point is detected as 255, record the coordinates of this point, and repeat this operation; after the scanning is completed, calculate the average value of the coordinates of all marked points to obtain the center point coordinates of the identification circle.

[0067] Further, slope stability prediction is used to predict the stability of the slope in the future for a period of time and send early warning information, so that protective measures can be taken in advance. Description of the drawings

[0068] Figure 1 It is a schematic diagram of the detection process of the present invention; Specific implementation manners

[0069] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0070] S1. Deploy a plurality of monitoring piles with specific marks at intervals on the slope platforms of each level of the high slope;

[0071] Among them, the monitoring pile uses a cuboid column, and the specific mark uses a red circular identifier;

[0072] S2. Deploy cameras on both sides of the high slope to monitor the real-time displacement of the high slope;

[0073] Further, calibrate the position of the camera, and use the Zhang Zhengyou calibration method; the specific steps are as follows:

[0074] (1) Prepare a Zhang-Zhengyou calibration standard checkerboard with a known size.

[0075] (2) Take pictures of the standard checkerboard from different angles.

[0076] (3) Detect the feature points in the image, such as the corner points of the standard checkerboard, to obtain the pixel coordinate values of the corner points of the standard checkerboard. Based on the known size of the checkerboard and the origin of the world coordinate system, obtain the physical coordinate values of the corner points of the standard checkerboard.

[0077] (4) Obtain the internal and external parameters of the camera.

[0078] (5) Obtain the distortion coefficients.

[0079] (6) The calibration is completed.

[0080] S3. When the camera is started, take pictures of the high slope. Perform image processing on the taken pictures (i.e., the pictures of the high slope) to obtain the coordinates of the center point of the red circular mark on the monitoring pile. The image processing steps are as follows:

[0081] S31. Image preprocessing, specific steps:

[0082] (1) Perform grayscale processing on the taken image using the maximum value method. The formula is as follows:

[0083] [f(i, j)] = [R′(i, j), G′(i, j), B′(i, j)]

[0084] R′(i, j) = G′(i, j) = B′(i, j) = max(R(i, j), G(i, j), B(i, j))

[0085] In the formula, i and j respectively represent the horizontal and vertical coordinates of each pixel point in the picture. [f(i, j)] represents the RGB value of the pixel point at coordinate (i, j) after grayscale processing using the maximum value method. R(i, j) and R'(i, j) respectively represent the red base components of the pixel point. G(i, j) and G'(i, j) respectively represent the green base components of the pixel point. B(i, j) and B'(i, j) respectively represent the blue base components of the pixel point.

[0086] (2) Image filtering. Median filtering is a non-linear filtering algorithm based on the gray-scale sorting of pixels in the image region. The gray-scale value of the filtering pixel is determined according to the gray-scale sorting of the pixels, and the median of the gray-scales in the pixel neighborhood is used as the new gray-scale value of the filtering pixel.

[0087] (3) Image enhancement: perform histogram equalization on the captured high-slope images to remove and weaken useless information and highlight useful information such as color, edges, and shapes. Histogram equalization is to transform the grayscale histogram of the original grayscale image into a uniformly distributed grayscale histogram, and correct and calculate the grayscale of the pixel points in the image according to the equalized histogram, so that the equalized image has information volume. By recalculating the grayscale values of the image points, the number of pixels within a certain grayscale level is roughly the same.

[0088] (4) Image binarization: use the following formula to divide the grayscale image of the captured image into two parts, background (greater than the threshold) and target (less than the threshold), according to the grayscale characteristics of the image. The larger the between-class variance between the background and the target, the greater the difference between the two parts that make up the image. The formula:

[0089]

[0090] In the formula, image(i,j) represents the average RGB value of the pixel point at the coordinate (i,j) of the unprocessed picture, output(i,j) represents the RGB value of the pixel point at the coordinate (i,j) of the processed picture, and t represents the threshold.

[0091] S32: Perform dilation and erosion operations on the picture.

[0092] Among them, erosion is a process of eliminating boundary points, eliminating noise, and shrinking the boundary inward, which can be used to eliminate small and meaningless objects. The specific steps are as follows:

[0093] Scan the original picture to find the first point with a pixel value of 1; move the origin of the pre-set structural element with a certain shape and origin position to this point; judge whether the pixel values within the range covered by the structural element are all 1. If so, the pixel value at the same position in the eroded image is 1; if at least one pixel value is 0, the pixel value at the same position in the eroded image is 0; repeat the erosion operation for all points with a pixel value of 1.

[0094] Among them, dilation is to merge the background points in contact with the target area into the target object A process of expanding the target boundary outward. Dilation can be used to fill some holes in the target area and eliminate small granular noises contained in the target area. The specific steps are as follows:

[0095] Scan the original image to find the first point with a pixel value of 0; move the origin of the structuring element with a preset shape and origin position to this point; determine whether there is at least one point with a pixel value of 1 within the range covered by the structuring element. If so, set the pixel value at the same position as the origin of the structuring element in the dilated image to 1. If all pixel values within this covered range are 0, set the pixel value at the same position in the dilated image to 0; repeat the dilation operation for all points with a value of 1;

[0096] S33. Perform edge detection on the picture to obtain the edge features of the red circular mark, and obtain the edge coordinate information of the red circular mark. After edge detection, the gray level of the edge of the circular mark is 255, and the gray level of other parts is 0; use the Roberts operator for edge detection, and this operator has the best detection effect in the horizontal and vertical directions;

[0097] S34. Perform center point detection on the picture after edge detection processing. The specific steps are as follows:

[0098] (1) Start detecting from the upper left corner of the picture after edge detection processing;

[0099] (2) Scan line by line. If all gray level values of a certain line are 0, then move to the next line and continue scanning;

[0100] (3) When scanning to a point with a gray level value not equal to 0, record the coordinates of this point, and starting from the coordinates of this point, detect its 8 neighborhoods in sequence. If the pixel of a certain point is detected as 255, record the coordinates of this point, and repeat this operation;

[0101] (4) After the scanning is completed, calculate the average value of the coordinates of all marked points to obtain the center point coordinates of the marked circle;

[0102] S35. Set the coordinates of the positioning mark point (x, y), and convert the center point coordinates into geodetic coordinates (M, N);

[0103]

[0104] In the formula, α represents the difference between the coordinates of the mark point and the geodetic coordinates in the x direction, and β represents the difference between the coordinates of the mark point and the geodetic coordinates in the y direction;

[0105] S36. Store the center point coordinates of the red circular marks on all monitoring piles into the storage module for comparison as historical data;

[0106] S4. During camera monitoring, perform image processing on the captured pictures, extract the shape features and color features of the red circular marks, identify the pictures with circular marks through the trained neural network, and perform edge detection and center point detection on the identified pictures with circular marks; compare the obtained center point coordinates with the historical data;

[0107] Among them, the neural network is trained by collecting multiple pictures containing the monitoring piles with red circular signs and pictures without monitoring piles to form a data set, and the data set is divided into a training set and a test set;

[0108] After normalizing the data of the training set, it is input into the neural network model. Through continuous forward propagation and error backpropagation, the result tends to be accurate. When the number of training times reaches the learning number T, the training process is stopped and the training result is saved to obtain the trained neural network model;

[0109] For the trained neural network model, the pictures of the normalized test set are input, and the recognized results are compared with the pictures of the test set. If the accuracy reaches u0, the training is completed; otherwise, the maximum number of training times is increased and retrained until the accuracy reaches u0 to obtain the trained neural network model;

[0110] If the daily vertical displacement or horizontal displacement is greater than Q1, then enter S5;

[0111] Among them, Q1 is set to 4mm;

[0112] S5. When the displacement of the high slope is detected, start the phase laser rangefinder installed on the camera to detect and verify the distance of the calibration points on the high slope; the distance calculation formula is:

[0113]

[0114] In the formula, c represents the speed of light, represents the phase delay generated by the light traveling back and forth once, and ω represents the modulation optical angular frequency;

[0115] If it is detected that the distance exceeds the threshold Q2 compared with the historical data, an alarm is immediately issued;

[0116] Among them, the threshold is set based on the actual measurement value of the engineer;

[0117] Otherwise, enter S6;

[0118] S6. Establish a slope prediction model to predict the slope stability, and judge whether early warning measures need to be taken for the slope according to the change of the slope of the cumulative displacement-time curve, that is, the tangent angle; specifically as follows:

[0119] After preprocessing and normalizing the warning of the monitored slope displacement data, input it into the slope prediction model to obtain the displacement situation of the slope in the future for a period of time. From the beginning of deformation to the final instability failure of the slope, the slope of the cumulative displacement-time curve is constantly changing. Therefore, the deformation evolution stage of the slope can be judged by the change of the slope of the cumulative displacement-time curve, that is, the tangent angle. Since the stretching of the horizontal and vertical coordinates will cause the change of the tangent angle, transformation processing is needed to make the dimensions of the coordinates consistent. The transformation method is to divide the cumulative displacement by the rate so that the vertical coordinate and the horizontal coordinate have the same time dimension. The formula is as follows:

[0120]

[0121] In the formula, T i is the vertical coordinate value at the i-th moment after dimension unification, S i is the displacement amount within a monitoring period, and v represents the displacement rate; further, the tangent angle is obtained as follows:

[0122]

[0123] In the formula, a represents the tangent angle, t i represents the monitoring moment, t i-1 is the moment before the monitoring moment, and T i-1 is the vertical coordinate value at the (i - 1)-th moment after dimension unification;

[0124] If a is less than or equal to 45°, the slope is in the initial deformation stage and no warning is required. Continue to monitor;

[0125] If a is greater than 45° and less than 85°, the slope is in the accelerated deformation stage, and the probability of landslide danger within the next week. Send a secondary warning;

[0126] If a is greater than or equal to 85° and less than 90°, the slope is in the high-acceleration deformation stage, and the probability of landslide danger within the next few hours. Send a primary warning;

[0127] S7. Set a siren alarm device near the highway slope site. Use siren alarm to remind and block the road section in time for the alarm information. Send reminder information to the monitoring equipment of the site management personnel, and take corresponding measures to deal with the slope warning information;

[0128] If a secondary warning is detected, it is necessary to reinforce the monitored slope to prevent landslides;

[0129] If a primary warning is detected, it is necessary to urgently block the road section to avoid vehicle accidents;

[0130] In use, although embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An integrated monitoring method for high-slope displacement of highways based on image processing, characterized in that: The following steps are involved: Red circular monitoring piles are installed on various platforms at high slopes; Install monitoring cameras and calibrate them; When the camera is started, the high slope pictures are processed to obtain the monitoring pile marking circle data and store them in the storage module as historical data; When the camera is monitoring, the high slope pictures are processed to obtain the monitoring pile marking data and compare them with the historical data; If the horizontal or vertical displacement between the detected monitoring pile identification circle center data and the historical data exceeds Q1, the laser rangefinder is started for verification; If the horizontal displacement or vertical displacement of the center data of the monitoring pile identification circle and the historical data does not exceed Q1, the slope stability prediction is carried out and an early warning message is sent based on the results; If the data measured by the laser rangefinder exceeds the threshold, an alarm message is sent; Process the received alarm information and warning information.

Images