Road slope deformation early warning method and system based on image recognition

Abstract

The present application relates to the field of slope monitoring, especially to a road slope deformation early warning method and system based on image recognition, the early warning method comprising: determining key analysis images of each target supervision area according to an initial risk coefficient; determining whether to perform regional correlation analysis or risk analysis according to a risk feature state; when performing regional correlation analysis, determining a correlation analysis area based on a geological interference index or a geological reference index; when performing risk analysis, determining a deformation risk coefficient of the target supervision area, and setting a risk assessment reliability coefficient according to a correlation risk difference parameter or a risk coverage parameter; determining an execution priority based on the deformation risk coefficient, and determining whether to perform execution optimization analysis on the target supervision area according to a regional risk difference index, which ensures the timeliness of the evaluation result of the slope deformation and improves the analysis efficiency of the evaluation process of the slope deformation.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of slope monitoring, in particular to a road slope deformation early warning method and system based on image recognition. BACKGROUND

[0002] By obtaining monitoring images of the road slope and completing image analysis, it is determined whether there are deformation risk features in the monitored road slope area, and the deformation of the road slope is evaluated in time to perform corresponding protection treatment on the road slope to avoid road accidents. In order to ensure the reliability of the evaluation result of the deformation of the road slope, the identification analysis results of multiple monitoring images are often used to make a judgment. However, in the actual evaluation process, the number of monitoring images of the road slope is too large, and continuous high-precision identification analysis of all monitoring images obtained for each road slope area can easily cause an excessive analysis burden in the analysis process of the deformation of the road slope, thereby affecting the actual analysis efficiency. Therefore, how to balance the reliability of the evaluation result and the analysis efficiency in the process of evaluating the deformation of the road slope based on image recognition is a problem to be solved in the field.

[0003] Chinese Patent Application Publication No. CN119380054A discloses a landslide deformation identification method and device based on image difference, relating to the technical field of landslide deformation identification. The method comprises: building a landslide test model, collecting multiple slope images; preprocessing the collected slope images, including: calculating parameter setting, displacement noise filtering and displacement value interpolation; analyzing and calculating the preprocessed slope image sequence data through LSPIV technology to obtain the displacement vector field or velocity vector field of the landslide surface in the slope image; and identifying the deformation and movement of the landslide body through a python-based algorithm. The present application realizes the identification and measurement of the deformation and movement of the landslide in the image sequence by processing the images in the landslide test model video, and compares the accuracy and feasibility with the actual deformation. However, the above-mentioned scheme has the following defects: it fails to effectively match and screen the slope images involved in the fine analysis process in the actual evaluation process, resulting in waste of computing resources in the actual evaluation process of the deformation of the slope, and thus it is difficult to balance the reliability of the evaluation result and the analysis efficiency in the actual analysis process. SUMMARY

[0004] Therefore, the present application provides a road slope deformation early warning method and system based on image recognition to overcome the problem that the prior art fails to effectively match and screen the slope images involved in the fine analysis process in the actual evaluation process, resulting in waste of computing resources in the actual evaluation process of the deformation of the slope, and thus it is difficult to balance the reliability of the evaluation result and the analysis efficiency in the actual analysis process.

[0005] To achieve the above object, the application provides a road slope deformation early warning method based on image recognition, comprising:

[0006] The target supervision route is divided to obtain a plurality of target supervision areas;

[0007] The initial risk analysis is performed on the monitoring and analysis images of each target supervision area to determine the initial risk coefficient of each monitoring and analysis image, and the key analysis image of each target supervision area is determined according to the initial risk coefficient;

[0008] The risk characteristic state of each target supervision area is determined according to the key distribution proportion index and the initial risk evaluation index, and whether to perform regional correlation analysis or risk analysis on each target supervision area is determined according to the risk characteristic state;

[0009] When the regional correlation analysis is performed, the reference risk distribution index of the target supervision area is used to determine the correlation analysis area based on the geological interference index or the geological comparison index;

[0010] When the risk analysis is performed, the deformation risk coefficient of the target supervision area is determined according to the slope deformation characteristics existing in the key analysis image, and the risk evaluation reliability coefficient is set according to the associated risk difference parameter or the risk coverage parameter based on the risk characteristic state;

[0011] The execution priority of each target supervision area is determined based on the deformation risk coefficient, and whether to perform execution optimization analysis on the target supervision area is determined according to the regional risk difference index to perform secondary adjustment on the execution priority.

[0012] Further, the key analysis image is the monitoring and analysis image of each target supervision area whose initial risk coefficient is greater than a preset initial risk coefficient;

[0013] The initial risk coefficient is determined according to the change risk proportion index and the reference change risk index of the monitoring and analysis image;

[0014] The initial risk coefficient has a positive correlation with the change risk proportion index and the reference change risk index, respectively.

[0015] Further, the risk characteristic state includes a first risk characteristic state and a second risk characteristic state;

[0016] If the key distribution proportion index of a target supervision area is greater than a preset key distribution proportion index or the initial risk evaluation index is greater than a preset initial risk evaluation index, it is determined that the target supervision area is in the first risk characteristic state;

[0017] If the key distribution proportion index of the target regulatory region is less than or equal to the preset key distribution proportion index and the initial risk evaluation index is less than or equal to the preset initial risk evaluation index, it is determined that the target regulatory region is in a second risk characteristic state.

[0018] Further, if the target regulatory region is in a first risk characteristic state, a regional correlation analysis is performed on the target regulatory region to determine the risk correlation parameters of other target regulatory regions in the target regulatory route to the target regulatory region.

[0019] The correlation analysis region is a target regulatory region with a risk correlation parameter greater than a preset risk correlation parameter.

[0020] Further, the risk correlation parameter is determined according to a reference risk distribution index of the target regulatory region in the first risk characteristic state, wherein,

[0021] If the reference risk distribution index of the target regulatory region is greater than a preset reference risk distribution index, the risk correlation parameters of each target regulatory region to the target regulatory region are determined according to a geological interference index.

[0022] If the reference risk distribution index of the target regulatory region is less than or equal to the preset reference risk distribution index, the risk correlation parameters of each target regulatory region to the target regulatory region are determined according to a geological interference index.

[0023] Further, if the target regulatory region is in a second risk characteristic state, or if the target regulatory region is in a first risk characteristic state and the regional correlation analysis is completed, a risk analysis is performed on the target regulatory region, wherein,

[0024] Deformation feature recognition is performed on each key analysis image to determine the deformation risk coefficient of the target regulatory region.

[0025] A verification strategy is determined according to the risk characteristic state of the target regulatory region to determine the risk assessment reliability coefficient of the target regulatory region.

[0026] Further, under the condition of risk assessment completion, the risk assessment reliability coefficient of each target regulatory region is set, wherein,

[0027] If the target regulatory region is in a first risk characteristic state, the risk assessment reliability coefficient of the target regulatory region is determined according to a correlation risk difference parameter.

[0028] If the target regulatory region is in a second risk characteristic state, the risk assessment reliability coefficient of the target regulatory region is determined according to a risk coverage parameter.

[0029] The risk assessment completion condition is that a deformation risk coefficient of the target supervision region is determined.

[0030] Further, if a regional risk difference index of the target supervision region is greater than a preset regional risk difference index, it is determined that the optimization analysis is performed for the target supervision region.

[0031] The regional risk difference index is determined according to a risk assessment reliable coefficient of each target supervision region in the associated evaluation range.

[0032] Further, when the optimization analysis is performed, the execution priority of each target supervision region in the associated evaluation range of the target supervision region is increased according to the risk association parameter and the risk assessment reliable coefficient.

[0033] The increase value of the execution priority is positively correlated with the risk association parameter and the risk assessment reliable coefficient.

[0034] The application also provides a road slope deformation early warning system based on image recognition, comprising:

[0035] An initial analysis module is used to perform initial risk analysis on each monitoring analysis image to determine an initial risk coefficient, and to determine a key analysis image of each target supervision region according to the initial risk coefficient.

[0036] A key evaluation module is connected with the initial analysis module and is used to determine a risk feature state of each target supervision region according to a key distribution proportion index and an initial risk evaluation index, and to determine whether to perform regional association analysis or risk analysis for each target supervision region according to the risk feature state.

[0037] An association analysis module is connected with the key evaluation module and is used to determine an associated analysis region based on a geological interference index or a geological contrast index according to a reference risk distribution index of the target supervision region.

[0038] A risk analysis module is connected with the key evaluation module and the association analysis module, respectively, and is used to determine a deformation risk coefficient of the target supervision region according to a slope deformation feature of the key analysis image, and to set a risk assessment reliable coefficient according to an associated risk difference parameter or a risk coverage parameter based on the risk feature state.

[0039] A regional execution module is connected with the risk analysis module and is used to determine an execution priority of each target supervision region based on the deformation risk coefficient, and to determine whether to perform optimization analysis for the target supervision region according to a regional risk difference index, so as to perform secondary adjustment on the execution priority.

[0040] Compared with the prior art, the beneficial effects of the present application are that, through initial risk analysis, the initial risk coefficient of the monitoring analysis image of each target supervision area is determined to determine the key analysis image that needs to be further analyzed, and through initial risk analysis, the risk of image feature change of the monitoring analysis image is avoided, excessive analysis of the monitoring analysis image that cannot be evaluated for the slope deformation condition is avoided, and the consumption of computing resources is avoided, the present application ensures the timeliness of the evaluation result of the slope deformation while improving the analysis efficiency of the evaluation process of the slope deformation.

[0041] Further, in the present application, the risk feature state of each target supervision area is determined according to the key distribution proportion index and the initial risk evaluation index, the distribution of the suspected slope deformation feature of each target supervision area is preliminarily judged through the key distribution proportion index and the initial risk evaluation index, and whether to perform regional correlation analysis on each target supervision area and whether to perform matching analysis on the key analysis image of each target supervision area is determined according to the risk feature state, that is, the influence degree of the target supervision area from the surrounding area and the influence degree on the surrounding area are judged, and then it is determined whether the expansion of the subsequent image matching range is needed, and the deformation evaluation of the surrounding target supervision area is analyzed cooperatively, which not only ensures the reliability of the deformation evaluation result of the target supervision area, but also avoids repeated analysis caused by sequential analysis.

[0042] Further, in the present application, for the target supervision area in the first risk feature state, the setting mode of the risk correlation parameter is determined according to the reference risk distribution index of the target supervision area, so as to determine the correlation analysis area, so that the determination result of the determined correlation analysis area is more consistent with the actual slope condition, and the effectiveness of the verification behavior of the subsequent evaluation result is ensured, and the present application ensures the timeliness of the evaluation result of the slope deformation while improving the analysis efficiency of the evaluation process of the slope deformation.

[0043] Further, in the present application, the setting mode of the priority matching parameter is determined according to the risk feature state of the target supervision area, so that the determined priority matching parameter can better conform to the actual situation of the target supervision area, and the effectiveness of the monitoring analysis image determined for participating in the slope deformation evaluation process according to the priority matching parameter is further ensured, the present application reduces the consumption of computing resources caused by excessive analysis of the monitoring analysis image that cannot be evaluated for the slope deformation condition, and ensures the reliability of the evaluation of the slope deformation condition.

[0044] Further, the present application determines whether to perform optimization analysis for the target regulatory area according to the regional risk difference index, so as to avoid the route overlap of the priority order set when verifying each target regulatory area, thereby reducing the efficiency of the verification process, and the execution priority of each target regulatory area is adjusted through the risk correlation parameter and the risk assessment reliability coefficient, so that the target regulatory areas with correlation are verified cooperatively, thereby ensuring the verification urgency of each target regulatory area and the execution efficiency of the actual verification process. BRIEF DESCRIPTION OF DRAWINGS

[0045] Figure 1 A schematic diagram of the road slope deformation early warning method based on image recognition of the present application is shown in the figure.

[0046] Figure 2 A flowchart for determining the risk characteristic state of each target regulatory area according to the key distribution proportion index and the initial risk evaluation index is shown in the figure.

[0047] Figure 3 A flowchart for determining whether to perform regional correlation analysis for each target regulatory area and whether to perform matching analysis on the key analysis image of each target regulatory area according to the risk characteristic state is shown in the figure.

[0048] Figure 4 A module connection diagram of the road slope deformation early warning system based on image recognition of the present application is shown in the figure. DETAILED DESCRIPTION

[0049] In order to make the purpose and advantages of the present application more clear and explicit, the present application will be further described below in combination with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present application, and do not limit the present application.

[0050] The preferred embodiments of the present application will be described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present application, and are not intended to limit the protection scope of the present application.

[0051] It should be noted that in the description of the present application, the terms "up", "down", "left", "right", "in", "out" and the like indicate the direction or positional relationship of the terms based on the direction or positional relationship shown in the drawings, which is only for the convenience of description, and does not indicate or imply that the device or element must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation of the present application.

[0052] In addition, it needs to be explained that, in the description of the present application, unless otherwise explicitly specified and limited, the terms "mounting", "connection", "linking" should be understood broadly, for example, can be fixed connection, can also be detachable connection, or integrally connected; can be mechanical connection, can also be electrical connection; can be directly connected, can also be indirectly connected through intermediate medium, can be internal communication of two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0053] Please refer to Figures 1 to 3 As shown in the drawings, the embodiment of the present application provides a road slope deformation early warning method based on image recognition, comprising:

[0054] S1, the target supervision route is divided, and a plurality of target supervision areas are obtained;

[0055] S1, the initial risk analysis is carried out on the monitoring and analysis image of each target supervision area, so as to determine the initial risk coefficient of each monitoring and analysis image, and the key analysis image of each target supervision area is determined according to the initial risk coefficient;

[0056] S2, according to the key distribution proportion index and the initial risk evaluation index, the risk characteristic state of each target supervision area is determined, and whether to carry out regional correlation analysis or risk analysis on each target supervision area is determined according to the risk characteristic state;

[0057] S3, when the regional correlation analysis is carried out, the reference risk distribution index of the target supervision area is determined to determine the correlation analysis area based on the geological interference index or the geological contrast index;

[0058] S4, when the risk analysis is carried out, the deformation risk coefficient of the target supervision area is determined according to the slope deformation characteristics existing in the key analysis image, and the risk evaluation reliability coefficient is set according to the associated risk difference parameter or the risk coverage parameter based on the risk characteristic state;

[0059] S5, the execution priority of each target supervision area is determined based on the deformation risk coefficient, and whether to carry out execution optimization analysis on the target supervision area is determined according to the regional risk difference index, so as to carry out secondary adjustment on the execution priority.

[0060] The application is used for the process of evaluating the deformation of the road side slope based on image recognition, and under the premise of ensuring the reliability of the evaluation result, the waste of computing resources in the evaluation process is reduced, the road route that needs to be evaluated for the deformation of the side slope is recorded as the target monitoring route, any side slope area corresponding to the target monitoring route is divided to obtain a plurality of target monitoring areas with the same area, the intersection of the straight line formed by the edge line of the target monitoring area obtained by division and the edge line of the road area corresponding to the target monitoring route close to the target monitoring area is recorded as the division intersection point of the target monitoring area during the division process of a single target monitoring area, the tangent line of the edge line of the road area corresponding to the target monitoring route based on the division intersection point is recorded as the intersection tangent line, the straight line formed by the edge line of the target monitoring area obtained by division is perpendicular to the intersection tangent line, a plurality of fixed monitoring points are provided for each target monitoring area in the application, each fixed monitoring point is provided with a fixed camera to obtain monitoring analysis images, the positions and image acquisition angles of different fixed monitoring points are different, the positions of the fixed monitoring points in each target monitoring area are not specifically limited in the application, but it is necessary to ensure that the camera can obtain more comprehensive image information as much as possible and avoid obstruction and interference as much as possible.

[0061] In the application, a plurality of side slope monitoring records are applied, each side slope monitoring record records at least one initial risk coefficient, a change risk parameter, a key distribution proportion index, an initial risk evaluation index, a risk correlation parameter and a regional risk difference index in the process of evaluating the deformation of the side slope of each target monitoring area in the target monitoring route, and each side slope monitoring record corresponds to a qualified mark, which records whether the reliability of the deformation evaluation process of the side slope area corresponding to the target monitoring route and the image analysis efficiency meet the user's demand, it can be understood that the user can determine whether the reliability of the deformation evaluation process of the side slope area corresponding to the target monitoring route and the image analysis efficiency meet the demand according to the self-set index, for example, the self-set index can be but is not limited to the evaluation effectiveness and the evaluation computing power consumption parameter, the evaluation effectiveness = the number of times that the deformation of the side slope is indeed determined after the field evaluation in the number of times that the deformation of the side slope is determined for the target monitoring area existing in the target monitoring route / the number of times that the deformation of the side slope is determined for the target monitoring area existing in the target monitoring route, and the evaluation computing power consumption parameter is the number of monitoring analysis images involved in the process of determining the deformation of the side slope.

[0062] Specifically, the key analysis image is a monitoring analysis image whose initial risk coefficient of each target monitoring area is greater than a preset initial risk coefficient;

[0063] The initial risk coefficient is determined according to the change risk proportion index of the monitoring analysis image and the reference change risk index.

[0064] The initial risk coefficient is positively correlated with the change risk proportion index and the reference change risk index, respectively.

[0065] In the present application, a cycle of image acquisition period is applied, the length of the image acquisition period can be determined by the user, the higher the reliability requirement of the user for the deformation evaluation process of the target monitoring route corresponding to the slope area, the shorter the length of the image acquisition period, the length of the image acquisition period is provided as 12h, at the starting time of each image acquisition period, the fixed monitoring point of each target monitoring area is used to obtain the monitoring analysis image of each target monitoring area, and initial risk analysis is performed on each monitoring analysis image to determine the initial risk coefficient of each monitoring analysis image.

[0066] For a single target supervision area, if the current time is the starting time of an image acquisition period, an initial risk analysis is performed on a plurality of monitoring analysis images obtained for the target supervision area at the time, each monitoring analysis image is divided into a plurality of rectangular areas with the same area, and is recorded as an analysis sub-area. The number of analysis sub-areas obtained by dividing each monitoring analysis image can be set by the user according to the actual working scene. The higher the user's requirements for the reliability of the deformation evaluation process of the target supervision route corresponding to the slope area and the image analysis efficiency, the greater the number of analysis sub-areas obtained by dividing each monitoring analysis image. For a single monitoring analysis image, the initial risk coefficient is the sum of the change risk proportion index and the reference change risk index. The change risk proportion index = the number of change risk areas contained in the monitoring analysis image / the number of analysis sub-areas contained in the monitoring analysis image. The reference change risk index = the maximum value of the change risk parameter of the change risk area contained in the monitoring risk image / the preset change risk parameter. The change risk area is an analysis sub-area with a change risk parameter greater than the preset change risk parameter. For a single analysis sub-area, the change risk parameter is the average value of the pixel difference proportion between the analysis sub-area and each change reference area. The analysis sub-area whose shooting range is consistent with the image corresponding to the analysis sub-area obtained for the target supervision area in the last image acquisition period of the current image acquisition period is recorded as the change reference area of the analysis sub-area. For any two analysis sub-areas, the pixel difference proportion = the number of difference pixels between the two analysis sub-areas / the number of pixels in the analysis sub-area. Two pixels with the same position between the two analysis sub-areas are recorded as a group of evaluation pixels. If the absolute value of the difference between the gray values of the two pixels in a group of evaluation pixels is greater than the preset gray difference value, the two pixels in the group of evaluation pixels are recorded as difference pixels. The resolution and image size of each monitoring analysis image obtained in the present application are consistent. How to determine the gray value of each pixel is known to those skilled in the art, and will not be described here.

[0067] The preset initial risk coefficient and the preset change risk parameter can be determined by the user according to the actual working scene. For example, the user can set according to the slope supervision record. The higher the user's requirement for the reliability of the deformation evaluation process of the target supervision route corresponding to the slope region and the image analysis efficiency, the greater the value of the preset initial risk coefficient, and the smaller the value of the preset change risk parameter. A method for determining the value of the preset initial risk coefficient is provided. The average value of the initial risk coefficient of the key analysis image in the slope supervision record that meets the user's requirement for the reliability of the deformation evaluation process of the target supervision route corresponding to the slope region and the image analysis efficiency is recorded as the preset initial risk image. A method for determining the value of the preset change risk parameter is provided. The minimum value of the change risk parameter of the change risk region in the slope supervision record that meets the user's requirement for the reliability of the deformation evaluation process of the target supervision route corresponding to the slope region and the image analysis efficiency is recorded as the preset change risk parameter.

[0068] Specifically, the risk feature state includes a first risk feature state and a second risk feature state.

[0069] If the key distribution proportion index of the target supervision region is greater than the preset key distribution proportion index or the initial risk evaluation index is greater than the preset initial risk evaluation index, it is determined that the target supervision region is in the first risk feature state.

[0070] If the key distribution proportion index of the target supervision region is less than or equal to the preset key distribution proportion index and the initial risk evaluation index is less than or equal to the preset initial risk evaluation index, it is determined that the target supervision region is in the second risk feature state.

[0071] Wherein, for the target supervision region, the key distribution proportion index = the number of fixed monitoring points corresponding to the key analysis image determined by the target supervision region in the current image acquisition period / the number of fixed monitoring points existing in the target supervision region, and the initial risk evaluation index = preset initial risk evaluation index / standard deviation of the initial risk coefficient of each key analysis image determined by the target supervision region in the current image acquisition period + average value of the initial risk coefficient of each key analysis image determined by the target supervision region in the current image acquisition period / preset initial risk coefficient;

[0072] The preset key distribution proportion index and the preset initial risk evaluation index can be determined by the user according to the actual working scene. For example, the user can set according to the slope supervision record. A preset key distribution index value method is provided. The slope supervision record of the target supervision area in the second risk feature state is recorded as a feature reference record. The maximum value of the key distribution proportion index of the target supervision area in the feature reference record that meets the reliability and image analysis efficiency requirements of the user for the deformation evaluation process of the slope area corresponding to the target supervision route is recorded as the preset key distribution proportion index. A preset initial risk evaluation index value method is provided. The minimum value of the initial risk evaluation index of the target supervision area in the feature reference record that meets the reliability and image analysis efficiency requirements of the user for the deformation evaluation process of the slope area corresponding to the target supervision route is recorded as the preset initial risk evaluation index.

[0073] Specifically, if there is a target supervision area in a first risk feature state, it is determined that the region correlation analysis is performed on the target supervision area to determine the risk correlation parameters of other target supervision areas in the target supervision route to the target supervision area.

[0074] The correlation analysis region is a target supervision area with a risk correlation parameter greater than a preset risk correlation parameter.

[0075] Specifically, the setting method of the risk correlation parameter is determined according to the reference risk distribution index of the target supervision area in the first risk feature state, wherein

[0076] If there is a target supervision area with a reference risk distribution index greater than a preset reference risk distribution index, the risk correlation parameters of each target supervision area to the target supervision area are determined according to the geological interference index.

[0077] If there is a target supervision area with a reference risk distribution index less than or equal to a preset reference risk distribution index, the risk correlation parameters of each target supervision area to the target supervision area are determined according to the geological interference index.

[0078] For a single target supervision area, if the target supervision area is in a first risk characteristic state, it indicates that the coverage of the image acquisition area corresponding to the key analysis image determined by the target supervision area in the current image acquisition period is relatively wide, or the initial risk coefficients of the key analysis images are generally large and the difference between the initial risk coefficients is small, indicating that the position distribution of the slope deformation features in the target supervision area in the current image acquisition period is relatively wide, and the formation of the slope deformation of the target supervision area is greatly affected by the deformation of the surrounding target supervision areas. Therefore, when selecting a matching execution image, the influence of other target supervision areas on the target supervision area needs to be evaluated to further expand the range of image matching, and the monitoring analysis images of the surrounding target supervision areas are combined.

[0079] For a single target supervision area, the reference risk distribution index is the average value of the change risk distribution indexes of each monitoring analysis image obtained by the target supervision area in the current image acquisition period. For a single monitoring analysis image, the change risk distribution index = the average value of the number of change risk regions existing in the distribution analysis range of each change risk region in the monitoring analysis image / the number of change risk regions existing in the monitoring analysis image. The risk correlation parameters between each target supervision area and the target supervision area in the correlation evaluation range of the target supervision area are detected. The shortest distance between any target supervision area and the target supervision area in the correlation evaluation range is less than a preset evaluation distance. The value of the preset evaluation distance can be set by the user according to the actual working scene. The higher the user's requirement for the reliability of the deformation evaluation process of the target supervision route corresponding to the slope area and the image analysis efficiency, the larger the value of the preset evaluation distance. It is provided that the preset evaluation distance is five times the length of the edge of the target supervision route existing in the target supervision area.

[0080] For a single target supervision area, if the reference risk distribution index of the target supervision area is greater than the preset reference risk distribution index, the change risk areas existing in each monitoring and analysis image obtained for the target supervision area in the current image acquisition period are more concentrated, thereby indicating that the deformation features possibly existing in the slope region corresponding to the target supervision area are more concentrated but involve a wide range, that is, the existing deformation features have a certain extension, the geological condition relationship between each target supervision area is analyzed to determine a risk correlation parameter, and the correlation analysis region determined in this way is used to evaluate the rationality of the time sequence of the deformation features generated by different target supervision areas. For any two target supervision areas, the geological interference index is the sum of the dip angle evaluation index and the runoff evaluation index, the dip angle evaluation index is the sum of the coincident rock layer proportion index and the coincident dip angle change index, the coincident rock layer proportion index = the number of rock layers existing in the above two target supervision areas / the number of different rock layers existing in the above two target supervision areas, the coincident dip angle change index is the average value of the dip angle change indexes of each rock layer existing in the above two target supervision areas, for a single rock layer existing in the above two target supervision areas, the dip angle change index = the absolute value of the difference between the reference rock layer dip angles of the rock layer in the above two target supervision areas / the average value of the reference rock layer dip angles of the rock layer in the above two target supervision areas, for a single target supervision area, the reference rock layer dip angle of any rock layer is the average value of the rock layer dip angles obtained for the rock layer by randomly selecting a plurality of points for the target supervision area, the inclination direction and inclination angle of each rock layer are measured by using a rock layer dip angle logging instrument, the inclination angle is recorded as the rock layer dip angle, the runoff evaluation index = the number of underground runoff existing in the above two target supervision areas / the number of different underground runoff existing in the above two target supervision areas, the determined geological interference index is normalized, the risk correlation parameter and the normalized geological interference index are in a positive correlation relationship, how to normalize is a content mastered by those skilled in the art, and will not be described here.

[0081] For a single target supervision area, if the reference risk distribution index of the target supervision area is less than or equal to the preset reference risk distribution index, it indicates that the distribution of the change risk area existing in each monitoring and analysis image obtained in the current image acquisition period for the target supervision area is relatively dispersed, and then it indicates that the distribution of the deformation feature existing in the slope region corresponding to the target supervision area is relatively dispersed. The similarity degree of the geological conditions between each target supervision area is analyzed to determine a risk correlation parameter. The correlation analysis region determined in this way is used to evaluate the rationality of the time sequence of the deformation features generated under similar conditions. For any two target supervision areas, the geological comparison index = 1 / average value of parameter difference indexes of each rock and soil property parameter, for a single rock and soil property parameter, the parameter difference index = difference between the values of the rock and soil property parameter of the above two target supervision areas / average value of the values of the rock and soil property parameter of the above two target supervision areas, the category of the rock and soil property parameter used to determine the geological comparison index includes but is not limited to: average particle size of rock and soil particles, particle size standard deviation of rock and soil particles, rock and soil density, and rock and soil porosity. The determined geological comparison index is normalized, and the risk correlation parameter and the normalized geological comparison index are in a positive correlation.

[0082] The values of the preset risk correlation parameter and the preset reference risk distribution index can be determined by the user according to the actual working scene. For example, the user can set it according to the slope supervision record. The higher the user's requirement for the reliability of the deformation evaluation process of the slope region corresponding to the target supervision route and the image analysis efficiency, the larger the value of the preset risk correlation parameter. A method for determining the value of the preset risk correlation parameter is provided. The minimum value of the risk correlation parameter of the correlation analysis region in the slope supervision record that meets the user's requirements for the reliability of the deformation evaluation process of the slope region corresponding to the target supervision route and the image analysis efficiency is taken as the preset risk correlation parameter. A method for determining the value of the preset reference risk distribution index is provided. The minimum value of the reference risk distribution index of each target supervision area determined according to the geological interference index in the slope supervision record that meets the user's requirements for the reliability of the deformation evaluation process of the slope region corresponding to the target supervision route and the image analysis efficiency is taken as the preset reference risk distribution index.

[0083] Specifically, if there is a target supervision area in the second type of risk feature state, or there is a target supervision area in the first type of risk feature state and the region correlation analysis is completed, wherein,

[0084] Deformation feature recognition is performed on each key analysis image to determine the deformation risk coefficient of the target supervision area;

[0085] A verification strategy is determined according to the risk feature state of the target supervision area to determine the risk evaluation reliability coefficient of the target supervision area.

[0086] Among them, the present application is set up for each type of slope deformation characteristics that may exist in the target regulatory area, and based on the set of each type of slope deformation characteristics, the key analysis image is identified to obtain the slope deformation characteristics existing in the key analysis image. For a single target regulatory area, the deformation risk coefficient = the number of slope deformation characteristics identified and determined for the key analysis image of the target regulatory area x (the number of categories of slope deformation characteristics identified and determined for the key analysis image of the target regulatory area / the number of categories of slope deformation characteristics set for the target regulatory route), how to identify the slope deformation characteristics existing in the obtained monitoring analysis image based on the set of slope deformation characteristics, to determine the feature similarity, which is mastered by those skilled in the art, and is not described here. The categories of slope deformation characteristics in the present application include but are not limited to: linear cracks, network cracks, shallow landslides, deep landslides, soil collapse, local bulge, large area bulge and settlement.

[0087] Specifically, under the condition of risk assessment completion, the risk assessment reliable coefficient of each target monitoring area is set, wherein,

[0088] If there is a target regulatory area in a risk characteristic state, the risk assessment reliable coefficient of the target regulatory area is determined according to the associated risk difference parameter;

[0089] If the target regulatory area is in a two-class risk characteristic state, the risk assessment reliable coefficient of the target regulatory area is determined according to the risk coverage parameter;

[0090] The risk assessment completion condition is that the deformation risk coefficient of the target regulatory area is determined.

[0091] Among them, for a single target regulatory area, if the target regulatory area is in a one-class risk characteristic state, it indicates that the slope deformation characteristics that may exist in the target regulatory area are relatively rich and relatively widely distributed. The influence of the collection quality and the external interference on the determination result of the deformation risk coefficient is relatively heavy in the evaluation process of the deformation risk coefficient determined for it, and it is necessary to evaluate whether the slope deformation of the associated analysis area is greatly different from it. The reliability degree of the determination of the deformation risk coefficient made for the target regulatory area is reflected from this side. The associated risk difference parameter = the standard deviation of the deformation risk coefficient of each associated analysis area of the target regulatory area / the average of the deformation risk coefficient of each associated analysis area of the target regulatory area. The risk assessment reliable coefficient and the associated risk difference parameter after normalization processing are in a negative correlation relationship;

[0092] For a single target supervision area, if the target supervision area is in the second type of risk characteristic state, the preliminary determination of the possible existing slope deformation characteristics of the target supervision area is relatively less, and the influence of the collection quality and external interference on the determination result of the deformation risk coefficient is relatively light. The reliability of the determination of the deformation risk coefficient of the target supervision area can be determined by the mutual confirmation between the slope deformation characteristics identified by the key analysis images of the target supervision area, the risk coverage parameter = the number of categories of slope deformation characteristics existing in each key analysis image of the target supervision area / the number of categories of different slope deformation characteristics existing in each key analysis image of the target supervision area, and the risk assessment reliability coefficient is negatively related to the risk coverage parameter after normalization.

[0093] Specifically, if the regional risk difference index of the target supervision area is greater than the preset regional risk difference index, it is determined that the optimization analysis is performed for the target supervision area.

[0094] The regional risk difference index is determined according to the deformation risk coefficients of the target supervision areas in the associated evaluation range.

[0095] Wherein, the execution priority of the corresponding target supervision area is determined according to the deformation risk coefficient of each target supervision area, the higher the execution priority, the more priority the regional verification is given, and the regional verification mode includes but is not limited to manual on-site verification and unmanned aerial vehicle inspection, for a single target supervision area, the execution priority is positively related to the deformation risk coefficient, the regional risk difference index = the average value of the difference between the deformation risk coefficient of the target supervision area and the deformation risk coefficients of the target supervision areas in the associated evaluation range / the deformation risk coefficient of the target supervision area, the value of the preset regional risk difference index can be determined by the user according to the actual working scene, for example, the user can set it according to the slope supervision record, the higher the user's requirement for the reliability of the deformation evaluation process of the target supervision route corresponding slope area and the image analysis efficiency, the smaller the value of the preset regional risk difference index, a method for determining the value of the preset regional risk difference index is provided, the slope supervision record for which the optimization analysis is performed is recorded as an optimization reference record, and the average value of the regional risk difference index of the target supervision area in the optimization reference record that meets the user's requirement for the reliability of the deformation evaluation process of the target supervision route corresponding slope area and the image analysis efficiency is recorded as the preset regional risk difference index.

[0096] Specifically, when performing the execution optimization analysis, the execution priority of each target regulatory region in the associated assessment range of the target regulatory region is adjusted in an increasing manner according to the risk association parameter and the risk assessment reliability coefficient.

[0097] The increasing value of the execution priority is in a positive correlation with the risk association parameter and the risk assessment reliability coefficient, respectively.

[0098] The increasing value of the execution priority is in a positive correlation with the optimization index for any target regulatory region in the associated assessment range of a single target regulatory region, and the optimization index is the product of the risk association parameter and the risk assessment reliability coefficient.

[0099] Referring to Figure 4 The present application also provides a road slope deformation early warning system based on image recognition, which comprises:

[0100] An initial analysis module is configured to perform initial risk analysis on each monitoring and analysis image to determine an initial risk coefficient, and determine key analysis images of each target regulatory region according to the initial risk coefficient.

[0101] A key evaluation module is connected to the initial analysis module and configured to determine a risk feature state of each target regulatory region according to a key distribution proportion index and an initial risk evaluation index, and determine whether to perform regional association analysis or risk analysis on each target regulatory region according to the risk feature state.

[0102] An association analysis module is connected to the key evaluation module and configured to determine an association analysis region based on a geological interference index or a geological reference index according to a reference risk distribution index of the target regulatory region.

[0103] A risk analysis module is connected to the key evaluation module and the association analysis module, and configured to determine a deformation risk coefficient of the target regulatory region according to a slope deformation feature existing in the key analysis image, and set a risk assessment reliability coefficient according to an associated risk difference parameter or a risk coverage parameter based on the risk feature state.

[0104] A regional execution module is connected to the risk analysis module and configured to determine an execution priority of each target regulatory region based on the deformation risk coefficient, and determine whether to perform execution optimization analysis on the target regulatory region according to a regional risk difference index to perform secondary adjustment on the execution priority.

[0105] The present application does not limit the specific structure of the initial analysis module, the key evaluation module, the correlation analysis module, the risk analysis module and the regional execution module, and the present application and each unit therein can be composed of a logic component, which includes a field programmable component, a computer or a microprocessor in the computer, but needs to be able to implement the above method.

[0106] So far, the technical solutions of the present application have been described in combination with the preferred embodiments shown in the drawings, but it is easy for those skilled in the art to understand that the protection scope of the present application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to the related technical features without departing from the principles of the present application, and the technical solutions after the changes or replacements will fall within the protection scope of the present application.

[0107] The above description is only the preferred embodiments of the present application and is not used to limit the present application; for those skilled in the art, the present application can have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for early warning of road slope deformation based on image recognition, characterized in that, include: The target regulatory routes are divided to obtain several target regulatory areas; An initial risk analysis is conducted on the monitoring and analysis images of each target regulatory area to determine the initial risk coefficient of each monitoring and analysis image, and the key analysis images of each target regulatory area are determined based on the initial risk coefficient. The risk characteristics of each target regulatory region are determined based on the key distribution ratio index and the initial risk assessment index, and regional correlation analysis or risk analysis is conducted for each target regulatory region based on the risk characteristics. When conducting regional correlation analysis, the correlation analysis area is determined based on the reference risk distribution index of the target regulatory area, or on the geological interference index or geological reference index. When conducting risk analysis, the deformation risk coefficient of the target regulatory area is determined based on the slope deformation characteristics present in the key analysis images, and the reliability coefficient of risk assessment is set based on the risk characteristic status according to the associated risk difference parameter or risk coverage parameter. The execution priority of each target regulatory region is determined based on the deformation risk coefficient, and the execution optimization analysis for the target regulatory region is determined based on the regional risk difference index, so as to make secondary adjustments to the execution priority.

2. The road slope deformation early warning method based on image recognition according to claim 1, characterized in that, The key analysis images are monitoring and analysis images of each target regulatory area where the initial risk coefficient is greater than the preset initial risk coefficient. The initial risk coefficient is determined based on the change risk ratio index of the monitored and analyzed images and the reference change risk index; The initial risk coefficient is positively correlated with both the change risk ratio index and the reference change risk index.

3. The road slope deformation early warning method based on image recognition according to claim 1, characterized in that, The risk characteristic states include a type I risk characteristic state and a type II risk characteristic state; If the key distribution ratio index of the target regulatory area is greater than the preset key distribution ratio index or the initial risk assessment index is greater than the preset initial risk assessment index, then the target regulatory area is determined to be in a risk characteristic state. If the key distribution ratio index of the target regulatory area is less than or equal to the preset key distribution ratio index and the initial risk assessment index is less than or equal to the preset initial risk assessment index, then the target regulatory area is determined to be in a Class II risk characteristic state.

4. The road slope deformation early warning method based on image recognition according to claim 3, characterized in that, If a target regulatory area is in a state of risk characteristics, then it is determined that a regional correlation analysis should be performed on the target regulatory area to determine the risk correlation parameters of other target regulatory areas within the target regulatory route to the target regulatory area. The correlation analysis area is the target regulatory area where the risk correlation parameter is greater than the preset risk correlation parameter.

5. The road slope deformation early warning method based on image recognition according to claim 4, characterized in that, The method for setting the risk-related parameters is determined based on the reference risk distribution index of the target regulatory area that is in a state of risk characteristic of a certain type, wherein, If the reference risk distribution index of a target regulatory area is greater than the preset reference risk distribution index, then the risk correlation parameters of each target regulatory area to that target regulatory area are determined according to the geological intervention index. If the reference risk distribution index of a target regulatory area is less than or equal to the preset reference risk distribution index, then the risk correlation parameters of each target regulatory area with respect to that target regulatory area are determined based on the geological comparison index.

6. The road slope deformation early warning method based on image recognition according to claim 5, characterized in that, If a target regulatory region is in a Class II risk characteristic state, or if a target regulatory region is in a Class I risk characteristic state and regional correlation analysis has been completed, then a risk analysis will be conducted for that target regulatory region. Deformation feature identification is performed on each key analysis image to determine the deformation risk coefficient of the target regulatory area; The verification strategy is determined based on the risk characteristics of the target regulatory area in order to determine the reliability coefficient of the risk assessment of the target regulatory area.

7. The road slope deformation early warning method based on image recognition according to claim 6, characterized in that, Assuming the risk assessment is complete, a reliability coefficient for the risk assessment is set for each target monitoring area. If a target regulatory area is in a state of risk characteristics, the risk assessment reliability coefficient of the target regulatory area shall be determined based on the associated risk difference parameter. If the target regulatory area is in a Class II risk characteristic state, the risk assessment reliability coefficient of the target regulatory area shall be determined according to the risk coverage parameter. The risk assessment is completed when the deformation risk coefficient of the target regulatory area is determined.

8. The road slope deformation early warning method based on image recognition according to claim 1, characterized in that, If the regional risk difference index of the target regulatory area is greater than the preset regional risk difference index, then it is determined that an execution optimization analysis should be performed for the target regulatory area. The regional risk difference index is determined based on the risk assessment reliability coefficient of each target regulatory region within the scope of the associated assessment.

9. The road slope deformation early warning method based on image recognition according to claim 8, characterized in that, When performing execution optimization analysis, the execution priority of each target regulatory region within the scope of the correlation assessment of the target regulatory region is increased and adjusted according to the risk correlation parameters and the risk assessment reliability coefficient. The increase in execution priority is positively correlated with both the risk-related parameter and the risk assessment reliability coefficient.

10. An early warning system applying the image recognition-based road slope deformation early warning method according to any one of claims 1 to 9, characterized in that, include: The initial analysis module is used to perform initial risk analysis on each monitoring and analysis image to determine the initial risk coefficient, and to determine the key analysis images for each target regulatory area based on the initial risk coefficient; The key assessment module, which is connected to the initial analysis module, is used to determine the risk characteristic status of each target regulatory area based on the key distribution ratio index and the initial risk assessment index, and to determine whether to conduct regional correlation analysis or risk analysis for each target regulatory area based on the risk characteristic status. The correlation analysis module, which is connected to the key assessment module, is used to determine the correlation analysis area based on the reference risk distribution index of the target regulatory area, or the geological interference index or geological comparison index. The risk analysis module is connected to the key assessment module and the correlation analysis module respectively. It is used to determine the deformation risk coefficient of the target regulatory area based on the slope deformation characteristics present in the key analysis image, and to set the risk assessment reliability coefficient based on the risk characteristic status according to the correlation risk difference parameter or risk coverage parameter. The regional execution module, which is connected to the risk analysis module, is used to determine the execution priority of each target regulatory region based on the deformation risk coefficient, and to determine whether to conduct execution optimization analysis for the target regulatory region based on the regional risk difference index, so as to make secondary adjustments to the execution priority.

Images