A project risk management method and system

Abstract

The present application relates to the field of building engineering digital management, and more particularly to a project risk management method and system, the method comprising: based on the real-time surface image of the wall and the reference image of the marked position which are periodically patrolled by the unmanned aerial vehicle; the real-time surface image of the wall is divided into at least one brightness area based on clustering analysis according to the ratio matrix of the pixel value of the real-time surface image and the reference image; for each brightness area, the pixel contrast is adjusted and the texture evaluation value is calculated according to each magnification coefficient starting from the initial magnification coefficient and increasing by a step, the risk coefficient of the brightness area is determined according to the texture gain change rate of the continuous texture evaluation value, and the crack risk of the building project is judged to complete the project risk management evaluation. The present application separates the noise crack by the ratio matrix and the profile coefficient clustering, and then locks the optimal magnification coefficient in a closed loop by the texture gain change rate driven by the gray run entropy value, highlights the high contrast of the crack, and suppresses the background artifacts.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of digital management of construction projects. In particular, it relates to a project risk management method and system. BACKGROUND

[0002] The essence of construction project risk management is to systematically identify, assess and resolve the uncertainty of quality, safety, progress and funding within the entire life cycle of the project at a controllable cost, ultimately ensuring that the building is delivered on time, with quality, within budget and long-term safe operation. By using digital means such as BIM, Internet of Things and drones, the uncertainty of safety and cost is identified, quantified and addressed in advance at the planning, construction and operation stages, and the project risk management of the building is carried out in order to identify and resolve the uncertainty of safety, quality, progress and funding in advance at a controllable cost within the entire life cycle, to avoid cracks and other diseases evolving into structural failure or huge maintenance, and to ensure personnel safety, prolong the life of the building and maximize investment efficiency.

[0003] Among them, building exterior wall cracks, as a common and significant structural disease, pose higher requirements for project risk management. As a common and significant structural disease, building exterior wall cracks can reduce bearing capacity, accelerate material degradation, cause leakage and energy consumption, and bring high maintenance and legal liability risks; therefore, accurate detection and quantitative assessment of cracks have become a key strategic decision in project risk management that determines safety margin and life cycle cost.

[0004] However, in actual operation, the shooting angle of the drone is unstable due to factors such as air flow and attitude control error, and the image is often accompanied by motion blur, uneven lighting and noise interference; if used directly for crack identification, wall texture and stains are easily misjudged as cracks, causing distortion of risk level and affecting subsequent maintenance decisions and project progress and cost control. SUMMARY

[0005] To solve the problem that the image is blurred, unevenly lit and noisy due to the influence of air flow and attitude error on the drone, and if used directly for crack identification, texture stains will be misjudged, causing distortion of risk level and interfering with maintenance decisions, the present application provides solutions in the following aspects.

[0006] In a first aspect, a project risk management method comprises: capturing real-time surface images of a wall and reference images of marked positions based on regular UAV cruising; dividing the real-time surface images of the wall into at least one brightness region based on clustering analysis on a ratio matrix of pixel values of the real-time surface images and the reference images; for each brightness region, increasing the magnification coefficient by a step from an initial magnification coefficient, adjusting the pixel contrast and calculating the texture evaluation value according to each magnification coefficient, determining whether to stop the growth of the magnification coefficient according to the texture gain change rate of the continuous texture evaluation values, and retaining the texture evaluation value corresponding to the previous step as the risk coefficient of the brightness region after the growth is terminated; judging whether the construction project has structural risks based on the risk coefficient of the wall, and completing the project risk management evaluation.

[0007] By constructing a pixel value ratio matrix of the real-time surface images and the reference images, and based on the adaptive clustering algorithm driven by the contour coefficient, the illumination noise and the crack signal are effectively separated in the space-gray joint domain; for each brightness region, the texture gain change rate is closed-loop controlled, that is, the texture gain change rate derived from the gray run matrix entropy value is calculated in real time during the process of increasing the magnification coefficient by a step, and the optimal magnification coefficient is automatically locked when the change rate becomes non-positive, so as to balance between the high-contrast highlighting of the crack and the low-distortion suppression of the background; finally, the locked texture evaluation value is mapped to a quantitative risk coefficient, which is used to determine the structural risks and complete the whole process, digitization and low false alarm project risk management from UAV inspection to maintenance decision.

[0008] Preferably, the ratio matrix of the pixel values comprises the following steps:

[0009] The real-time surface images and the reference images are aligned, the ratio of the pixel values of each corresponding position pixel point in the aligned real-time surface images and reference images is calculated, and the ratio matrix is obtained.

[0010] Preferably, the real-time surface images of the wall are divided into at least one brightness region based on clustering analysis, and the steps comprise:

[0011] The ratio matrix is clustered to obtain a plurality of clustering clusters, the contour coefficients between the clustering clusters are calculated, and the number of clustering clusters is determined based on the contour coefficients, wherein one clustering cluster corresponds to one brightness region;

[0012] In response to the contour coefficient being greater than the contour coefficient threshold, the number of clustering clusters is increased by 1, and the contour coefficients of the adjusted clustering clusters are recalculated until the contour coefficient is less than or equal to the contour coefficient threshold, then the clustering is completed, and the clustering result is output.

[0013] Preferably, the calculation method of the contour coefficient comprises:

[0014] Taking any one of the cluster as a target cluster, calculating the average of the absolute deviation of all pixel ratio in the target cluster from the average of all ratios in the cluster, and taking 1 minus the average as the dispersion degree of the target cluster, and selecting the maximum dispersion degree in all clusters as the profile coefficient.

[0015] Preferably, the process of determining whether to stop the increase of the amplification coefficient comprises:

[0016] Taking the average of the brightness region as a reference, calculating the adjusted pixel value of each pixel in the brightness region, replacing the original pixel value of the brightness region with the adjusted pixel value to obtain a marking matrix, and traversing all pixels in the marking matrix according to the direction and displacement to obtain a gray run matrix, and selecting the maximum value of the entropy value of the gray run matrix in each direction as the texture evaluation value.

[0017] In response to the texture evaluation value being greater than or equal to a preset threshold, the above steps are iterated until the texture evaluation value is less than the preset threshold, and then the amplification coefficient is stopped from increasing.

[0018] By adaptively adjusting the pixel value based on the average of the brightness region to generate a marking matrix, and then traversing in multiple directions to construct a gray run matrix, and then taking the maximum value of the entropy value in each direction as the texture evaluation value, the complexity of the crack texture is quantified; when the texture evaluation value is not lower than the preset threshold, the amplification coefficient is iterated, and the iteration is terminated when the texture evaluation value is lower than the threshold, so that the optimal enhancement intensity is dynamically approximated, and over-amplification of noise and artifacts is avoided, and the accuracy and robustness of subsequent crack identification are ensured.

[0019] Preferably, the calculation method of the adjusted pixel value comprises:

[0020] Taking the pixel value corresponding to any one of the pixels in the brightness region as a target pixel value, calculating the relative deviation of the target pixel value from the average of the pixel value sequence of all pixels in the brightness region, calculating the ratio between the relative deviation and the average of the pixel value sequence of all pixels in the brightness region for normalization, and taking the result after normalization and amplification by the amplification coefficient and adding 1 to form a gain factor, and taking the result after multiplying the gain factor by the target pixel value to the integer part of to obtain the adjusted pixel value after contrast enhancement.

[0021] By calculating the normalized deviation of the gray value of any pixel in the brightness region from the average of the region, and generating a gain factor weighted by the amplification coefficient, and then limiting the amplitude and taking the integer part, local adaptive contrast enhancement is realized, the crack highlight feature is highlighted, the background texture is suppressed, and gray overflow is avoided, so that the crack detectability is improved while the overall image fidelity is maintained.

[0022] Preferably, the grayscale run-length matrix is ​​constructed as follows: several directions are preset, and all pixels are traversed in each direction in row or column order. If the grayscale value of the current pixel is... and along Continuous direction The grayscale value of each pixel is Then record one trip. After the traversal is complete, all directions are obtained. Sets, construct a two-dimensional matrix , representing gray levels The column represents the run length. The collected travel itineraries The frequency of occurrence is filled into the corresponding positions in the two-dimensional matrix to obtain the grayscale run-length matrix.

[0023] Preferably, the risk coefficient is calculated in the following ways:

[0024] Based on the increasing sequence of magnification coefficients, texture evaluation values ​​are obtained sequentially according to the calculation method of adjusting pixel values. With any texture evaluation value as the target value, the ratio of the target value to the previous texture evaluation value and the difference between the previous texture evaluation value and the ratio of the two previous texture evaluation values ​​of the target value are calculated to obtain the risk coefficient of the target value.

[0025] By driving continuous sampling of texture evaluation values ​​through an increasing sequence of amplification coefficients, and constructing a risk coefficient based on the ratio difference between adjacent three values, the marginal change inflection point of texture gain is captured in real time during the iteration process. When the response value changes from positive to negative, the increment is immediately terminated and the optimal amplification coefficient is locked to avoid noise amplification or crack edge distortion caused by over-enhancement, thus achieving a balance between adaptive contrast adjustment and robust crack recognition.

[0026] Preferably, the determination of whether a building project has structural risks includes:

[0027] When the risk coefficient is less than or equal to the preset risk threshold, there are no wall cracks; conversely, when it is greater than the preset wind direction threshold, wall cracks appear.

[0028] Secondly, a project risk management system includes a processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement the aforementioned project risk management method.

[0029] The present invention has the following effects:

[0030] 1. The application constructs a pixel value ratio matrix of real-time surface image and reference image, and separates the illumination noise and crack signal in the space-gray joint domain based on the adaptive k-means clustering algorithm driven by the contour coefficient, so that the crack detection false alarm rate is reduced, and the false crack marking caused by motion blur, uneven exposure or texture stains is significantly inhibited.

[0031] 2. The application introduces the gray run matrix entropy value as the texture complexity measure in each brightness area, and quantifies the marginal benefit of crack details and background texture in the enhancement process by taking the texture gain change rate as the feedback quantity, and implementing the closed-loop control of successive incremental-termination on the amplification coefficient, so that when the texture gain change rate becomes non-positive, the previous gear amplification coefficient is locked as the optimal coefficient, so that the background noise and overshoot artifacts are effectively inhibited while the high contrast of the crack is highlighted, and the balance between adaptive enhancement and robustness is achieved. BRIEF DESCRIPTION OF DRAWINGS

[0032] Figure 1 It is a method flow chart of steps S1-S4 in the project risk management method of the embodiment of the application.

[0033] Figure 2 It is a structural block diagram of the project risk management system of the embodiment of the application. DETAILED DESCRIPTION

[0034] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are part of the embodiments of the application, rather than all the embodiments of the application.

[0035] Specific scenario: cracks will appear on the building body for a long time, and the cracks on the building body are mainly caused by the imbalance of multiple stresses: uneven settlement of the foundation causes the wall to bear shear force, and inclined cracks, horizontal cracks between windows, or vertical cracks are prone to occur on the soft soil foundation; the difference between thermal expansion and contraction of concrete and masonry materials caused by temperature changes, especially the main tensile stress between the top wall and the roof panel caused by temperature difference, will form an eight-shaped or horizontal crack when the tensile strength exceeds the material tensile strength; the shrinkage stress caused by the dry shrinkage deformation of concrete itself as a non-homogeneous brittle material and the hydration heat of cement (such as high water-cement ratio, insufficient curing, etc.) will cause through cracks or diagonal cracks at the corners of the floor.

[0036] Reference Figure 1 A project risk management method includes steps S1-S4, and specifically as follows:

[0037] S1: Based on the real-time surface image of the wall body and the reference image of the marked position photographed by the regular cruise of the unmanned aerial vehicle.

[0038] The unmanned aerial vehicle is preset with a flight route by a ground station, and continuously photographs the wall surface at a fixed cycle. The real-time image is returned by the on-board camera. The first flight or the image without cracks is recorded by GNSS / RTK and written into EXIF as a "reference image of the marked position" and stored in the database. The real-time surface image is obtained on the same flight path every time, so as to realize one-to-one correspondence between the real-time image and the reference image.

[0039] S2: The real-time surface image of the wall is divided into at least one brightness area based on cluster analysis according to the ratio matrix of the pixel values of the real-time surface image and the reference image.

[0040] The ratio matrix of the pixel values includes the following steps:

[0041] The real-time surface image and the reference image are aligned, the ratio of the pixel values of each corresponding position pixel point in the aligned real-time surface image and the reference image is calculated, and the ratio matrix is obtained.

[0042] For example, in this embodiment, the alignment of the reference image and the real-time image is completed by OpenCV. It can also be achieved by mutual information or phase correlation registration algorithm, SIFT / SUR feature point matching and RANSAC transformation matrix estimation, deep learning homography network, or any technical means that can realize the alignment function.

[0043] The sub-pixel level registration is performed on the real-time surface image and the reference image of the marked position to eliminate the geometric misalignment caused by the attitude drift of the unmanned aerial vehicle or the difference in the shooting angle. After the registration is completed, the ratio of the real-time gray value and the reference gray value is calculated pixel by pixel to obtain the dimensionless ratio matrix, so as to convert the absolute illumination difference into the relative ratio difference, thereby providing the data input reflecting only the crack characteristics and not affected by the illumination fluctuation for the subsequent cluster analysis.

[0044] The ratio matrix is clustered to obtain a plurality of cluster clusters, the contour coefficients between the cluster clusters are calculated, and the number of cluster clusters is determined based on the contour coefficients, wherein one cluster cluster corresponds to one brightness area.

[0045] In response to the contour coefficient being greater than the contour coefficient threshold, the number of cluster clusters is increased by 1, and the contour coefficient of the adjusted cluster cluster is recalculated, until the contour coefficient is less than or equal to the contour coefficient threshold, then the clustering is completed, and the clustering result is output.

[0046] For example, the preset contour coefficient threshold is 0.6, which can be adjusted according to the specific situation.

[0047] The calculation method of the contour coefficient includes:

[0048] Taking any cluster as the target cluster, calculate the average of the absolute deviations between the ratios of all pixels in the target cluster and the mean of all ratios in the cluster. Subtract the average from 1 to obtain the dispersion of the target cluster, and select the maximum dispersion among all clusters as the contour coefficient.

[0049] Specifically, the contour coefficients satisfy the following polynomial:

[0050] ;

[0051] ;

[0052] In the formula, Represents the profile coefficient. Indicates the first The degree of dispersion of each cluster, Indicates the first In the cluster, the th The ratio of each location point Indicates the first The mean of all ratios in each cluster Indicates the first Each cluster contains the total number of locations.

[0053] In other words, the smaller the contour coefficient, the more concentrated the pixel ratio within the cluster is at the mean, the more compact the brightness partition is, and the better it represents the same lighting conditions, thus effectively suppressing noise and texture interference. By adaptively determining the optimal number of clusters without manually setting thresholds, a clear and reliable brightness region segmentation basis is provided for subsequent crack identification.

[0054] S3: For each brightness area, increase the magnification factor step by step from the initial magnification factor, adjust the pixel contrast and calculate the texture evaluation value according to each magnification factor, and determine whether to stop the increase of the magnification factor based on the texture gain change rate of the continuous texture evaluation values. After stopping the increase, retain the texture evaluation value corresponding to the previous value as the risk factor of the brightness area.

[0055] Using the average value of the brightness region as a benchmark, the adjusted pixel value of each pixel in the brightness region is calculated. The original pixel value of the brightness region is replaced with the adjusted pixel value to obtain the marking matrix. The marking matrix is ​​then traversed through all pixels according to direction and displacement to obtain the grayscale run-length matrix. The maximum value of the entropy of the grayscale run-length matrix in each direction is selected as the texture evaluation value.

[0056] When the texture evaluation value is greater than or equal to a preset threshold, the above steps continue to be iterated until the texture evaluation value is less than the preset threshold, at which point the amplification factor growth stops.

[0057] For example, the preset threshold is 0, but it can be adjusted according to specific circumstances.

[0058] The methods for calculating adjusted pixel values ​​include:

[0059] Using the pixel value corresponding to any pixel within the brightness region as the target pixel value, calculate the relative deviation between the target pixel value and the mean of the pixel value sequence of all pixels within the brightness region. Calculate the ratio between this relative deviation and the mean of the pixel value sequence of all pixels within the brightness region for normalization. Amplify the normalized result using an amplification factor and add 1 to form a gain factor. Multiply the gain factor by the target pixel value and round the result using a range limiting function. This yields the adjusted pixel values ​​after contrast enhancement.

[0060] Specifically, the pixel values ​​are adjusted to satisfy the following relationship:

[0061] ;

[0062] In the formula, express Adjust the pixel value of the position point. Indicates the range is The floor function, This is the magnification factor. express The original pixel value of the location point, express The sequence of pixel values ​​for all pixels within the brightness region of the location point. This represents the mean function.

[0063] Within the brightness area, As a normalization benchmark, when the relative deviation is positive, pixel grayscale is amplified to highlight the bright features of potential cracks; when the relative deviation is negative, grayscale is suppressed to weaken background texture. The initial value of the amplification factor is set to 1, and... Within the range As the step size increases, the larger the magnification factor, the more significant the difference in grayscale between the crack and the background, resulting in higher detection sensitivity. Conversely, the smaller the magnification factor, the less noise amplification and the stronger the robustness. The final crack determination requires a secondary verification using texture entropy and contour coefficients to eliminate interference from uneven lighting or stains, achieving a balance between enhancement effect and reliability.

[0064] By normalizing pixel deviations based on the local brightness average, and then weighting the gain using an amplification factor, dark areas become darker and bright areas become brighter, before finally limiting the gain. This achieves low-distortion, high-contrast enhancement of crack details.

[0065] The grayscale run-length matrix is ​​constructed as follows: several directions are preset, and all pixels are traversed in each direction in row or column order. If the grayscale value of the current pixel is... and along Continuous direction The grayscale value of each pixel is Then record one trip. After the traversal is complete, all directions are obtained. Sets, construct a two-dimensional matrix , representing gray levels The column represents the run length. The collected travel itineraries The frequency of occurrence is filled into the corresponding positions in the two-dimensional matrix to obtain the grayscale run-length matrix.

[0066] For example, several directions are preset as follows: 0°, 45°, 90°, and 135°, which can be adjusted according to specific circumstances.

[0067] The methods for calculating the risk factor include:

[0068] Based on the increasing sequence of magnification coefficients, texture evaluation values ​​are obtained sequentially according to the calculation method of adjusting pixel values. With any texture evaluation value as the target value, the ratio of the target value to the previous texture evaluation value and the difference between the previous texture evaluation value and the ratio of the two previous texture evaluation values ​​of the target value are calculated to obtain the risk coefficient of the target value.

[0069] Specifically, the risk coefficient satisfies the following relationship:

[0070] ;

[0071] In the formula, Indicates the risk coefficient. Indicates the first Each texture evaluation value, Indicates the first Each texture evaluation value, Indicates the first Each texture evaluation value.

[0072] The marginal benefit of texture gain is monitored using a risk coefficient. When the risk coefficient is less than or equal to 0, it indicates that increasing the amplification factor no longer brings effective texture gain, and the system immediately terminates the iteration to avoid over-enhancement and noise amplification. After termination, the texture evaluation value corresponding to the previous effective amplification factor is directly assigned as the risk coefficient of the brightness area, completing the transition from pixel-level enhancement to region-level risk quantization, thereby providing reliable input for judgment.

[0073] S4: Determine whether there are structural risks in a building project based on the risk coefficient of the walls, and complete the project risk management assessment.

[0074] When the risk coefficient is less than or equal to the preset risk threshold, it is determined that the wall does not have a wall crack; otherwise, when the risk coefficient is greater than the preset risk threshold, it is determined that the wall has a wall crack.

[0075] In this embodiment, the risk threshold is 0.7, which can be adjusted according to specific conditions.

[0076] The present application also provides a project risk management system. Figure 2 As shown in the figure, the system comprises a processor and a memory, the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a project risk management method according to the first aspect of the present application is realized. The system also comprises a communication bus and a communication interface and other components familiar to those skilled in the art, the setting and function of which are known in the art, and therefore will not be described here.

[0077] It should be pointed out that for those skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are all within the protection scope of the present application. Therefore, the protection scope of the present application patent should be subject to the appended claims.

Claims

1. A project risk management method characterized by, The method comprises the steps of: capturing real-time surface images of the wall and reference images of the marked positions based on regular patrol of a UAV; dividing the real-time surface images of the wall into at least one brightness area based on cluster analysis according to a ratio matrix of pixel values of the real-time surface images and the reference images; for each brightness area, increasing the initial magnification coefficient by a step, adjusting the pixel values according to each magnification coefficient, calculating a texture evaluation value based on the adjusted pixel values, determining whether to stop the increase of the magnification coefficient according to the texture gain change rate of the continuous texture evaluation values, and retaining the risk coefficient of the brightness area as the difference between the ratio of the texture evaluation value corresponding to the previous step and the texture evaluation value of the previous two steps after the increase is terminated; wherein, taking the mean value of the brightness area as a reference, calculating the adjusted pixel value of each pixel in the brightness area, replacing the original pixel value of the brightness area with the adjusted pixel value to obtain a marking matrix, and traversing all pixels according to the direction and displacement to obtain a gray run length matrix, and selecting the maximum value of the entropy values of the gray run length matrix in each direction as the texture evaluation value; Taking a pixel value corresponding to any pixel point in the brightness region as a target pixel value, a relative deviation between the target pixel value and a mean value of a pixel value sequence of all pixel points in the brightness region is calculated, a ratio between the relative deviation and the mean value of the pixel value sequence of all pixel points in the brightness region is calculated, and the ratio is used for normalization processing. After the normalized result is amplified by a magnification coefficient and then added by 1, a gain factor is formed. The gain factor is multiplied by the target pixel value, and the result is rounded to by a value range limiting function, to obtain an adjusted pixel value after contrast enhancement. judging whether the building project has a structural risk based on the risk coefficient of the wall, and completing the project risk management evaluation.

2. The project risk management method of claim 1, wherein, The method comprises the steps of: aligning the real-time surface images and the reference images, calculating the ratio of the pixel values of each corresponding position pixel point in the aligned real-time surface images and reference images to obtain the ratio matrix.

3. The project risk management method of claim 1, wherein, The method comprises the steps of: clustering the ratio matrix to obtain a plurality of cluster clusters, calculating the contour coefficient between each cluster cluster, and judging the number of cluster clusters based on the contour coefficient, wherein one cluster cluster corresponds to one brightness area; in response to the contour coefficient being greater than the contour coefficient threshold, the number of cluster clusters is increased by 1, and the contour coefficient of the adjusted cluster cluster is recalculated, until the contour coefficient is less than or equal to the contour coefficient threshold, then the clustering is completed, and the clustering result is output.

4. The project risk management method of claim 3, wherein, The calculation method of the contour coefficient comprises: taking any cluster cluster as a target cluster cluster, calculating the average value of the absolute deviation of all pixel point ratios in the target cluster cluster from the average value of all ratios in the cluster cluster, taking 1 minus the average value as the dispersion degree of the target cluster cluster, and selecting the maximum dispersion degree in all cluster clusters as the contour coefficient.

5. The project risk management method of claim 1, wherein, The process of determining whether to stop the increase of the magnification coefficient comprises: in response to the texture evaluation value being greater than or equal to the preset threshold, the above steps are iterated until the texture evaluation value is less than the preset threshold, then the increase of the magnification coefficient is stopped.

6. The method of claim 1, wherein, The grayscale run-length matrix is ​​constructed as follows: several directions are preset, and all pixels are traversed in each direction in row or column order. If the grayscale value of the current pixel is... and along Continuous direction The grayscale value of each pixel is Then record one trip. After the traversal is complete, all directions are obtained. Sets, construct a two-dimensional matrix , representing gray levels The column represents the run length. The collected travel itineraries The frequency of occurrence is filled into the corresponding positions in the two-dimensional matrix to obtain the grayscale run-length matrix.

7. The method of claim 1, wherein, The calculation method of the risk coefficient comprises: according to the increasing sequence of the magnification coefficient, obtaining the texture evaluation value according to the calculation method of the adjusted pixel value in sequence, taking any texture evaluation value as a target value, calculating the difference between the ratio of the target value and the texture evaluation value before the target value and the ratio between the texture evaluation value before the target value and the texture evaluation value of the previous two steps, and obtaining the risk coefficient of the target value.

8. The method of claim 1, wherein, The method comprises the steps of: When the risk coefficient is less than or equal to the preset risk threshold, it is determined that the wall does not have a wall crack; otherwise, when the risk coefficient is greater than the preset risk threshold, it is determined that the wall has a wall crack.

9. A project risk management system, characterized by The application relates to a method for managing project risks, comprising the following steps: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the project risk management method according to any one of claims 1-8 is realized.

Images