Road production loss assessment method and system

Abstract

The invention relates to the technical field of loss assessment, in particular to a road property loss assessment method and system, and the method comprises the following steps: carrying out the regular inspection of a road property on a target road, so as to collect a road property damage image, and at the same time, collecting the indirect economic loss case data caused by the damage of the road property; extracting road production loss characteristics in the road production damage image by using a machine learning algorithm, and evaluating first road production loss by using a direct loss evaluation model in combination with the road production monitoring data set; according to the indirect economic loss case data, evaluating second road production loss by adopting an indirect loss evaluation strategy of road production damage; and combining the first road production loss and the second road production loss to obtain the total road production loss of the target road, and repairing the damaged road production in time. According to the method, the loss of the value of the road production and the indirect economic loss caused by the damage of the road production are considered, so that the road production loss can be comprehensively and accurately evaluated.

Description

Technical Field

[0001] The present invention relates to the technical field of loss assessment, and in particular to a method and system for assessing highway property losses. Background Art

[0002] During the operation of highways, damage to highway property often occurs due to the influence of various factors such as vehicle load and natural environment. In order to ensure highway traffic safety, maintain the integrity of highway facilities and ensure highway traffic capacity, it is necessary to repair the highway in a timely manner, and it is very necessary to accurately assess the degree of damage to highway property and the required repair costs.

[0003] At present, there are some methods for evaluating the loss of highway property, such as the replacement cost method, the market value method and the income present value method. Among them, the replacement cost method evaluates the loss based on the cost required to repurchase or rebuild highway facilities with the same or similar functions; the market value method evaluates the loss based on the transaction price of similar highway property facilities in the market; the income present value method evaluates the loss by evaluating the potential loss of income caused by the damage to highway facilities. However, these methods are mainly aimed at the evaluation of direct economic losses, such as repair costs, equipment replacement costs and highway tolls, etc., ignoring the loss of the value of highway property itself, and not taking into account the direct and indirect economic losses caused by the damage to highway property at the same time, resulting in the evaluation results of the existing highway property loss evaluation methods being not comprehensive and accurate. As the status of highways in social and economic development continues to improve, the indirect economic losses caused by the damage to highway property are becoming increasingly prominent.

[0004] Therefore, it is necessary to develop a new method that can comprehensively and accurately assess highway property losses to overcome the shortcomings of existing technologies. Summary of the invention

[0005] In view of the defects in the prior art, the present invention provides a method and system for evaluating highway property losses.

[0006] In order to achieve the above-mentioned purpose, in the first aspect, the present invention provides a method for assessing the loss of highway property, the method comprising the following steps: conducting regular inspections of highway property on a target highway to collect images of highway property damage, and at the same time collecting case data of indirect economic losses caused by highway property damage; using a machine learning algorithm to extract the characteristics of highway property loss in the images of highway property damage, and then using a direct loss assessment model in combination with a highway property monitoring data set to assess the first highway property loss; based on the indirect economic loss case data, using an indirect loss assessment strategy for highway property damage to assess the second highway property loss; combining the first highway property loss and the second highway property loss to obtain the total highway property loss of the target highway, and repairing the damaged highway property in a timely manner. The present invention takes into account the loss of the value of the highway property itself and the indirect economic losses caused by the damage to the highway property, and can thus comprehensively and accurately assess the loss of highway property.

[0007] Optionally, the indirect economic loss case data includes the road property loss characteristic data, direct loss amount, traffic flow data, ambient temperature data, road surface vibration speed data and indirect loss amount, and the road property loss characteristic data, the direct loss amount, the traffic flow data, the ambient temperature data and the road surface vibration speed data correspond to five indirect loss assessment indicators of comprehensive loss characteristics, direct loss, traffic flow, ambient temperature and road surface vibration speed in sequence; The method of regularly inspecting road properties on the target road to collect images of road property damage and collecting case data of indirect economic losses caused by road property damage includes the following steps: Conduct regular inspections on road assets on the target highway to collect original images of road asset damage; Preprocessing the original image of the highway damage to obtain a clear image, and marking the acquisition position and acquisition time on the clear image to obtain the highway damage image; Query case highways where road property is damaged, and collect case data on indirect economic losses caused by road property damage on the case highways.

[0008] Furthermore, the present invention provides data support for accurate assessment of direct losses by preprocessing images and marking time and location information, and can provide a data basis for more accurate assessment of indirect economic losses by collecting case data on indirect economic losses.

[0009] Optionally, the extracting road property loss features in the highway road property damage image using a machine learning algorithm, and then evaluating the first road property loss using a direct loss assessment model in combination with a highway road property monitoring data set comprises the following steps: Collect historical road damage images to build a road damage dataset; Based on the road property damage data set, a road property loss identification model is constructed using a machine learning algorithm; Extracting road property loss features in the highway road property damage image using the road property loss recognition model, wherein the road property loss features include road property type and loss level; According to the road property loss characteristics and in combination with the highway road property monitoring data set, the direct loss assessment model is used to assess the first road property loss.

[0010] Furthermore, the present invention extracts road property types and loss levels based on machine learning algorithms, which can not only improve the accuracy of direct loss assessment, but also improve the efficiency of assessment.

[0011] Optionally, the step of constructing a road property loss identification model based on the road property damage dataset using a machine learning algorithm comprises the following steps: The VGG-16 network in Faster R-CNN is replaced with ResNet-50, and a recursive feature pyramid is introduced between the ResNet-50 network and the region proposal network and the region of interest pooling layer to obtain an image recognition network. The image recognition network is trained and verified using a road property damage dataset, thereby obtaining the road property loss recognition model.

[0012] Furthermore, the present invention uses ResNet-50 and recursive feature pyramid to improve Faster R-CNN to improve the recognition accuracy of Faster R-CNN for road property loss characteristics, thereby improving the accuracy of direct loss assessment.

[0013] Optionally, the highway monitoring dataset is constructed using all historical highway road damage images of the target highway; The step of using the direct loss assessment model to assess the first road property loss based on the road property loss characteristics and in combination with the highway road property monitoring data set comprises the following steps: According to the highway property monitoring data set, determining the last repair time of the highway property in the highway property damage image, and then determining the value attenuation coefficient of the highway property; The first road property loss is assessed using the direct loss assessment model based on the road property loss characteristics and the value attenuation coefficient.

[0014] Furthermore, the present invention determines whether the road property has been repaired based on the highway road property monitoring data set to determine the value attenuation coefficient of the road property, and then uses the direct loss assessment model to assess the direct loss, which can improve the accuracy and reliability of the assessment.

[0015] Optionally, the direct loss assessment model satisfies the following relationship:

[0016] in, is the first road property loss; I is the number of road property types; J is the number of loss levels; K is the number of road properties with the jth loss level under the ith road property type; The amount required to repair the kth road property of the jth loss level under the ith road property type; It is the value attenuation coefficient corresponding to the kth road property in the jth loss level of the i-th road property type.

[0017] Furthermore, the direct loss assessment model improves the accuracy of the assessment by introducing a value attenuation coefficient to take into account the loss of the value of the highway property itself. The model is simple to calculate and can improve the efficiency of the assessment.

[0018] Optionally, the step of evaluating the second road property loss using an indirect loss evaluation strategy for road property damage according to the indirect economic loss case data comprises the following steps: Constructing a case data set using the indirect economic loss case data of a plurality of the case highways; Acquire comprehensive loss characteristic data of different case roads according to the case data set, and then construct a first data set by combining the direct loss amount, traffic flow data, ambient temperature data, and road surface vibration speed data of the case roads; According to the first data set, a plurality of principal components are obtained using a principal component analysis method; For each set of indirect loss assessment indicator data in the first data set, a corresponding set of principal component values ​​is calculated, and the principal component values ​​and the corresponding indirect loss amounts are combined into a set of queryable case data; Construct a queryable case dataset using multiple sets of queryable case data; The indirect loss assessment index data of the target highway is obtained, and a plurality of reference cases are searched in the queryable case data set, and then the indirect economic loss assessment model is used to assess the second road property loss.

[0019] Furthermore, after constructing the first data set, the present invention uses principal component analysis to perform data dimensionality reduction to remove the coupling between indicators, and then constructs a queryable case data set, and then uses the indirect economic loss assessment model to assess the indirect losses, which is conducive to simplifying the assessment process and improving the accuracy of the assessment.

[0020] Optionally, the comprehensive loss characteristic satisfies the following relationship:

[0021] in, is the value of the comprehensive loss characteristic data, I is the number of road property types, J is the number of loss levels, is the number of road properties of the ith road property type, and K is the number of road properties with the jth loss level under the ith road property type.

[0022] Furthermore, the comprehensive loss characteristic data is calculated based on the road property type and its loss level, which can comprehensively reflect the damage to the highway property and calculate it to facilitate the indirect loss assessment.

[0023] Optionally, the step of obtaining the indirect loss assessment index data of the target highway, searching for multiple reference cases in the queryable case data set, and then using the indirect economic loss assessment model to assess the second road property loss comprises the following steps: Obtain indirect loss assessment index data of the target highway, and then calculate a set of principal component values, and record the set of principal component values ​​as a set of target principal component values; For each of the principal components, a principal component value closest to the target principal component value is found in the queryable case data set, and the principal component value is recorded as a reference principal component value; Taking a group of queryable case data where the reference principal component value is located as the referenceable case, and then obtaining a plurality of the referenceable cases; The cumulative variance contribution rate of each principal component is used as the principal component weight, and the principal component weight is normalized to obtain a normalized weight; According to the referenceable case and the normalized weight, the indirect economic loss assessment model is used to assess the second road property loss, and the indirect economic loss assessment model satisfies the following relationship:

[0024] in, is the second path loss, N is the number of the principal components, is the target principal component value of the nth principal component, is the reference principal component value of the nth principal component, is the normalized weight of the nth principal component, The indirect loss amount of the case where the reference principal component value of the nth principal component is located can be used as a reference.

[0025] Furthermore, the present invention determines reference cases based on the principal component values, determines normalized weights based on the cumulative variance contribution rate, and then uses the indirect economic loss assessment model to assess the indirect losses, which can improve the objectivity of the assessment, thereby improving the accuracy of the assessment, and also improve the efficiency of the assessment.

[0026] In the second aspect, the present invention provides a highway property loss assessment system, the highway property loss assessment system comprising: a data acquisition device, a data output device, a processor and a storage device, the storage device comprising a computer-readable storage medium, the computer-readable storage medium storing a computer program, the computer program comprising program instructions, the program instructions when executed by the processor enable the processor to implement a highway property loss assessment method provided by the present invention. The system is mutually compatible with the method provided by the present invention, which can not only stably run the method of the present invention and improve the efficiency of highway property loss assessment, but also enhance the practicality of the method provided by the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0028] Figure 1 A schematic diagram of a process flow of a highway property loss assessment method according to an embodiment of the present invention; Figure 2 A simplified structural diagram of an image recognition network according to an embodiment of the present invention; Figure 3 A schematic diagram of the framework of a highway property loss assessment system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0029] The specific embodiments of the present invention will be described in detail below. It should be noted that the embodiments described herein are only for illustration and are not intended to limit the present invention. In the following description, a large number of specific details are set forth in order to provide a thorough understanding of the present invention. However, it is obvious to those of ordinary skill in the art that these specific details do not need to be adopted to implement the present invention. In other examples, in order to avoid confusing the present invention, known circuits, software or methods are not specifically described.

[0030] Throughout the specification, references to "one embodiment," "an embodiment," "an example," or "an example" mean that a particular feature, structure, or characteristic described in conjunction with the embodiment or example is included in at least one embodiment of the present invention. Therefore, the phrases "in one embodiment," "in an embodiment," "an example," or "an example" appearing in various places throughout the specification do not necessarily all refer to the same embodiment or example. In addition, particular features, structures, or characteristics may be combined in one or more embodiments or examples in any suitable combination and / or subcombination. In addition, it should be understood by those of ordinary skill in the art that the figures provided herein are for illustrative purposes and that the figures are not necessarily drawn to scale.

[0031] It should be noted in advance that, in an optional embodiment, except for independent explanations, the same symbols or letters appearing in all formulas have the same meanings and values.

[0032] In an alternative embodiment, see Figure 1 The present invention provides a method for assessing highway property losses, the method comprising the following steps: S1. Conduct regular inspections on road properties on the target highway to collect images of road property damage, and collect case data on indirect economic losses caused by road property damage.

[0033] Wherein, step S1 specifically includes the following steps: S11. Perform regular inspections on road properties on the target road to collect original images of road property damage.

[0034] Specifically, in this embodiment, a drone carrying a high-definition camera is used to perform regular inspections on road properties on the target road, and to collect original images of road property damage. The models of the drone and the camera can be selected according to actual conditions.

[0035] S12, preprocessing the original image of the highway damage to obtain a clear image, and marking the acquisition position and acquisition time on the clear image to obtain the highway damage image.

[0036] Specifically, in this embodiment, preprocessing the original image of highway damage includes image deblurring and image noise reduction to obtain a clear image.

[0037] Furthermore, the acquisition location and acquisition time are marked on the clear image to obtain a damaged image of the highway property. The purpose of marking the acquisition location and acquisition time is to facilitate tracing the repair location and repair time after the road property is repaired, and it is also conducive to accurately assessing the direct economic losses.

[0038] S13. Query case highways with damaged road properties, and collect case data on indirect economic losses caused by the damaged road properties on the case highways.

[0039] Specifically, in this embodiment, case roads and corresponding indirect economic loss case data are obtained by searching on the Internet or consulting newspapers and periodicals, or real-time collection is performed for different roads. If indirect economic loss case data is collected in real time for different roads, the input-output model can be used to predict the amount of indirect losses. Although it is more complicated to use the input-output model to predict the amount of indirect losses, considering that using the input-output model to predict indirect economic losses is a routine operation for evaluating indirect economic losses, it will not be described in detail here.

[0040] More specifically, the indirect economic loss case data include road property loss characteristic data, direct loss amount, traffic flow data, ambient temperature data, pavement vibration speed data and indirect loss amount. The road property loss characteristic data, direct loss amount, traffic flow data, ambient temperature data and pavement vibration speed data correspond to the five indirect loss assessment indicators of comprehensive loss characteristics, direct loss, traffic flow, ambient temperature and pavement vibration speed respectively. These five indirect loss assessment indicators can reflect the characteristics of the highway more comprehensively and help improve the accuracy of indirect loss assessment.

[0041] Furthermore, the traffic flow data is the average daily traffic flow of the case highway in the quarter when the road property damage inspection was carried out; the ambient temperature data is the average daily ambient temperature of the case highway in the quarter when the road property damage inspection was carried out; and the road surface vibration speed data is the average daily vibration speed of the road surface in the quarter when the road property damage inspection was carried out. Because there are many types of road properties on a highway, for the same case highway, although there is only one data for road property loss characteristic data, direct loss amount, traffic flow data, ambient temperature data, and road surface vibration speed data, multiple sets of road property loss characteristic data can be collected.

[0042] S2. Use a machine learning algorithm to extract road property loss features in the highway road property damage image, and then use a direct loss assessment model in combination with the highway road property monitoring data set to assess the first road property loss.

[0043] The first road property loss is the direct economic loss. Step S2 specifically includes the following steps: S21. Collect historical road damage images to construct a road damage dataset.

[0044] Specifically, in this embodiment, the historical road damage images can be obtained by searching on the Internet, reading newspapers and periodicals, or by field collecting on different roads. After collecting the historical road damage images, it is necessary to mark the road damage type and loss level on the historical road damage images to facilitate subsequent model training. After marking the road damage type and loss level on the historical road damage images, the historical road damage images can be used to construct a road damage dataset.

[0045] More specifically, the road property type is the name of the damaged road property, such as the road surface, shoulder, guardrail, traffic sign, greening facilities and drainage facilities, etc. The road property type can be represented by different numbers. The loss level is a quantitative assessment of the degree of damage to the road property, which includes: minor losses, which have little impact on the function of the road property and do not require immediate repair or only simple repairs, and can be represented by the number 1; moderate losses, which have a certain impact on the function of the road property and need to be repaired as soon as possible to restore the road's traffic capacity, can be represented by the number 2; severe losses, which have a significant impact on the function of the road property and may endanger the safety of road users, and immediate measures need to be taken to repair it.

[0046] S22. Based on the road property damage dataset, a road property loss identification model is constructed using a machine learning algorithm.

[0047] Wherein, step S22 specifically includes the following steps: S221. Replace the VGG-16 network in Faster R-CNN with ResNet-50, and introduce a recursive feature pyramid between the ResNet-50 network and the region proposal network and the region of interest pooling layer to obtain an image recognition network.

[0048] Specifically, in this embodiment, Faster R-CNN includes a backbone network, a region candidate network, a region of interest pooling layer, and classification and position regression. The original Faster R-CNN uses the VGG-16 network as the backbone network, which has problems of deep network degradation and high training error rate. Therefore, this embodiment uses ResNet-50 to replace the VGG-16 network in Faster R-CNN, and then introduces a residual structure in Faster R-CNN to solve this problem.

[0049] In addition, in Faster R-CNN, whether it is the VGG-16 network or ResNet-50, the size of the feature map will gradually decrease with the convolution process, which will eventually make it difficult for Faster R-CNN to identify the tiny features of road property damage, and ultimately lead to the inability to accurately identify the type of road property and the level of loss. Therefore, in order to solve the problem of feature map scale reduction in Faster R-CNN, this embodiment introduces a recursive feature pyramid between the ResNet-50 network and the region candidate network of Faster R-CNN. Specifically, the output of the first residual block of each stage of ResNet-50 is modified using the information fed back by the recursive feature pyramid, and the output of the recursive feature pyramid is used as the input of the region candidate network and the pooling layer of the region of interest. After the modification, the output of the first residual block of each stage of ResNet-50 satisfies the following relationship:

[0050] in, Indicates that a convolution with a convolution kernel size of 1×1 is performed; To indicate a convolution with a kernel size of 1×1 and a stride of 2; is the input of the first residual block in each stage of ResNet-50; R is the information fed back by the recursive feature pyramid.

[0051] More specifically, ResNet-50 has 5 stages, stage0 to stage4. The resulting structure of the image recognition network is shown below: Figure 2 shown.

[0052] S222. Use the road property damage data set to train and verify the image recognition network, and then obtain the road property loss recognition model.

[0053] Specifically, in this embodiment, the road property damage data set is divided into a training set and a verification set to complete the training and verification of the image recognition network, and finally obtain a road property loss recognition model.

[0054] S23. Use the road property loss identification model to extract road property loss features in the highway road property damage image, where the road property loss features include road property type and loss level.

[0055] Specifically, in this embodiment, the present invention uses a road property loss identification model to accurately extract road property types and loss levels, thereby improving the accuracy and efficiency of direct loss assessment.

[0056] S24. According to the road property loss characteristics and in combination with the highway road property monitoring data set, the direct loss assessment model is used to assess the first road property loss.

[0057] Among them, this embodiment determines whether the road property has been repaired based on the highway property monitoring data set to determine the value decay coefficient of the road property, and then uses the direct loss assessment model to assess the direct loss, which can improve the accuracy and reliability of the assessment. Step S24 specifically includes the following steps: S241. Determine the last repair time of the road property in the highway property damage image based on the highway property monitoring data set, and then determine the value attenuation coefficient of the road property.

[0058] Specifically, in this embodiment, the highway monitoring data set is constructed using all historical highway property damage images of the target highway. Since the historical highway property damage images in the highway property monitoring data set are marked with the acquisition location and acquisition time, it is only necessary to compare the highway property damage image with the historical highway property damage image in the highway property monitoring data set to determine the last repair time of the road property in the highway property damage image, and then determine the value decay coefficient of the road property. The value decay coefficient satisfies the following relationship:

[0059] in, is the value decay coefficient, T is the expected use time of the road property, It is the time interval from the last repair of the road property to the current time. T needs to evaluate the maintenance degree of the road property based on the designed service life of the road property.

[0060] S242. According to the road property loss characteristics and the value attenuation coefficient, use the direct loss assessment model to assess the first road property loss.

[0061] Specifically, in this embodiment, the direct loss assessment model satisfies the following relationship:

[0062] in, is the first road property loss; I is the number of road property types; J is the number of loss levels; K is the number of road properties with the jth loss level under the ith road property type; is the amount required to repair the kth road property in the jth loss level under the ith road property type; is the value decay coefficient corresponding to the kth road property in the jth loss level under the ith road property type. The direct loss assessment model introduces the value decay coefficient to consider the loss of the value of the highway property itself, which improves the accuracy of the direct economic loss assessment. The model is simple to calculate and can improve the efficiency of the direct economic loss assessment.

[0063] S3. Based on the indirect economic loss case data, the second road property loss is assessed using the indirect loss assessment strategy for highway property damage.

[0064] Wherein, step S3 specifically includes the following steps: S31. Construct a case data set using the indirect economic loss case data of a plurality of the case highways.

[0065] Specifically, in this embodiment, as described in step S13, since for the same case highway, although there is only one data for road property loss characteristic data, direct loss amount, traffic flow data, ambient temperature data, and road surface vibration speed data, multiple groups of road property loss characteristic data can be collected, wherein one group of road property loss characteristic data includes a road property type and a digital representation of a loss level. Therefore, a group of indirect economic loss case data in the case data set should include a direct loss amount, traffic flow data, ambient temperature data, road surface vibration speed data, and multiple groups of road property loss characteristic data of a case highway.

[0066] S32. Obtain comprehensive loss characteristic data of different case highways according to the case data set, and then construct a first data set in combination with the direct loss amount, traffic flow data, ambient temperature data, and road surface vibration speed data of the case highways.

[0067] Specifically, in this embodiment, for any case road, its comprehensive loss characteristics satisfy the following relationship:

[0068] in, is the value of the comprehensive loss characteristic, I is the number of road property types, J is the number of loss levels, is the number of road properties of the ith road property type, and K is the number of road properties with the jth loss level under the ith road property type. The comprehensive loss characteristic data is calculated based on the road property type and its loss level, which can comprehensively reflect the damage to highway road properties and calculate them to facilitate subsequent indirect loss assessment.

[0069] Furthermore, after calculating the comprehensive loss characteristic data of the case highway, the comprehensive loss characteristic data, direct loss amount, traffic flow data, ambient temperature data, and road surface vibration speed data corresponding to the case highway are used as a group of data in the first data set, thereby constructing the first data set.

[0070] S33. According to the first data set, use principal component analysis to obtain multiple principal components.

[0071] Specifically, in this embodiment, in the first data set, there is a coupling relationship between the indirect loss assessment indicators, so this embodiment uses the principal component analysis method to decouple the indicators, obtain multiple principal components, and also obtain the cumulative variance contribution rate of each principal component. Each principal component satisfies the following relationship:

[0072] in, is the principal component value of the nth principal component, is the coefficient of the rth indirect loss assessment index when calculating the principal component value of the nth principal component, is the indirect loss assessment indicator data of the rth indirect loss assessment indicator.

[0073] Furthermore, since the principal component analysis method is a prior art, it will not be described in detail here.

[0074] S34. For each set of indirect loss assessment index data in the first data set, a corresponding set of principal component values ​​is calculated, and together with the corresponding indirect loss amount, a set of queryable case data is formed.

[0075] S35. Use multiple sets of queryable case data to construct a queryable case data set.

[0076] S36. Obtain indirect loss assessment index data of the target highway, and query multiple reference cases in the queryable case data set, and then use the indirect economic loss assessment model to assess the second road property loss.

[0077] In this embodiment, it is considered that if the indirect loss assessment index data is directly used to assess the indirect economic losses, without using the machine learning algorithm, there is a complex relationship between the various indirect loss assessment indicators, and the weight of each indirect loss assessment indicator for the indirect loss amount is difficult to determine. If the machine learning algorithm is used to assess the indirect loss amount, the complexity and cost of the loss assessment will increase. Therefore, this embodiment determines the reference case based on the principal component value and constructs an indirect economic loss assessment model, and in the indirect economic loss assessment model, the normalized weight is determined based on the cumulative variance contribution rate of the principal component, which makes the weight more objective, can improve the objectivity of the indirect economic loss assessment, and then improve the accuracy of the indirect economic loss assessment, and the simple indirect economic loss assessment model can also improve the efficiency of the assessment.

[0078] Step S36 specifically includes the following steps: S361. Obtain indirect loss assessment index data of the target highway, and then calculate a set of principal component values, and record the set of principal component values ​​as a set of target principal component values.

[0079] S362. For each principal component, find a principal component value closest to the target principal component value in the queryable case data set, and record the principal component value as a reference principal component value.

[0080] Specifically, in this embodiment, for any principal component, there is a principal component value in the queryable case data set that is closest to the target principal component value. In this embodiment, this principal component value is recorded as the reference principal component value.

[0081] S363: Take a group of queryable case data containing the reference principal component values ​​as the referenceable cases, and then obtain a plurality of referenceable cases.

[0082] Specifically, in this embodiment, since the reference principal component values ​​of each principal component usually come from different groups of queryable case data, a plurality of reference cases are generally obtained in the end.

[0083] S364: taking the cumulative variance contribution rate of each principal component as the principal component weight, and normalizing the principal component weight to obtain a normalized weight.

[0084] Specifically, in this embodiment, the cumulative variance contribution rate of each principal component is recorded as , then the normalized weights satisfy the following relationship:

[0085] in, is the normalized weight of the mth principal component, is the cumulative variance contribution rate of the mth principal component, and M is the number of principal components.

[0086] S365. According to the referenceable case and the normalized weight, use the indirect economic loss assessment model to assess the second road property loss.

[0087] Specifically, in this embodiment, the indirect economic loss assessment model satisfies the following relationship:

[0088] in, is the second path loss, N is the number of the principal components, is the target principal component value of the nth principal component, is the reference principal component value of the nth principal component, is the normalized weight of the nth principal component, The indirect loss amount of the case where the reference principal component value of the nth principal component is located can be used as a reference.

[0089] S4. Combining the first road property loss and the second road property loss, obtain the total road property loss of the target highway, and repair the damaged road property in a timely manner.

[0090] Specifically, in this embodiment, the total loss of highway property is the sum of the first road property loss and the second road property loss. After the total loss of highway property is obtained, the damaged highway property needs to be repaired in time. On the one hand, it can increase the service life of the highway, and on the other hand, it can provide a basis for the subsequent accurate assessment of the total loss of highway property.

[0091] It should be noted that, in some cases, the actions described in the specification can be performed in a different order and still achieve the desired results. In this embodiment, the order of steps given is only to make the embodiment appear clearer and easier to explain, rather than to limit it.

[0092] In an alternative embodiment, see Figure 3 The present invention also provides a highway property loss assessment system, which includes: a data acquisition device 1, a data output device 2, a processor 3 and a storage 4, wherein the storage 4 includes a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by the processor 3, the processor 3 implements the contents described in steps S1 to S4.

[0093] Specifically, in this embodiment, the data collection device 1 includes a drone and a digital display screen, wherein the drone is used to carry a high-definition camera to conduct regular inspections of road properties on the target road and collect original images of road property damage, and the digital display screen is used to manually input indirect economic loss case data. The data output device 2 also includes a digital display screen for displaying the first road property loss, the second road property loss and the total road property loss.

[0094] In summary, firstly, this method improves Faster R-CNN to accurately extract the characteristics of road property losses, and then considers the loss of the value of highway property itself to construct a direct loss assessment model, thereby realizing a rapid and accurate assessment of the direct economic losses caused by damage to highway property, namely, the first road property loss. Secondly, this method is based on the principal component analysis method to determine the normalized weights of reference cases and principal components, and constructs an indirect economic loss assessment model, thereby quickly and accurately assessing the indirect economic losses caused by damage to highway property, namely, the second road property loss. Finally, this method uses the first road property loss and the second road property loss to obtain the total loss of highway property, thereby comprehensively and accurately assessing the loss of highway property. In addition, this system and this method are compatible with each other, which can not only stably operate this method and improve the efficiency of highway property loss assessment, but also enhance the practicality of the method provided by the present invention.

[0095] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.

Claims

1. A method for assessing highway property losses, characterized in that: The steps include: Conduct regular inspections of road assets on target roads to collect images of road asset damage, and collect case data on indirect economic losses caused by road asset damage; Using a machine learning algorithm to extract road property loss features in the highway road property damage image, and then using a direct loss assessment model in combination with a highway road property monitoring data set to assess the first road property loss; Based on the indirect economic loss case data, the second road property loss is assessed using the indirect loss assessment strategy for road property damage; The total road property loss of the target highway is obtained by combining the first road property loss and the second road property loss, and the damaged road property is repaired in a timely manner.

2. A method for assessing highway property losses according to claim 1, characterized in that: The indirect economic loss case data includes the road property loss characteristic data, direct loss amount, traffic flow data, ambient temperature data, road surface vibration speed data and indirect loss amount, and the road property loss characteristic data, direct loss amount, traffic flow data, ambient temperature data and road surface vibration speed data correspond to the five indirect loss assessment indicators of comprehensive loss characteristics, direct loss, traffic flow, ambient temperature and road surface vibration speed in sequence; The method of regularly inspecting road properties on the target road to collect images of road property damage and collecting case data of indirect economic losses caused by road property damage includes the following steps: Conduct regular inspections on road assets on the target highway to collect original images of road asset damage; Preprocessing the original image of the highway damage to obtain a clear image, and marking the acquisition position and acquisition time on the clear image to obtain the highway damage image; Query case highways where road property is damaged, and collect case data on indirect economic losses caused by road property damage on the case highways.

3. A method for assessing highway property losses according to claim 2, characterized in that: The extracting of the road property loss features in the road property damage image by using a machine learning algorithm, and then using a direct loss assessment model to assess the first road property loss in combination with the road property monitoring data set comprises the following steps: Collect historical road damage images to build a road damage dataset; Based on the road property damage data set, a road property loss identification model is constructed using a machine learning algorithm; Extracting road property loss features in the highway road property damage image using the road property loss recognition model, wherein the road property loss features include road property type and loss level; According to the road property loss characteristics and in combination with the highway road property monitoring data set, the direct loss assessment model is used to assess the first road property loss.

4. A method for assessing highway property losses according to claim 3, characterized in that: The method of constructing a road property loss identification model based on the road property damage data set using a machine learning algorithm comprises the following steps: The VGG-16 network in Faster R-CNN is replaced with ResNet-50, and a recursive feature pyramid is introduced between the ResNet-50 network and the region proposal network and the region of interest pooling layer to obtain an image recognition network. The image recognition network is trained and verified using a road property damage dataset, thereby obtaining the road property loss recognition model.

5. A method for assessing highway property losses according to claim 3, characterized in that: The highway monitoring dataset is constructed using all historical highway road damage images of the target highway; The step of using the direct loss assessment model to assess the first road property loss based on the road property loss characteristics and in combination with the highway road property monitoring data set comprises the following steps: According to the highway property monitoring data set, determining the last repair time of the highway property in the highway property damage image, and then determining the value attenuation coefficient of the highway property; The first road property loss is assessed using the direct loss assessment model based on the road property loss characteristics and the value attenuation coefficient.

6. A method for assessing highway property losses according to claim 3, characterized in that: The direct loss assessment model satisfies the following relationship: ， in, is the first road property loss; I is the number of road property types; J is the number of loss levels; K is the number of road properties with the jth loss level under the ith road property type; The amount required to repair the kth road property of the jth loss level under the ith road property type; It is the value attenuation coefficient corresponding to the kth road property in the jth loss level of the i-th road property type.

7. A method for assessing highway property losses according to claim 3, characterized in that: The method of evaluating the second road property loss using the indirect loss evaluation strategy of road property damage according to the indirect economic loss case data includes the following steps: Constructing a case data set using the indirect economic loss case data of a plurality of the case highways; Acquire comprehensive loss characteristic data of different case roads according to the case data set, and then construct a first data set by combining the direct loss amount, traffic flow data, ambient temperature data, and road surface vibration speed data of the case roads; According to the first data set, a plurality of principal components are obtained using a principal component analysis method; For each set of indirect loss assessment indicator data in the first data set, a corresponding set of principal component values ​​is calculated, and the principal component values ​​and the corresponding indirect loss amounts are combined into a set of queryable case data; Construct a queryable case dataset using multiple sets of queryable case data; The indirect loss assessment index data of the target highway is obtained, and a plurality of reference cases are searched in the queryable case data set, and then the indirect economic loss assessment model is used to assess the second road property loss.

8. A method for assessing highway property losses according to claim 7, characterized in that: The comprehensive loss characteristics satisfy the following relationship: ， in, is the comprehensive loss characteristic data, I is the number of road property types, J is the number of loss levels, is the number of road properties of the ith road property type, and K is the number of road properties with the jth loss level under the ith road property type.

9. A method for assessing highway property losses according to claim 7, characterized in that: The step of obtaining the indirect loss assessment index data of the target highway, searching for multiple reference cases in the queryable case data set, and then using the indirect economic loss assessment model to assess the second road property loss includes the following steps: Obtain indirect loss assessment index data of the target highway, and then calculate a set of principal component values, and record the set of principal component values ​​as a set of target principal component values; For each of the principal components, a principal component value closest to the target principal component value is found in the queryable case data set, and the principal component value is recorded as a reference principal component value; Taking a group of queryable case data where the reference principal component value is located as the referenceable case, and then obtaining a plurality of the referenceable cases; The cumulative variance contribution rate of each principal component is used as the principal component weight, and the principal component weight is normalized to obtain a normalized weight; According to the referenceable case and the normalized weight, the indirect economic loss assessment model is used to assess the second road property loss, and the indirect economic loss assessment model satisfies the following relationship: ， in, is the second path loss, N is the number of the principal components, is the target principal component value of the nth principal component, is the reference principal component value of the nth principal component, is the normalized weight of the nth principal component, The indirect loss amount of the case where the reference principal component value of the nth principal component is located can be used as a reference.

10. A highway property loss assessment system, characterized in that: The highway property loss assessment system comprises: a data acquisition device, a data output device, a processor and a storage device, wherein the storage device comprises a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, wherein the computer program comprises program instructions, and when the program instructions are executed by the processor, the processor implements a highway property loss assessment method as described in any one of claims 1 to 9.

Images