Traffic noise prediction method, device and equipment and storage medium

Abstract

The application relates to the technical field of data processing, and discloses a traffic noise prediction method, device and equipment and a storage medium, which comprises the following steps: acquiring traffic data of a target area; analyzing the traffic data to obtain spatial structure data and corresponding traffic flow data of the target area; extracting features from the spatial structure data to obtain traffic features of the target area; and predicting traffic noise of the target area based on the traffic features and the traffic flow data to obtain a traffic noise prediction result of the target area. The method can accurately predict traffic noise in a timely manner.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a method, apparatus, device, and storage medium for predicting traffic noise. Background Technology

[0002] In recent years, with increasing emphasis on environmental protection, green building has become a new development trend. Vibrations from traffic noise can damage building structures, especially in older buildings where the damage is more pronounced. Therefore, reducing a building's sensitivity to traffic noise requires accurate prediction of traffic noise levels around the building. Summary of the Invention

[0003] In view of this, the present invention provides a traffic noise prediction method, apparatus, device and storage medium to solve the problem of accurate traffic noise prediction.

[0004] In a first aspect, the present invention provides a traffic noise prediction method, the method comprising: Obtain traffic data for the target area; By analyzing traffic data, spatial structure data and corresponding traffic flow data of the target area can be obtained; Feature extraction is performed on spatial structure data to obtain traffic characteristics of the target area; Traffic noise in the target area is predicted based on traffic characteristics and traffic flow data, and the traffic noise prediction results for the target area are obtained.

[0005] The traffic noise prediction method provided in this invention acquires traffic data of a target area, analyzes the traffic data to obtain spatial structure data and corresponding traffic flow data of the target area; extracts features from the spatial structure data to obtain traffic characteristics of the target area, thereby improving the accuracy of traffic noise prediction; and predicts traffic noise in the target area based on traffic characteristics and traffic flow data to obtain traffic noise prediction results for the target area. This allows for reminders to building designers based on the traffic noise prediction results, and also improves the efficiency of users in developing solutions to reduce the impact of traffic noise on building structures within the target area.

[0006] In some optional implementations, feature extraction is performed on the spatial structure data to obtain traffic features of the target area, including: The spatial structure data is sampled for the locations of traffic roads and building structures to obtain the corresponding noise location sampling sequence and building location sampling sequence. Calculate the distance between the noise location sampling sequence and the building location sampling sequence to obtain the noise distance sequence; By splicing the noise distance sequence, the noise distance features of the target area are obtained. Traffic features include noise distance features.

[0007] The traffic noise prediction method provided in this embodiment of the invention obtains the corresponding noise location sampling sequence and building location sampling sequence by sampling the location of traffic roads and building structures in spatial structure data, and calculates the distance between the noise location sampling sequence and the building location sampling sequence to obtain the traffic characteristics of the noise distance characteristics of the target area.

[0008] In some optional implementations, the distance between the noise location sampling sequence and the building location sampling sequence is calculated to obtain a noise distance sequence, including: Calculate the distance between each noise location sampling point in the noise location sampling sequence and each building location sampling point in the building location sampling sequence to obtain the sampling distance sequence; The distances in the sampling distance sequence are compared with the target distance to obtain the distance comparison results; Based on the distance comparison results, sampling distances that are greater than the target distance in the sampling distance sequence are filtered out to obtain a noisy distance sequence.

[0009] The traffic noise prediction method provided in this embodiment of the invention obtains a sampling distance sequence by calculating the distance between each noise location sampling point in the noise location sampling sequence and each building location sampling point in the building location sampling sequence, and then filtering out the sampling distances in the sampling distance sequence that are greater than the target distance to obtain a noise distance sequence, thereby improving the reliability of noise distance feature acquisition.

[0010] In some optional implementations, feature extraction is performed on the spatial structure data to obtain traffic features of the target area, including: Based on spatial structure data, triangular facets of the target area are determined. These triangular facets are used to characterize the positional relationships between building structures and traffic roads within the target area. Based on beam projection information, the intersection information of each building location sampling point and the triangular facet in the building location sampling sequence is determined; Based on the intersection information, the thickness of the building structure in the target direction is calculated to obtain the geometric occlusion features between the traffic road and the building structure. The traffic features include the geometric occlusion features.

[0011] The traffic noise prediction method provided in this invention determines triangular facets representing the positional relationship between building structures and traffic roads within a target area based on spatial structure data. Then, based on beam projection information, it determines the intersection information between each building location sampling point in the building location sampling sequence and the triangular facets. Based on the intersection information, it calculates the thickness of the building structure in the target direction, obtaining traffic characteristics representing the geometric occlusion features between traffic roads and building structures. This improves the accuracy of traffic noise prediction in the target area and provides a basis for determining the impact of traffic noise on building structures.

[0012] In some optional implementations, traffic noise in the target area is predicted based on traffic characteristics and traffic flow data to obtain traffic noise prediction results for the target area, including: Traffic characteristics and traffic flow data are input into the noise prediction model to obtain noise prediction results; The traffic noise prediction results for the target area are determined based on the noise prediction results.

[0013] The traffic noise prediction method provided in this invention predicts traffic noise in a target area by inputting traffic characteristics and traffic flow data into a noise prediction model, thereby improving the accuracy of traffic noise prediction.

[0014] In some optional implementations, traffic noise in the target area is predicted based on traffic characteristics and traffic flow data to obtain traffic noise prediction results for the target area, including: The noise distance characteristics and traffic flow data are input into the feature extraction unit of the noise prediction model to obtain the corresponding first and second features. After fusing the first feature and the second feature, the result is input into the prediction unit in the noise prediction model to obtain the first noise prediction result. The traffic noise prediction result for the target area is determined based on the first noise prediction result.

[0015] The traffic noise prediction method provided in this embodiment of the invention obtains corresponding first and second features by inputting noise distance features and traffic flow data into the feature extraction unit of the noise prediction model, and then fusing the first and second features and inputting them into the prediction unit of the noise prediction model to obtain a first noise prediction result, thereby further improving the accuracy of traffic noise prediction.

[0016] In some optional implementations, traffic noise in the target area is predicted based on traffic characteristics to obtain traffic noise prediction results for the target area, including: Traffic flow data and geometric occlusion features are input into the feature extraction unit of the noise prediction model to obtain the corresponding second and third features. After fusing the second and third features, the results are input into the prediction unit of the noise prediction model to obtain the second noise prediction result. The traffic noise prediction results for the target area are determined based on the second noise prediction results.

[0017] The traffic noise prediction method provided in this embodiment of the invention obtains corresponding second and third features by inputting traffic flow data and geometric occlusion features into the feature extraction unit of the noise prediction model, and then fused the second and third features and input them into the prediction unit of the noise prediction model to obtain the second noise prediction result, thereby further improving the accuracy of traffic noise prediction.

[0018] In some alternative implementations, the noise prediction model is built on a deep learning neural network.

[0019] In some optional implementations, feature extraction is performed on the spatial structure data to obtain traffic features of the target area, including: Based on the noise location sampling sequence, the noise decibel value of each building location sampling point in the building location sampling sequence is calculated to obtain the noise decibel sequence; The noise decibel sequence is spliced ​​to obtain the noise decibel feature, and the traffic feature includes the noise decibel feature.

[0020] The traffic noise prediction method provided in this invention calculates the noise decibel value of each building location sampling point in the building location sampling sequence based on the noise location sampling sequence, and obtains a noise decibel sequence. The noise decibel sequence is then spliced ​​together to obtain the noise decibel feature in the traffic features, which provides the necessary conditions for improving the accuracy of the noise prediction model in predicting traffic noise.

[0021] In some alternative implementations, obtaining a noise prediction model includes: The noise decibel characteristics and traffic flow data, along with noise distance characteristics or geometric occlusion characteristics, are input into the initial prediction model to obtain the initial prediction results; Calculate the loss value between the initial prediction result and the target prediction result; The model parameters of the initial prediction model are updated based on the loss value to obtain the noise prediction model.

[0022] The traffic noise prediction method provided in this invention inputs noise decibel features and traffic flow data, along with noise distance features or geometric occlusion features, into an initial prediction model to obtain an initial prediction result. The model parameters of the initial prediction model are then updated based on the loss value, thereby obtaining a noise prediction model with high prediction accuracy and efficiency.

[0023] In a second aspect, the present invention provides a traffic noise prediction device, the device comprising: The data acquisition module is used to acquire traffic data for the target area; The data analysis module is used to analyze traffic data to obtain spatial structure data and corresponding traffic flow data for the target area. The feature extraction module is used to extract features from spatial structure data to obtain traffic features of the target area; The noise prediction module is used to predict traffic noise in the target area based on traffic characteristics and traffic flow data, and obtain the traffic noise prediction results for the target area.

[0024] The traffic noise prediction device provided in this invention acquires traffic data of a target area, analyzes the traffic data to obtain spatial structure data and corresponding traffic flow data of the target area; extracts features from the spatial structure data to obtain traffic features of the target area, thereby improving the accuracy of traffic noise prediction; and predicts traffic noise in the target area based on traffic features and traffic flow data to obtain traffic noise prediction results for the target area. This allows for reminders to building designers based on the traffic noise prediction results, and also improves the efficiency of users in developing solutions to reduce the impact of traffic noise on building structures within the target area.

[0025] Thirdly, the present invention provides a computer device, comprising: The memory and the processor are interconnected and communicate with each other. The memory stores computer instructions, and the processor executes the computer instructions to perform the traffic noise prediction method of the first aspect or any of its corresponding embodiments described above. Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to perform the traffic noise prediction method of the first aspect or any corresponding embodiment thereof.

[0026] It should be noted that the corresponding beneficial effects of the traffic noise prediction device, computer equipment, and computer-readable storage medium provided in the embodiments of the present invention can be found in the description of the corresponding beneficial effects of the traffic noise prediction method above, and will not be repeated here. Attached Figure Description

[0027] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0028] Figure 1This is a schematic flowchart of the traffic noise prediction method according to an embodiment of the present invention; Figure 2 This is a diagram illustrating the effect of traffic noise prediction according to an embodiment of the present invention; Figure 3 This is another schematic flowchart of the traffic noise prediction method according to an embodiment of the present invention; Figure 4 This is another flowchart illustrating the traffic noise prediction method according to an embodiment of the present invention; Figure 5 This is a structural block diagram of the traffic noise prediction device according to an embodiment of the present invention; Figure 6 This is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] In traffic noise prediction technologies, wave equations and geometric acoustics are primarily employed. The wave equation, or wave propagation equation, is an important partial differential equation that typically describes all types of waves, such as sound waves, light waves, and water waves. Geometric acoustics, similar to geometric optics, mainly studies the propagation of energy in straight lines when the wavelength is very small (compared to the scale of space or objects), then calculates the energy attenuation step by step to obtain the decibel value. However, parallel computation based on the wave equation is difficult and has a low degree of parallelism, while methods based on geometric acoustics have low computational efficiency and lack products capable of high-parallel computation on GPUs (Graphics Processing Units). These factors severely impact the efficiency of building structures during the conceptual design phase, which requires repeated modifications to the design scheme.

[0031] Based on this, embodiments of the present invention provide a traffic noise prediction method, apparatus, computer equipment, and storage medium. By acquiring traffic data of a target area and analyzing the traffic data, spatial structure data and corresponding traffic flow data of the target area are obtained. By extracting features from the spatial structure data, traffic characteristics of the target area are obtained, thereby improving the accuracy of traffic noise prediction. By predicting traffic noise in the target area based on traffic characteristics and traffic flow data, a traffic noise prediction result for the target area is obtained. This allows for reminders to building designers based on the traffic noise prediction result, and also improves the efficiency of users in developing solutions to reduce the impact of traffic noise on building structures within the target area.

[0032] This embodiment provides a traffic noise prediction method. Figure 1 This is a flowchart of a traffic noise prediction method according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps: Step S101: Obtain traffic data for the target area.

[0033] The traffic data for the target area can include building structure data, road data, and corresponding traffic flow data. Building structure data can include building structure information and location information; road data can include road type information and location information; and traffic flow information can include road traffic volume information. Building structure data and road data can be obtained from a regional map of the target area, while traffic flow data can be obtained from the target area's traffic management system. The content of the traffic data for the target area can also be added to or removed based on actual needs; for example, the traffic data can include ground data of the target area. The traffic data for the target area can also be obtained from OpenStreetMap.

[0034] Step S102: Analyze the traffic data to obtain the spatial structure data and corresponding traffic flow data of the target area.

[0035] Among these methods, analyzing traffic data can involve data cleaning and filtering to obtain spatial structure data for the target area.

[0036] In some optional implementations, the spatial structure data of the target area can be obtained through 3D reconstruction based on the traffic data of the target area, thereby obtaining spatial structure data including building structure information and traffic road information. The traffic flow data of the target area can be extracted by defining a high-frequency travel interval as a feature extraction interval based on the traffic road information, traffic characteristics, and people's travel habits of the target area. Here, people's travel habits are not limited to driving or honking habits. Traffic road information is not limited to the location or type of traffic roads. Traffic characteristics are not limited to vehicle flow level or vehicle type.

[0037] Specifically, the maximum and minimum passing speeds within the feature extraction interval can be obtained; the time corresponding to the minimum passing speed is taken as the start time of feature extraction, and the time corresponding to the maximum passing speed is taken as the end time of feature extraction. Traffic data segments are extracted from the traffic data to obtain traffic flow data.

[0038] Step S103: Extract features from the spatial structure data to obtain the traffic features of the target area.

[0039] When extracting features from spatial structure data, one can extract only the noise distance feature of the target area and use it as the traffic feature of the target area; alternatively, one can extract only the geometric occlusion feature of the target area and use it as the traffic feature; another option is to extract only the noise decibel feature of the target area and use it as the traffic feature; yet another is to combine the noise distance feature, noise decibel feature, and geometric occlusion feature in pairs as the traffic feature of the target area; and still another is to combine the noise distance feature, noise decibel feature, and geometric occlusion feature as the traffic feature of the target area. Using the combination of noise distance feature, noise decibel feature, and geometric occlusion feature as the traffic feature of the target area for initial prediction model training results in a more accurate noise prediction model.

[0040] In some optional implementations, when extracting features from spatial structure data to obtain traffic features of the target area, the spatial structure data can be sampled for the locations of traffic roads and building structures to obtain corresponding noise location sampling sequences and building location sampling sequences; the distance between the noise location sampling sequences and the building location sampling sequences can be calculated to obtain a noise distance sequence; the noise distance sequences can be concatenated to obtain the noise distance features of the target area, and the traffic features include the noise distance features.

[0041] Specifically, traffic roads and building structures in the target area are obtained based on spatial structure data. The traffic roads and building structures are sampled at target intervals to obtain noise location sampling sequences corresponding to traffic roads and building location sampling sequences corresponding to building structures. Then, the distance between each noise location sampling point in the noise location sampling sequence and each building location sampling point in the building location sampling sequence is calculated, and the obtained sampling distance sequences are spliced ​​together to obtain the noise distance features in the traffic features of the target area.

[0042] The traffic noise prediction method provided in this embodiment of the invention obtains the corresponding noise location sampling sequence and building location sampling sequence by sampling the location of traffic roads and building structures in spatial structure data, and calculates the distance between the noise location sampling sequence and the building location sampling sequence to obtain the traffic characteristics of the noise distance characteristics of the target area.

[0043] In some optional implementations, when calculating the distance between the noise location sampling sequence and the building location sampling sequence to obtain the noise distance sequence, the distance between each noise location sampling point in the noise location sampling sequence and each building location sampling point in the building location sampling sequence is first calculated to obtain the sampling distance sequence; each sampling distance in the sampling distance sequence is compared with the target distance to obtain the distance comparison result; based on the distance comparison result, the sampling distances in the sampling distance sequence that are greater than the target distance are filtered out to obtain the noise distance sequence.

[0044] Specifically, the Euclidean distance between each noise location sampling point in the noise location sampling sequence and each building location sampling point in the building location sampling sequence is calculated to obtain the sampling distance sequence; then, each sampling distance in the sampling distance sequence is compared with the target distance, and sampling distances in the sampling distance sequence that are greater than the target distance are filtered out to obtain the noise distance sequence.

[0045] The traffic noise prediction method provided in this embodiment of the invention obtains a sampling distance sequence by calculating the distance between each noise location sampling point in the noise location sampling sequence and each building location sampling point in the building location sampling sequence, and then filtering out the sampling distances in the sampling distance sequence that are greater than the target distance to obtain a noise distance sequence, thereby improving the reliability of noise distance feature acquisition.

[0046] In some optional implementations, when extracting features from spatial structure data to obtain traffic features of the target area, triangular facets of the target area can be determined based on the spatial structure data. These triangular facets are used to characterize the positional relationship between building structures and traffic roads within the target area. Based on beam projection information, the intersection information between each building location sampling point in the building location sampling sequence and the triangular facets is determined. Based on the intersection information, the thickness of the building structure in the target direction is calculated to obtain the geometric occlusion features between traffic roads and building structures. The traffic features include geometric occlusion features.

[0047] Specifically, spatial structure data is analyzed to obtain the three-dimensional coordinates of building structures and roads within the target area. These coordinates are then transformed into triangular facets represented by coordinate vertices and indices to characterize the positional relationships between building structures and roads within the target area. Next, the building structures are sampled at target intervals to obtain a corresponding building position sampling sequence. The Euclidean distance between each noise position sampling point in the noise position sampling sequence and each building position sampling point in the building position sampling sequence is calculated to obtain a sampling distance sequence. Each sampling distance in the sampling distance sequence is compared with the target distance, and sampling distances greater than the target distance are filtered out, resulting in a target building position sampling sequence where the sampling distance is less than or equal to the target distance. Finally, based on beam projection information, the intersection information between each building position sampling point in the target building position sampling sequence and the triangular facets is determined. Based on this intersection information, the thickness of the building structure in the target direction is calculated, yielding the geometric occlusion features in the traffic characteristics between roads and building structures.

[0048] Please see Figure 2 This diagram illustrates the acquisition of beam projection information, specifically using ray tracing technology to determine the intersection points of each building location sampling point with the triangular facet in the target building location sampling sequence. In determining these intersection points using ray tracing, reflection and diffraction of light rays can be ignored, retaining only the path of maximum light energy—straight-line propagation. Gray dots represent sampling points of the noise source (traffic roads), and black lines represent the light propagation paths between the retained sampling points and the measured point (building structure). The sampling interval can be set based on the speed of the entire noise prediction model; sparser sampling points result in faster prediction speeds but reduced accuracy. A threshold can also be set to determine which sampling distances to retain (filtering out sampling distances greater than the target distance in the sampling distance sequence).

[0049] Furthermore, spatial structure data is analyzed to obtain the three-dimensional coordinates of building structures, roads, and the ground within the target area. These coordinates are then transformed into triangular patches represented by coordinate vertices and indices to characterize the positional relationships between building structures and roads within the target area. Based on requirements, equally spaced sampling is performed on building structures, roads, and the ground. This yields noise location sampling sequences corresponding to roads, building location sampling sequences corresponding to building structures, and ground location sampling sequences corresponding to the ground. The Euclidean distances between each noise location sampling point in the noise location sampling sequence and each ground location sampling point in the ground location sampling sequence and each building location sampling point in the building location sampling sequence are calculated. The first and second sampling distance sequences are obtained. Each sampling distance in the first and second sampling distance sequences is compared with the corresponding target distance, and sampling distances greater than the target distance in the sampling distance sequence are filtered out, thereby obtaining the target building location sampling sequence and the target ground location sampling sequence that are less than or equal to the target distance. Then, based on the beam projection information, the intersection information of each sampling point in the target building location sampling sequence and the target ground location sampling sequence with the triangular facet is determined, thereby obtaining the noise distance feature in the traffic features of the target area. Finally, based on the intersection information, the thickness of the building structure in the target direction is calculated, thereby obtaining the geometric occlusion feature in the traffic features between the traffic road and the building structure.

[0050] The traffic noise prediction method provided in this invention determines triangular facets representing the positional relationship between building structures and traffic roads within a target area based on spatial structure data. Then, based on beam projection information, it determines the intersection information between each building location sampling point in the building location sampling sequence and the triangular facets. Based on the intersection information, it calculates the thickness of the building structure in the target direction, obtaining traffic characteristics representing the geometric occlusion features between traffic roads and building structures. This improves the accuracy of traffic noise prediction in the target area and provides a basis for determining the impact of traffic noise on building structures.

[0051] Step S104: Based on traffic characteristics and traffic flow data, predict the traffic noise in the target area to obtain the traffic noise prediction result for the target area.

[0052] The noise prediction model for predicting traffic noise in a target area based on traffic characteristics and traffic flow data is built on a machine learning model. This machine learning model is not limited to machine regression learning models. When using machine regression learning models, it can be, but is not limited to, Gaussian process regression models, support vector machine models, correlation vector machine models, multiple linear regression models, multi-order polynomial models, random forest models, etc.

[0053] In some optional implementations, when predicting traffic noise in the target area based on traffic characteristics and traffic flow data, the traffic characteristics and traffic flow data can be input into the noise prediction model to obtain the noise prediction result; the traffic noise prediction result of the target area is determined based on the noise prediction result.

[0054] Of course, noise prediction models are not limited to the Gaussian process regression model mentioned above. They can also be support vector machine models, correlation vector machine models, multiple linear regression models, multi-order polynomial models, random forest models, etc. There are no restrictions here.

[0055] In some optional implementations, when extracting features from spatial structure data to obtain traffic features of the target area, the spatial structure data can also be sampled for the locations of traffic roads and building structures to obtain corresponding noise location sampling sequences and building location sampling sequences; based on the noise location sampling sequences, the noise decibel values ​​of each building location sampling point in the building location sampling sequence are calculated to obtain a noise decibel sequence; the noise decibel sequences are spliced ​​together to obtain noise decibel features, and the traffic features include noise decibel features.

[0056] Traffic features include noise distance features and geometric occlusion features. When obtaining a noise prediction model, noise decibel features and traffic flow data, along with noise distance features or geometric occlusion features, can be input into the initial prediction model to obtain the initial prediction result. The loss value between the initial prediction result and the target prediction result is calculated. The model parameters of the initial prediction model are updated based on the loss value to obtain the noise prediction model.

[0057] The traffic noise prediction method provided in this invention calculates the noise decibel value of each building location sampling point in the building location sampling sequence based on the noise location sampling sequence, and obtains a noise decibel sequence. The noise decibel sequence is then spliced ​​to obtain the noise decibel feature in the traffic features, providing the necessary conditions for improving the accuracy of the noise prediction model in predicting traffic noise. By inputting the noise decibel feature and traffic flow data, along with noise distance features or geometric occlusion features, into the initial prediction model, an initial prediction result is obtained. The model parameters of the initial prediction model are then updated based on the loss value, thereby obtaining a noise prediction model with high prediction accuracy and efficiency.

[0058] The traffic noise prediction method provided in this embodiment acquires traffic data of the target area, analyzes the traffic data to obtain spatial structure data and corresponding traffic flow data of the target area; extracts features from the spatial structure data to obtain traffic characteristics of the target area, thereby improving the accuracy of traffic noise prediction; and predicts traffic noise in the target area based on traffic characteristics and traffic flow data to obtain traffic noise prediction results for the target area. This allows for reminders to building designers based on the traffic noise prediction results, and also improves the efficiency of users in formulating plans to reduce the impact of traffic noise on building structures within the target area.

[0059] This embodiment provides a traffic noise prediction method. Figure 3 This is a flowchart of a traffic noise prediction method according to an embodiment of the present invention, such as... Figure 3 As shown, the process includes the following steps: Step S301: Obtain traffic data for the target area.

[0060] Please see details Figure 1 Step S101 of the illustrated embodiment will not be described again here.

[0061] Step S302: Analyze the traffic data to obtain the spatial structure data and corresponding traffic flow data of the target area.

[0062] Please see details Figure 1 Step S102 of the illustrated embodiment will not be described again here.

[0063] Step S303: Extract features from the spatial structure data to obtain the traffic features of the target area.

[0064] Please see details Figure 1 Step S103 of the illustrated embodiment will not be described again here.

[0065] Step S304: Based on traffic characteristics and traffic flow data, predict the traffic noise in the target area to obtain the traffic noise prediction result for the target area.

[0066] Among these traffic characteristics are noise distance characteristics.

[0067] Specifically, step S304 above includes: Step S3041: Input the noise distance features and traffic flow data into the feature extraction unit in the noise prediction model to obtain the corresponding first feature and second feature.

[0068] In the noise prediction model, both noise distance features and traffic flow data have their own feature extraction units. That is, the feature extraction unit corresponding to the noise distance features is a feature extraction unit adapted to the noise features, and the feature extraction unit corresponding to the traffic flow data is a feature extraction unit adapted to the traffic flow.

[0069] The noise prediction model is built upon deep learning neural networks. Its feature extraction units include Convolutional Neural Networks (CNNs), batch normalization layers, and nonlinear activation functions. CNNs are used to replace nonlinearly connected layers for feature extraction, reducing the number of parameters. Batch normalization layers prevent mean shift in the output of convolutional layers. Nonlinear activation layers increase the nonlinearity of the model. In other examples, the feature extraction units in the noise prediction model can be adjusted based on noise distance features and attributes of traffic flow data.

[0070] Step S3042: After fusing the first feature and the second feature, input them into the prediction unit in the noise prediction model to obtain the first noise prediction result.

[0071] Specifically, the first and second features can be concatenated first, and then the concatenated features can be input into the prediction unit (such as a regressor) of the noise prediction model for linear regression (i.e., linear summation) to obtain the first noise prediction result. Alternatively, the first and second features can be input into the prediction unit of the noise prediction model for feature concatenation before performing sequential linear regression calculations to obtain the first noise prediction result. Before concatenating the first and second features, they can also be input into the attention unit of the noise prediction model to determine their importance (weights) before feature concatenation.

[0072] Step S3043: Determine the traffic noise prediction result for the target area based on the first noise prediction result.

[0073] The traffic noise prediction method provided in this embodiment of the invention obtains corresponding first and second features by inputting noise distance features and traffic flow data into the feature extraction unit of the noise prediction model, and then fusing the first and second features and inputting them into the prediction unit of the noise prediction model to obtain a first noise prediction result, thereby further improving the accuracy of traffic noise prediction.

[0074] This embodiment provides a traffic noise prediction method. Figure 4 This is a flowchart of a traffic noise prediction method according to an embodiment of the present invention, such as... Figure 4 As shown, the process includes the following steps: Step S401: Obtain traffic data for the target area.

[0075] Please see details Figure 1 Step S101 of the illustrated embodiment will not be described again here.

[0076] Step S402: Analyze the traffic data to obtain the spatial structure data and corresponding traffic flow data of the target area.

[0077] Please see details Figure 1 Step S102 of the illustrated embodiment will not be described again here.

[0078] Step S403: Extract features from the spatial structure data to obtain the traffic features of the target area.

[0079] Please see details Figure 1 Step S103 of the illustrated embodiment will not be described again here.

[0080] Step S404: Based on traffic characteristics and traffic flow data, predict the traffic noise in the target area to obtain the traffic noise prediction result for the target area.

[0081] Among these traffic features are geometric occlusion features.

[0082] Specifically, step S404 above includes: Step S4041: Input traffic flow data and geometric occlusion features into the feature extraction unit in the noise prediction model to obtain the corresponding second and third features.

[0083] In the noise prediction model, both geometric occlusion features and traffic flow data have their own feature extraction units. That is, the feature extraction unit corresponding to geometric occlusion features is a feature extraction unit adapted to ray tracing, and the feature extraction unit corresponding to traffic flow data is a feature extraction unit adapted to traffic flow.

[0084] The noise prediction model is built upon deep learning neural networks. Its feature extraction units include convolutional neural networks (CNNs), batch normalization layers, and non-linear activation functions. CNNs are used to extract features instead of non-linear connection layers, reducing the number of parameters; batch normalization layers prevent mean shift in the output of convolutional layers; and non-linear activation layers increase the non-linearity of the model. In other examples, the feature extraction units in the noise prediction model can be adjusted based on noise distance features and attributes of traffic flow data.

[0085] Step S4042: After fusing the second feature and the third feature, input them into the prediction unit in the noise prediction model to obtain the second noise prediction result.

[0086] Specifically, the second and third features can be concatenated first, and then the concatenated features can be input into the prediction unit of the noise prediction model for linear regression (i.e., linear summation) to obtain the second noise prediction result. Alternatively, the second and third features can be input into the prediction unit of the noise prediction model for feature concatenation, and then linear regression calculation can be performed sequentially to obtain the second noise prediction result.

[0087] Step S4043: Determine the traffic noise prediction result for the target area based on the second noise prediction result.

[0088] The traffic noise prediction method provided in this embodiment of the invention obtains corresponding second and third features by inputting traffic flow data and geometric occlusion features into the feature extraction unit of the noise prediction model, and then fused the second and third features and input them into the prediction unit of the noise prediction model to obtain the second noise prediction result, thereby further improving the accuracy of traffic noise prediction.

[0089] In some optional implementations, when predicting traffic noise in the target area based on traffic features and traffic flow data, and obtaining the traffic noise prediction result for the target area, noise distance features, geometric occlusion features, and traffic flow data can also be input into the feature extraction unit in the noise prediction model to obtain the corresponding first feature, second feature, and third feature; after fusing the first feature, second feature, and third feature, they are input into the prediction unit in the noise prediction model to obtain the third noise prediction result; and the traffic noise prediction result for the target area is determined based on the third noise prediction result.

[0090] This embodiment also provides a traffic noise prediction device for implementing the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0091] The following is a detailed explanation of the power regulation process provided in the embodiments of this application using a specific example.

[0092] Step 1: Obtain traffic data for the target area.

[0093] The traffic data for the target area may include building structure data, road data, and corresponding traffic flow data.

[0094] Step two involves analyzing the traffic data to obtain the spatial structure data and corresponding traffic flow data of the target area.

[0095] The spatial structure data of the target area can be obtained by performing 3D reconstruction based on the traffic data of the target area, thus obtaining spatial structure data including building structure information and traffic road information. The traffic flow data of the target area can be extracted by defining a high-frequency travel interval as a feature extraction interval based on the traffic road information, traffic characteristics, and people's travel habits of the target area.

[0096] Step 3: Extract features from the spatial structure data to obtain the traffic features of the target area.

[0097] Specifically, spatial structure data is analyzed to obtain the three-dimensional coordinates of building structures, roads, and the ground within the target area. These three-dimensional coordinates are then transformed into triangular patches represented by coordinate vertices and coordinate indices to characterize the positional relationships between building structures and roads within the target area. Based on requirements, the building structures, roads, and ground are sampled at equal intervals to obtain noise location sampling sequences corresponding to roads, building location sampling sequences corresponding to building structures, and ground location sampling sequences corresponding to the ground. The data types of these noise location sampling sequences, building location sampling sequences, and ground location sampling sequences are then converted into GPU-acceptable Tensor data types, and the data is copied to the GPU device. The Euclidean distances between each noise location sampling point in the noise location sampling sequence and each ground location sampling point in the ground location sampling sequence and each building location sampling point in the building location sampling sequence are calculated to obtain the corresponding first and second sampling distance sequences. Each sampling distance in the first and second sampling distance sequences is compared with its corresponding target distance, and sampling distances greater than the target distance are filtered out. This results in target building location sampling sequences and target ground location sampling sequences with sampling distances less than or equal to the target distance. The data of these target building location sampling sequences and target ground location sampling sequences are copied to the CPU and converted to Python array format data types. Using ray tracing technology, based on beam projection information, the intersection information between each sampling point in the target building location sampling sequence and the triangular facet is determined, thus obtaining the noise distance characteristics in the traffic features of the target area. Finally, based on the intersection information, the thickness of the building structure in the target direction is calculated to obtain the geometric occlusion characteristics in the traffic features between the road and the building structure.

[0098] Step 4: Based on traffic characteristics and traffic flow data, predict the traffic noise in the target area to obtain the traffic noise prediction results for the target area.

[0099] Specifically, the PyTorch neural network framework and ONNX are used to convert the noise prediction model into ONNX format files. TensorRT is used to accelerate the conversion of the ONNX files into 32-bit floating-point .engine format files. The target building location sampling sequences, target ground location sampling sequences, and geometric occlusion features are converted to 16-bit floating-point precision and copied to the GPU along with traffic flow data. TensorRT loads the .engine file, starts the TensorRT inference engine, and performs traffic noise prediction for the target area.

[0100] This embodiment provides a traffic noise prediction device, such as... Figure 5 As shown, it includes: Data acquisition module 501 is used to acquire traffic data for the target area; The data analysis module 502 is used to analyze traffic data to obtain spatial structure data and corresponding traffic flow data of the target area.

[0101] The feature extraction module 503 is used to extract features from spatial structure data to obtain traffic features of the target area.

[0102] The noise prediction module 504 is used to predict traffic noise in the target area based on traffic characteristics and traffic flow data, and obtain the traffic noise prediction result of the target area.

[0103] In some alternative implementations, the feature extraction module 503 includes: The first sampling unit is used to sample the location of traffic roads and building structures in the spatial structure data to obtain the corresponding noise location sampling sequence and building location sampling sequence.

[0104] The first calculation unit is used to calculate the distance between the noise location sampling sequence and the building location sampling sequence to obtain the noise distance sequence.

[0105] The first splicing unit is used to splice the noise distance sequence to obtain the noise distance features of the target area. Traffic features include noise distance features.

[0106] In some alternative implementations, the first computing unit includes: The distance calculation subunit is used to calculate the distance between each noise location sampling point in the noise location sampling sequence and each building location sampling point in the building location sampling sequence, thus obtaining the sampling distance sequence.

[0107] The distance comparison subunit is used to compare each sampled distance in the sampled distance sequence with the target distance to obtain the distance comparison result.

[0108] The distance filtering subunit is used to filter out the sampling distances in the sampling distance sequence that are greater than the target distance based on the distance comparison results, so as to obtain a noisy distance sequence.

[0109] In some alternative implementations, the feature extraction module 503 includes: The triangular facet determination unit is used to determine the triangular facets of a target area based on spatial structure data. The triangular facets are used to characterize the positional relationships between building structures and traffic roads within the target area.

[0110] The intersection point determination unit is used to determine the intersection point information between the sampling points of each building location and the triangular facet in the target area based on the beam projection information.

[0111] The feature extraction unit is used to calculate the thickness of the building structure in the target direction based on the intersection information, and obtain the geometric occlusion features between the traffic road and the building structure. The traffic features include geometric occlusion features.

[0112] In some alternative implementations, the noise prediction module 504 includes: The data calculation unit is used to input traffic characteristics and traffic flow data into the noise prediction model to obtain noise prediction results; the noise prediction model is built based on a deep learning neural network.

[0113] The result determination unit is used to determine the traffic noise prediction result of the target area based on the noise prediction result.

[0114] In some alternative implementations, the noise prediction module 504 includes: The first extraction unit is used to input noise distance features and traffic flow data into the feature extraction unit in the noise prediction model to obtain the corresponding first feature and second feature.

[0115] The first prediction unit is used to fuse the first feature and the second feature and then input them into the prediction unit in the noise prediction model to obtain the first noise prediction result.

[0116] The first determining unit is used to determine the traffic noise prediction result of the target area based on the first noise prediction result.

[0117] In some alternative implementations, the noise prediction module 504 includes: The second extraction unit is used to input traffic flow data and geometric occlusion features into the feature extraction unit in the noise prediction model to obtain the corresponding second and third features.

[0118] The second prediction unit is used to fuse the second and third features and input them into the prediction unit of the noise prediction model to obtain the second noise prediction result.

[0119] The second determining unit is used to determine the traffic noise prediction result of the target area based on the second noise prediction result.

[0120] In some alternative implementations, the feature extraction module 503 includes: The location sampling unit is used to sample the location of traffic roads and building structures in the spatial structure data to obtain the corresponding noise location sampling sequence and building location sampling sequence.

[0121] The noise calculation unit is used to calculate the noise decibel value of each building location sampling point in the building location sampling sequence based on the noise location sampling sequence, and obtain the noise decibel sequence.

[0122] The feature splicing unit is used to splice the noise decibel sequence to obtain noise decibel features. Traffic features include noise decibel features.

[0123] In some optional implementations, noise decibel characteristics and traffic flow data, along with noise distance characteristics or geometric occlusion characteristics, can be input into the initial prediction model to obtain the initial prediction result; the loss value between the initial prediction result and the target prediction result can be calculated; and the model parameters of the initial prediction model can be updated based on the loss value to obtain the noise prediction model.

[0124] The traffic noise prediction device provided in this embodiment acquires traffic data of a target area, analyzes the traffic data to obtain spatial structure data and corresponding traffic flow data of the target area; extracts features from the spatial structure data to obtain traffic characteristics of the target area, thereby improving the accuracy of traffic noise prediction; and predicts traffic noise in the target area based on traffic characteristics and traffic flow data to obtain traffic noise prediction results for the target area. This allows for reminders to building designers based on the traffic noise prediction results, and also improves the efficiency of users in developing solutions to reduce the impact of traffic noise on building structures within the target area.

[0125] Further functional descriptions of the above modules and units are the same as those in the corresponding embodiments described above, and will not be repeated here.

[0126] In this embodiment, the traffic noise prediction device is presented in the form of a functional unit. Here, a unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0127] This invention also provides a computer device having the above-described features. Figure 5 The traffic noise prediction device shown.

[0128] Please see Figure 6 , Figure 6 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of the present invention, such as... Figure 6 As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). Figure 6 Take a processor 10 as an example.

[0129] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0130] The memory 20 stores instructions executable by at least one processor 10 to cause at least one processor 10 to perform the method shown in the above embodiments.

[0131] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0132] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0133] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0134] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded over a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0135] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A traffic noise prediction method, characterized in that, The method includes: Obtain traffic data for the target area; The traffic data is analyzed to obtain the spatial structure data of the target area and the corresponding traffic flow data; Feature extraction is performed on the spatial structure data to obtain the traffic features of the target area; the traffic features include noise decibel features, noise distance features, and at least one of geometric occlusion features, wherein the geometric occlusion features are used to characterize the occlusion range and degree of the building structure on the traffic road in the target direction; Predicting traffic noise in the target area based on the traffic characteristics and traffic flow data to obtain a traffic noise prediction result for the target area includes: inputting the traffic characteristics and traffic flow data into a noise prediction model to obtain a noise prediction result; and determining the traffic noise prediction result for the target area based on the noise prediction result. Obtaining the noise prediction model includes: inputting the noise decibel feature and the traffic flow data, along with the noise distance feature or the geometric occlusion feature, into an initial prediction model to obtain an initial prediction result; calculating the loss value between the initial prediction result and the target prediction result; and updating the model parameters of the initial prediction model based on the loss value to obtain the noise prediction model.

2. The method according to claim 1, characterized in that, The step of extracting features from the spatial structure data to obtain the traffic features of the target area includes: The spatial structure data is sampled for the locations of traffic roads and building structures to obtain corresponding noise location sampling sequences and building location sampling sequences. Calculate the distance between the noise location sampling sequence and the building location sampling sequence to obtain the noise distance sequence; The noise distance sequence is spliced ​​together to obtain the noise distance features of the target area, and the traffic features include the noise distance features.

3. The method according to claim 2, characterized in that, The step of calculating the distance between the noise location sampling sequence and the building location sampling sequence to obtain the noise distance sequence includes: Calculate the distance between each noise location sampling point in the noise location sampling sequence and each building location sampling point in the building location sampling sequence to obtain the sampling distance sequence; The sampling distances in the sampling distance sequence are compared with the target distance to obtain the distance comparison results; Based on the distance comparison results, the sampling distances in the sampling distance sequence that are greater than the target distance are filtered out to obtain the noise distance sequence.

4. The method according to claim 1, characterized in that, The step of extracting features from the spatial structure data to obtain the traffic features of the target area includes: Based on the spatial structure data, triangular facets of the target area are determined, and the triangular facets are used to characterize the positional relationships between building structures and traffic roads within the target area. Based on the beam projection information, the intersection information of each building location sampling point in the target area with the triangular facet is determined; Based on the intersection information, the thickness of the building structure in the target direction is calculated to obtain the geometric occlusion features between the traffic road and the building structure, and the traffic features include the geometric occlusion features.

5. The method according to claim 1, characterized in that, The traffic features include noise distance features. The step of predicting traffic noise in the target area based on the traffic features and the traffic flow data to obtain the traffic noise prediction result for the target area includes: The noise distance feature and the traffic flow data are input into the feature extraction unit of the noise prediction model to obtain the corresponding first feature and second feature. After fusing the first feature and the second feature, the result is input into the prediction unit of the noise prediction model to obtain the first noise prediction result. Based on the first noise prediction result, the traffic noise prediction result for the target area is determined.

6. The method according to claim 5, characterized in that, The traffic features include geometric occlusion features. The step of predicting traffic noise in the target area based on these traffic features to obtain a traffic noise prediction result for the target area includes: The traffic flow data and the geometric occlusion features are input into the feature extraction unit of the noise prediction model to obtain the corresponding second and third features. After fusing the second feature and the third feature, the result is input into the prediction unit of the noise prediction model to obtain the second noise prediction result. The traffic noise prediction result for the target area is determined based on the second noise prediction result.

7. The method according to claim 4, characterized in that, The noise prediction model is built based on a deep learning neural network.

8. The method according to claim 1, characterized in that, The step of extracting features from the spatial structure data to obtain the traffic features of the target area includes: The spatial structure data is sampled for the locations of traffic roads and building structures to obtain corresponding noise location sampling sequences and building location sampling sequences. Based on the noise location sampling sequence, the noise decibel value of each building location sampling point in the building location sampling sequence is calculated to obtain the noise decibel sequence; The noise decibel sequence is spliced ​​together to obtain noise decibel features, and the traffic features include the noise decibel features.

9. A traffic noise prediction device, characterized in that, The device includes: The data acquisition module is used to acquire traffic data for the target area; The data analysis module is used to analyze the traffic data to obtain the spatial structure data of the target area and the corresponding traffic flow data. A feature extraction module is used to extract features from the spatial structure data to obtain traffic features of the target area. The traffic features include noise decibel features, noise distance features, and at least one of geometric occlusion features. The geometric occlusion features characterize the occlusion range and degree of the building structure on the traffic road in the target direction. A noise prediction model is obtained by: inputting the noise decibel features and the traffic flow data, along with the noise distance features or the geometric occlusion features, into an initial prediction model to obtain an initial prediction result; calculating the loss value between the initial prediction result and the target prediction result; and updating the model parameters of the initial prediction model based on the loss value to obtain the noise prediction model. The noise prediction module is used to predict the traffic noise of the target area based on the traffic characteristics and the traffic flow data, and to obtain the traffic noise prediction result of the target area.

10. A computer device, characterized in that, include: A memory and a processor are communicatively connected, the memory storing computer instructions, and the processor executing the computer instructions to perform the traffic noise prediction method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to perform the traffic noise prediction method according to any one of claims 1 to 8.

Images