A vehicle route planning method, device and equipment based on environment perception

Abstract

The present application relates to the technical field of route planning, in particular to a vehicle route planning method, device and equipment based on environment perception. The present application firstly identifies the physical element information and emotion perception data of each road section from the street view image of the road section, then comprehensively evaluates each road section according to the physical element information and emotion perception data of each road section to obtain a comprehensive score, and finally intelligently selects the connected target road sections from each road section according to the comprehensive score, the length of the road section and the number of passengers who need to ride on each road section, and plans the driving route of the vehicle from the starting point to the ending point using the target road sections. From the above analysis, it can be seen that the present application not only considers the physical elements of the road section, but also fully considers the emotion perception data of the road section when planning the route, so that the route planned by the present application can meet the actual needs of the passengers.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of route planning, in particular to a vehicle route planning method, device and equipment based on environment perception. BACKGROUND

[0002] A vehicle travels from a starting point to an ending point along different routes, and the street views on different routes bring different experiences to passengers. The experience is associated with the street visual quality, which not only reflects in the street physical conditions such as vegetation coverage on both sides of the road, the openness of the field of view, etc., but also includes the emotional perception brought by the street view: environmental aesthetics, historical and cultural characteristics, etc. These factors together constitute the "perceived quality" of the street, which directly affects the perception and experience of tourists. However, the prior art only considers the street physical conditions when planning the driving route of the vehicle, and ignores the emotional perception brought by the street view to the passengers, so that the route planned by the prior art is difficult to meet the actual needs of passengers.

[0003] In summary, the route planned by the prior art is difficult to meet the actual needs of passengers.

[0004] Therefore, the prior art still needs to be improved and improved. SUMMARY

[0005] To solve the above technical problems, the present application provides a vehicle route planning method, device and equipment based on environment perception, which solves the problem that the route planned by the prior art is difficult to meet the actual needs of passengers.

[0006] To achieve the above purpose, the present application adopts the following technical solutions:

[0007] In a first aspect, the present application provides a vehicle route planning method based on environment perception, which comprises:

[0008] Obtaining each road segment between the starting point and the ending point, and obtaining the street view image of each road segment, identifying the physical element information and emotional perception data on the street view image, the physical element information being used to represent the object information contained in the environment where the road segment is located, and the emotional perception data being used to represent the emotional response that can be triggered by the environment where the road segment is located;

[0009] According to the physical element information and the emotional perception data of each road segment, a comprehensive score of each road segment is obtained;

[0010] Obtaining the length of each of the road segments, obtaining the number of passengers of each of the road segments that need to ride a vehicle, and screening out connected target road segments from each of the road segments according to the length of each of the road segments, the number of passengers of each of the road segments, and the comprehensive score of each of the road segments, and planning a driving route of the vehicle from the starting point to the ending point according to a path where the target road segments are located.

[0011] In an implementation manner, the number of sky pixel points, the number of vegetation pixel points, and the number of building pixel points on the street view image are identified by a semantic segmentation model as the physical element information; the semantic segmentation model comprises a backbone network, a dilated convolution, an ASPP module, and a decoder connected in sequence.

[0012] In an implementation manner, a vitality score, a beauty score, a safety score, a richness score, a boredom score, and a depression score corresponding to the street view image are determined by an image classification model as the emotion perception data; the image classification model comprises a convolution layer, a residual module, a global average pooling layer, and a full connection layer connected in sequence.

[0013] In an implementation manner, the comprehensive score of each of the road segments is obtained according to the physical element information and the emotion perception data of each of the road segments, comprising:

[0014] A positive physical index and a negative physical index in the physical element information are determined, the positive physical index is a physical index that can enhance passenger experience, and the negative physical index is a physical index that reduces passenger experience.

[0015] A positive emotion perception index and a negative emotion perception index in the emotion perception data are determined, the positive emotion perception index is an emotion index that enhances passenger experience, and the negative emotion perception index is an emotion index that reduces passenger experience.

[0016] The comprehensive score of each of the road segments is obtained according to the positive emotion perception index, the negative emotion perception index, the positive physical index, and the negative physical index of each of the road segments.

[0017] In an implementation manner, the distance between the geographical coordinates of the street view image and the road geographical coordinates of the road segment is less than a threshold value.

[0018] In an implementation manner, the connected target road segments are screened out from each of the road segments according to the length of each of the road segments, the number of passengers of each of the road segments, and the comprehensive score of each of the road segments, comprising:

[0019] dividing each of the road segments into a plurality of groups, each of the road segments in each of the groups having connectivity;

[0020] establishing a first function of each of the groups according to the passenger quantity of the road segments in each of the groups, and establishing a second function of each of the groups according to the comprehensive score of the road segments in each of the groups;

[0021] obtaining a capacity limit of the vehicle, obtaining a straight line length from the starting point to the ending point, and establishing a constraint condition of each of the groups according to the capacity limit and the straight line length;

[0022] screening a target group from each of the groups according to the constraint condition, the first function and the second function of each of the groups, and taking each of the road segments in the target group as a target road segment.

[0023] In an implementation manner, the screening of the target road segments having connectivity from the road segments according to the length of each of the road segments, the passenger quantity of each of the road segments and the comprehensive score of each of the road segments comprises:

[0024] constructing a route optimization model according to the length of each of the road segments, the passenger quantity of each of the road segments and the comprehensive score of each of the road segments;

[0025] applying a non-dominated sorting algorithm to the route optimization model to screen the target road segments having connectivity from the road segments.

[0026] In a second aspect, an embodiment of the present application further provides a vehicle route planning device based on environmental perception, wherein the device comprises the following components:

[0027] an image analysis module, configured to obtain each of road segments between a starting point and an ending point, and obtain a street view image of each of the road segments, identify physical element information and emotional perception data on the street view image, the physical element information being used to represent object information contained in an environment where the road segment is located, and the emotional perception data being used to represent emotional reactions that can be triggered by the environment where the road segment is located;

[0028] a scoring module, configured to obtain a comprehensive score of each of the road segments according to the physical element information and the emotional perception data of each of the road segments;

[0029] The planning module is configured to acquire lengths of the road segments, acquire passenger numbers of the road segments that need to be taken by the vehicle, and select target road segments that are connected from the road segments according to the lengths of the road segments, the passenger numbers of the road segments and the comprehensive scores of the road segments, and plan a driving route of the vehicle from the starting point to the ending point according to a path in which the target road segments are located.

[0030] In a third aspect, an embodiment of the present application further provides a terminal device, wherein the terminal device comprises a memory, a processor, and an environment-aware vehicle route planning program stored in the memory and capable of running on the processor, and the processor implements steps of the environment-aware vehicle route planning method when executing the environment-aware vehicle route planning program.

[0031] In a fourth aspect, an embodiment of the present application further provides a computer readable storage medium, wherein the computer readable storage medium stores an environment-aware vehicle route planning program, and the processor implements steps of the environment-aware vehicle route planning method when executing the environment-aware vehicle route planning program.

[0032] Beneficial effects: The present application firstly identifies physical element information and emotion perception data of each road segment from a street view image of the road segment, then comprehensively evaluates each road segment according to the physical element information and the emotion perception data of each road segment to obtain a comprehensive score, and finally intelligently selects target road segments that are connected from each road segment according to the comprehensive score, the length of the road segment and the passenger number of the vehicle that needs to be taken on each road segment, and plans a driving route of the vehicle from the starting point to the ending point using the target road segments. As can be known from the above analysis, the present application not only considers physical elements of the road segment but also fully considers emotion perception data of the road segment when planning the route, so that the route planned by the present application can meet actual needs of passengers. BRIEF DESCRIPTION OF DRAWINGS

[0033] Figure 1 is a whole flowchart of the present application;

[0034] Figure 2 is a schematic diagram of a road segment and a route in an embodiment of the present application;

[0035] Figure 3 is a structure diagram of an environment-aware vehicle route planning device provided by the present application;

[0036] Figure 4 is an internal structure principle block diagram of a terminal device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0037] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments and accompanying drawings. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0038] Research has found that as vehicles travel along different routes from their starting point to their destination, the streetscapes along those routes provide passengers with varying experiences. This experience is linked to the visual quality of the streets, which encompasses not only physical conditions such as vegetation cover and visibility, but also the emotional impact of the streetscape: environmental aesthetics, historical and cultural characteristics, and more. These factors collectively constitute the "perceived quality" of the street, directly influencing tourists' perceptions and experiences. However, current technology, when planning vehicle routes, only considers the physical conditions of the streets, neglecting the emotional impact of the streetscape on passengers, resulting in routes that fail to meet their actual needs.

[0039] To address the aforementioned technical problems, this invention provides a vehicle route planning method, apparatus, and device based on environmental perception, solving the problem that routes planned by existing technologies are difficult to meet the actual needs of passengers. Specifically, the invention acquires various road segments from the starting point to the destination, along with street view images of each segment. It identifies physical element information and emotion perception data from the street view images. The physical element information characterizes the objects contained in the environment of the road segment, while the emotion perception data characterizes the emotional responses evoked by the environment. Based on the physical element information and emotion perception data of each road segment, a comprehensive score is obtained for each segment. The length of each road segment and the number of passengers required to travel on each segment are acquired. Based on the length of each road segment, the number of passengers on each segment, and the comprehensive score of each segment, connected target road segments are selected from the various road segments. Finally, the travel route from the starting point to the destination is planned based on the path of the target road segment.

[0040] For example, such as Figure 2 As shown, the vehicle needs to travel from the starting point S to the ending point T. The route from starting point S to ending point T includes: route SABT (consisting of road segments SA, AB, and BT); route SDEBT (consisting of road segments SD, DE, EB, and BT); route SEBT (consisting of road segments SE, EB, and BT); and route SEFGHT (consisting of road segments SE, EF, FG, GH, and HT). Figure 2There are other routes, not listed here. The present application collects street view images of the above 11 sections, a total of 11 street view images, obtains physical element information and emotional perception data of each section according to the street view image of each section, and then obtains the comprehensive score of each section according to the physical element information and the emotional perception data. Then, according to the length of each section, the number of passengers of each section who need to ride a vehicle, and the comprehensive score, the target section connected from the above 11 sections is selected. Since the connected sections include a section group composed of section SA, section AB, and section BT, the connected sections also include a section group composed of section SD, section DE, section EB, and section BT, the connected sections also include a section group composed of section SE, section EB, and section BT, and the connected sections also include a section group composed of section SE, section EF, section FG, section GH, and section HT. Therefore, the target section connected from the 11 sections is selected, that is, one of the above four groups of sections is selected as the target section. If the selected target section is the section group composed of section SA, section AB, and section BT, the vehicle is controlled to travel along the route SABT.

[0041] The vehicle route planning method based on environmental perception of the embodiment can be applied to a terminal device, which can be a terminal product with image acquisition and image processing functions, such as a car machine system, etc. In this embodiment, as shown in Figure 1 The vehicle route planning method based on environmental perception specifically includes the following steps:

[0042] S100, obtaining each section between the starting point and the ending point, and obtaining the street view image of each section, identifying the physical element information and emotional perception data on the street view image, the physical element information being used to represent the object information contained in the environment where the section is located, and the emotional perception data being used to represent the emotional response that can be triggered by the environment where the section is located;

[0043] The physical element information and the emotional perception data both belong to the environmental perception data.

[0044] S200, obtaining the comprehensive score of each section according to the physical element information and the emotional perception data of each section;

[0045] S300, obtaining the length of each section, obtaining the number of passengers of each section who need to ride a vehicle, and selecting a target section connected from each section according to the length of each section, the number of passengers of each section who need to ride a vehicle, and the comprehensive score of each section, and planning the driving route of the vehicle from the starting point to the ending point according to the path where the target section is located. The vehicle is a tourist bus.

[0046] In an embodiment, step S100 identifies sky pixels S, vegetation pixels V, and building pixels B on the street view image by a semantic segmentation model (semantic segmentation model DeepLab v3), and counts the number Num of sky pixels S, the number Num of vegetation pixels V, and the number Num of building pixels B. S V B The semantic segmentation model comprises a backbone network, a dilated convolution, an ASPP module, and a decoder connected in sequence.

[0047] The backbone network is ResNet-101, and the backbone network is configured to extract multi-scale features of the street view image. The dilated convolution can expand the receptive field while preserving high-resolution features, and is configured to process different scales of street view physical elements. The ASPP module (Atrous Spatial Pyramid Pooling) is configured to perform multi-scale processing on the feature map output by the dilated convolution using convolution kernels of multiple dilation rates, to generate a comprehensive feature map with rich context information. The decoder is configured to segment the street view image according to the comprehensive feature map at a pixel level, to obtain sky pixels, vegetation pixels, and building pixels. The sky pixels are counted to obtain the number S of sky pixels, the vegetation pixels are counted to obtain the number V of vegetation pixels, and the building pixels are counted to obtain the number B of building pixels.

[0048] The image classification model (the image classification model adopts a ResNet architecture) is configured to determine a vitality score Score E , a beauty score Score b , a safety score Score s , a richness score Score W , a boredom score Score Bo , and a depression score Score D The image classification model comprises a convolution layer, a residual module, a global average pooling layer, and a fully connected layer connected in sequence.

[0049] The convolution layer is a 7*7 convolution kernel, and the convolution layer is configured to extract preliminary features of the street view image. The residual module is stacked by multiple basic residual modules (Basic Block or Bottleneck Block). Each module includes a convolution layer, a batch normalization (BatchNorm), and a ReLU activation function, and is connected by a skip connection to enhance gradient propagation. The global average pooling layer (Global Average Pooling) performs global pooling on the feature map, and compresses the spatial dimension H*W to 1 dimension. The fully connected layer (Fully Connected Layer) outputs scores of six emotional dimensions, that is, the fully connected layer outputs a vitality score Score​​E , beauty score Score b , security score Score s , richness score Score W , boredom score Score Bo , and depression score Score D .

[0050] The semantic segmentation model and the image classification model of this embodiment are models after training, wherein the training steps of the model include obtaining original street view images, preprocessing the original street view images to obtain sample street view images, constructing a physical perception sample dataset, constructing an emotional perception sample dataset, and training the model.

[0051] The obtaining of the original street view images and the preprocessing of the original street view images include the following specific steps:

[0052] The road network of the sample area is divided according to a preset interval d, so as to ensure that the sampling points can comprehensively cover the key areas and capture the typical features of the road segments. For each road segment on the road network, the starting point coordinates are set as (x0, y0), and the end point coordinates are set as (x n , y n , then the sampling point coordinates along the road segment can be calculated by interpolation:

[0053]

[0054] Wherein i is the sampling point serial number, and n is the total number of sampling points predetermined for the road segment, which is determined by the length of the road segment and d.

[0055] The API usage permission is applied, and after the application, an API key (API_KEY) can be obtained, that is, the authorization is obtained, then the API request parameters are configured, the request parameters include but are not limited to the longitude and latitude coordinates of the target position, the angle and resolution of the required image, etc., then the original local street view images of different angles at each sampling point position on the same road segment meeting the request parameters are obtained through the API, then the original local street view images of different angles at each sampling point position are spliced to obtain the original street view image of the road segment, and the original street view image is a panoramic image, wherein the splicing needs to use image processing technology to identify and match the overlapping areas between different images, and the seamless panoramic view is generated by automatic splicing through algorithm.

[0056] After generating the original street view image, preprocessing operations such as verification and cleaning are performed to obtain sample street view images. Next, the sample street view images need to be stored. This embodiment uses a structured database model to store the sample street view images. When the database model is a non-relational database (such as MongoDB), a unique identifier (ID) needs to be assigned to each sample street view image, and its location information, shooting time, and other metadata should be stored in relevant tables for subsequent querying of the sample street view images.

[0057] This embodiment constructs a physical sensing sample dataset, including the following specific steps:

[0058] Sky pixels, vegetation pixels, and building pixels in the sample street view image are labeled, with label 1 representing sky pixels, label 2 representing vegetation pixels, label 3 representing building pixels, and label 0 representing other pixels, to form a label matrix M. The row and column positions in M ​​represent the positions of the pixels in the sample street view image, and the element values ​​at the row and column positions represent the physical feature types of the sample street view image. The sample street view image and the label matrix M are used as the physical perception sample dataset.

[0059] This embodiment trains a semantic segmentation model, including the following specific steps:

[0060] Sample street view image I i The label matrix M is input into the semantic segmentation model, and the semantic segmentation model outputs a segmentation probability matrix. The element value in the value represents the corresponding I. i The probability that a pixel belongs to category k (k∈{1,2,3}) is based on And M calculate the cross-extraction loss L of the semantic segmentation model seg :

[0061]

[0062] N is I i The total number of pixels is used to calculate the cross-picking loss L. seg Then, based on the cross-cutting loss L seg Adjust the parameters of the semantic segmentation model, and then perform the next iteration of training until the cross-loss L is reached. seg Training terminates when the set conditions are met or the set number of iterations is reached, resulting in the trained semantic segmentation model.

[0063] This embodiment constructs an emotion-perceived sample dataset, including the following specific steps:

[0064] The sample street view images are labeled with vitality score, beauty score, safety score, richness score, boredom score, and depression score to obtain the score label Score′.E Score' b Score' s Score' W Score' Bo Score' D E represents vitality, b represents beauty, s represents safety, W represents wealth, Bo represents boredom, and D represents depression. Each label is determined by the average score of multi-dimensional evaluation. To ensure the fairness and consistency of the scoring process, multiple rounds of scoring process may be implemented, and periodic evaluation training and scoring standard review may be conducted to ensure that each evaluator has a high degree of consistency in understanding and applying the standards. The sample street view images and their corresponding Score' E Score' b Score' s Score' W Score' Bo Score' D are used as emotion perception sample data sets.

[0065] This embodiment trains an image classification model, including the following specific steps:

[0066] The sample street view images are input into the image classification model, and the image classification model outputs a training score Score" i According to the score label Score' i and the training score Score" i , a loss function L is calculated. emo

[0067]

[0068] The parameters of the image classification model are adjusted according to the loss function L emo , and the next iteration training is performed until the loss function L emo reaches a set condition or the number of iterations reaches a set number of times, and the training is terminated, obtaining the trained image classification model.

[0069] In example two, based on example one, when collecting street view images of each road segment, the street view images are provided with geographical coordinates of their shooting positions, i.e., each street view image has corresponding shooting geographical coordinates. GIS data (such as road center line shapefile or GeoJSON format) is used to obtain spatial geometric information (spatial geometric information includes road geographical coordinates) of the road segment. The shooting geographical coordinates of each street view image are projected onto the corresponding road segment using the nearest distance method, and if the distance between the shooting geographical coordinates and the road geographical coordinates is less than a threshold value, the street view image is considered to belong to the road segment, wherein the threshold value is 10 meters. ​

[0070] After the street view images of each road segment are collected, the semantic segmentation model and the image classification model trained in Embodiment One are respectively applied to the street view images to obtain the physical element information and the emotional perception data of each road segment.

[0071] In Embodiment Three based on Embodiment One or Embodiment Two, step S200 comprises the following specific steps S201, S202 and S203.

[0072] S201, determine the positive physical indicators and the negative physical indicators in the physical element information, wherein the positive physical indicators are the physical indicators that can enhance the passenger experience, and the negative physical indicators are the physical indicators that reduce the passenger experience.

[0073] The physical element information comprises the number Num S of sky pixel points S, the number Num V of vegetation pixel points V, and the number Num B of building pixel points B.

[0074] Since buildings and vegetation can increase visual aesthetics and improve overall perception quality, the number Num B of building pixel points B, the number Num V of vegetation pixel points V are positive physical indicators. Too much sky may represent a vacant street or lack of coverage, reflecting a low-quality street experience, so the lower the number Num S of sky pixel points S is, the better it is, and thus the number Num S of sky pixel points S is a negative physical indicator.

[0075] S202, determine the positive emotional perception indicators and the negative emotional perception indicators in the emotional perception data, wherein the positive emotional perception indicators are the emotional indicators that enhance the passenger experience, and the negative emotional perception indicators are the emotional indicators that reduce the passenger experience.

[0076] The emotional perception data comprises the vitality score Score E , the beauty score Score b , the safety score Score s , the richness score Score W , the boredom score Score Bo , and the depression score Score D , wherein the value range of Score E , Score b , Score s , Score W , Score Bo , and Score Bo is 1-10.

[0077] Since the vitality, beauty, safety, and richness dimensions directly represent the attractiveness, safety, and beauty of the street, and are the core of the preferences of tourists, the higher the value, the better, the vitality score Score E , the beauty score Score b , the safety score Score s , and the richness score Score W are positive emotional perception indicators. Boredom is a negative emotional dimension, and the lower the value, the better. Depression is also a negative emotional dimension, and the lower the value, the better. Therefore, the boredom score Score Bo and the depression score Score D are negative emotional perception indicators.

[0078] S203, according to the positive emotional perception indicators, the negative emotional perception indicators, the positive physical indicators, and the negative physical indicators of each road segment, obtaining a comprehensive score of each road segment.

[0079] For the same road segment, a plurality of street view images are collected, and the number of sky pixel points S Num S , the number of vegetation pixel points V Num V , the number of building pixel points B Num B , the vitality score Score E , the beauty score Score b , the safety score Score s , the richness score Score W , the boredom score Score Bo , and the depression score Score D of each street view image are calculated, and then the mean values of the nine indicators are calculated using a plurality of street view images to obtain the mean value of the number of sky pixel points S , the mean value of the number of vegetation pixel points V , the mean value of the number of building pixel points B , the mean value of the vitality , the mean value of the beauty , the mean value of the safety , the mean value of the richness , the mean value of the boredom , and the mean value of the depression are calculated. Then the nine mean values are normalized respectively to obtain the normalized number of sky pixel points χ norm,S , the normalized number of vegetation pixel points χ norm,V , the normalized number of building pixel points χ norm,B , the normalized value of vitality χ norm,E , the normalized value of beauty χ norm,b , the normalized value of safety χ norm,s, the normalized value χ norm,W , the normalized value χ norm,Bo , the normalized value χ norm,D .

[0080] wherein

[0081] including χ min is the minimum value of the jth index, χ max is the maximum value of the jth index.

[0082] the comprehensive score Q of the kth road segment k :

[0083]

[0084] w j′ is the weight of the j'th positive index, χ norm,j′ is the normalized value of the j'th positive index, χ norm,j′ including the normalized number χ norm,V , the normalized number χ norm,B , the normalized value χ norm,E , the normalized value χ norm,b , the normalized value χ norm,s , the normalized value χ norm,W ; w j″ is the weight of the j''th negative index, χ norm,j″ is the normalized value of the j''th negative index, χ norm,j″ including the normalized value χ norm,Bo , the normalized value χ norm,D , the normalized number χ norm,S .

[0085] In an embodiment four, the embodiment four is based on the embodiment one or the embodiment two or the embodiment three, and in the embodiment four, the step S300 comprises the following specific steps S301-S304:

[0086] S301, dividing each of the road segments into a plurality of groups, and each of the road segments in each of the groups has connectivity.

[0087] Dividing all the road segments covered from the start point to the end point into a plurality of groups, and the path formed by all the road segments in each of the groups can allow a vehicle to travel from the start point to the end point.

[0088] S302, establishing a first function of each of the groups according to the passenger number of the road segments in each of the groups; and establishing a second function of each of the groups according to the comprehensive score Q of the road segments in each of the groups.k , and a second function of each of the groups is established.

[0089] the first function DS of the mth group m : wherein d mk is the number of passengers corresponding to the kth road section in the mth group, that is, the number of passengers who need to take the vehicle on the kth road section, S m is the total number of road sections in the mth group. The second function VQ m of the mth group is:

[0090] S303, obtaining the capacity limit C of the vehicle (the capacity limit C is the maximum number of people that the vehicle can accommodate within a safe range), obtaining the straight-line length L from the starting point to the ending point sp , and establishing a constraint condition of each of the groups according to the capacity limit C and the straight-line length L sp .

[0091] S304, screening a target group from each of the groups according to the constraint condition, the first function and the second function of each of the groups, and taking each of the road sections in the target group as a target road section.

[0092] A preselected group is screened from all groups, and the preselected group satisfies the constraint condition, that is, the sum of the lengths of all road sections in the preselected group is less than or equal to aL sp , and the sum of the passenger numbers corresponding to all road sections in the preselected group is less than or equal to the capacity limit C, wherein a is a constant greater than or equal to 1.

[0093] A target group is further screened from the preselected group, wherein the first function DS of the target group takes the maximum value among all groups, and the second function VQ m of the target group also takes the maximum value.

[0094] In the embodiment based on the first embodiment or the second embodiment or the third embodiment, step S300 includes the following specific steps: a route optimization model is constructed according to the length of each of the road sections, the passenger number of each of the road sections and the comprehensive score of each of the road sections; and a non-dominated sorting algorithm is applied to the route optimization model to screen connected target road sections from each of the road sections.

[0095] The objective function of the embodiment: Max DS =∑ (i,j)∈A d ij x ij

[0096]

[0097] Constraints:

[0098]

[0099] In the formula, DS represents the level of passenger flow satisfaction, VQI is the visual quality index value, and d ij Let C be the number of passengers needing to take a vehicle on the route between station i and station j, C be the vehicle capacity limit, A be the set of all stations, and Q be the number of passengers needing to take a vehicle on the route between station i and station j. ij x ij The comprehensive score for the road segment between station i and station j, l ij Let L be the length of the road segment between the i-th station and the j-th station, where α is a constant greater than or equal to 1. sp x is the straight line length from the starting point to the said ending point. ij Let x be a binary decision variable for the road segment between station i and station j. ij The value can be either 0 or 1.

[0100] This embodiment calculates the road segment that maximizes both DS and VQI under constraints. This road segment is the target road segment, which is calculated under constraints for which x values ​​are maximized. ij Taking the value 1 will maximize both DS and VQI. ij The corresponding road segment is the target road segment. Therefore, solving for the target road segment becomes solving for x under constraints that maximizes both DS and VQI. ij The value of .

[0101] This embodiment uses NSGA-II to solve x. ij The value is determined by the following specific steps:

[0102] Initialization: Randomly generate a population, where each individual represents a vehicle route design. The stations along the route are represented as a permutation of station numbers (e.g., [A, B, C, D]). The population size is N (typically 50-200).

[0103] Fitness calculation: Calculate the target value for each individual based on DS and VQI.

[0104] Non-dominated ranking: Stratify by dominance relationship and calculate crowding distance. Individual A dominates individual B if A is no worse than B in all objectives (DS). A ≥DS B VQI A ≥VQI B A is superior to B in at least one objective. Crowding distance calculation: For individuals in each layer, crowding is calculated based on the distribution of objective values.

[0105] Crowding Distance =∑ k (f k+1 -f k-1 ) / (f max -f min )

[0106] f k+1 is the target value of the next individual of the current individual i on the target k.f k-1 is the target value of the last individual of the current individual i on the target k.f max and f min represent the maximum and minimum values of all individuals in the population on the target k, respectively. Individuals with higher crowding distance are preferred to be preserved to maintain the diversity of the population.

[0107] Genetic operations: generate the next generation through selection, crossover and mutation. Through tournament selection: compare the non-dominated layers of two individuals, preferentially select the individual with lower layer (better solution). Randomly select two parents, exchange part of the site sequence, and ensure that the generated offspring sequence is legal. Mutation site sequence mutation: randomly exchange two sites in the site sequence.

[0108] Stopping condition: reach the maximum number of generations or the Pareto front converges, at which time the value of x ij is obtained.

[0109] This embodiment also uses historical passenger flow data to identify high-demand lines and combines the distribution of tourist hotspots to select lines that can cover important scenic spots and carry larger passenger flow. After identifying the lines to be optimized, the above multi-objective optimization algorithm is used to optimize these lines. Finally, the results need to be evaluated to verify whether the effect meets the expectation. The optimization effect can be quantified by comparing the passenger flow satisfaction and passenger experience (such as visual quality) changes before and after optimization. In addition, if the optimization effect of some lines does not meet the standard, the reasons need to be further analyzed, and adjustments and iterative optimization need to be made for the problem link.

[0110] In summary, the present application comprehensively analyzes the quality of street environment through the fusion of physical perception and emotional perception dual-layer evaluation framework. At the physical perception level, key physical features of the street are extracted based on computer vision technology using large-scale street view images, such as vegetation coverage, sky index, etc. At the emotional perception level, an emotional perception system is constructed based on a deep learning model, which classifies and scores the emotional features of street view images to explore the potential impact of street environment on users' psychology and behavior, including whether it is relaxing or depressing, thereby achieving comprehensive evaluation in both physical and emotional dimensions. Meanwhile, the present application proposes a solution based on a multi-objective heuristic optimization algorithm for vehicle route optimization, effectively balancing the multi-dimensional needs in street visual quality evaluation. By introducing advanced optimization methods such as the non-dominated sorting genetic algorithm (NSGA-II), the model jointly optimizes public transportation efficiency and street visual indicators to balance the needs of each dimension, ensuring the robustness and applicability of the model in practical applications. The present application has achieved breakthroughs in theory and application in the fields of street visual quality evaluation, urban traffic optimization, and public space design.

[0111] The present embodiment also provides a vehicle route planning device based on environmental perception, as shown in Figure 3 The device includes the following components:

[0112] An image analysis module 01 is used to obtain each road segment between the starting point and the ending point, and obtain street view images of each road segment, identify physical element information and emotional perception data on the street view images, the physical element information is used to represent the object information contained in the environment where the road segment is located, and the emotional perception data is used to represent the emotional response that can be triggered by the environment where the road segment is located;

[0113] A scoring module 02 is used to obtain the comprehensive score of each road segment according to the physical element information and the emotional perception data of each road segment;

[0114] A planning module 03 is used to obtain the length of each road segment, obtain the number of passengers who need to ride vehicles on each road segment, and select connected target road segments from each road segment according to the length of each road segment, the number of passengers on each road segment, and the comprehensive score of each road segment, and plan the driving route of the vehicle from the starting point to the ending point according to the path where the target road segment is located.

[0115] Based on the above embodiment, the present application also provides a terminal device, and the principle block diagram thereof can be as shown in Figure 4As shown in the figure. The terminal device includes a processor, a memory, a network interface, a display screen connected through a system bus. Among them, the processor of the terminal device is used to provide computing and control capability. The memory of the terminal device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The network interface of the terminal device is used to communicate with the external terminal through the network connection. The computer program is executed by the processor to implement an environment-aware vehicle route planning method. The display screen of the terminal device can be a liquid crystal display screen or an electronic ink display screen.

[0116] Those skilled in the art can understand that, Figure 4 The person skilled in the art can understand that,

[0117] In one embodiment, a terminal device is provided, which includes a memory, a processor, and an environment-aware vehicle route planning program stored in the memory and executable on the processor. When the processor executes the environment-aware vehicle route planning program, the following operation instructions are implemented:

[0118] Obtain each road segment between the starting point and the ending point, and obtain the street view image of each road segment, identify the physical element information and the emotional perception data on the street view image, the physical element information is used to represent the object information contained in the environment of the road segment, and the emotional perception data is used to represent the emotional response that can be triggered by the environment of the road segment;

[0119] According to the physical element information and the emotional perception data of each road segment, obtain the comprehensive score of each road segment;

[0120] Obtain the length of each road segment, obtain the passenger number of each road segment that needs to be taken by the vehicle, and according to the length of each road segment, the passenger number of each road segment and the comprehensive score of each road segment, select the target road segment from each road segment, and plan the driving route of the vehicle from the starting point to the ending point according to the path where the target road segment is located.

[0121] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, storage, database or other medium used in each embodiment of the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0122] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An environment perception based vehicle route planning method, characterized by, The method comprises: obtaining each road segment between a starting point and an ending point, and obtaining a street view image of each road segment, identifying physical element information and emotional perception data on the street view image, the physical element information being used to represent object information contained in an environment where the road segment is located, and the emotional perception data being used to represent emotional reactions that can be triggered by the environment where the road segment is located; obtaining a comprehensive score of each road segment according to the physical element information and the emotional perception data of each road segment; obtaining the length of each road segment, obtaining the number of passengers who need to ride a vehicle on each road segment, and selecting a connected target road segment from each road segment according to the length of each road segment, the number of passengers on each road segment, and the comprehensive score of each road segment, and planning a driving route of the vehicle from the starting point to the ending point according to the path where the target road segment is located. 2.The environment perception based vehicle route planning method of claim 1, wherein, The number of sky pixel points, the number of vegetation pixel points, and the number of building pixel points on the street view image are identified by a semantic segmentation model, and the number of sky pixel points, the number of vegetation pixel points, and the number of building pixel points are used as the physical element information; the semantic segmentation model comprises a backbone network, a dilated convolution, an ASPP module, and a decoder connected in sequence. 3.The environment perception based vehicle route planning method of claim 1, wherein, The vitality score, the beauty score, the safety score, the richness score, the boredom score, and the depression score corresponding to the street view image are determined by an image classification model, and the vitality score, the beauty score, the safety score, the richness score, the boredom score, and the depression score are used as the emotional perception data; the image classification model comprises a convolution layer, a residual module, a global average pooling layer, and a full connection layer connected in sequence. 4.The environment-aware based vehicle route planning method of claim 1, wherein, The comprehensive score of each road segment is obtained according to the physical element information and the emotional perception data of each road segment, which comprises: determining positive physical indicators and negative physical indicators in the physical element information, the positive physical indicators being physical indicators that can enhance passenger experience, and the negative physical indicators being physical indicators that can reduce passenger experience; determining positive emotional perception indicators and negative emotional perception indicators in the emotional perception data, the positive emotional perception indicators being emotional indicators that can enhance passenger experience, and the negative emotional perception indicators being emotional indicators that can reduce passenger experience; obtaining the comprehensive score of each road segment according to the positive emotional perception indicators, the negative emotional perception indicators, the positive physical indicators, and the negative physical indicators of each road segment.

5. The environment-aware based vehicle route planning method according to any one of claims 1-4, wherein, The distance between the shooting geographic coordinates of the street view image and the road geographic coordinates of the road segment is less than a threshold value. 6.The environment perception based vehicle route planning method of claim 1, wherein, The connected target road segment is selected from each road segment according to the length of each road segment, the number of passengers on each road segment, and the comprehensive score of each road segment, which comprises: dividing each road segment into a plurality of groups, and each road segment in each group having connectivity; establishing a first function of each of the groups according to the passenger number of the road sections in each of the groups; and establishing a second function of each of the groups according to the comprehensive score of the road sections in each of the groups; obtaining a capacity limit of the vehicle, obtaining a straight line length from the starting point to the ending point, and establishing a constraint condition of each of the groups according to the capacity limit and the straight line length; screening a target group from each of the groups according to the constraint condition, the first function and the second function of each of the groups, and taking each of the road sections in the target group as a target road section. 7.The environment perception based vehicle route planning method of claim 1, wherein, The screening of the connected target road sections from each of the road sections according to the length of each of the road sections, the passenger number of each of the road sections and the comprehensive score of each of the road sections comprises: constructing a route optimization model according to the length of each of the road sections, the passenger number of each of the road sections and the comprehensive score of each of the road sections; applying a non-dominated sorting algorithm to the route optimization model to screen the connected target road sections from each of the road sections.

8. An environment perception based vehicle route planning device, characterized by, The device comprises the following components: an image analysis module, configured to obtain each road section between a starting point and an ending point, and obtain a street view image of each road section, identify physical element information and emotional perception data on the street view image, the physical element information being used to represent object information contained in an environment where the road section is located, and the emotional perception data being used to represent emotional reactions that can be triggered by the environment where the road section is located; a scoring module, configured to obtain a comprehensive score of each road section according to the physical element information and the emotional perception data of each road section; a planning module, configured to obtain a length of each road section, obtain a passenger number of a vehicle that needs to be taken on each road section, screen connected target road sections from each of the road sections according to the length of each of the road sections, the passenger number of each of the road sections and the comprehensive score of each of the road sections, and plan a driving route of the vehicle from the starting point to the ending point according to a path where the target road sections are located.

9. A terminal device, comprising: The terminal device comprises a memory, a processor and an environment perception based vehicle route planning program stored in the memory and executable on the processor, and the processor implements the steps of the environment perception based vehicle route planning method according to any one of claims 1-7 when executing the environment perception based vehicle route planning program.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores an environment perception based vehicle route planning program, and the processor implements the steps of the environment perception based vehicle route planning method according to any one of claims 1-7 when executing the environment perception based vehicle route planning program.

Images