Parking lot identification method, and training method, device and equipment of parking lot identification model

Abstract

The invention provides a parking lot identification method, and a training method, device and equipment of a parking lot identification model, and relates to the technical field of artificial intelligence, in particular to the technical field of automatic driving, intelligent traffic and machine learning. According to the specific implementation scheme, parking track points corresponding to a parking lot are determined from driving tracks of different vehicles; clustering the parking track points to obtain a parking point cluster corresponding to the parking lot; performing feature extraction on the stay point cluster to obtain a cluster feature of the stay point cluster; and determining the identification type of the parking lot according to the cluster features. According to the technical scheme, the accuracy of parking lot identification can be improved.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, and more specifically to the field of autonomous driving, intelligent transportation and machine learning. Background Art

[0002] With the continuous development of intelligent navigation technology, more and more drivers are looking for parking lots through navigation products. In navigation products, parking lots, as one of the important service facilities, play a very important role in improving the driver's travel experience and alleviating urban traffic pressure. However, the construction of parking lot data for navigation products also faces problems such as missing data, low data quality, and untimely data updates. Therefore, strengthening the construction of parking lot data and improving the quality and timeliness of parking lot information are of vital importance to the development of intelligent navigation products. Summary of the invention

[0003] The present invention provides a parking lot recognition method, a parking lot recognition model training method, a device and equipment.

[0004] According to one aspect of the present disclosure, a parking lot identification method is provided, the method comprising:

[0005] Determine the stop trajectory points corresponding to the parking lot from the driving trajectories of different vehicles;

[0006] Clustering the stop trajectory points to obtain a stop point cluster corresponding to the parking lot;

[0007] Extracting features from the stay point cluster to obtain cluster features of the stay point cluster;

[0008] The identification type of the parking lot is determined according to the cluster features.

[0009] According to another aspect of the present disclosure, a method for training a parking lot recognition model is provided, the method comprising:

[0010] Determine the stop trajectory points corresponding to the parking lot from the driving trajectories of different vehicles;

[0011] Clustering the stop trajectory points to obtain a stop point cluster corresponding to the parking lot;

[0012] Extracting features from the stay point cluster to obtain cluster features of the stay point cluster;

[0013] Determining the identification type of the parking lot according to the cluster features;

[0014] The parking lot recognition model is trained according to the recognition type and the label data of the stay point cluster.

[0015] According to another aspect of the present disclosure, a parking lot identification device is provided, the device comprising:

[0016] A first stop track point determination module, used to determine the stop track point corresponding to the parking lot from the driving tracks of different vehicles;

[0017] A first stop point cluster determination module, used for clustering the stop trajectory points to obtain a stop point cluster corresponding to the parking lot;

[0018] A first cluster feature determination module is used to extract features from the stay point cluster to obtain cluster features of the stay point cluster;

[0019] The first recognition category determination module is used to determine the recognition type of the parking lot according to the cluster features.

[0020] According to another aspect of the present disclosure, a training device for a parking lot recognition model is provided, comprising:

[0021] A second stop track point determination module is used to determine the stop track point corresponding to the parking lot from the driving tracks of different vehicles;

[0022] A second stop point cluster determination module is used to cluster the stop trajectory points to obtain a stop point cluster corresponding to the parking lot;

[0023] A second cluster feature determination module is used to extract features from the stay point cluster to obtain cluster features of the stay point cluster;

[0024] A second identification category determination module, used to determine the identification type of the parking lot according to the cluster features;

[0025] The model training module is used to train the parking lot recognition model according to the recognition type and the label data of the stop point cluster.

[0026] According to another aspect of the present disclosure, there is provided an electronic device, the electronic device comprising:

[0027] At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the parking lot recognition method or the parking lot recognition model training method provided by the present disclosure.

[0028] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable a computer to execute the parking lot recognition method or the parking lot recognition model training method provided by the present disclosure.

[0029] According to another aspect of the present disclosure, a computer program product is provided, including a computer program, which, when executed by a processor, implements the parking lot recognition method or the parking lot recognition model training method provided in the present disclosure.

[0030] According to the technology disclosed in the present invention, the accuracy of parking lot recognition can be improved.

[0031] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0033] Figure 1 is a flow chart of a parking lot identification method provided according to an embodiment of the present disclosure;

[0034] Figure 2 is a flowchart of another parking lot identification method provided according to an embodiment of the present disclosure;

[0035] Figure 3 is a flowchart of a method for training a parking lot recognition model according to an embodiment of the present disclosure;

[0036] Figure 4 is a structural schematic diagram of another parking lot identification device provided according to an embodiment of the present disclosure;

[0037] Figure 5 It is a structural schematic diagram of a parking lot recognition model training device provided according to an embodiment of the present disclosure;

[0038] Figure 6 The present invention is a block diagram of an electronic device for implementing the parking lot recognition method or the parking lot recognition model training method according to the embodiment of the present disclosure. DETAILED DESCRIPTION

[0039] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0040] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0041] In addition, it should be noted that in the technical solution of the present invention, the collection, storage, use, processing, transmission, provision and disclosure of relevant data such as driving trajectories are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0042] By analyzing the positioning effects of different vehicles in the basement, it is found that the positioning results of vehicles of different models when parking are roughly divided into two types: the basement entrance or the actual parking position, and the positioning difference is strongly correlated with the model. Low-end vehicles (with lower production costs and relatively poor hardware and software such as positioning chips and inertial navigation) tend to gather at the basement entrance, while high-end vehicles with higher costs can be positioned at the actual parking position. Therefore, the present disclosure provides a method for determining the type of parking lot based on the vehicle's driving trajectory.

[0043] Figure 1 1 is a flow chart of a parking lot identification method provided according to an embodiment of the present disclosure. The method is applicable to the situation of how to accurately and quickly identify parking lots in navigation scenarios. The method can be executed by a parking lot identification device, which can be implemented in software and / or hardware and can be integrated into an electronic device that carries the parking lot identification function, such as a server. Figure 1 As shown, the parking lot identification method of this embodiment may include:

[0044] S101, determining a stop trajectory point corresponding to a parking lot from driving trajectories of different vehicles.

[0045] In this embodiment, the driving trajectory refers to the driving trajectory of the vehicle obtained from different vehicle terminals; optionally, the driving trajectory includes multiple trajectory points, each trajectory point includes information such as the terminal identification of the vehicle terminal, vehicle model, instantaneous speed, location and time.

[0046] The so-called stop trajectory point refers to the trajectory point in the driving trajectory that keeps the same position in a continuous time.

[0047] An optional method is to obtain the driving trajectory of the vehicle entering the parking lot from different vehicle terminals; then, for each vehicle's driving trajectory, the trajectory points belonging to the same set period in the driving trajectory are sorted in chronological order to obtain the sorted trajectory points; and then, for each sorted trajectory point, determine whether the position of the trajectory point changes in a continuous time period. If not, the trajectory point is used as a stop trajectory point.

[0048] S102, clustering the stop trajectory points to obtain a stop point cluster corresponding to the parking lot.

[0049] In this embodiment, the stay point cluster refers to a cluster formed by clustering multiple stay points.

[0050] Specifically, the stop trajectory points determined to enter the parking lot within a set time period can be obtained, and the stop trajectory points can be clustered based on a preset clustering method to obtain a stop point cluster corresponding to the parking lot. It should be noted that the preset clustering method is not specifically limited in this embodiment and can be selected by those skilled in the art according to actual needs, for example, it can be density-based spatial clustering of applications with noise (Density-Based Spatial Clustering of Applications with Noise, DBSCAN), or hierarchical density-based spatial clustering of applications with noise (Hierarchical Density-Based Spatial Clustering of Applications with Noise, HDBSCAN).

[0051] S103, extracting features from the stay point cluster to obtain cluster features of the stay point cluster.

[0052] In this embodiment, cluster features refer to the features of the stop point cluster, which can be represented by a vector or a matrix. Optionally, the cluster features include at least one of vehicle type distribution features, cluster area features, the number of stop points, and dispersion distance features. Among them, the vehicle type distribution feature refers to the feature obtained by encoding the model of the vehicle corresponding to the stop point cluster in the stop point cluster, which can be represented by a vector or a matrix. The so-called cluster area feature refers to the area feature of the outline of the stop point cluster. The so-called number of stop points refers to the number of stop points in the stop point cluster. The so-called dispersion distance feature refers to the feature used to characterize the distribution of stop points in the stop point cluster. It can be understood that the cluster features contain features of different dimensions, which can characterize the stop point cluster in multiple dimensions, laying the foundation for subsequent parking lot identification.

[0053] Specifically, based on the feature extraction model, feature extraction can be performed on the stay point cluster to obtain the cluster features of the stay point cluster. It should be noted that the feature extraction model can be obtained based on a machine learning algorithm.

[0054] An optional method is to extract the vehicle type distribution characteristics, cluster area characteristics, number of stay points and dispersion distance characteristics of the stay point cluster respectively, and then select any one of the vehicle type distribution characteristics, cluster area characteristics, number of stay points and dispersion distance characteristics as the cluster feature of the stay point cluster.

[0055] An optional method is to extract the vehicle type distribution characteristics, cluster area characteristics, number of stay points and dispersion distance characteristics of the stay point cluster respectively, and then select any two features from the vehicle type distribution characteristics, cluster area characteristics, number of stay points and dispersion distance characteristics, and fuse these two features to obtain the cluster characteristics of the stay point cluster.

[0056] An optional method is to extract the vehicle type distribution characteristics, cluster area characteristics, number of stay points and dispersion distance characteristics of the stay point cluster respectively, and then select any three characteristics from the vehicle type distribution characteristics, cluster area characteristics, number of stay points and dispersion distance characteristics, and fuse these three characteristics to obtain the cluster characteristics of the stay point cluster.

[0057] An optional method is to extract the vehicle type distribution characteristics, cluster area characteristics, number of stop points and dispersion distance characteristics of the stop point cluster respectively, and then fuse these four characteristics to obtain the cluster characteristics of the stop point cluster.

[0058] S104: Determine the identification type of the parking lot according to the cluster features.

[0059] In this embodiment, the identification type refers to the predicted type of parking lot; optionally, the identification type includes a ground parking lot, an underground parking lot, or an underground parking lot entrance.

[0060] Specifically, the parking lot type can be predicted based on the parking lot identification model according to the cluster features of the stop point cluster to obtain the identification type of the parking lot. It should be noted that the parking lot identification model can be pre-trained based on a machine learning algorithm.

[0061] The technical solution provided by the embodiment of the present disclosure determines the stop trajectory points corresponding to the parking lot from the driving trajectories of different vehicles, then clusters the stop trajectory points to obtain the stop point clusters corresponding to the parking lot, and then extracts features from the stop point clusters to obtain cluster features of the stop point clusters, and finally determines the identification type of the parking lot based on the cluster features. Compared with the prior art method of obtaining parking lot data through field collection or third-party cooperation, which is costly and difficult to implement on a large scale, the above technical solution extracts features after clustering the driving trajectories of different vehicles to identify the parking lot type, and mines parking lot data based only on vehicle driving trajectories, which reduces costs while also improving the accuracy of parking lot identification.

[0062] On the basis of the above embodiments, as an optional method of the present disclosure, determining the stop trajectory points corresponding to the parking lot from the driving trajectories of different vehicles includes: determining continuous trajectory points corresponding to the parking lot from the driving trajectories of different vehicles; and determining whether the continuous trajectory points are the stop trajectory points corresponding to the parking lot according to the instantaneous speed and position of the continuous trajectory points.

[0063] Among them, continuous trajectory points refer to trajectory points that are continuous in time and associated in position in the driving trajectory.

[0064] Specifically, the driving trajectory points of the same vehicle entering the parking lot in the same time period are sorted in chronological order to obtain continuous trajectory points; then for each continuous trajectory point, it is determined whether the instantaneous speed of the continuous trajectory point is 0 for a continuous set time period, and the position corresponding to the continuous trajectory point has almost no change, that is, the position change range is within the set range, then the continuous trajectory point is determined to be the parking trajectory point corresponding to the parking lot. It should be noted that the set time period can be set by a technician in this field according to actual conditions, such as 1 hour; the set range can be set by a technician in this field according to actual conditions, such as 5 meters.

[0065] It can be understood that the stop trajectory point is determined based on the trajectory point state of the driving trajectory, that is, the speed and position, so that the determination of the stop trajectory point is more accurate, thereby providing a guarantee for the subsequent parking lot identification.

[0066] Figure 2 is a flow chart of another parking lot identification method provided according to an embodiment of the present disclosure; based on the above embodiment, this embodiment further optimizes "extracting features from the stay point clusters to obtain cluster features of the stay point clusters" and provides an optional method. Figure 2 As shown, the parking lot identification method of this embodiment may include:

[0067] S201, determining a stop trajectory point corresponding to a parking lot from driving trajectories of different vehicles.

[0068] S202, clustering the stop trajectory points to obtain a stop point cluster corresponding to the parking lot.

[0069] S203, encoding the vehicle models corresponding to the stop trajectory points in the stop point cluster to obtain the vehicle model distribution characteristics of the stop point cluster.

[0070] Specifically, since different vehicles have a model, first determine the vehicle model corresponding to each stop trajectory point in the stop point cluster, then encode the vehicle model corresponding to each stop trajectory point based on the bag-of-words model to obtain the vehicle model distribution characteristics of the stop point cluster, and then use the vehicle model distribution characteristics of the stop point cluster as the cluster characteristics of the stop point cluster. For example, the top 1000 models in the vehicle data can be encoded into a 1000-dimensional vector, and the data corresponding to the vehicle model of each stop point in the stop point cluster can be counted, and the corresponding vector can be generated according to certain rules as the vehicle model distribution characteristics of the stop point cluster, and then the vehicle model distribution characteristics of the stop point cluster can be used as the cluster characteristics of the stop point cluster.

[0071] It should be noted that, through practical analysis, the positioning results of different vehicles when parking in underground parking lots can be roughly divided into two types: the entrance of the underground parking lot or the actual parking position, and the positioning difference is strongly correlated with the vehicle model. Since low-end vehicles have lower production costs and poor hardware and software such as positioning chips and inertial navigation, the positioning in underground parking lots is often the entrance of the underground parking lot, rather than the actual parking position of the vehicle; while high-end vehicles with higher costs can locate the actual parking position of the vehicle. Therefore, introducing the vehicle model corresponding to the stop trajectory point to determine the vehicle model distribution characteristics of the stop point cluster can increase the differences between parking lots, thereby enabling more accurate identification of the parking lot type.

[0072] S204: Determine the identification type of the parking lot according to the cluster features.

[0073] The technical solution provided by the embodiment of the present disclosure determines the stop trajectory points corresponding to the parking lot from the driving trajectories of different vehicles, then clusters the stop trajectory points to obtain the stop point cluster corresponding to the parking lot, and then encodes the vehicle models corresponding to the stop trajectory points in the stop point cluster to obtain the vehicle model distribution characteristics of the stop point cluster, and finally determines the identification type of the parking lot based on the cluster characteristics. The above technical solution introduces the vehicle model distribution characteristics to identify the identification type of the parking lot, and can fully learn the distinctive features of the stop point cluster, thereby providing a basis for the identification of the parking lot.

[0074] On the basis of the above embodiments, as an optional method of the present disclosure, feature extraction is performed on the stay point cluster to obtain cluster features of the stay point cluster, including: performing point-to-surface transformation on the stay trajectory points in the stay point cluster to obtain a cluster contour corresponding to the stay point cluster; determining the area of ​​the cluster contour, and using the area as the area feature of the stay point cluster.

[0075] The cluster contour refers to the corresponding contour of the stay point cluster, that is, the shape of the stay point cluster in the two-dimensional space, which can be a polygon.

[0076] Specifically, based on the point-to-surface transformation algorithm, the stop trajectory points in the stop point cluster can be transformed into a point-to-surface transformation, that is, the stop trajectory points in the stop point cluster are processed into a polygonal outline, and the polygonal outline is used as the cluster outline corresponding to the stop point cluster; wherein, in this embodiment, the point-to-surface transformation algorithm is not specifically limited, for example, the point-to-surface transformation algorithm can be a convex hull algorithm. Afterwards, the area of ​​the cluster outline is calculated, and the area is used as the area feature of the stop point cluster. Furthermore, the area can also be vectorized, and the processed result is used as the area feature of the stop point cluster.

[0077] It is understandable that, since different parking lots have different shapes and areas, by determining the cluster outline of the stay points and determining their area characteristics, the shape and area characteristics of the stay point cluster can be vividly characterized, thereby providing a basis for subsequent parking lot identification.

[0078] On the basis of the above embodiments, as an optional method of the present disclosure, feature extraction is performed on the stay point cluster to obtain cluster features of the stay point cluster, including: determining the cluster center of the stay point cluster; determining the distance between the stay trajectory point in the stay point cluster and the cluster center, and determining the dispersion distance feature of the stay point cluster based on the distance.

[0079] The cluster center refers to the centroid of the stop point cluster; to some extent, it can also be understood as the central position of the parking lot.

[0080] Specifically, the point at the center of the stay point cluster is used as the cluster center; then, for each stay trajectory point in the stay point cluster, the distance between the stay trajectory point and the cluster center is determined; then the mean of the distance between each stay trajectory point and the center of the cluster is calculated, and the mean is used as the dispersion distance feature of the stay point cluster.

[0081] It can be understood that since different vehicles have different positions in the parking lot, by determining the distance of each stop trajectory point relative to the cluster center of the stop point cluster, the characteristics of the stop point cluster can be more richly characterized from the distance dimension, thereby more vividly representing the characteristics of the parking lot and providing a basis for subsequent parking lot identification.

[0082] Based on the above embodiment, as an optional method of the present disclosure, after identifying the identification type of the parking lot according to the cluster features, the method further includes: if the identification type of the parking lot is an underground parking lot entrance, determining the parking lot entrance position according to the trajectory point position information of the stop trajectory point in the stop point cluster. Further, the parking lot entrance position is used for navigation guidance.

[0083] The track position information refers to the position information of the stop track point, which can be the coordinates (x, y) of a two-dimensional plane.

[0084] Specifically, if the identification type of the parking lot is identified as an underground parking entrance, the coordinate mean of each stop trajectory point in the stop point cluster is calculated, and the coordinate mean is used as the entrance position of the parking lot.

[0085] It should be noted that the basement (short for underground parking lot) is a relatively complex scene in the parking lot. Because basements are often distributed in complex high-rise buildings, the production of parking lot data will be affected by many unfavorable factors such as positioning interference and hidden location. Therefore, the production of the basement entrance is difficult, and there is a lack of high-quality and large-scale production methods. The current construction of underground parking entrance data mainly relies on the following methods: 1) Field collection, that is, by hiring professionals to collect data in the field, obtain the image and location information of the parking lot, and then extract the basic information such as the entrance and parking location of the parking lot through internal processing, as well as detailed information such as the number of parking spaces and charging standards. 2) Third-party cooperation, that is, through cooperation with parking service providers (such as Speed ​​Parking, Dashiyun Parking, etc.) or government agencies, some information about the parking lot is obtained, and then some data processing is performed to obtain applicable parking lot data.

[0086] It is understandable that, since the entrance position of an underground parking lot is difficult to determine, the present disclosure locates the entrance position of an underground parking lot through the vehicle driving trajectory, and can accurately determine the entrance position of the parking lot, thereby being used for navigation guidance of the parking lot.

[0087] Figure 3 1 is a flow chart of a method for training a parking lot recognition model according to an embodiment of the present disclosure. This embodiment is applicable to the situation of how to accurately and quickly recognize parking lots in navigation scenarios. The method can be executed by a training device for a parking lot recognition model, which can be implemented in software and / or hardware and can be integrated into an electronic device that carries the training function of the parking lot recognition model, such as a server. Figure 3 As shown, the training method of the parking lot recognition model of this embodiment may include:

[0088] S301, determining a stop trajectory point corresponding to a parking lot from driving trajectories of different vehicles.

[0089] In this embodiment, the driving trajectory refers to the driving trajectory of the vehicle obtained from different vehicle terminals; optionally, the driving trajectory includes multiple trajectory points, each trajectory point includes information such as the terminal identification of the vehicle terminal, vehicle model, instantaneous speed, location and time.

[0090] The so-called stop trajectory point refers to the trajectory point in the driving trajectory that keeps the same position in a continuous time.

[0091] An optional method is to obtain the driving trajectory of the vehicle entering the parking lot from different vehicle terminals; then, for each vehicle's driving trajectory, the trajectory points belonging to the same set period in the driving trajectory are sorted in chronological order to obtain the sorted trajectory points; and then, for each sorted trajectory point, determine whether the position of the trajectory point changes in a continuous time period. If not, the trajectory point is used as a stop trajectory point.

[0092] Another optional method is to determine the continuous trajectory points corresponding to the parking lot from the driving trajectories of different vehicles; and determine whether the continuous trajectory points are the stop trajectory points corresponding to the parking lot based on the instantaneous speed and position of the continuous trajectory points.

[0093] S302, clustering the stop trajectory points to obtain a stop point cluster corresponding to the parking lot.

[0094] In this embodiment, the stay point cluster refers to a cluster formed by clustering multiple stay points.

[0095] Specifically, the stop trajectory points determined to enter the parking lot within a set time period can be obtained, and the stop trajectory points can be clustered based on a preset clustering method to obtain a stop point cluster corresponding to the parking lot. It should be noted that the preset clustering method is not specifically limited in this embodiment and can be selected by those skilled in the art according to actual needs, for example, it can be density-based spatial clustering of applications with noise (Density-Based Spatial Clustering of Applications with Noise, DBSCAN), or hierarchical density-based spatial clustering of applications with noise (Hierarchical Density-Based Spatial Clustering of Applications with Noise, HDBSCAN).

[0096] S303: extract features from the stay point cluster to obtain cluster features of the stay point cluster.

[0097] In this embodiment, the cluster feature refers to the feature of the stay point cluster, which can be represented by a vector or a matrix. Optionally, the cluster feature includes at least one of the vehicle model distribution feature, cluster area feature, number of stay points and dispersion distance feature. Among them, the vehicle model distribution feature refers to the feature obtained by encoding the model of the vehicle corresponding to the stay point cluster in the stay point cluster, which can be represented by a vector or a matrix. The so-called cluster area feature refers to the area feature of the outline of the stay point cluster. The so-called number of stay points refers to the number of stay points in the stay point cluster. The so-called dispersion distance feature refers to the feature used to characterize the distribution of stay points in the stay point cluster.

[0098] Specifically, based on the feature extraction model, feature extraction can be performed on the stay point cluster to obtain the cluster features of the stay point cluster. It should be noted that the feature extraction model can be obtained based on a machine learning algorithm.

[0099] An optional method is to extract the vehicle type distribution characteristics, cluster area characteristics, number of stay points and dispersion distance characteristics of the stay point cluster respectively, and then select any one of the vehicle type distribution characteristics, cluster area characteristics, number of stay points and dispersion distance characteristics as the cluster feature of the stay point cluster.

[0100] An optional method is to extract the vehicle type distribution characteristics, cluster area characteristics, number of stay points and dispersion distance characteristics of the stay point cluster respectively, and then select any two features from the vehicle type distribution characteristics, cluster area characteristics, number of stay points and dispersion distance characteristics, and fuse these two features to obtain the cluster characteristics of the stay point cluster.

[0101] An optional method is to extract the vehicle type distribution characteristics, cluster area characteristics, number of stay points and dispersion distance characteristics of the stay point cluster respectively, and then select any three characteristics from the vehicle type distribution characteristics, cluster area characteristics, number of stay points and dispersion distance characteristics, and fuse these three characteristics to obtain the cluster characteristics of the stay point cluster.

[0102] An optional method is to extract the vehicle type distribution characteristics, cluster area characteristics, number of stop points and dispersion distance characteristics of the stop point cluster respectively, and then fuse these four characteristics to obtain the cluster characteristics of the stop point cluster.

[0103] It should be noted that the specific implementation method for determining the vehicle type distribution characteristics, cluster area characteristics, number of stay points and dispersion distance characteristics is as described in the above embodiment, and will not be repeated in this embodiment.

[0104] S304: Determine the identification type of the parking lot according to the cluster features.

[0105] Specifically, the cluster features can be input into a parking lot recognition model, and the recognition type of the parking lot can be obtained through model prediction.

[0106] S305: Training a parking lot recognition model according to the recognition type and label data of the stop point cluster.

[0107] In this embodiment, the label data of the stay point cluster refers to the type annotation data of the stay point cluster. The parking lot recognition model refers to a model for predicting parking lots based on the cluster features of the stay point cluster. Optionally, the parking lot recognition model can be a classification recognition model based on machine learning, such as a support vector machine (SVM) or an iterative decision tree (GBDT) model.

[0108] Specifically, the training loss can be determined based on the preset loss function and the label data of the recognition type and the stop point cluster, and the parking lot recognition model can be iteratively trained based on the training loss until the training stop condition is met and the model training is stopped. It should be noted that the training stop condition is that the training loss is stable within a set range or the number of iterations reaches a set number, wherein the set range and the set number can be set by those skilled in the art according to actual conditions.

[0109] The technical solution provided by the embodiment of the present disclosure determines the stop trajectory points corresponding to the parking lot from the driving trajectories of different vehicles, then clusters the stop trajectory points to obtain the stop point clusters corresponding to the parking lot, and then extracts features from the stop point clusters to obtain cluster features of the stop point clusters, and determines the identification type of the parking lot based on the cluster features, and finally trains the parking lot identification model based on the identification type and the label data of the stop point clusters. Compared with the prior art method of obtaining parking lot data through field collection or third-party cooperation, which is costly and difficult to implement on a large scale, the above technical solution extracts features after clustering the driving trajectories of different vehicles to identify the parking lot type, and mines parking lot data based only on vehicle driving trajectories, thereby reducing costs while improving the accuracy of parking lot identification.

[0110] Figure 4 is a schematic diagram of the structure of another parking lot identification device provided according to an embodiment of the present disclosure. The method is applicable to the situation of how to accurately and quickly identify parking lots in navigation scenarios. The device can be implemented in software and / or hardware, and can be integrated into an electronic device that carries the parking lot identification function, such as a server. Figure 4 As shown, the parking lot identification device of this embodiment may include:

[0111] A first stop track point determination module 401 is used to determine the stop track point corresponding to the parking lot from the driving tracks of different vehicles;

[0112] A first stop point cluster determination module 402, for clustering the stop trajectory points to obtain a stop point cluster corresponding to the parking lot;

[0113] The first cluster feature determination module 403 is used to extract features from the stay point cluster to obtain cluster features of the stay point cluster;

[0114] The first recognition category determination module 404 is used to determine the recognition type of the parking lot according to the cluster features.

[0115] The technical solution provided by the embodiment of the present disclosure determines the stop trajectory points corresponding to the parking lot from the driving trajectories of different vehicles, then clusters the stop trajectory points to obtain the stop point clusters corresponding to the parking lot, and then extracts features from the stop point clusters to obtain cluster features of the stop point clusters, and finally determines the identification type of the parking lot based on the cluster features. Compared with the prior art method of obtaining parking lot data through field collection or third-party cooperation, which is costly and difficult to implement on a large scale, the above technical solution extracts features after clustering the driving trajectories of different vehicles to identify the parking lot type, and mines parking lot data based only on vehicle driving trajectories, which reduces costs while also improving the accuracy of parking lot identification.

[0116] Furthermore, the first stop trajectory point determination module 401 is specifically used for:

[0117] Determine continuous trajectory points corresponding to the parking lot from the driving trajectories of different vehicles;

[0118] According to the instantaneous speed and position of the continuous trajectory point, it is determined whether the continuous trajectory point is a stop trajectory point corresponding to the parking lot.

[0119] Furthermore, the cluster feature includes at least one of a vehicle type distribution feature, a cluster area feature, a number of stay points, and a dispersion distance feature.

[0120] Furthermore, the first cluster feature determination module 403 is specifically used for:

[0121] The vehicle models corresponding to the stop trajectory points in the stop point cluster are encoded to obtain the vehicle model distribution characteristics of the stop point cluster.

[0122] Furthermore, the first cluster feature determination module 403 is specifically used for:

[0123] Perform point-to-surface transformation on the stop trajectory points in the stop point cluster to obtain the cluster outline corresponding to the stop point cluster;

[0124] The area of ​​the cluster outline is determined and used as the area feature of the stay point cluster.

[0125] Furthermore, the first cluster feature determination module 403 is specifically used for:

[0126] Determine the cluster center of the stay point cluster;

[0127] The distance between the stay trajectory points in the stay point cluster and the cluster center is determined, and the dispersion distance characteristics of the stay point cluster are determined based on the distance.

[0128] Further, the identification type includes a ground parking lot, an underground parking lot, or an underground parking lot entrance.

[0129] Furthermore, the device also includes:

[0130] The entrance position determination module is used to determine the entrance position of the parking lot according to the trajectory point position information of the stop trajectory point in the stop point cluster if the identification type of the parking lot is the entrance of the underground parking lot.

[0131] Furthermore, the entrance location of the parking lot is used for navigation guidance.

[0132] Figure 5 1 is a schematic diagram of a parking lot recognition model training device according to an embodiment of the present disclosure. This embodiment is applicable to the situation of how to accurately and quickly recognize parking lots in navigation scenarios. The device can be implemented in software and / or hardware, and can be integrated into an electronic device that carries the parking lot recognition model training function, such as a server. Figure 5 As shown, the training device of the parking lot recognition model of this embodiment may include:

[0133] A second stop track point determination module 501 is used to determine the stop track point corresponding to the parking lot from the driving tracks of different vehicles;

[0134] A second stop point cluster determination module 502 is used to cluster the stop trajectory points to obtain a stop point cluster corresponding to the parking lot;

[0135] The second cluster feature determination module 503 is used to extract features from the stay point cluster to obtain cluster features of the stay point cluster;

[0136] A second identification category determination module 504, used to determine the identification type of the parking lot according to the cluster features;

[0137] The model training module 505 is used to train the parking lot recognition model according to the label data of the recognition type and the stay point cluster.

[0138] The technical solution provided by the embodiment of the present disclosure determines the stop trajectory points corresponding to the parking lot from the driving trajectories of different vehicles, then clusters the stop trajectory points to obtain the stop point clusters corresponding to the parking lot, and then extracts features from the stop point clusters to obtain cluster features of the stop point clusters, and determines the identification type of the parking lot based on the cluster features, and finally trains the parking lot identification model based on the identification type and the label data of the stop point clusters. Compared with the prior art method of obtaining parking lot data through field collection or third-party cooperation, which is costly and difficult to implement on a large scale, the above technical solution extracts features after clustering the driving trajectories of different vehicles to identify the parking lot type, and mines parking lot data based only on vehicle driving trajectories, thereby reducing costs while improving the accuracy of parking lot identification.

[0139] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0140] Figure 6 The present invention is a block diagram of an electronic device for implementing the parking lot recognition method or the parking lot recognition model training method according to the embodiment of the present disclosure. Figure 6 A schematic block diagram of an example electronic device 600 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0141] like Figure 6 As shown, the electronic device 600 includes a computing unit 601, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the electronic device 600 can also be stored. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0142] Multiple components in the electronic device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the electronic device 600 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0143] The computing unit 601 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 601 performs the various methods and processes described above, such as a parking lot recognition method or a parking lot recognition model training method. For example, in some embodiments, the parking lot recognition method or the parking lot recognition model training method may be implemented as a computer software program, which is tangibly included in a machine-readable medium, such as a storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 600 via the ROM 602 and / or the communication unit 609. When the computer program is loaded into the RAM 603 and executed by the computing unit 601, one or more steps of the parking lot recognition method or the parking lot recognition model training method described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to execute the parking lot recognition method or the parking lot recognition model training method in any other appropriate manner (for example, by means of firmware).

[0144] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0145] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0146] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0147] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0148] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0149] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0150] Artificial intelligence is a discipline that studies how to use computers to simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, planning, etc.). It includes both hardware-level and software-level technologies. Artificial intelligence hardware technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, and big data processing; artificial intelligence software technologies mainly include computer vision technology, speech recognition technology, natural language processing technology, machine learning / deep learning technology, big data processing technology, knowledge graph technology, and other major directions.

[0151] Cloud computing refers to a technology system that uses network access to elastically scalable shared physical or virtual resource pools. Resources can include servers, operating systems, networks, software, applications, and storage devices, and can be deployed and managed on demand and in a self-service manner. Cloud computing technology can provide efficient and powerful data processing capabilities for technical applications such as artificial intelligence and blockchain, as well as model training.

[0152] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0153] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A parking lot identification method, comprising: Determine the stop trajectory points corresponding to the parking lot from the driving trajectories of different vehicles; Clustering the stop trajectory points to obtain a stop point cluster corresponding to the parking lot; Extracting features from the stay point cluster to obtain cluster features of the stay point cluster; The identification type of the parking lot is determined according to the cluster features.

2. The method according to claim 1, wherein: The step of determining the stop track points corresponding to the parking lot from the driving tracks of different vehicles includes: Determine continuous trajectory points corresponding to the parking lot from the driving trajectories of different vehicles; According to the instantaneous speed and position of the continuous trajectory point, it is determined whether the continuous trajectory point is a stop trajectory point corresponding to the parking lot.

3. The method according to claim 1, wherein: The cluster feature includes at least one of a vehicle type distribution feature, a cluster area feature, a number of stay points, and a dispersion distance feature.

4. The method according to claim 3, wherein: The extracting features of the stay point cluster to obtain cluster features of the stay point cluster includes: The vehicle models corresponding to the stop trajectory points in the stop point cluster are encoded to obtain the vehicle model distribution characteristics of the stop point cluster.

5. The method according to claim 3, wherein: The extracting features of the stay point cluster to obtain cluster features of the stay point cluster includes: Performing point-to-surface transformation on the stop trajectory points in the stop point cluster to obtain a cluster contour corresponding to the stop point cluster; An area of ​​the cluster outline is determined, and the area is used as an area feature of the stay point cluster.

6. The method according to claim 3, wherein: The extracting features of the stay point cluster to obtain cluster features of the stay point cluster includes: Determining a cluster center of the stay point cluster; The distance between the stay trajectory points in the stay point cluster and the cluster center is determined, and the dispersion distance feature of the stay point cluster is determined according to the distance.

7. The method according to any one of claims 1 to 6, wherein: The identification type includes a ground parking lot, an underground parking lot or an underground parking lot entrance.

8. The method according to claim 1, wherein: After identifying the type of the parking lot according to the cluster features, the method further includes: If the identification type of the parking lot is an underground parking lot entrance, the entrance position of the parking lot is determined according to the trajectory point position information of the stop trajectory points in the stop point cluster.

9. The method according to claim 8, wherein: The entrance position of the parking lot is used for navigation guidance.

10. A method for training a parking lot recognition model, comprising: Determine the stop trajectory points corresponding to the parking lot from the driving trajectories of different vehicles; Clustering the stop trajectory points to obtain a stop point cluster corresponding to the parking lot; Extracting features from the stay point cluster to obtain cluster features of the stay point cluster; Determining the identification type of the parking lot according to the cluster features; The parking lot recognition model is trained according to the recognition type and the label data of the stay point cluster.

11. A parking lot identification device, comprising: A first stop track point determination module, used to determine the stop track point corresponding to the parking lot from the driving tracks of different vehicles; A first stop point cluster determination module, used for clustering the stop trajectory points to obtain a stop point cluster corresponding to the parking lot; A first cluster feature determination module is used to extract features from the stay point cluster to obtain cluster features of the stay point cluster; The first recognition category determination module is used to determine the recognition type of the parking lot according to the cluster features.

12. The device according to claim 11, wherein The first stop trajectory point determination module is specifically used for: Determine continuous trajectory points corresponding to the parking lot from the driving trajectories of different vehicles; According to the instantaneous speed and position of the continuous trajectory point, it is determined whether the continuous trajectory point is a stop trajectory point corresponding to the parking lot.

13. The device according to claim 11, wherein: The cluster feature includes at least one of a vehicle type distribution feature, a cluster area feature, a number of stay points, and a dispersion distance feature.

14. The device according to claim 13, wherein: The first cluster feature determination module is specifically used for: The vehicle models corresponding to the stop trajectory points in the stop point cluster are encoded to obtain the vehicle model distribution characteristics of the stop point cluster.

15. The device according to claim 13, wherein: The first cluster feature determination module is specifically used for: Performing point-to-surface transformation on the stop trajectory points in the stop point cluster to obtain a cluster contour corresponding to the stop point cluster; An area of ​​the cluster outline is determined, and the area is used as an area feature of the stay point cluster.

16. The device according to claim 13, wherein: The first cluster feature determination module is specifically used for: Determining a cluster center of the stay point cluster; The distance between the stay trajectory points in the stay point cluster and the cluster center is determined, and the dispersion distance feature of the stay point cluster is determined according to the distance.

17. The device according to any one of claims 11 to 16, wherein: The identification type includes a ground parking lot, an underground parking lot or an underground parking lot entrance.

18. The apparatus according to claim 11, further comprising: The entrance position determination module is used to determine the entrance position of the parking lot according to the trajectory point position information of the stop trajectory points in the stop point cluster if the identification type of the parking lot is the entrance of an underground parking lot.

19. The device according to claim 18, wherein: The entrance position of the parking lot is used for navigation guidance.

20. A training device for a parking lot recognition model, comprising: A second stop track point determination module is used to determine the stop track point corresponding to the parking lot from the driving tracks of different vehicles; A second stop point cluster determination module is used to cluster the stop trajectory points to obtain a stop point cluster corresponding to the parking lot; A second cluster feature determination module is used to extract features from the stay point cluster to obtain cluster features of the stay point cluster; A second identification category determination module, used to determine the identification type of the parking lot according to the cluster features; The model training module is used to train the parking lot recognition model according to the recognition type and the label data of the stop point cluster.

21. An electronic device, comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the parking lot recognition method described in any one of claims 1 to 9, or the parking lot recognition model training method described in claim 10.

22. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to enable a computer to execute the parking lot recognition method according to any one of claims 1 to 9, or the parking lot recognition model training method according to claim 10.

23. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements the parking lot recognition method according to any one of claims 1 to 9, or the parking lot recognition model training method according to claim 10.

Images