Parking space quick positioning method and system for target segmentation

Abstract

The application relates to the technical field of intelligent parking, and provides a parking space rapid positioning method and system for target segmentation. The method comprises the following steps: receiving a first parking request; identifying a parking lot based on parking expected information to obtain a target parking lot; obtaining a parking area image set by interacting with the target parking lot; pre-constructing a parking identification subnetwork; synchronizing the parking area image set to the parking identification subnetwork to identify empty parking spaces and obtain an empty parking space coordinate set; performing parking path analysis according to the empty parking space coordinate set and real-time position information to obtain a parking time consumption information set; serializing the parking time consumption information set to obtain a target parking space; and sending the target parking space to the real-time parking for parking guidance. The application solves the technical problems of difficulty in searching for a parking space and long parking time consumption in the prior art, and achieves the technical effects of rapidly positioning a parking space and saving time cost.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent parking, in particular to a parking space rapid positioning method and system for target segmentation. BACKGROUND

[0002] Image segmentation is a classic problem in computer vision research, and has become a hot spot in the field of image understanding. Image segmentation is the first step of image analysis, the basis of computer vision, an important part of image understanding, and one of the most difficult problems in image processing. Image segmentation refers to dividing an image into several mutually exclusive regions according to gray, color, spatial texture, geometric shape and other characteristics, so that these characteristics show consistency or similarity within the same region, and show obvious differences between different regions. Simply put, it is to separate the target from the background in an image. For a gray image, the pixels within a region generally have gray similarity, and the gray is generally discontinuous on the boundary of the region. Due to the popularity of vehicles today, parking spaces are becoming less and less, increasing the difficulty of finding parking spaces.

[0003] In summary, the prior art has the problems of difficulty in finding parking spaces and long parking time. SUMMARY

[0004] Therefore, it is necessary to provide a parking space rapid positioning method and system for target segmentation, which can quickly position a parking space and save time cost.

[0005] In a first aspect, the present application provides a parking space rapid positioning method for target segmentation, which comprises: receiving a first parking request, wherein the first parking request is a parking request of a real-time parking user, and the first parking request comprises request generation time, parking expectation information and real-time position information; identifying a parking lot based on the parking expectation information to obtain a target parking lot; obtaining a parking area image set by interacting with the target parking lot, wherein each parking area image in the parking area image set is specifically identified by image acquisition position; pre-constructing a parking identification sub-network, wherein the parking identification sub-network comprises a first-level parking space identification module and a second-level state identification module; synchronizing the parking area image set to the parking identification sub-network for empty parking space identification to obtain an empty parking space coordinate set; performing parking path analysis according to the empty parking space coordinate set and the real-time position information to obtain a parking time consumption information set; serializing the parking time consumption information set to obtain a target parking space, and sending the target parking space to the real-time parking for parking guidance, wherein the target parking space is identified by a target parking path.

[0006] In a second aspect, the application further provides a parking space quick positioning system for target segmentation, comprising: a parking request receiving module, configured to receive a first parking request, wherein the first parking request is a parking request of a real-time parking user, and the first parking request comprises request generation time, parking expectation information and real-time position information; a target parking lot obtaining module, configured to identify a parking lot based on the parking expectation information, and obtain a target parking lot; a parking area image set obtaining module, configured to obtain a parking area image set from the target parking lot, wherein each parking area image in the parking area image set is labeled with a specific image collection position; a parking identification sub-network pre-construction module, configured to pre-construct a parking identification sub-network, wherein the parking identification sub-network comprises a first-level parking space identification module and a second-level state identification module; a vacant parking space coordinate set obtaining module, configured to synchronize the parking area image set to the parking identification sub-network to identify vacant parking spaces, and obtain a vacant parking space coordinate set; a parking time consumption information set obtaining module, configured to analyze a parking path based on the vacant parking space coordinate set and the real-time position information, and obtain a parking time consumption information set; and a parking guidance module, configured to serialize the parking time consumption information set to obtain a target parking space, and send the target parking space to the real-time parking user for parking guidance, wherein the target parking space is labeled with a target parking path.

[0007] The one or more technical solutions described above in the embodiments of the application have at least one or more of the following technical effects:

[0008] First, a first parking request is received, wherein the first parking request is a parking request of a real-time parking user, and the first parking request includes request generation time, parking expectation information, and real-time position information; second, a parking lot is identified based on the parking expectation information to obtain a target parking lot; next, a parking area image set is obtained by interacting with the target parking lot, wherein each parking area image in the parking area image set is specifically identified by an image collection position; then, a parking identification sub-network is pre-constructed, wherein the parking identification sub-network includes a first-level parking space identification module and a second-level state identification module; again, the parking area image set is synchronized to the parking identification sub-network to identify a vacant parking space to obtain a vacant parking space coordinate set; then, parking path analysis is performed according to the vacant parking space coordinate set and the real-time position information to obtain a parking time consumption information set; finally, the parking time consumption information set is serialized to obtain a target parking space, and the target parking space is sent to the real-time parking for parking guidance, wherein the target parking space is identified by a target parking path. The present application solves the technical problems of difficulty in finding a parking space and long parking time in the prior art, and achieves the technical effects of quickly positioning a parking space and saving time cost.

[0009] The above description is only a summary of the technical solutions of the present application. In order to enable one skilled in the art to better understand the technical means of the present application, the present application can be implemented according to the content of the specification, and in order to enable the above and other purposes, features and advantages of the present application to be more obvious and easy to understand, the following specific embodiments of the present application are described. BRIEF DESCRIPTION OF DRAWINGS

[0010] Figure 1 A flowchart of a parking space quick positioning method for target segmentation in an embodiment;

[0011] Figure 2 A flowchart of a parking space quick positioning method for target segmentation in an embodiment;

[0012] Figure 3 A structural block diagram of a parking space quick positioning system for target segmentation in an embodiment.

[0013] Marked for parking request receiving module 11, target parking lot acquisition module 12, parking area image set acquisition module 13, parking identification sub-network pre-construction module 14, vacant parking space coordinate set acquisition module 15, parking time consumption information set acquisition module 16, parking guidance module 17. DETAILED DESCRIPTION

[0014] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not to limit the present application.

[0015] As shown in Figure 1 The present application provides a parking space rapid positioning method for target segmentation, characterized in that the method comprises:

[0016] receiving a first parking request, wherein the first parking request is a parking request of a real-time parking user, and the first parking request includes request generation time, parking expectation information and real-time location information;

[0017] Image segmentation is a technology and process of dividing an image into several specific regions with unique properties and proposing a target of interest. It is a key step from image processing to image analysis. From a mathematical point of view, image segmentation is a process of dividing a digital image into mutually disjoint regions. The process of image segmentation is also a labeling process, that is, pixels belonging to the same region are assigned the same number. The target segmentation in the present application refers to a method of segmenting an image according to different target states; the parking space rapid positioning refers to quickly finding a vacant parking space in all target parking spaces for parking operation. By providing the parking space rapid positioning method for target segmentation, the accuracy of the monitoring and analysis results is improved, and the intervention ability of the monitoring results on dangerous driving of the driver is improved.

[0018] The parking user refers to a vehicle driver who needs to perform parking operation, and the parking request refers to a parking requirement proposed by the vehicle according to its own parking demand, denoted as a first parking request. The first parking request includes request generation time, parking expectation information and real-time location information. The request generation time refers to the time when the parking user issues the parking request, the parking expectation information refers to the demand of the parking user for the parking space, such as the distance of the parking space, and the real-time location information refers to the location of the vehicle when the parking request is issued. By receiving the first parking request, the demand of the parking user for the parking space and the location of the vehicle are understood, which provides a data basis for subsequently obtaining information of the target parking lot and making parking guidance.

[0019] Based on the parking expectation information, a target parking lot is identified;

[0020] The parking lot identification refers to that in the vicinity of the parking user, there are multiple parking lots, and the parking lot most consistent with the parking expectation information is selected as the target parking lot for research. By identifying the parking lot in combination with the parking expectation information, the target parking lot is obtained, which provides support for subsequent analysis of the vacant parking space in the target parking lot.

[0021] interacting with the target parking lot to obtain a parking area image set, wherein each parking area image in the parking area image set is associated with a specific image collection position;

[0022] The parking area image set refers to the parking space image information obtained by various monitoring means in the target parking area. The parking space image information is integrated to obtain the parking area image set. Each parking area image in the parking area image set has specific collection position information of the parking area image. The parking area image set is obtained to lay the foundation for subsequent boundary segmentation of the multiple sample parking space images.

[0023] Pre-constructing a parking recognition sub-network, wherein the parking recognition sub-network includes a first-level parking space recognition module and a second-level state recognition module;

[0024] The parking recognition sub-network is a module for recognizing the parking space and the state of the parking area in the parking lot. The parking recognition sub-network includes a first-level parking space recognition module and a second-level state recognition module. The parking area image set is input into the parking recognition sub-network. First, the parking space recognition module is used to recognize the parking space in the parking lot, such as the parking area. The state recognition module is used to determine the use of the parking space in the parking lot by analyzing the output data of the parking space recognition module to obtain the used parking space and the empty parking space. By pre-constructing the parking recognition sub-network, the empty parking space and its coordinates are determined to provide support for subsequent search for empty parking space and determination of parking guidance.

[0025] Interacting to obtain multiple sample parking space images;

[0026] Boundary segmentation is performed on the multiple sample parking space images to obtain multiple sample parking boundary images and multiple sample parking area images;

[0027] Boundary pixel feature extraction is performed on the multiple sample parking boundary images to obtain a sample boundary feature set;

[0028] The first-level parking space recognition module is constructed based on the sample boundary feature set;

[0029] The multiple sample parking area images are used for identification training of the second-level state recognition module;

[0030] The first-level parking space recognition module and the second-level state recognition module constitute the parking recognition sub-network.

[0031] The parking area image set is interacted to select multiple parking area images as sample parking space images. The boundary segmentation refers to boundary-based segmentation, that is, the segmentation of the image is completed by searching the boundary between different regions. Then, the multiple sample parking space boundary images and the multiple sample parking area images are obtained by segmenting the multiple sample parking space boundary images according to the boundary segmentation technology. The sample parking space boundary image refers to the boundary image of the parking space in the parking lot, that is, the parking frame. The sample parking area image refers to the image of the parking space in the parking lot, that is, the part enclosed by the parking frame. The pixel refers to the basic color element and the basic coding of its gray scale. The pixel is the basic unit of digital image. The size of the image resolution is usually represented in pixels per inch. The boundary pixel feature set is obtained by extracting the boundary pixel features according to the pixel features of the sample parking space boundary image. The boundary pixel feature set is the pixel feature set of the parking frame. The first-level parking space recognition module is constructed through the boundary pixel feature set. Then, the second-level state recognition module is trained according to the multiple sample parking area images. The sample parking area images are input into the second-level state recognition module, and the parking condition of the parking area is output. The output parking condition of the parking area is compared with the corresponding parking condition of the sample parking area image. When the error of the output parking condition of the parking area is negligible, the training of the second-level state recognition module is completed. The first-level parking space recognition module and the second-level state recognition module constitute the parking recognition sub-network. The parking recognition sub-network is constructed to contribute to the subsequent search for the best parking path.

[0032] The first-level parking space recognition module includes a parking image division sub-module and a parking boundary connection sub-module.

[0033] The boundary division scale and the pixel feature density threshold are pre-constructed.

[0034] The boundary division scale and the sample boundary feature set are synchronized to the parking image division sub-module.

[0035] The pixel feature density threshold is synchronized to the parking boundary connection sub-module.

[0036] The first parking space recognition module comprises a parking image division submodule and a parking boundary connection submodule. The parking image division submodule is a module for dividing the parking image. The parking boundary connection submodule is a module for connecting the boundaries of the parking area to determine the parking frame. A pre-constructed boundary division scale is used to determine the parking image division according to the pixel value of the a×a square block, where a is a positive number. A pixel feature density threshold is a data set by the staff. The parking position of the parking space is determined according to the pixel feature density threshold. The boundary division scale and the sample boundary feature set are synchronized to the parking image division submodule to obtain the internal features of the parking frame. The pixel feature density threshold is synchronized to the parking boundary connection submodule to determine the complete area of the parking frame. By constructing the first parking space recognition module, the parking frame and parking space in the parking lot can be better recognized, which lays the foundation for subsequent judgment of the parking state.

[0037] The parking state of the plurality of sample parking area images is identified to obtain a plurality of identified parking area images.

[0038] An image division scale is pre-constructed, and the plurality of identified parking area images are preprocessed using the image division scale to obtain a plurality of sample division images. Each sample division image corresponds to an identified parking area image, and each sample division image has a parking state identification.

[0039] The pixel point RGB value of the plurality of sample division images is called and the pixel point RGB fluctuation interval is analyzed to obtain a plurality of sample fluctuation thresholds.

[0040] The plurality of sample fluctuation thresholds are classified based on the parking state identification to obtain a parking fluctuation threshold group and an empty position fluctuation threshold group.

[0041] The parking fluctuation threshold group is serialized to obtain a parking threshold determinator.

[0042] The second state recognition module comprises an image division execution submodule and a parking state judgment module.

[0043] The image division scale is synchronized to the image division submodule, and the parking threshold determinator is synchronized to the parking state judgment module.

[0044] The parking state of the plurality of sample parking area images is identified, that is, whether the parking space in the plurality of sample parking area images is parked or not is labeled, to obtain a plurality of identified parking area images; a pre-constructed image division scale is constructed, that is, a square of a x a is constructed as an image division scale, a is a positive number, the plurality of identified parking area images are divided using the image division scale, and the plurality of identified parking area images are divided into squares of the above specification, denoted as a plurality of groups of sample division images, wherein each group of sample division images corresponds to an identified parking area image, and each group of sample division images has a parking state identification; the RGB color mode is an industry standard color, which obtains various colors through the changes of the red, green and blue color channels and the superposition between them, that is, RGB represents the colors of the three channels of red, green and blue, this standard almost includes all the colors that can be perceived by human vision, and is one of the most widely used color systems, the pixel point RGB value of the plurality of groups of sample division images is called, that is, the color of the pixel point of the plurality of groups of sample division images is called, and the color variation interval in the plurality of groups of sample division images is analyzed to obtain a plurality of sample fluctuation thresholds, wherein the sample fluctuation threshold refers to the fluctuation range of the pixel point RGB value; the plurality of sample fluctuation thresholds are classified according to the parking state identification, for example, the image pixel points of the used parking space will have the color of the vehicle, which is different from the color of the ground and the parking frame, and the image pixel points of the empty parking space only have the color of the ground and the parking frame, to obtain a parking fluctuation threshold group and an empty space fluctuation threshold group; the parking fluctuation threshold group is sequenced, that is, the parking fluctuation threshold group is sorted according to size to obtain a parking threshold judge, which refers to a judge for judging the use of the parking space; the secondary state recognition module includes an image division execution submodule and a parking state judgment module, wherein the image division execution submodule is a module for preprocessing the plurality of identified parking area images using the image division scale, and the parking state judgment module refers to a module for judging the parking state according to the pixel color. The secondary state recognition module includes an image division execution submodule and a parking state judgment module, and the construction method of the module is analyzed to lay the foundation for subsequent quick parking route searching.

[0045] The parking area image set is synchronized to the parking recognition subnetwork for empty parking space recognition to obtain an empty parking space coordinate set;

[0046] The parking area image set is input into the parking recognition subnetwork for empty parking space recognition, that is, the empty parking space in the parking area image set is obtained, which is identified by the position information in the parking area image set, to obtain the empty parking space coordinate set. The empty parking space coordinate set is obtained to lay the foundation for subsequent acquisition of a parking time consumption information set.

[0047] According to the free parking space coordinate set and the real-time position information, a parking path analysis is performed to obtain a parking time consumption information set;

[0048] The parking path refers to the distance from the real-time position information to the target parking lot and the distance inside the target parking lot to the free parking space. Since the distance from the real-time position information to the target parking lot is equal, only the path from the entrance of the target parking lot to the free parking space is analyzed, and all time consumptions from the real-time position information to the free parking space coordinates are obtained according to the path and the parking speed constraint. Through the parking time consumption information set, a contribution is made for subsequent obtaining of the best route

[0049] The parking path distribution and the parking speed constraint of the target parking lot are obtained interactively;

[0050] Based on the real-time position information, a target parking starting point is located in the parking path distribution;

[0051] Based on the free parking space coordinate set and the target parking starting point, a parking path fitting is performed on the parking path distribution to obtain a parking route set;

[0052] According to the parking speed constraint and the parking route set, the parking time consumption information set is generated.

[0053] The parking path distribution refers to the road in the target parking lot from the entrance to the target parking space. The parking speed constraint refers to a constraint that the maximum value of the parking speed in the target parking lot is afraid to cause an accident, that is, the speed of the vehicle in the target parking lot cannot exceed the maximum value. Based on the real-time position information, a target parking starting point is located in the parking path distribution. The entrance closest to the target parking lot entrance of the target vehicle is found based on the real-time position information, and the entrance is taken as the target parking starting point. All routes from the target parking starting point to the free parking space are obtained by traversing the parking path distribution based on the free parking space coordinate set and the target parking starting point, and the routes are integrated to obtain a parking route set. According to the parking speed constraint and the parking route set, the time consumption of the routes in the parking route set is obtained, and the parking time consumption information set is generated. Through the parking time consumption information set, a contribution is made for subsequent obtaining of the best route.

[0054] As shown in Figure 2 , a preset parking space empty monitoring window is provided;

[0055] According to the request generation time, the monitoring period of the parking space empty monitoring window is optimized to obtain a second parking monitoring node;

[0056] Based on the second parking monitoring node, the remaining parking spaces in the target parking lot are updated to obtain a parking space dynamic update result.

[0057] The parking space vacancy monitoring window is a secondary supervision on the vacancy parking space in the vacancy parking space coordinate set. Since the target vehicle is still a distance away from the target parking lot after sending a parking request, the vacancy parking space coordinate set may have other vehicles parked in the process of driving to the target parking lot. Therefore, the vacancy parking space needs to be monitored. The monitoring period of the parking space vacancy monitoring window is optimized according to the request generation time, that is, the target vehicle sends a parking request, and the target parking space is obtained. A period of time is monitored again, and the monitoring result is a second parking monitoring node. Based on the second parking monitoring node, the remaining parking spaces in the target parking lot are updated to obtain a parking space dynamic update result. The vacancy parking space is found from the parking space dynamic update result for the above technical analysis. Through the monitoring and updating of the vacancy parking space, better service is provided for the driver.

[0058] The parking time information set is serialized to obtain a target parking space, and the target parking space is sent to the real-time parking for parking guidance. The target parking space is identified by a target parking path.

[0059] The parking time information in the parking time information set is sorted according to the size of the data, and the corresponding parking space with the minimum time consumption is selected as a target parking space. The target parking space is sent to the real-time parking for parking guidance. The target parking space is identified by a target parking path. Through sorting the parking time information, the parking space corresponding to the parking route with the least time is found as the target parking position. The application solves the technical problems of difficulty in finding a parking space and long parking time in the prior art, and achieves the technical effects of quickly positioning a parking space and saving time cost.

[0060] As shown in Figure 3 The application also provides a parking space quick positioning system for target segmentation, which comprises:

[0061] A parking request receiving module 11 is configured to receive a first parking request. The first parking request is a parking request of a real-time parking user. The first parking request comprises a request generation time, parking expectation information, and real-time location information.

[0062] A target parking lot obtaining module 12 is configured to identify a parking lot based on the parking expectation information to obtain a target parking lot.

[0063] A parking area image set obtaining module 13 is configured to obtain a parking area image set from the target parking lot, wherein each parking area image in the parking area image set is associated with a specific image collection position;

[0064] A parking recognition sub-network pre-construction module 14 is configured to pre-construct a parking recognition sub-network, wherein the parking recognition sub-network comprises a first-level parking space recognition module and a second-level state recognition module;

[0065] A vacant parking space coordinate set obtaining module 15 is configured to synchronize the parking area image set to the parking recognition sub-network to identify vacant parking spaces and obtain a vacant parking space coordinate set;

[0066] A parking time consumption information set obtaining module 16 is configured to analyze a parking path based on the vacant parking space coordinate set and the real-time position information, and obtain a parking time consumption information set;

[0067] A parking guidance module 17 is configured to serialize the parking time consumption information set to obtain a target parking space, and send the target parking space to the real-time parking for parking guidance, wherein the target parking space is identified by a target parking path.

[0068] Further, the embodiments of the present application further comprise:

[0069] A sample parking space image interaction module is configured to interact to obtain a plurality of sample parking space images;

[0070] A boundary segmentation module is configured to perform boundary segmentation on the plurality of sample parking space images to obtain a plurality of sample parking space boundary images and a plurality of sample parking area images;

[0071] A sample boundary feature set obtaining module is configured to extract boundary pixel features from the plurality of sample parking space boundary images to obtain a sample boundary feature set;

[0072] A parking space recognition module is configured to construct the first-level parking space recognition module based on the sample boundary feature set;

[0073] A state recognition module is configured to perform recognition training of the second-level state recognition module by using the plurality of sample parking area images;

[0074] A parking recognition sub-network construction module is configured to construct the parking recognition sub-network by using the first-level parking space recognition module and the second-level state recognition module.

[0075] Further, the embodiments of the present application further include:

[0076] The parking space recognition module comprises a module, and the parking space recognition module comprises a module for the first-level parking space recognition module to include a parking image division sub-module and a parking boundary connection sub-module.

[0077] The feature density threshold value construction module is used for pre-constructing a boundary division scale and a pixel feature density threshold value.

[0078] The sub-module division module is used for synchronizing the boundary division scale and the sample boundary feature set to the parking image division sub-module.

[0079] The parking boundary connection sub-module is used for synchronizing the pixel feature density threshold value to the parking boundary connection sub-module.

[0080] Further, the embodiments of the present application further include:

[0081] The identified parking area image obtaining module is used for performing parking state identification on the multiple sample parking area images to obtain multiple identified parking area images.

[0082] The image division scale pre-construction module is used for pre-constructing an image division scale, and pre-processing the multiple identified parking area images by using the image division scale to obtain multiple sets of sample division images, wherein each set of sample division images corresponds to an identified parking area image, and each set of sample division images has a parking state identification.

[0083] The sample fluctuation threshold value obtaining module is used for performing pixel point RGB value calling and pixel point RGB fluctuation interval analysis on the multiple sets of sample division images to obtain multiple sample fluctuation threshold values.

[0084] The sample fluctuation threshold value classification module is used for classifying the multiple sample fluctuation threshold values based on the parking state identification to obtain a parking fluctuation threshold value group and an empty space fluctuation threshold value group.

[0085] The parking threshold value judge obtaining module is used for serializing the parking fluctuation threshold value group to obtain a parking threshold value judge.

[0086] The state recognition module is used for the second-level state recognition module to include an image division execution sub-module and a parking state judgment module.

[0087] A parking state judgment module is configured to synchronize the image division scale to the image division sub-module and synchronize the parking threshold judgment device to the parking state judgment module.

[0088] Further, the embodiments of the application further include:

[0089] A parking path distribution module is configured to interactively obtain the parking path distribution and the parking speed constraint of the target parking lot.

[0090] A target parking starting point obtaining module is configured to obtain a target parking starting point based on the real-time position information in the parking path distribution.

[0091] A parking route set obtaining module is configured to perform parking path fitting based on the empty parking space coordinate set and the target parking starting point to obtain a parking route set.

[0092] A parking time consumption information set generating module is configured to generate the parking time consumption information set according to the parking speed constraint and the parking route set.

[0093] Further, the embodiments of the application further include:

[0094] A parking space empty monitoring window preset module is configured to preset a parking space empty monitoring window.

[0095] A monitoring cycle optimization module is configured to optimize the monitoring cycle of the parking space empty monitoring window according to the request generation time to obtain a second parking monitoring node.

[0096] A parking space dynamic update result obtaining module is configured to perform remaining parking space identification update on the target parking lot based on the second parking monitoring node to obtain a parking space dynamic update result.

[0097] For specific embodiments of the parking space rapid positioning system for target segmentation, refer to the embodiments of the parking space rapid positioning method for target segmentation described above, which will not be repeated here. The above modules can be embedded in or independent of the processor in the computer device in hardware form, or stored in the memory in the computer device in software form, so that the processor can call and execute the operations corresponding to the above modules.

[0098] Any combination of the technical features in the above embodiments can be made, and for the sake of brevity, not all possible combinations are described above, however, as long as the combination of the technical features does not exist in contradiction, it shall be considered within the scope of the present disclosure.

[0099] The above embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it shall not be understood as a limitation on the patent scope of the present application. It shall be pointed out that, for ordinary skilled persons in the art, several modifications and improvements can be made without departing from the concept of the present application, and these shall be within the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the appended claims.

Claims

1. A method for fast positioning of parking space for target segmentation, characterized in that, The method comprises: receiving a first parking request, wherein the first parking request is a parking request of a real-time parking user, the first parking request comprises request generation time, parking expectation information and real-time location information; based on the parking expectation information, a parking lot is identified to obtain a target parking lot; interacting with the target parking lot to obtain a set of parking area images, wherein each parking area image in the set of parking area images is specifically identified by image acquisition location; a pre-constructed parking identification sub-network, wherein the parking identification sub-network comprises a first-level parking space identification module and a second-level state identification module, and the step comprises: interacting to obtain a plurality of sample parking space images; boundary segmentation is performed on the plurality of sample parking space images to obtain a plurality of sample parking space boundary images and a plurality of sample parking area images; boundary pixel feature extraction is performed on the plurality of sample parking space boundary images to obtain a sample boundary feature set; based on the sample boundary feature set, the first-level parking space identification module is constructed, and the step comprises: the first-level parking space identification module comprises a parking image division sub-module and a parking boundary connection sub-module; a boundary division scale and a pixel feature density threshold value are pre-constructed; the boundary division scale and the sample boundary feature set are synchronized to the parking image division sub-module; the pixel feature density threshold value is synchronized to the parking boundary connection sub-module; the plurality of sample parking area images are used for identification training of the second-level state identification module, and the step further comprises: parking state identification is performed on the plurality of sample parking area images to obtain a plurality of identified parking area images; an image division scale is pre-constructed, and the image division scale is used for pre-processing of the plurality of identified parking area images to obtain a plurality of sample division images, wherein each group of sample division images corresponds to an identified parking area image, and each group of sample division images has parking state identification; pixel point RGB value calling and pixel point RGB fluctuation interval analysis are performed on the plurality of sample division images to obtain a plurality of sample fluctuation threshold values; based on parking state identification, the plurality of sample fluctuation threshold values are classified to obtain a parking fluctuation threshold value group and an empty space fluctuation threshold value group; the parking fluctuation threshold value group is serialized to obtain a parking threshold value judge; the second-level state identification module comprises an image division execution sub-module and a parking state judgment module; the image division scale is synchronized to the image division sub-module, and the parking threshold value judge is synchronized to the parking state judgment module; the first-level parking space identification module and the second-level state identification module constitute the parking identification sub-network; the set of parking area images is synchronized to the parking identification sub-network for empty parking space identification to obtain an empty parking space coordinate set; based on the empty parking space coordinate set and the real-time location information, parking path analysis is performed to obtain a parking time consumption information set; the parking time consumption information set is serialized to obtain a target parking space, and the target parking space is sent to the real-time parking for parking guidance, wherein the target parking space is identified by a target parking path.

2. The method of claim 1, wherein, The method further includes performing parking path analysis based on the set of available parking space coordinates and the real-time location information to obtain a set of parking time information. The parking path distribution and parking speed constraints of the target parking lot are obtained interactively. Based on the real-time location information, the target parking starting point is obtained by locating the parking path distribution. Based on the set of coordinates of the available parking spaces and the target parking starting point, the parking path distribution is traversed to fit the parking path and obtain the parking route set; The parking time information set is generated based on the parking speed constraint and the parking route set.

3. The method of claim 1, wherein, The method further includes performing parking path analysis based on the set of available parking space coordinates and the real-time location information to obtain a set of parking time information. Preset berth availability monitoring window; The monitoring cycle of the parking space availability monitoring window is optimized based on the request generation time to obtain a second parking monitoring node; Based on the second parking monitoring node, the remaining parking spaces in the target parking lot are identified and updated to obtain dynamic update results for parking spaces.

4. A parking space quick positioning system for target segmentation, for performing the method of claim 1, characterized in that, The system includes: A parking request receiving module is used to receive a first parking request, wherein the first parking request is a parking request from a real-time parking user, and the first parking request includes the request generation time, parking expectation information, and real-time location information; A target parking lot acquisition module is used to identify parking lots based on the parking expectation information and obtain target parking lots. A parking area image set acquisition module is used to interact with the target parking lot to obtain a parking area image set, wherein each parking area image in the parking area image set has a specific image acquisition location identifier. A parking recognition subnetwork pre-construction module is used to pre-construct a parking recognition subnetwork, wherein the parking recognition subnetwork includes a primary parking space recognition module and a secondary status recognition module; A module for obtaining the coordinate set of available parking spaces is used to synchronize the parking area image set to the parking recognition sub-network to identify available parking spaces and obtain the coordinate set of available parking spaces. A parking time information set acquisition module is used to perform parking path analysis based on the available parking space coordinate set and the real-time location information to obtain a parking time information set. A parking guidance module is used to serialize the parking time information set to obtain the target parking space, and send the target parking space to the real-time parking system for parking guidance. The target parking space is identified by a target parking path.

Images