A parking lot parking flow monitoring method and system

Abstract

The application provides a parking flow monitoring method and system for a parking lot, relates to the technical field of data processing, and obtains a historical parking quantity information set by taking a parking flow monitoring window as a monitoring period, monitoring the parking flow of the parking lot by using a parking flow monitoring subnetwork, and traversing the historical parking quantity information set based on a parking saturation threshold to obtain K parking saturation nodes; obtains K historical node traffic information based on the interaction of the K parking saturation nodes and a traffic information collection area; and optimizes the parking space of the parking lot based on the K historical node traffic information. The technical problem that the parking space design of the parking lot is unreasonable in the prior art, leading to tight parking for users and unsatisfied parking demand, is solved. The technical effects of improving the adaptability and relevance of the parking space quantity design of the parking lot and the parking demand of the area where the parking lot is located and improving the parking experience of the parking users around the parking lot are achieved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a parking lot traffic monitoring method and system. Background Technology

[0002] Currently, some commercial areas face a severe shortage of parking spaces due to insufficient consideration of customer traffic and parking demand. This is especially true on weekends and holidays, when drivers often have to search for available spaces, sometimes resulting in queues. This situation exacerbates parking difficulties and negatively impacts the operational efficiency of shopping malls.

[0003] In summary, existing technologies suffer from unreasonable parking space design, leading to parking shortages and unmet parking needs. Summary of the Invention

[0004] This application provides a parking flow monitoring method and system for parking lots, which is used to address the technical problem in the prior art where unreasonable parking space design in parking lots leads to parking shortages and unmet parking needs.

[0005] In view of the above problems, this application provides a parking flow monitoring method and system for parking lots.

[0006] The first aspect of this application provides a parking flow monitoring method for a parking lot. The method includes: pre-constructing a parking flow monitoring sub-network, wherein the parking flow monitoring sub-network is used to dynamically monitor the parking volume of a target parking lot; pre-setting a parking flow monitoring window, wherein the parking flow monitoring window is used to control the operation cycle of the parking flow monitoring sub-network; using the parking flow monitoring window as the monitoring cycle, using the parking flow monitoring sub-network to monitor the parking flow of the target parking lot and obtain a historical parking quantity information set, wherein each historical parking quantity information in the historical parking quantity information set has a collection time node identifier; pre-setting a parking saturation threshold, and traversing the historical parking quantity information set based on the parking saturation threshold to obtain K parking saturation nodes; pre-setting traffic flow information collection constraints, and constructing a traffic flow information collection area based on the traffic flow information collection constraints and the target parking lot; interacting with the traffic flow information collection area based on the K parking saturation nodes to obtain K historical node traffic flow information; and optimizing parking spaces in the target parking lot based on the K historical node traffic flow information.

[0007] A second aspect of this application provides a parking lot traffic monitoring system, the system comprising: a monitoring model construction module for pre-constructing a parking traffic monitoring sub-network, wherein the parking traffic monitoring sub-network is used to dynamically monitor the parking volume of a target parking lot; a monitoring window construction module for pre-setting a parking traffic monitoring window, wherein the parking traffic monitoring window is used to control the operating cycle of the parking traffic monitoring sub-network; and a parking data acquisition module for monitoring the parking traffic of the target parking lot using the parking traffic monitoring sub-network with the parking traffic monitoring window as the monitoring cycle, and obtaining a set of historical parking quantity information, wherein the historical parking quantity information... Each historical parking quantity information in the set has a collection time node identifier; a saturation node determination module is used to preset a parking saturation threshold and traverse the historical parking quantity information set based on the parking saturation threshold to obtain K parking saturation nodes; a collection area setting module is used to preset traffic flow information collection constraints and construct a traffic flow information collection area based on the traffic flow information collection constraints and the target parking lot; a node traffic flow collection module is used to interact with the traffic flow information collection area based on the K parking saturation nodes to obtain K historical node traffic flow information; a parking space optimization execution module is used to optimize the parking spaces of the target parking lot based on the K historical node traffic flow information.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] The method provided in this application embodiment pre-constructs a parking flow monitoring sub-network, wherein the parking flow monitoring sub-network is used to dynamically monitor the parking volume of a target parking lot; a preset parking flow monitoring window is preset, wherein the parking flow monitoring window is used to control the operation cycle of the parking flow monitoring sub-network; using the parking flow monitoring window as the monitoring cycle, the parking flow monitoring sub-network is used to monitor the parking flow of the target parking lot to obtain a set of historical parking quantity information, wherein each historical parking quantity information in the set of historical parking quantity information has a collection time node identifier; a preset parking saturation threshold is preset, and the set of historical parking quantity information is traversed based on the parking saturation threshold to obtain K parking saturation nodes; a preset traffic flow information collection constraint is preset, and a traffic flow information collection area is constructed based on the traffic flow information collection constraint and the target parking lot; the traffic flow information is obtained by interacting with the traffic flow information collection area based on the K parking saturation nodes; and parking space optimization of the target parking lot is performed based on the K historical node traffic flow information. This technology achieves the goal of improving the compatibility and relevance between the number of parking spaces in a parking lot and the parking demand in the area, thereby enhancing the parking experience for users around the parking lot. Attached Figure Description

[0010] Figure 1 A schematic diagram of a parking lot traffic monitoring method provided in this application;

[0011] Figure 2 A schematic diagram illustrating the process of constructing a parking flow monitoring sub-network in a parking lot parking flow monitoring method provided in this application;

[0012] Figure 3 A flowchart illustrating the process of obtaining historical node traffic flow information in a parking lot traffic flow monitoring method provided in this application;

[0013] Figure 4 This application provides a schematic diagram of the structure of a parking lot traffic monitoring system.

[0014] Figure labeling: 1. Monitoring model construction module; 2. Monitoring window construction module; 3. Parking data acquisition module; 4. Saturation node determination module; 5. Acquisition area setting module; 6. Node traffic flow acquisition module; 7. Parking space optimization execution module. Detailed Implementation

[0015] This application provides a parking lot traffic flow monitoring method and system to address the technical problem in existing parking lots where unreasonable parking space design leads to parking shortages and unmet parking needs. It achieves the technical effect of improving the fit and correlation between the number of parking spaces in a parking lot and the parking demand in the surrounding area, thereby enhancing the parking experience for users in the vicinity of the parking lot.

[0016] The acquisition, storage, use, and processing of data in the technical solution of this invention all comply with relevant regulations.

[0017] The technical solutions of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. It should be understood that the present invention is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. It should also be noted that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings, not all of them.

[0018] Example 1

[0019] like Figure 1 As shown, this application provides a method for monitoring parking flow in a parking lot, the method comprising:

[0020] A100: Pre-built parking flow monitoring sub-network, wherein the parking flow monitoring sub-network is used to dynamically monitor the number of cars parked in the target parking lot;

[0021] In one embodiment, such as Figure 2 As shown, a pre-constructed parking flow monitoring sub-network is used to dynamically monitor the parking volume of the target parking lot. Step A100 of the method provided in this application further includes:

[0022] A110: The parking flow monitoring subnetwork includes an inlet information monitoring subnetwork and an outlet information monitoring subnetwork;

[0023] A120: The target parking lot is pre-installed with a first image acquisition device and a second image acquisition device. The first image acquisition device is used to acquire the entrance image information of the target parking lot and transmit the entrance image information to the entrance information monitoring sub-network. The second image acquisition device is used to acquire the exit image information of the target parking lot and transmit the exit image information to the exit information monitoring sub-network.

[0024] A130: The import information monitoring subnetwork identifies the import image information to obtain a set of import vehicle information;

[0025] A140: The export information monitoring sub-network identifies the export image information to obtain a set of export vehicle information;

[0026] A150: Real-time parking volume monitoring results are calculated based on the imported vehicle information set and the exported vehicle information set.

[0027] In one embodiment, the import information monitoring subnetwork identifies the import image information to obtain an import vehicle information set. Step A130 of the method provided in this application further includes:

[0028] A131: The import information monitoring sub-network includes a video frame splitting execution module and an entry vehicle identification and counting module, wherein the video frame splitting execution module is used to perform frame splitting processing on the import image information;

[0029] A132: Obtain the set of sample vehicle entry images;

[0030] A133: Perform semantic segmentation on the set of sample vehicle entry images to obtain a set of sample vehicle entry image segmentation results;

[0031] A134: The vehicle entry recognition and counting module is constructed based on a CNN neural network, and the recognition accuracy of the vehicle entry recognition and counting module is optimized using the sample vehicle entry image set and the sample vehicle entry image segmentation result set.

[0032] Specifically, in this embodiment, the target parking lot is a parking lot with an indefinite number of parking spaces located in a busy urban area with good traffic, and the real-time parking volume information of the target parking lot is obtained through the pre-constructed parking flow monitoring sub-network.

[0033] The entrance of the target parking lot is equipped with a first image acquisition device, which is used to acquire image information of the entrance of the target parking lot. The exit of the target parking lot is equipped with a second image acquisition device, which is used to acquire image information of the exit of the target parking lot.

[0034] The parking flow monitoring subnetwork includes an entrance information monitoring subnetwork and an exit information monitoring subnetwork. The entrance information monitoring subnetwork is communicatively connected to the first image acquisition device. After obtaining the entrance image information transmitted in real time by the first image acquisition device, it performs vehicle entry identification and vehicle count based on the entrance image information to obtain the entrance vehicle information set. The exit information monitoring subnetwork is communicatively connected to the second image acquisition device. After obtaining the exit image information transmitted in real time by the second image acquisition device, it performs vehicle exit identification and vehicle count based on the parking garage image information to obtain the exit vehicle information set. Finally, by subtracting the entrance vehicle information set from the entrance vehicle information set, the real-time parking volume monitoring result, which is highly immediacy and represents the number of vehicles parked in the target parking lot, can be obtained.

[0035] Since the import information monitoring subnetwork identifies the import image information to obtain an import vehicle information set, and the export information monitoring subnetwork identifies the export image information to obtain an export vehicle information set, the image information recognition methods of the import information monitoring subnetwork and the export information monitoring subnetwork are consistent. Therefore, this embodiment takes the method of the import information monitoring subnetwork identifying the import image information to obtain the import vehicle information set as an example to illustrate the detailed parameters of the construction and operation methods of the import information monitoring subnetwork and the export information monitoring subnetwork.

[0036] Specifically, in this embodiment, the import information monitoring sub-network includes a video frame splitting execution module and a vehicle entry recognition and counting module. The video frame splitting execution module is used to capture and split the import image information every 10 seconds to split the import image information into multiple intermittent image frames with a 10-second interval, so as to identify whether a car has entered the target parking lot based on the vehicle entry recognition and counting module.

[0037] The vehicle entry recognition and counting module is preferably constructed based on a CNN neural network. The input information of the vehicle entry recognition and counting module is a single image, and the output result is whether a vehicle exists in the single image.

[0038] The training process of the vehicle entry recognition and counting module is as follows: a set of sample vehicle entry images is collected, and semantic segmentation is performed on the sample vehicle entry image set by manual analysis to identify the vehicle identifiers in the sample vehicle entry images and obtain a set of sample vehicle entry image segmentation results.

[0039] The set of sample vehicle entry images and the set of sample vehicle entry image segmentation results are divided into a training set, a test set, and a validation set for training, testing, and validation of the vehicle entry recognition and counting module, thereby improving its recognition accuracy. When the vehicle entry recognition and counting module detects a vehicle entering, it increments the vehicle entry information set by one entry count. Similarly, the same method is used to construct the vehicle exit recognition and counting module.

[0040] This embodiment improves the accuracy of obtaining parking vehicle counts for the target parking lot by constructing a parking flow monitoring sub-network.

[0041] A200: Preset parking flow monitoring window, wherein the parking flow monitoring window is used to control the operating cycle of the parking flow monitoring sub-network;

[0042] A300: Using the parking flow monitoring window as the monitoring period, the parking flow monitoring sub-network is used to monitor the parking flow of the target parking lot and obtain a set of historical parking quantity information, wherein each piece of historical parking quantity information in the set of historical parking quantity information has a collection time node identifier.

[0043] Specifically, it should be understood that the number of vehicles parked in the target parking lot is dynamically changing. Continuous collection of parking information from the target parking lot results in data that lacks specificity. Therefore, this embodiment pre-defines a parking flow monitoring window to manage the operating cycle of the parking flow monitoring sub-network. For example, the parking flow monitoring window can be a window that collects and retrieves the real-time parking volume monitoring results output by the parking flow monitoring sub-network at 15-minute intervals. It should be noted that the parking flow monitoring sub-network is constantly running.

[0044] Using the parking flow monitoring window as the monitoring period, the parking flow monitoring sub-network is used to monitor the parking flow of the target parking lot and obtain a set of historical parking quantity information. Each historical parking quantity information in the set of historical parking quantity information has a collection time node identifier. The time span between two historical parking quantity information with adjacent collection times in the set of historical parking quantity information is the parking flow monitoring window.

[0045] A400: A preset parking saturation threshold is established, and the historical parking quantity information set is traversed based on the parking saturation threshold to obtain K parking saturation nodes.

[0046] Specifically, in this embodiment, the parking saturation threshold is the upper limit of parking spaces in the target parking lot that cannot accommodate new cars. Based on the parking saturation threshold, the historical parking quantity information set is traversed to obtain K parking saturation nodes. The parking saturation nodes are the time nodes when the number of vehicles parked in the historical target parking lot reaches the parking saturation threshold.

[0047] A500: Preset traffic flow information collection constraints, and construct a traffic flow information collection area based on the traffic flow information collection constraints and the target parking lot;

[0048] Specifically, in this embodiment, based on the parking information of the target parking lot, the traffic flow information of the roads around the target parking lot that can access the target parking lot is further obtained. The purpose of obtaining the traffic flow information is to obtain the correlation between the traffic flow data and the parking information in the target parking lot, so as to optimize the number of parking spaces in the target parking lot according to the traffic flow information around the target parking lot, so that the number of parking spaces in the target parking lot meets the parking demand of the area where the target parking lot is located.

[0049] A preset traffic flow information collection constraint is established, which limits the area for collecting traffic flow information around the target parking lot. For example, the constraint has a radius of 1 km. The target location information of the target parking lot is obtained interactively. A circle is drawn with the target location information as the center and the traffic flow information collection constraint as the radius to obtain the traffic flow information collection area. This area is the range for collecting traffic flow information around the target parking lot.

[0050] A600: Based on the interaction of the K parking saturation nodes with the traffic flow information collection area, obtain K historical node traffic flow information;

[0051] In one embodiment, such as Figure 3 As shown, based on the interaction of the K parking saturation nodes with the traffic flow information collection area, K historical node traffic flow information is obtained. The method step A600 provided in this application further includes:

[0052] A610: Interact to obtain the target location information of the target parking lot;

[0053] A620: Divide the region according to the target location information and the traffic flow information collection constraints to obtain the traffic flow information collection area;

[0054] A630: Based on the K parking saturation nodes, the first parking saturation node and the second parking saturation node are extracted;

[0055] A640: Construct a traffic flow information collection threshold based on the first parking saturation node and the second parking saturation node;

[0056] A650: Interact with the traffic management cloud platform of the traffic flow information collection area using the traffic flow information collection threshold as a constraint to obtain a historical traffic flow information set;

[0057] A660: Use the K parking saturation nodes to traverse the historical traffic flow information to obtain the traffic flow information of the K historical nodes.

[0058] Specifically, in this embodiment, a first parking saturation node and a second parking saturation node are extracted based on the K parking saturation nodes. The first parking saturation node is the parking saturation node that is furthest from the current time among the K parking saturation nodes, and conversely, the second parking saturation node is the parking saturation node that is closest to the current time among the K parking saturation nodes.

[0059] A vehicle flow information collection threshold is constructed based on the first parking saturation node and the second parking saturation node. The vehicle flow information collection threshold serves as a time span constraint for subsequent collection of historical vehicle flow information in the vehicle flow information collection area.

[0060] The traffic flow data of the traffic flow information collection area is recorded in real time by the road transaction management cloud platform. Based on this, this embodiment uses the traffic flow information collection threshold as a constraint to interact with the traffic management cloud platform of the traffic flow information collection area to obtain the historical traffic flow information set. The historical traffic flow information set stores the real-time traffic flow information of the traffic flow information collection area within the time span constrained by the traffic flow information collection threshold.

[0061] The historical traffic flow information is traversed using the K parking saturation nodes to obtain the traffic flow information of the K historical nodes. The traffic flow information of the K historical nodes is the traffic flow data of the roads around the target parking lot at the same time as the K parking saturation nodes.

[0062] This embodiment achieves the technical effect of obtaining synchronized vehicle information and traffic flow information of the target parking lot, providing highly reliable and available data for subsequent analysis of the actual parking space demand of the target parking lot.

[0063] A700: Optimize parking spaces in the target parking lot based on the traffic flow information from the K historical nodes.

[0064] In one embodiment, the parking space optimization of the target parking lot is performed based on traffic flow information from K historical nodes. Prior to this, the method step A700 provided in this application further includes:

[0065] A711: Preset parking availability features, and based on the parking availability features, traverse the historical parking quantity information set to obtain M parking availability nodes and M used parking space quantity information, wherein the M parking availability nodes and the M used parking space quantity information are mapped one-to-one.

[0066] A712: Based on the M parking availability nodes, traverse the historical traffic flow information set to obtain traffic flow information for the M historical nodes;

[0067] A713: Pre-construct a used parking space prediction sub-network, and use the M used parking space quantity information and the M historical node traffic flow information to adjust the prediction accuracy of the used parking space prediction sub-network.

[0068] In one embodiment, the method step A700 of this application further includes optimizing parking spaces in the target parking lot based on the traffic flow information of the K historical nodes:

[0069] A721: Serialize the traffic flow information of the K historical nodes to obtain the extreme values ​​of the historical traffic flow information;

[0070] A722: Preset information splitting rules are used to split the extreme values ​​of the historical traffic flow information to obtain multiple sets of split vehicle information extreme values;

[0071] A723: Input the extreme values ​​of the multiple sets of split vehicle information into the used parking space prediction subnetwork to obtain multiple sets of used parking space prediction results;

[0072] A724: Summing the multiple sets of predicted used parking spaces yields the target parking space demand.

[0073] A725: Optimize parking spaces in the target parking lot based on the target parking space demand.

[0074] Specifically, it should be understood that when a parking lot has ample parking spaces, there is a correlation between the occupancy of parking spaces in the parking lot and the traffic flow data around the parking lot. However, when the parking lot is full, this correlation disappears.

[0075] Based on this, this embodiment establishes a correlation between the parking space occupancy status in the parking lot and the traffic flow data around the parking lot, which is used to characterize the parking space occupancy status in the parking lot. The parking space occupancy feature is affected by the number of parking spaces in the parking lot and the traffic flow density of the area where the parking lot is located. This embodiment does not impose specific numerical limitations on the parking space occupancy feature.

[0076] This embodiment traverses the historical parking quantity information set based on the parking availability feature to obtain M parking availability nodes and M used parking space quantity information, wherein the M parking availability nodes and the M used parking space quantity information are mapped one-to-one. The parking availability node is the time node in the target parking lot where the parking space occupancy has not reached the point where the target parking lot cannot accommodate new vehicles, and the corresponding used parking space quantity information is the specific number of parking spaces in use.

[0077] Based on the M parking availability nodes, the historical traffic flow information set is traversed to obtain M historical node traffic flow information, which is the traffic flow data of the roads surrounding the target parking lot corresponding to the parking availability node.

[0078] A pre-constructed used parking space prediction subnetwork is built based on a backpropagation neural network. The input data of the used parking space prediction subnetwork is the traffic flow information around the target parking lot, and the output result is the number of parking spaces occupied in the target parking lot. The prediction accuracy of the used parking space prediction subnetwork is adjusted using the M used parking space quantity information and the M historical node traffic flow information.

[0079] Based on the constructed sub-network for predicting used parking spaces, this embodiment serializes the traffic flow information of the K historical nodes to obtain the extreme values ​​of historical traffic flow information. The new extreme values ​​of historical traffic flow are the maximum traffic flow values. A preset information splitting rule is used, for example, to divide the extreme values ​​of historical traffic flow information into three equal parts. Based on this information splitting rule, the extreme values ​​of historical traffic flow information are split to obtain multiple sets of split vehicle information extreme values.

[0080] The extreme values ​​of the multiple sets of split vehicle information are input into the used parking space prediction subnetwork to obtain multiple sets of used parking space prediction results; the multiple sets of used parking space prediction results are summed to obtain the target parking space demand; the current number of parking spaces in the target parking lot is expanded based on the target parking space demand.

[0081] This embodiment achieves the technical effect of accurately estimating the number of parked vehicles in a target parking lot after obtaining traffic flow information around the target parking lot by constructing a used parking space prediction subnetwork. Simultaneously, this embodiment reduces the numerical value by splitting the data using preset information splitting rules, thus preventing the actual obtained data from exceeding the constructed data of the used parking space prediction subnetwork, which would reduce the output accuracy of the used parking space prediction subnetwork. This results in a highly reliable estimate of the number of parking spaces that should be planned for the target parking lot.

[0082] Example 2

[0083] Based on the same inventive concept as the parking flow monitoring method in the foregoing embodiments, such as Figure 4 As shown, this application provides a parking lot traffic monitoring system, wherein the system includes:

[0084] Monitoring model construction module 1 is used to pre-build a parking flow monitoring sub-network, wherein the parking flow monitoring sub-network is used to dynamically monitor the parking volume of the target parking lot;

[0085] Monitoring window construction module 2 is used to preset parking flow monitoring window, wherein the parking flow monitoring window is used to control the operation cycle of the parking flow monitoring sub-network;

[0086] The parking data acquisition module 3 is used to monitor the parking flow of the target parking lot using the parking flow monitoring window as the monitoring period and the parking flow monitoring sub-network to obtain a set of historical parking quantity information, wherein each piece of historical parking quantity information in the set of historical parking quantity information has a collection time node identifier.

[0087] The saturation node determination module 4 is used to preset the parking saturation threshold and traverse the historical parking quantity information set based on the parking saturation threshold to obtain K parking saturation nodes.

[0088] The data collection area setting module 5 is used to preset traffic flow information collection constraints and construct a traffic flow information collection area based on the traffic flow information collection constraints and the target parking lot.

[0089] The node traffic flow acquisition module 6 is used to obtain K historical node traffic flow information by interacting with the traffic flow information acquisition area based on the K parking saturation nodes;

[0090] The parking space optimization execution module 7 is used to optimize the parking spaces of the target parking lot based on the traffic flow information of the K historical nodes.

[0091] In one embodiment, the monitoring model construction module 1 further includes:

[0092] The parking flow monitoring subnetwork includes an inbound information monitoring subnetwork and an outbound information monitoring subnetwork;

[0093] The target parking lot is pre-installed with a first image acquisition device and a second image acquisition device. The first image acquisition device is used to acquire the entrance image information of the target parking lot and transmit the entrance image information to the entrance information monitoring sub-network. The second image acquisition device is used to acquire the exit image information of the target parking lot and transmit the exit image information to the exit information monitoring sub-network.

[0094] The import information monitoring subnetwork identifies the import image information to obtain a set of import vehicle information;

[0095] The export information monitoring subnetwork identifies the export image information to obtain a set of export vehicle information;

[0096] The real-time parking volume monitoring results are calculated based on the imported vehicle information set and the exported vehicle information set.

[0097] In one embodiment, the monitoring model construction module 1 further includes:

[0098] The import information monitoring subnetwork includes a video frame splitting execution module and an entry vehicle identification and counting module, wherein the video frame splitting execution module is used to perform frame splitting processing on the import image information;

[0099] Obtain a set of sample vehicle entry images;

[0100] Semantic segmentation is performed on the set of sample vehicle entry images to obtain a set of sample vehicle entry image segmentation results;

[0101] The vehicle entry recognition and counting module is constructed based on a CNN neural network, and the recognition accuracy of the vehicle entry recognition and counting module is optimized using the sample vehicle entry image set and the sample vehicle entry image segmentation result set.

[0102] In one embodiment, the node traffic flow acquisition module 6 further includes:

[0103] Interactively obtain the target location information of the target parking lot;

[0104] The target location information and the traffic flow information collection constraints are used to divide the area and obtain the traffic flow information collection area.

[0105] The first parking saturation node and the second parking saturation node are extracted based on the K parking saturation nodes;

[0106] A threshold for collecting traffic flow information is constructed based on the first parking saturation node and the second parking saturation node;

[0107] Using the traffic flow information collection threshold as a constraint, interact with the traffic management cloud platform of the traffic flow information collection area to obtain a historical traffic flow information set;

[0108] The historical traffic flow information is obtained by traversing the K parking saturation nodes.

[0109] In one embodiment, the parking space optimization execution module 7 further includes:

[0110] A parking availability feature is preset, and the historical parking quantity information set is traversed based on the parking availability feature to obtain M parking availability nodes and M used parking space quantity information, wherein the M parking availability nodes and the M used parking space quantity information are mapped one-to-one.

[0111] Based on the M parking availability nodes, traverse the historical traffic flow information set to obtain traffic flow information for the M historical nodes;

[0112] A pre-constructed sub-network for predicting used parking spaces is used, and the prediction accuracy of the sub-network is adjusted using the M used parking space quantity information and the M historical node traffic flow information.

[0113] In one embodiment, the parking space optimization execution module 7 further includes:

[0114] Serialize the traffic flow information of the K historical nodes to obtain the extreme values ​​of the historical traffic flow information;

[0115] The preset information splitting rules are used to split the extreme values ​​of the historical traffic flow information to obtain multiple sets of split vehicle information extreme values.

[0116] The extreme values ​​of the multiple sets of split vehicle information are input into the used parking space prediction subnetwork to obtain multiple sets of used parking space prediction results.

[0117] The multiple sets of predicted used parking spaces are summed to obtain the target parking space demand.

[0118] The parking spaces in the target parking lot are optimized based on the target parking space demand.

[0119] In summary, any of the methods or steps described above can be stored as computer instructions or programs in various types of computer memory, and the computer instructions or programs can be recognized by various types of computer processors to implement any of the above methods or steps.

[0120] Based on the above specific embodiments of the present invention, any improvements and modifications made to the present invention by those skilled in the art without departing from the principle of the present invention shall fall within the patent protection scope of the present invention.

Claims

1. A method for monitoring parking flow in a parking lot, characterized in that, The method includes: A pre-built parking flow monitoring sub-network is used to dynamically monitor the number of cars parked in the target parking lot. A preset parking flow monitoring window is provided, wherein the parking flow monitoring window is used to control the operating cycle of the parking flow monitoring sub-network; Using the parking flow monitoring window as the monitoring period, the parking flow monitoring sub-network is used to monitor the parking flow of the target parking lot and obtain a set of historical parking quantity information. Each piece of historical parking quantity information in the set of historical parking quantity information has a collection time node identifier. A parking saturation threshold is preset, and the historical parking quantity information set is traversed based on the parking saturation threshold to obtain K parking saturation nodes; Preset traffic flow information collection constraints, and construct a traffic flow information collection area based on the traffic flow information collection constraints and the target parking lot; Based on the interaction of the K parking saturation nodes with the traffic flow information collection area, the traffic flow information of K historical nodes is obtained. This step also includes interactively obtaining the target location information of the target parking lot. The target location information and the traffic flow information collection constraints are used to divide the area and obtain the traffic flow information collection area. The first parking saturation node and the second parking saturation node are extracted based on the K parking saturation nodes; A threshold for collecting traffic flow information is constructed based on the first parking saturation node and the second parking saturation node; Using the traffic flow information collection threshold as a constraint, interact with the traffic management cloud platform of the traffic flow information collection area to obtain a historical traffic flow information set; The historical traffic flow information is obtained by traversing the K parking saturation nodes; Based on the traffic flow information from the K historical nodes, the method optimizes parking spaces in the target parking lot. Prior to this, the method also includes: A parking availability feature is preset, and the historical parking quantity information set is traversed based on the parking availability feature to obtain M parking availability nodes and M used parking space quantity information, wherein the M parking availability nodes and the M used parking space quantity information are mapped one-to-one. Based on the M parking availability nodes, traverse the historical traffic flow information set to obtain traffic flow information for the M historical nodes; A pre-constructed sub-network for predicting used parking spaces is used, and the prediction accuracy of the sub-network is adjusted using the M used parking space quantity information and the M historical node traffic flow information. The method further includes optimizing parking spaces in the target parking lot based on the traffic flow information from the K historical nodes: Serialize the traffic flow information of the K historical nodes to obtain the extreme values ​​of the historical traffic flow information; The preset information splitting rules are used to split the extreme values ​​of the historical traffic flow information to obtain multiple sets of split vehicle information extreme values. The extreme values ​​of the multiple sets of split vehicle information are input into the used parking space prediction subnetwork to obtain multiple sets of used parking space prediction results. The multiple sets of predicted used parking spaces are summed to obtain the target parking space demand. The parking spaces in the target parking lot are optimized based on the target parking space demand.

2. The method as described in claim 1, characterized in that, A pre-constructed parking flow monitoring sub-network is provided, wherein the parking flow monitoring sub-network is used to dynamically monitor the parking volume of the target parking lot, and the method further includes: The parking flow monitoring subnetwork includes an inbound information monitoring subnetwork and an outbound information monitoring subnetwork; The target parking lot is pre-installed with a first image acquisition device and a second image acquisition device. The first image acquisition device is used to acquire the entrance image information of the target parking lot and transmit the entrance image information to the entrance information monitoring sub-network. The second image acquisition device is used to acquire the exit image information of the target parking lot and transmit the exit image information to the exit information monitoring sub-network. The import information monitoring subnetwork identifies the import image information to obtain a set of import vehicle information; The export information monitoring subnetwork identifies the export image information to obtain a set of export vehicle information; The real-time parking volume monitoring results are calculated based on the imported vehicle information set and the exported vehicle information set.

3. The method as described in claim 2, characterized in that, The import information monitoring subnetwork identifies the import image information to obtain an import vehicle information set, and the method further includes: The import information monitoring subnetwork includes a video frame splitting execution module and an entry vehicle identification and counting module, wherein the video frame splitting execution module is used to perform frame splitting processing on the import image information; Obtain a set of sample vehicle entry images; Semantic segmentation is performed on the set of sample vehicle entry images to obtain a set of sample vehicle entry image segmentation results; The vehicle entry recognition and counting module is constructed based on a CNN neural network, and the recognition accuracy of the vehicle entry recognition and counting module is optimized using the sample vehicle entry image set and the sample vehicle entry image segmentation result set.

4. A parking lot traffic monitoring system, used to execute the method described in any one of claims 1 to 3, characterized in that, The system includes: A monitoring model construction module is used to pre-build a parking flow monitoring sub-network, wherein the parking flow monitoring sub-network is used to dynamically monitor the parking volume of the target parking lot; A monitoring window construction module is used to preset a parking flow monitoring window, wherein the parking flow monitoring window is used to control the operating cycle of the parking flow monitoring sub-network; The parking data acquisition module is used to monitor the parking flow of the target parking lot using the parking flow monitoring window as the monitoring period and the parking flow monitoring sub-network to obtain a set of historical parking quantity information, wherein each piece of historical parking quantity information in the set of historical parking quantity information has a collection time node identifier. The saturation node determination module is used to preset a parking saturation threshold and, based on the parking saturation threshold, traverse the historical parking quantity information set to obtain K parking saturation nodes. The data collection area setting module is used to preset traffic flow information collection constraints and construct a traffic flow information collection area based on the traffic flow information collection constraints and the target parking lot. The node traffic flow acquisition module is used to obtain K historical node traffic flow information by interacting with the traffic flow information acquisition area based on the K parking saturation nodes. The parking space optimization execution module is used to optimize the parking spaces of the target parking lot based on the traffic flow information of the K historical nodes.

Images