A method and system for determining roadside parking abnormal occlusion

Abstract

The application discloses a roadside parking abnormal shielding judgment method and system, and belongs to the intelligent transportation field, which comprises the following steps: vehicle start-stop detection is carried out based on an image to determine a vehicle to be judged; parking type judgment is carried out on a target section to determine a target section type; fixed shielding targets and dynamic shielding targets are determined; the fixed shielding targets and the dynamic shielding targets are taken as detection directions, and abnormal shielding detection is carried out on the vehicle to be judged based on the target section type to obtain a target identification result, wherein the detection requirements of a temporary parking section, a long parking section and a no-parking section are different; and side parking management is carried out on the vehicle to be judged based on the target identification result. The application solves the technical problems of low parking efficiency and great parking safety hazards in the prior art, and achieves the technical effects of improving the efficiency and safety of roadside parking.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent transportation, in particular to a method and system for determining abnormal occlusion of roadside parking. BACKGROUND

[0002] With the acceleration of urbanization, the traffic problem in the city is becoming more and more prominent, and the parking space has become one of the important problems in the field of urban traffic. At present, the situation of roadside parking lot is complex, for example, when the parking space is small, it is easy to cause scratches with surrounding cars or road stones, and a long time is needed for parking operation, or the situation of not being able to stop in the parking space occurs; for example, in the case of heavy traffic on the roadside, due to the visual blind area, the problem of hitting pedestrians may occur. Therefore, in the parking process, there are technical problems of long parking time and large parking safety hazards. SUMMARY

[0003] The embodiments of the present application provide a method and system for determining abnormal occlusion of roadside parking, to solve the technical problems of low parking efficiency and large parking safety hazards in the prior art, and achieve the technical effects of improving the efficiency and safety of roadside parking.

[0004] In a first aspect, the embodiments of the present application provide a method for determining abnormal occlusion of roadside parking, wherein the determining method comprises: performing vehicle start-stop detection based on an image to determine a vehicle to be determined; performing parking type determination on a target section to determine a target section type, wherein the target section type includes a temporary stop section and a long stop section, a no-parking section; determining a fixed occlusion target and a dynamic occlusion target, wherein the fixed occlusion target refers to a real occlusion, and the dynamic occlusion target refers to a visual occlusion; taking the fixed occlusion target and the dynamic occlusion target as a detection direction, performing abnormal occlusion detection on the vehicle to be determined based on the target section type, and obtaining a target recognition result, wherein the detection requirements of the temporary stop section, the long stop section and the no-parking section are different; based on the target recognition result, performing side parking management on the vehicle to be determined.

[0005] In another aspect, the embodiment of the present application also provides a determination system for abnormal occlusion of roadside parking, wherein the determination system comprises: a vehicle start-stop detection module, which performs vehicle start-stop detection based on an image to determine a vehicle to be determined; a road section type determination module, which is configured to determine a parking type of a target road section to determine a target road section type, wherein the target road section type comprises a temporary stop road section, a long stop road section and a no-parking road section; a determination of occlusion target module, which is configured to determine a fixed occlusion target and a dynamic occlusion target, wherein the fixed occlusion target refers to a real occlusion, and the dynamic occlusion target refers to a visual occlusion; an acquisition target identification module, which is configured to take the fixed occlusion target and the dynamic occlusion target as a detection direction, perform abnormal occlusion detection on the vehicle to be determined based on the target road section type, and acquire a target identification result, wherein the detection requirements of the temporary stop road section, the long stop road section and the no-parking road section are different; and a side parking management module, which is configured to perform side parking management on the vehicle to be determined based on the target identification result.

[0006] The one or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0007] Since the computer vision technology is used to identify the occlusion and different road section types are set to different detection methods for side parking management, the technical problems of low parking efficiency and large parking safety hazards in the prior art are effectively solved, and the technical effects of improving the efficiency and safety of roadside parking are achieved. BRIEF DESCRIPTION OF DRAWINGS

[0008] In order to more clearly illustrate the technical solutions of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only exemplary, and other drawings can be obtained by those skilled in the art without creative labor on the basis of the provided drawings.

[0009] Figure 1 A flowchart of a determination method for abnormal occlusion of roadside parking provided by the embodiment of the present application;

[0010] Figure 2 A flowchart of acquiring a target identification result in a determination method for abnormal occlusion of roadside parking provided by the embodiment of the present application;

[0011] Figure 3 A flowchart of determining a first target identification result in a determination method for abnormal occlusion of roadside parking provided by the embodiment of the present application;

[0012] Figure 4A structural schematic diagram of a determination system of abnormal occlusion of roadside parking provided by an embodiment of the present application.

[0013] Explanation of reference signs: vehicle start-stop detection module 11; determination road section type module 12; determination occlusion target module 13; acquisition target identification module 14; side parking management module 15. DETAILED DESCRIPTION

[0014] The embodiment of the present application provides a determination method and system of abnormal occlusion of roadside parking, solves the technical problems of low parking efficiency and large parking safety hazard in the prior art, and achieves the technical effects of improving the efficiency and safety of roadside parking.

[0015] Hereinafter, the technical solutions in the present application will be described clearly and completely with reference to the drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application. In addition, it should be noted that, for the convenience of description, only parts related to the present application are shown in the drawings, rather than all.

[0016] Embodiment one

[0017] As shown in the drawings, the present application provides a determination method of abnormal occlusion of roadside parking, which comprises the following steps: Figure 1

[0018] Step S1000: vehicle start-stop detection based on an image, to determine a vehicle to be determined.

[0019] Specifically, the image is a technology for sensing and detecting objects by using geomagnetic field changes, which can be used for vehicle detection, personnel positioning and the like. Data acquisition and processing are performed through a geomagnetic sensor, and through processing and analysis of geomagnetic field data, detection and positioning of target objects are achieved.

[0020] ​Firstly, the geomagnetic sensor is installed on the ground of each parking space in the parking lot to ensure the stability and accuracy of the sensor. Secondly, the geomagnetic signal of each parking space is collected and processed to calculate the geomagnetic signal threshold of the parking space as the reference value of the parking space. Then, the signal change of the geomagnetic sensor is monitored in real time, and if the signal exceeds the threshold, it indicates that a new vehicle is parked or a vehicle is leaving the parking space. Finally, according to the change of the geomagnetic signal and the detection rule, the start and stop state of the vehicle is judged to determine the vehicle to be judged. The vehicle to be judged refers to the information of the vehicle moving or parked in the detection range of the parking space detected by the geomagnetic sensor, including the size of the vehicle, the height of the chassis, the blind area range, etc. Through the image, the parking space information and the vehicle information are fully understood to provide technical support for subsequent parking control.

[0021] Step S2000: determining the parking type of the target section to determine the type of the target section, wherein the type of the target section includes the pause section and the long parking section, the no parking section;

[0022] Specifically, the target section refers to the section where the vehicle to be judged is about to park. When the vehicle is detected to be about to park, the judgment system locates the vehicle by combining GPS, calls the city parking plan, determines the type of the target section, and if there is no clear target section parking plan, calls the specific information of the target section, analyzes the traffic flow, section width, pedestrian flow, plan and surrounding traffic facilities of the target section by big data processing technology, and determines the type of the target section.

[0023] The target section includes the pause section, the long parking section and the no parking section. The pause section refers to the section that can be parked for a short time, and this type of section generally does not allow long-term parking, and this section has no corresponding parking space marking line. The long parking section refers to the section suitable for long-term parking, and this type of section sets a relatively long parking time limit, and the section is marked with a corresponding parking space marking line. The no parking section refers to the section that does not allow parking, such as bus lane, fire access, etc. By determining and classifying the target section, the parking is managed according to the type of the section to improve the parking efficiency.

[0024] Step S3000: determining the fixed occlusion target and the dynamic occlusion target, wherein the fixed occlusion target refers to the real occlusion, and the dynamic occlusion target refers to the visual occlusion;

[0025] Specifically, the real occlusion refers to the situation that in real life, some objects are occluded by other objects due to the position and relative position of the actual objects. The fixed occlusion target refers to the relative position relationship between objects in real life, which causes some objects to be occluded by other objects, which is a kind of real occlusion. For example, other parked vehicles, trees, traffic signs and other objects.

[0026] Visual occlusion refers to a situation in which, due to occlusion of images or videos, part or all of an object cannot be correctly detected and recognized by a computer. Dynamic occluded target refers to a situation in which, due to occlusion of a moving object, part or all of an occluded target cannot be correctly detected and recognized in a video or image sequence, which is a kind of visual occlusion. For example, in the process of camera recognition, part of the object may be occluded by other objects, resulting in that the object in the image cannot be completely detected or recognized.

[0027] The type and relative position relationship of the occluded object are learned in advance by using a machine learning algorithm. Through analysis and processing of known occluded object data, feature data is obtained, based on which it is determined whether the occluded object in front of the vehicle-mounted camera belongs to a fixed occluded target or a dynamic occluded target. For example, for vehicle-type occluded objects, a machine learning algorithm, a convolutional neural network, and other deep learning algorithms and an optical flow method are used to determine whether the occlusion exists by learning and extracting vehicle features; for pedestrian-type occluded objects, a stereo vision method and an optical flow method are used to determine whether the occlusion exists by calculating depth information and motion direction; for static obstacle-type occluded objects (such as buildings and street lamp posts), a laser radar scanning method and a stereo vision method are used to determine whether the occlusion exists by analyzing point cloud data, depth information, and shape features of the obstacle.

[0028] The fixed occluded target is used as a real occlusion, and the dynamic occluded target is used as a visual occlusion. The occlusion observed by the naked eye is dataized and processed by an algorithm to improve the accuracy of occluded object determination and thus improve the efficiency and safety of roadside parking.

[0029] Step S4000: The fixed occluded target and the dynamic occluded target are used as detection directions, and abnormal occlusion detection is performed on the vehicle to be determined based on the target road section type to obtain a target recognition result, wherein the detection requirements of the temporary parking road section, the long parking road section, and the no-parking road section are different.

[0030] Specifically, for the fixed occluded target, 3D imaging is performed in real time to eliminate the interference of the fixed occluded object and improve the target recognition accuracy. At the same time, a machine learning and deep learning technology is used to train a system algorithm to improve the recognition ability of the system. For the dynamic occluded target, a target tracking technology is used to track and recognize the moving target. At the same time, an infrared sensing technology is used to analyze the infrared radiation of the vehicle to realize tracking and recognition of the vehicle target.

[0031] For the detection needs of different target road types, combined with different detection methods, the shielding detection is carried out, the relatively sensitive real-time detection method is used in the temporary stopping road section, the fixed shielding target and the dynamic shielding target are detected at the same time, the detection of the fixed shielding and the abnormal shielding is used in the long-stopping road section, and the vehicle is directly locked on the no-stopping road section. Different detection methods are used for different types of road sections, and the classification detection improves the detection efficiency.

[0032] Step S5000: based on the target recognition result, the side parking management of the vehicle to be judged is performed.

[0033] Specifically, the target recognition result is the result of image recognition of the vehicle and the parking space on different types of road sections fed back to the system, and the judgment system formulates a corresponding side parking management scheme according to the target feedback result to manage the vehicle to be judged.

[0034] A management model is established through historical management data, the target recognition result is input into the management model, the side parking management scheme is automatically generated, the scheme is audited and tested by the traffic dispatch personnel, and then the corresponding dispatching instruction is generated for management. The accuracy and reliability of the target recognition result are continuously improved by establishing the model, and the parking effect and safety are improved.

[0035] Further, as shown in Figure 2 The embodiment of the present application further comprises:

[0036] Step S4100: when the target road type is a temporary stopping road section, the fixed shielding target and the dynamic shielding target are detected and recognized to determine a first target recognition result;

[0037] Step S4200: when the target road type is a long-stopping road section, the fixed shielding target is detected and recognized to determine a second target recognition result;

[0038] Step S4300: when the target road type is a no-stopping road section, the vehicle to be judged is locked as a third target recognition result;

[0039] Step S4400: the first target recognition result, the second target recognition result and the third target recognition result are used as the detection branch of the target recognition result.

[0040] Specifically, the parking section cannot be parked for a long time, does not contain parking space markings, cannot be detected for parking space markings, and needs to detect fixed and dynamic occlusion targets. The video data of the parking section is analyzed and processed using computer vision technology, target detection is performed, and then the occlusion is determined. According to the data set labeled in the early stage, the features of the target are extracted and recognized and classified by the SVM algorithm, and the accurate occlusion recognition result is obtained as the first target recognition result.

[0041] The long parking section is parked for a long time, the target is relatively static, and unlike the parking section, the vehicle flow is large, only the fixed occlusion target is detected. The video data of the long parking section is analyzed and processed using computer vision technology, target detection is performed, and then the occlusion is determined. The features of the target are extracted and recognized by SVM, and the recognition result of the fixed occlusion target is obtained as the second target recognition result.

[0042] The no parking section prohibits parking of the vehicle, and any parked vehicle needs to be detected and recognized to determine whether the vehicle needs to be punished. After recognizing that a vehicle enters the no parking section, the vehicle target is locked, and information such as the license plate number, vehicle model, and stay time of the vehicle is recorded, and the no parking section is determined according to the third target recognition result.

[0043] The parking section, the long parking section, and the no parking section have different requirements for vehicle recognition, and different detection methods are used for different sections. The branch recognition method improves the accuracy and efficiency of target recognition, better adapts to actual needs, improves the intelligent level of the system, and improves the parking efficiency.

[0044] Further, as shown in Figure 3 The embodiment of the present application further comprises:

[0045] Step S4110: Collecting speed limit on the target section;

[0046] Step S4120: Calculating a preset safety distance based on the speed limit of the section, wherein the preset safety distance is the maximum driving distance required from the time node of discovering the vehicle to be judged to completing lane changing or emergency stopping;

[0047] Step S4130: Determining whether the target visibility meets the preset safety distance, wherein the target visibility is the initial identification position of the vehicle to be judged and the section distance of the vehicle to be judged;

[0048] Step S4140: If not, calculate the safety distance difference, and feed back the safety distance difference and the vehicle to be judged to the urban intelligent parking system.

[0049] Specifically, the vehicle is positioned by the vehicle-mounted GPS system, and traffic control information of the road section is obtained. The information is obtained through the urban traffic management system and related platforms. When it is determined through the image that the vehicle to be judged needs to stop, the maximum form distance required for the vehicle to stop is the preset safety distance. The speed of the vehicle is controlled within the speed limit of the route. The reaction distance is determined through the neural network algorithm. The deceleration distance is determined through the simulation mode. Then, the preset safety distance is obtained.

[0050] The target sight distance refers to the actual distance between the vehicle to be judged and the parking space. The driver's line of sight can be simulated based on map data, environmental models, etc. to calculate the farthest distance that the driver can see and recognize. First, a road map is established, and map data is used to establish the geometric characteristics of the road and lane information, including road length, slope, curvature, signs and markings, etc. Second, an environmental model is created, and environmental data such as light, rain, fog, road conditions, etc. are collected during vehicle travel. Different parameters and conditions are set according to different environmental factors. Finally, simulation calculation is performed to calculate the road information that the driver can see and recognize and the distance to the target parking space in front.

[0051] The target sight distance and the preset safety distance are compared. The target sight distance is the distance between the vehicle to be judged and the parking space from the initial recognition position of the vehicle to be judged. The time point at this position is the starting point of the reaction time zone. If the distance between the vehicle to be judged and the target parking space is not sufficient to ensure safe parking, the safety distance difference is calculated based on the preset safety distance and the target sight distance.

[0052] By calculating the safety distance difference, it can be determined whether the distance between the current vehicle to be judged and the fixed or dynamic blocking target in front is safe. The safety distance difference and the vehicle to be judged are fed back to the urban intelligent parking system for further operation, reducing the misjudgment of parking and improving the intelligence of parking, thereby improving the parking efficiency.

[0053] Further, the embodiments of the present application also include:

[0054] Step S4121: determining the recognition reaction time zone and calculating the reaction distance, wherein the recognition reaction time zone is based on the uniform speed driving time zone of the road section speed limit;

[0055] Step S4122: taking the road section speed limit as the initial speed to determine the deceleration distance, wherein the deceleration distance is the driving distance when the initial speed is reduced to 0;

[0056] Step S4123: determining the preset safety distance based on the reaction distance and the deceleration distance.

[0057] Specifically, identifying the reaction time zone refers to a time period from a time node when the vehicle to be determined needs to perform a parking operation to a time period when the vehicle starts to react to deceleration after the system issues an instruction. The reaction time zone is determined based on a neural network algorithm, which is based on the processing manner of the human nervous system for reaction events, and a prediction result of the reaction time is generated by modeling the connection relationship and dynamic response between neurons. The neural network has multiple nodes and connected neural networks to simulate the process of human information processing and reaction time, and the reaction time is determined by continuously learning the parking information of the vehicle owner and optimizing the algorithm. Then, according to the road section speed limit and the identified reaction time zone, the reaction distance is determined, and the road section speed limit is the maximum speed of the vehicle on the road section, and the reaction distance obtained according to the speed is the maximum reaction distance.

[0058] The deceleration distance refers to the driving distance required for the vehicle to be determined to start deceleration to stop. When calculating the deceleration distance, the real-time speed of the driving vehicle, the road section speed limit, and the braking performance of the driving vehicle need to be considered, and the deceleration distance is obtained by simulation. First, a vehicle dynamics model is established to describe the physical motion of the vehicle during driving. The model includes a vehicle mass point motion model and a vehicle suspension system model. Second, a brake control strategy is set, which refers to how to control the braking of the vehicle during driving, including torque control strategy, hydraulic control strategy, etc. Then, brake parameters are set according to historical brake records, which are determined according to road conditions and brake performance. Finally, the deceleration distance is obtained by simulation, which calculates the deceleration distance of the vehicle by simulating the dynamic changes during driving. During simulation, parameters and strategies are continuously adjusted to improve the accuracy of simulation.

[0059] The reaction distance and the deceleration distance obtained according to the actual situation are added to obtain the preset safety distance. The reaction time is determined by the neural network algorithm, and the deceleration distance is determined by simulation, which greatly improves the accuracy of the preset safety distance and improves the safety of roadside parking.

[0060] Further, the embodiments of the present application also include:

[0061] Step S4210: identifying the fixed occlusion target based on the curb machine, and determining a target recognition result;

[0062] Step S4220: if the target recognition result has abnormal occlusion, locking the abnormal occlusion target;

[0063] Step S4230: feeding the abnormal occlusion target and the vehicle to be determined to the urban intelligent parking system, and managing the side parking of the vehicle to be determined.

[0064] Specifically, the curb machine is installed on the parking space, and the information on the parking space is collected in real time by using the lens. When there is no other vehicle on the parking space, the parking space is collected when the vehicle to be determined is to be parked, and the photos of whether there are fixed shielding targets on the parking space are collected, for example, there are other non-motor vehicles on the parking space, the vehicles on the adjacent parking spaces exceed the parking line, etc. The curb machine sends the measured data to the microchip mounted thereon, wherein the photo of the parking space without a parked vehicle is set as a preset photo, and whether there is a foreign shielding on the parking space is determined by the SSIM algorithm. When there is a foreign shielding, the SSIM algorithm outputs an abnormal result, locks the abnormal shielding target, and the curb machine feeds back the abnormal shielding target and the vehicle to be determined to the urban intelligent parking system. Through the identification and judgment of the fixed shielding target, the information is fed back to the urban intelligent parking system, the management of the vehicle is improved, the parking process is optimized, and the parking efficiency is improved.

[0065] Further, the embodiment of the present application also includes:

[0066] Step S4231: generating a personnel dispatching instruction based on the urban intelligent parking system, wherein the dispatching execution information is additional output information;

[0067] Step S4232: positioning the on-duty personnel on the target section, taking the shortest distance as the response target, and screening the personnel to be dispatched.

[0068] Step S4233: sending the personnel dispatching instruction and the dispatching execution information to the mobile terminal of the personnel to be dispatched, and managing the vehicle to be determined by side parking.

[0069] Specifically, the urban intelligent parking system collects vehicle data and parking space information through urban parking information collectors (such as curb machines, cameras, etc.), including vehicle type, parking duration, parking location, and parking space remaining condition. Based on this information, the target and specific operation to be dispatched are determined according to the actual situation, for example, increasing the number of on-duty personnel in places with large personnel flow, and transporting parked vehicles on forbidden parking sections. After determining the dispatching target and specific operation, the corresponding personnel dispatching instruction is generated according to the region, vehicle type, personnel, etc., including dispatching task, on-duty personnel, operation method, execution information, etc. The dispatching instruction includes automatically generated instructions and manually generated instructions. Among them, the automatically generated instructions determine which parking spaces need to be dispatched by applying clustering algorithms, and determine the optimal solution of the dispatching quantity by using regression analysis. Finally, the automatically generated instructions and manually generated instructions are summarized into dispatching execution information.

[0070] After generating the personnel mobilization instruction, the system takes the target road section positioning as the center, positions the on-duty personnel, screens the on-duty personnel with the distance of the on-duty personnel to the target road section as the screening condition, and the personnel who meet the requirements enter the personnel dispatching instruction. After screening out the personnel who meet the requirements of the dispatching task, stop screening, send the personnel dispatching instruction and dispatching execution information to the mobile terminal of the personnel to be dispatched, and manage the side parking of the vehicle, including towing, notifying the vehicle owner, cleaning the obstruction, etc.

[0071] By collecting parking information, processing parking information, generating corresponding dispatching mode according to parking information, improving the efficiency of parking, and cleaning up hidden dangers in a timely manner, the safety of parking is improved.

[0072] In summary, a method for determining abnormal obstruction of roadside parking has the following technical effects:

[0073] Based on the image, the vehicle start-stop detection is performed to determine the vehicle to be determined, and the state of the vehicle is accurately judged to provide technical support for subsequent calculation of the target visibility and the preset safety distance according to the vehicle state. The target road section is determined by the parking type determination, and the target road section type is determined, wherein the target road section type includes the pause road section and the long parking road section, the no parking road section, different side parking management schemes are set according to different road section types, and the accuracy of parking management is improved. The fixed obstruction target and the dynamic obstruction target are determined, wherein the fixed obstruction target refers to the real obstruction, and the dynamic obstruction target refers to the visual obstruction. The obstruction targets are classified and respectively refer to the real obstruction and the visual obstruction, the obstruction data is simplified, and the time complexity of the algorithm is reduced. The fixed obstruction target and the dynamic obstruction target are used as the detection direction, the abnormal obstruction detection of the vehicle to be determined is performed based on the target road section type, and the target recognition result is obtained, wherein the detection requirements of the pause road section, the long parking road section and the no parking road section are different, the target recognition result is accurately and efficiently calculated by measuring and analyzing various factors affecting parking. Based on the target recognition result, the side parking management of the vehicle to be determined is performed, the target recognition result determined based on various factors can accurately reflect the parking condition, the system judges the parking management based on this, and the parking efficiency and safety are improved.

[0074] Embodiment two

[0075] Based on the same inventive concept as the method for determining abnormal obstruction of roadside parking in the foregoing embodiments, as Figure 4 shown, the present application also provides a system for determining abnormal obstruction of roadside parking, wherein the system comprises:

[0076] A vehicle start-stop detection module, the vehicle detection module performs vehicle start-stop detection based on images to determine the vehicle to be determined;

[0077] A determination road section type module is configured to determine a parking type of a target road section, and determine a target road section type, wherein the target road section type comprises a pause road section, a long parking road section and a no parking road section;

[0078] A determination shelter target module is configured to determine a fixed shelter target and a dynamic shelter target, wherein the fixed shelter target refers to a real shelter, and the dynamic shelter target refers to a visual shelter;

[0079] An acquisition target identification module is configured to take the fixed shelter target and the dynamic shelter target as a detection direction, perform an abnormal shelter detection on the vehicle to be determined based on the target road section type, and acquire a target identification result, wherein the detection requirements of the pause road section, the long parking road section and the no parking road section are different;

[0080] A side parking management module is configured to perform a side parking management on the vehicle to be determined based on the target identification result.

[0081] Further, the embodiment of the application further comprises:

[0082] A first target identification module is configured to perform a detection identification on the fixed shelter target and the dynamic shelter target when the target road section type is the pause road section, and determine a first target identification result;

[0083] A second target identification module is configured to perform a detection identification on the fixed shelter target when the target road section type is the long parking road section, and determine a second target identification result;

[0084] A third target identification module is configured to lock the vehicle to be determined as a third target identification result when the target road section type is the no parking road section;

[0085] A target identification result module is configured to take the first target identification result, the second target identification result and the third target identification result as a detection branch of the target identification result.

[0086] Further, the embodiment of the application further comprises:

[0087] A speed limit collection module is configured to collect a speed limit of the target road section;

[0088] A safety distance module is configured to calculate a preset safety distance based on a road section speed limit, wherein the preset safety distance is a maximum driving distance required from a time node of finding the vehicle to be determined to completing a lane changing or an emergency stop;

[0089] A target sight distance module is configured to determine whether a target sight distance meets a preset safety distance, wherein the target sight distance is a distance between a first identification position of the vehicle to be determined and a road section distance of the vehicle to be determined.

[0090] A safety distance difference value module is configured to calculate a safety distance difference value if the target sight distance does not meet the preset safety distance, and feed the safety distance difference value and the vehicle to be determined to the urban intelligent parking system.

[0091] Further, the embodiment of the present application further comprises:

[0092] A reaction distance module is configured to determine an identification reaction time zone and calculate a reaction distance, wherein the identification reaction time zone is a uniform speed driving time zone based on the road section speed limit.

[0093] A deceleration distance module is configured to determine a deceleration distance by taking the road section speed limit as an initial speed, wherein the deceleration distance is a driving distance when the initial speed is reduced to 0.

[0094] A preset safety distance module is configured to determine the preset safety distance based on the reaction distance and the deceleration distance.

[0095] Further, the embodiment of the present application further comprises:

[0096] A target identification module is configured to identify the fixed shielding target based on a curb machine, and determine a target identification result.

[0097] An abnormal shielding target module is configured to lock an abnormal shielding target if the target identification result is abnormal.

[0098] A feedback system module is configured to feed the abnormal shielding target and the vehicle to be determined to the urban intelligent parking system, and manage side parking of the vehicle to be determined.

[0099] Further, the embodiment of the present application further comprises:

[0100] An instruction generation module is configured to generate a personnel dispatching instruction based on the urban intelligent parking system, wherein dispatching execution information is additional output information.

[0101] A personnel positioning module is configured to position on-duty personnel on the target road section, take the shortest distance as a response target, and screen personnel to be dispatched.

[0102] A feedback terminal module is configured to send the personnel dispatching instruction and the dispatching execution information to a mobile terminal of the personnel to be dispatched, and to manage side parking of the vehicle to be judged.

[0103] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the difference from other embodiments. The foregoing Figure 1 The determination method and specific examples of the abnormal occlusion of the roadside parking in the first embodiment are also applicable to the determination system of the abnormal occlusion of the roadside parking in the present embodiment. The determination system of the abnormal occlusion of the roadside parking in the present embodiment can be clearly understood by the skilled in the art through the foregoing detailed description of the determination method of the abnormal occlusion of the roadside parking. Therefore, for the sake of brevity of the specification, the determination system of the abnormal occlusion of the roadside parking in the present embodiment will not be described in detail again, and the relevant part can be referred to the method part description.

[0104] The above description of the disclosed embodiments enables a person skilled in the art to implement or use the present application. Various modifications to the embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for determining abnormal occlusion of a roadside parking, characterized by, The method comprises: Based on the image, the vehicle start-stop detection is carried out, and the vehicle to be judged is determined; The parking type of the target section is determined, and the target section type is determined, wherein the target section type includes the pause section and the long parking section, the no parking section; Determine the fixed occlusion target and the dynamic occlusion target, wherein the fixed occlusion target refers to the real occlusion, and the dynamic occlusion target refers to the situation that part or all of the occluded target cannot be correctly detected and recognized due to the occlusion of moving objects in the video or image sequence; The fixed occlusion target and the dynamic occlusion target are taken as the detection direction, the target recognition result is obtained by carrying out abnormal occlusion detection on the vehicle to be judged based on the target section type, wherein the detection requirements of the pause section, the long parking section and the no parking section are different; Based on the target recognition result, the side parking management of the vehicle to be judged is carried out; The fixed occlusion target and the dynamic occlusion target are taken as the detection direction, the target recognition result is obtained by carrying out abnormal occlusion detection on the vehicle to be judged based on the target section type, comprising: When the target section type is the pause section, the fixed occlusion target and the dynamic occlusion target are detected and identified to determine the first target recognition result; When the target section type is the long parking section, the fixed occlusion target is detected and identified to determine the second target recognition result; When the target section type is the no parking section, the vehicle to be judged is locked as the third target recognition result; The first target recognition result, the second target recognition result and the third target recognition result are taken as the detection branch of the target recognition result.

2. The method of claim 1, wherein, The determination of the first target recognition result comprises: Collecting the speed limit of the target section; Based on the section speed limit, the preset safety distance is calculated, wherein the preset safety distance is the maximum driving distance required from the time node of discovering the vehicle to be judged to completing lane changing or emergency stopping; Judge whether the target visual distance meets the preset safety distance, wherein the target visual distance is the distance between the initial identification position of the vehicle to be judged and the parking space; If not, calculate the safety distance difference value, and feed back the safety distance difference value and the vehicle to be judged to the urban intelligent parking system.

3. The method of claim 2, wherein, The preset safety distance is calculated based on the section speed limit, comprising: Determine the identification reaction time zone and calculate the reaction distance, wherein the identification reaction time zone is the uniform speed driving time zone based on the section speed limit; Taking the section speed limit as the initial speed, the deceleration distance is determined, wherein the deceleration distance is the driving distance when the initial speed is reduced to 0; Based on the reaction distance and the deceleration distance, the preset safety distance is determined.

4. The method of claim 2, wherein, The determination of the second target recognition result comprises: Based on the curb machine, the fixed occlusion target is identified to determine the target recognition result; If the target recognition result has abnormal occlusion, the abnormal occlusion target is locked; The abnormal occlusion target and the vehicle to be judged are fed back to the urban intelligent parking system, and the side parking management of the vehicle to be judged is carried out.

5. The method of claim 4, wherein, Comprise: The personnel dispatching instruction is generated based on the urban intelligent parking system, wherein the dispatching execution information is additional output information; The target section is positioned by the on-duty personnel, and the shortest distance is taken as a response target to screen the personnel to be dispatched; The personnel dispatching instruction and the dispatching execution information are sent to a mobile terminal of the personnel to be dispatched to manage side parking of the vehicle to be determined.

6. A system for determining abnormal occlusion of a roadside parking, characterized by, The system comprises: A vehicle start-stop detection module that detects vehicle start-stop based on images to determine a vehicle to be determined; A section type determination module that determines a parking type of a target section to determine a target section type, wherein the target section type comprises a pause section and a long-stopping section, and a no-parking section; A determination of an occlusion target module that determines a fixed occlusion target and a dynamic occlusion target, wherein the fixed occlusion target refers to a real occlusion, and the dynamic occlusion target refers to a situation in which part or all of an occluded target cannot be correctly detected and recognized due to the occlusion of a moving object in a video or image sequence; An acquisition target identification module that takes the fixed occlusion target and the dynamic occlusion target as a detection direction, performs abnormal occlusion detection on the vehicle to be determined based on the target section type, and acquires a target identification result, wherein the detection requirements of the pause section, the long-stopping section, and the no-parking section are different; The target identification module also takes the fixed occlusion target and the dynamic occlusion target as a detection direction, performs abnormal occlusion detection on the vehicle to be determined based on the target section type, and acquires a target identification result, comprising: When the target section type is a pause section, the fixed occlusion target and the dynamic occlusion target are detected and identified to determine a first target identification result; When the target section type is a long-stopping section, the fixed occlusion target is detected and identified to determine a second target identification result; When the target section type is a no-parking section, the vehicle to be determined is locked as a third target identification result; The first target identification result, the second target identification result, and the third target identification result are taken as detection branches of the target identification result; A side parking management module that manages side parking of the vehicle to be determined based on the target identification result.

Images