CN112348817A - Parking space identification method and device, vehicle-mounted terminal and storage medium - Google Patents
Parking space identification method and device, vehicle-mounted terminal and storage medium Download PDFInfo
- Publication number
- CN112348817A CN112348817A CN202110021253.7A CN202110021253A CN112348817A CN 112348817 A CN112348817 A CN 112348817A CN 202110021253 A CN202110021253 A CN 202110021253A CN 112348817 A CN112348817 A CN 112348817A
- Authority
- CN
- China
- Prior art keywords
- parking space
- point
- points
- angular
- parking
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/64—Analysis of geometric attributes of convexity or concavity
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20112—Image segmentation details
- G06T2207/20164—Salient point detection; Corner detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30248—Vehicle exterior or interior
- G06T2207/30252—Vehicle exterior; Vicinity of vehicle
- G06T2207/30264—Parking
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Probability & Statistics with Applications (AREA)
- Quality & Reliability (AREA)
- Geometry (AREA)
- Traffic Control Systems (AREA)
Abstract
The application relates to the technical field of automatic driving, and provides a parking space identification method and device, a vehicle-mounted terminal and a storage medium. This application uses parking stall angular point as the check point of density cluster, because adjacent parking stall angular point is connected by the parking stall line, namely, there is the check point of car position line between adjacent parking stall angular point, consequently, can reach the check point that can confirm that each check point is adjacent based on the density, obtain the respective adjacent parking stall angular point of parking stall angular point, utilize adjacent parking stall angular point to pass through the characteristic of parking stall line connection, distortion parking stall/special-shaped parking stall in the image can be discerned, improve parking stall identification accuracy. Moreover, each parking space angular point is used as a check point of density clustering, so that when a parking space line near the parking space angular point is shielded, and under the condition that fewer detection points of a parking space line exist, a parking space corresponding to the parking space angular point can be identified, the parking space inspection total rate is improved, and the missing inspection of the parking space is avoided.
Description
Technical Field
The application relates to the technical field of automatic driving, in particular to a parking space identification method and device, a vehicle-mounted terminal and a storage medium.
Background
With the development of the automatic driving technology, an automatic parking technology has appeared. The automatic parking is divided into three stages of parking space sensing, decision planning and control, wherein the parking space sensing is the basis of the automatic parking, whether the parking space can be accurately extracted or not is judged, and whether the automatic parking task is accurately finished or not is judged. In the parking stall perception scheme that traditional technique provided, generally utilize the matching of parking stall template to confirm the parking stall, distortion parking stall/special-shaped parking stall in this kind of mode is difficult to discern the image, and parking stall discernment precision is lower.
Disclosure of Invention
In view of the above, it is necessary to provide a parking space recognition method, device, vehicle-mounted terminal, and storage medium for solving the above technical problems.
A parking space identification method, the method comprising:
acquiring detection points of all parking spaces obtained by carrying out parking space detection on the parking space map;
taking each parking space angular point in each parking space detection point as a check point of density clustering, and determining respective adjacent check points of each check point based on density to obtain respective adjacent parking space angular points of each parking space angular point;
and determining the parking spaces in the parking space map based on the parking space angular points and the parking space angular points adjacent to the parking space angular points.
A parking space recognition device, the device comprising:
the parking space detection point acquisition module is used for acquiring each parking space detection point obtained by carrying out parking space detection on the parking space map;
the density clustering module is used for taking each parking space angular point in each parking space detection point as a check point of density clustering, determining respective adjacent check points of each check point based on density, and obtaining respective adjacent parking space angular points of each parking space angular point;
and the parking space determining module is used for determining the parking spaces in the parking space map based on the parking space angular points and the parking space angular points adjacent to the parking space angular points.
An in-vehicle terminal comprising a memory and a processor, the memory storing a computer program, the processor performing the method of:
acquiring detection points of all parking spaces obtained by carrying out parking space detection on the parking space map;
taking each parking space angular point in each parking space detection point as a check point of density clustering, and determining respective adjacent check points of each check point based on density to obtain respective adjacent parking space angular points of each parking space angular point;
and determining the parking spaces in the parking space map based on the parking space angular points and the parking space angular points adjacent to the parking space angular points.
A computer-readable storage medium, on which a computer program is stored, the computer program being executable by a processor to perform the method of:
acquiring detection points of all parking spaces obtained by carrying out parking space detection on the parking space map;
taking each parking space angular point in each parking space detection point as a check point of density clustering, and determining respective adjacent check points of each check point based on density to obtain respective adjacent parking space angular points of each parking space angular point;
and determining the parking spaces in the parking space map based on the parking space angular points and the parking space angular points adjacent to the parking space angular points.
After acquiring each parking space detection point obtained by detecting the parking space of the parking space map, the parking space identification method, the device, the vehicle-mounted terminal and the storage medium take each parking space angular point in each parking space detection point as a check point of density clustering, and can determine the respective adjacent check point of each check point based on the density to obtain the respective adjacent parking space angular point of each parking space angular point; and determining the parking spaces in the parking space map based on the parking space angular points and the parking space angular points adjacent to the parking space angular points. Therefore, the parking space angular points are used as the check points of density clustering, and the adjacent parking space angular points are connected by the parking space lines, namely, the check points of the parking space lines exist between the adjacent parking space angular points, so that the check points adjacent to each check point can be determined based on the density, the parking space angular points adjacent to each parking space angular point are obtained, the characteristic that the adjacent parking space angular points are connected through the parking space lines is utilized, the distorted parking spaces/special-shaped parking spaces in the images can be identified, and the parking space identification precision is improved. Moreover, each parking space angular point is used as a check point of density clustering, so that when a parking space line near the parking space angular point is shielded, and under the condition that fewer detection points of a parking space line exist, a parking space corresponding to the parking space angular point can be identified, the parking space inspection total rate is improved, and the missing inspection of the parking space is avoided.
Drawings
FIG. 1 is a diagram of an exemplary embodiment of a parking space recognition method;
FIG. 2 is a diagram of an exemplary embodiment of a parking space recognition method;
FIG. 3 is a flow chart illustrating a parking space recognition method according to an embodiment;
FIG. 4 is a diagram of an exemplary implementation of a parking space recognition method;
fig. 5(a) is a schematic flow chart of a parking space identification method in an embodiment;
FIG. 5(b) is a schematic diagram of a disabled vehicle group in one embodiment;
FIG. 6 is a flowchart illustrating a parking space recognition method according to an embodiment;
fig. 7(a) to 7(c) are diagrams illustrating parking space recognition results in one embodiment;
FIG. 8 is a block diagram of an embodiment of a parking space recognition device;
fig. 9 is an internal configuration diagram of the in-vehicle terminal in one embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
Reference in the specification to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the specification. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is explicitly and implicitly understood by one skilled in the art that the embodiments described herein can be combined with other embodiments.
The parking space identification method can be applied to computer equipment such as a vehicle-mounted terminal and a remote server, and the following description is given by taking the vehicle-mounted terminal as an example.
The application provides a parking space identification method which can be applied to computer equipment such as a vehicle-mounted terminal and a remote server. The following description is made with reference to fig. 1 to fig. 3, and taking as an example that the method is applied to a vehicle-mounted terminal, and includes the following steps:
step S301, acquiring detection points of each parking space obtained by detecting the parking spaces of the parking space map;
step S302, taking each parking space angular point in each parking space detection point as a check point of density clustering, and determining the respective adjacent check point of each check point based on density to obtain the respective adjacent parking space angular point of each parking space angular point;
as shown in fig. 1 and fig. 2, after the parking space detection is performed on the parking space map, the obtained parking space detection points may include parking space angle points 101a to 101c and a point 102 on a parking space line (hereinafter, referred to as a parking space line point).
Because two adjacent parking space angular points are connected through a parking space line in an actual parking space, when the adjacent relation between the parking space angular points in the parking space detection points is determined, the adjacent relation can be determined by using the density; as shown in fig. 1, a plurality of parking space line points 102 are distributed between the parking space angle point 101a and the parking space angle point 101b, so that after density clustering is performed on a plurality of parking space detection points by using the parking space angle point 101a as a density clustering check point, it can be determined that the parking space angle point 101b is reachable by the density of the parking space angle point 101a, and the parking space angle point 101a and the parking space angle point 101b have an adjacent relationship.
In the density clustering kernel point selection mode, if the number of parking space detection points distributed in the neighborhood epsilon of a certain parking space detection point is greater than or equal to the threshold value minPts, for example, the parking space angular point 101d in FIG. 2 can be selected as a kernel point; however, for the parking space angular point 101c in fig. 2, since the parking space line near the parking space angular point 101c is shielded, the number of the parking space detection points distributed in the epsilon in the neighborhood of the parking space angular point 101c is smaller than the threshold value minPts, so the parking space angular point 101c cannot be selected as the core point of the density cluster, and the situation that the parking space corresponding to the parking space angular point 101c is missed to be detected exists. Therefore, in the above steps, each parking space angular point is used as a check point of density clustering, so that the missing detection of the parking space corresponding to the parking space angular point 101c can be avoided.
Step S303, determining the parking spaces in the parking space map based on the parking space angular points and the parking space angular points adjacent to the parking space angular points.
In the parking space identification method, the parking space angular points are taken as the check points of density clustering, and the adjacent parking space angular points are connected by parking space lines, namely, a check point of a parking space line exists between the adjacent parking space angular points, so that the adjacent check points of each check point can be determined based on the density, the respective adjacent parking space angular points of the parking space angular points are obtained, and the distorted/abnormal parking spaces in the image can be identified by utilizing the characteristic that the adjacent parking space angular points are connected by the parking space lines, thereby improving the parking space identification precision. Moreover, each parking space angular point is used as a check point of density clustering, so that when a parking space line near the parking space angular point is shielded, and under the condition that fewer detection points of a parking space line exist, a parking space corresponding to the parking space angular point can be identified, the parking space inspection total rate is improved, and the missing inspection of the parking space is avoided.
In some scenarios, in the process of performing density clustering based on each core point, the termination condition that the density can be reached may be that the number of the detected parking space points distributed in the neighborhood epsilon where the detected parking space points do not satisfy the detected parking space points is greater than or equal to a threshold value minPts (the detected parking space points that do not satisfy the termination condition may be referred to as edge points); however, as shown in fig. 4, based on the termination condition, when the parking space angular point 101e is a core point for density clustering, finding the parking space line point 102a stops, and at this time, even if the parking space angular point 101f is adjacent to the parking space angular point 101e in reality, the parking space angular point 101f cannot be found because the parking space line point is missing due to the middle part of the parking space line being shielded, resulting in low parking space identification accuracy.
Therefore, the parking space identification accuracy is further improved, and in one embodiment provided by the application, the first core point is found according to the reachable termination condition of the density of each core point. That is to say, when the parking space angular point 101e performs density clustering for the core point, after finding the parking space line point 102a, the range is still expanded to continue finding and find the parking space line point 102b, and then a first parking space angular point 101f is found, and then the continuing finding of the next parking space angular point along the parking space angular point 101f is stopped, that is, even if the density of the parking space line point 102c can be reached by the parking space angular point 101e, the finding of the parking space angular point from the parking space angular point 101e to the parking space angular point 101f is stopped at this time, and the parking space angular point 101b is used as the parking space angular point adjacent to the parking space angular point 101 a. It can be understood that the search of the parking space angle point in the direction from the parking space angle point 101e to the parking space angle point 101f does not affect the search of the parking space angle point in the direction from the parking space angle point to the parking space angle point 101 g.
Further, the in-vehicle terminal may use a pre-constructed Density Clustering model, which may be a DBSCAN model (Density-Based Spatial Clustering of Applications with Noise), when executing the step S302. Specifically, the in-vehicle terminal may perform the steps of: acquiring a pre-constructed density clustering model; the density clustering model selects the corner points of each parking space as the core points in the process of performing density clustering on the detection points of each parking space, and finds the first core point according to the termination condition that the density of each core point can reach; inputting the parking space detection points into the density clustering model, so that the density clustering model performs density clustering on the parking space detection points according to the kernel point selection condition and the termination condition that the density of each kernel point can reach, and outputs a point set of each kernel point; and taking other core points in the point set of each core point as core points adjacent to each core point to obtain parking space angle points adjacent to each parking space angle point.
Exemplarily, the point set of the parking space angle point 101e includes a parking space angle point 101f and a parking space angle point 101g, and at this time, the parking space angle point 101f and the parking space angle point 101g may be used as the parking space angle point adjacent to the parking space angle point 101 e.
In one embodiment, the parking lot map comprises a set of handicapped parking lots; each parking space detection point comprises a plurality of parking space detection points which form the incomplete parking space group; each parking space angular point comprises a plurality of parking space angular points forming the incomplete parking space group.
As shown in fig. 5(a), when executing step S303, the in-vehicle terminal may further execute the following steps: determining a convex hull formed by each parking space angular point and the parking space angular points adjacent to the parking space angular point based on the parking space angular point and the parking space angular point adjacent to the parking space angular point (step S501); determining parking space angular points in a convex hull and straight lines where the parking space angular points adjacent to the parking space angular points in the convex hull are located; the parking space angular points in the convex hull are parking space detection points which are positioned in the convex hull and form the incomplete parking space group; taking the intersection point of the straight line and the convex hull as a parking space auxiliary point (step S502); determining a finished vehicle position group corresponding to the incomplete vehicle position group in the vehicle bitmap based on the parking space auxiliary points and available parking space angular points; the available parking space angular points are parking space angular points in the plurality of parking space detection points forming the incomplete parking space group except the parking space angular points in the convex hull.
As shown in fig. 5(a), after the vehicle-mounted terminal executes steps S301 to S302, the obtained adjacent relationship between the parking space corner points ABCDEF is: a- > (B, C), B- > (A), C- > (A, D, E), D- > (C), E- > (C, F) and F- > (E). It can be seen that B, D and F both have only one adjacent parking space angle point, and therefore, the parking spaces formed by the parking space angle points ABCD and the parking spaces formed by the parking space angle points CDEF belong to the incomplete parking spaces. The parking space group formed by the parking spaces (i) and (ii) can be called a defective parking space group, and in the defective parking space group, the parking spaces (i) and (ii) share the same defective part (a parking space line CD). In some embodiments, the handicapped parking space group comprises at least one handicapped parking space and a complete parking space adjacent to the handicapped parking space; as shown in fig. 5(b), the incomplete parking space group may include incomplete parking spaces formed by parking space angular points ABCD and complete parking spaces formed by parking space angular points HIEF, and the complete parking spaces are adjacent to the incomplete parking spaces. In addition, the number of the parking spaces which can be included in the incomplete parking space group can be three or more than three.
The vehicle-mounted terminal performs convex hull fitting on the parking space angular points ABCDEF to obtain corresponding convex hulls, and the parking space angular points D are located in the convex hulls; and then, the vehicle-mounted terminal takes the intersection G of the straight line where the parking space line CD is located and the convex hull as a parking space auxiliary point, and obtains a complete parking space formed by a parking space angular point ABCG and a complete parking space formed by a parking space angular point CGEF based on the parking space auxiliary point G and the parking space angular point ABCEF.
It is thus clear that under the condition that the detection error obtains parking stall angular point D, the above-mentioned embodiment determines the parking stall auxiliary point based on convex closure fitting and the nodical of corresponding straight line and convex closure, supplements incomplete parking stall, obtains complete parking stall, can further improve the precision of parking stall discernment.
Further, when the vehicle-mounted terminal determines the complete vehicle position group corresponding to the incomplete vehicle position group in the vehicle position map based on the parking space auxiliary point and the available parking space angular point, the vehicle-mounted terminal may further perform the following steps: determining a first target parking space angular point in the available parking space angular points; the first target parking space angular point is a parking space angular point adjacent to the parking space angular point in the convex hull; taking the first target parking space angular point as a parking space angular point adjacent to the parking space auxiliary point; determining a second target parking space angular point in the available parking space angular points; the second target parking space angular point is a parking space angular point with the number of adjacent parking space angular points in the available parking space angular points being only one; taking the second target parking space angular point as a parking space angular point adjacent to the parking space auxiliary point; and obtaining a finished vehicle position group corresponding to the incomplete vehicle position group in the vehicle position diagram based on the adjacent vehicle position angular points of the vehicle position auxiliary points and the available vehicle position angular points.
Exemplarily, the adjacent relationship of the above-mentioned parking space corner points ABCDEF is: a- > (B, C), B- > (A), C- > (A, D, E), D- > (C), E- > (C, F) and F- > (E). After the parking space auxiliary point G is determined, updating the adjacent relation to obtain an updated parking space angular point adjacent relation as follows: a- > (B, C), B- > (A, G), C- > (A, G, E), G- > (B, C, F), E- > (C, F), F- > (E, G). Based on the updated adjacent relationship of the parking space angular points, a complete parking space formed by the parking space angular points ABCG and a complete parking space formed by the parking space angular points CGEF can be obtained.
In one embodiment, when executing step S301, the in-vehicle terminal may execute the following steps: acquiring a ring view of the periphery of a vehicle; intercepting an image of a parking space detection effective area in the ring view to obtain the parking space map; inputting the parking space map into a pre-constructed neural network so that the neural network performs parking space detection on the parking space map and outputs the detection points of the parking spaces; and acquiring the detection points of the parking spaces output by the neural network.
Therefore, in the mode, the parking space detection is carried out only by inputting the effective parking space detection area into the neural network, so that the parking space detection speed can be increased, and the calculation resources are saved.
In order to better understand the above method, an application example of the parking space identification method of the present application is described in detail below with reference to fig. 6. In the application example, the parking spaces are segmented through the deep neural network to obtain a plurality of parking space detection points (including parking space line points and parking space angular points), the adjacency relation of the parking space angular points is obtained through the improved DBSCAN model, then the convex hull fitting algorithm is adopted to obtain the convex hull of the parking space angular points, the incomplete parking space lines are supplemented through the convex hull, the adjacency relation is updated, and finally the parking spaces are extracted through the adjacency relation.
Specifically, the application example may include the steps of:
step S601, ring view input:
and calibrating the four fisheye cameras by using computer equipment, and splicing a ring view with a left coverage range and a right coverage range of 6m according to calibration parameters.
Step S602, image preprocessing:
the computer device intercepts the 6m ring view, because most of the 6m ring view is invalid regions, the invalid regions need to be deleted, intercepts the interested regions with valid parking space detection, and adjusts the size of the interested regions to the input image size specified by the neural network.
Step S603, extracting parking space angular points and parking space lines:
and the computer equipment inputs the parking space map obtained by preprocessing into a neural network, the neural network outputs a plurality of parking space detection points, and the parking space detection points comprise parking space angular points and parking space line points.
Step S604, establishing a parking space corner adjacency relation (equivalent to a parking space corner adjacency relation):
the step can be realized by using a DBSCAN model, and the process of determining the adjacent relation of the parking space angular points by the DBSCAN model can be realized by selecting a check point from a plurality of parking space detection points, adding the parking space detection points meeting epsilon and minPts conditions into a point set of the check point from the check point until the encountered parking space detection points are edge points, and traversing all the parking space detection points to obtain the point set of each parking space detection point.
Further, the core point selection condition of the DBSCAN model may be adjusted as follows: taking the parking space angular points as check points; the end condition that the density can reach is to find the first parking space angular point. At this time, the computer device may select one parking space angular point from the multiple candidate parking space angular points as a core point, and use the adjusted DBSCAN model to use another parking space angular point of the same kind as the parking space angular point as an adjacent parking space angular point of the parking space angular point until all candidate parking space angular points are traversed, and establish an adjacency relation between the parking space angular points.
Step S605, supplementing the incomplete parking space:
the computer equipment can perform convex hull fitting on all parking space angular points to obtain a minimum convex hull, then extend the parking space lines in the minimum convex hull to the convex hull to obtain intersection points of the extended parking space lines and the convex hull, and take the intersection points as parking space auxiliary points.
Step S606, updating the adjacent relation of the parking space angular points:
and replacing the parking space angular points in the convex hull by using the parking space auxiliary points, and updating the adjacent relation of the corresponding parking space angular points.
Step S607, parking space extraction:
and extracting the parking spaces by utilizing the updated adjacent relation of the parking space angular points to obtain complete parking spaces formed by the parking space angular points ABCG and complete parking spaces formed by the parking space angular points CGEF.
Step S608, outputting the parking space information:
and the computer equipment outputs the extracted parking space information to a subsequent planning and control module to complete the automatic parking task.
Fig. 7(a) to 7(c) show parking spaces 701 to 705 identified by the above method, and it can be seen that, in this application example, based on the sensing result of deep learning, a special-shaped parking space and a partially worn parking space are extracted, which can adapt to complex working conditions and improve the parking space identification accuracy.
It should be understood that, although the steps in the flowcharts of fig. 1 to 6 are shown in sequence as indicated by the arrows, the steps are not necessarily performed in sequence as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 1 to 6 may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of performing the steps or stages is not necessarily sequential, but may be performed alternately or alternately with other steps or at least some of the other steps or stages.
In one embodiment, as shown in fig. 8, there is provided a parking space recognition device including:
a parking space detection point acquisition module 801, configured to acquire parking space detection points obtained by performing parking space detection on the parking space map;
a density clustering module 802, configured to use each parking space angular point in each parking space detection point as a check point of density clustering, and determine, based on density, a respective adjacent check point of each check point, so as to obtain a respective adjacent parking space angular point of each parking space angular point;
and a parking space determining module 803, configured to determine a parking space in the parking space map based on each parking space angular point and the respective adjacent parking space angular point of each parking space angular point.
In one embodiment, the achievable termination condition for the density of each core point is finding the first core point.
In an embodiment, the density clustering module 802 is further configured to obtain a pre-constructed density clustering model; the density clustering model selects the corner points of each parking space as the core points in the process of performing density clustering on the detection points of each parking space, and finds the first core point according to the termination condition that the density of each core point can reach; inputting the parking space detection points into the density clustering model, so that the density clustering model performs density clustering on the parking space detection points according to the kernel point selection condition and the termination condition that the density of each kernel point can reach, and outputs a point set of each kernel point; and taking other core points in the point set of each core point as core points adjacent to each core point to obtain parking space angle points adjacent to each parking space angle point.
In one embodiment, the parking lot map comprises a set of handicapped parking lots; each parking space detection point comprises a plurality of parking space detection points which form the incomplete parking space group; each parking space angular point comprises a plurality of parking space angular points forming the incomplete parking space group. The parking space determining module 803 is further configured to determine, based on each parking space angular point and the respective adjacent parking space angular point of each parking space angular point, a convex hull formed by each parking space angular point and the respective adjacent parking space angular point of each parking space angular point; determining parking space angular points in a convex hull and straight lines where the parking space angular points adjacent to the parking space angular points in the convex hull are located; the parking space angular points in the convex hull are parking space detection points which are positioned in the convex hull and form the incomplete parking space group; taking the intersection point of the straight line and the convex hull as a parking space auxiliary point; determining a finished vehicle position group corresponding to the incomplete vehicle position group in the vehicle bitmap based on the parking space auxiliary points and available parking space angular points; the available parking space angular points are parking space angular points in the plurality of parking space detection points forming the incomplete parking space group except the parking space angular points in the convex hull.
In an embodiment, the parking space determining module 803 is further configured to determine a first target parking space angle point in the available parking space angle points; the first target parking space angular point is a parking space angular point adjacent to the parking space angular point in the convex hull; taking the first target parking space angular point as a parking space angular point adjacent to the parking space auxiliary point; determining a second target parking space angular point in the available parking space angular points; the second target parking space angular point is a parking space angular point with the number of adjacent parking space angular points in the available parking space angular points being only one; taking the second target parking space angular point as a parking space angular point adjacent to the parking space auxiliary point; and obtaining a finished vehicle position group corresponding to the incomplete vehicle position group in the vehicle position diagram based on the adjacent vehicle position angular points of the vehicle position auxiliary points and the available vehicle position angular points.
In one embodiment, the defective parking space group at least comprises two defective parking spaces sharing the same defective part; and/or the incomplete parking space group at least comprises one incomplete parking space and a complete parking space adjacent to the incomplete parking space.
In one embodiment, the parking space detection point acquisition module 801 is further configured to acquire a surrounding view of the periphery of the vehicle; intercepting an image of a parking space detection effective area in the ring view to obtain the parking space map; inputting the parking space map into a pre-constructed neural network so that the neural network performs parking space detection on the parking space map and outputs the detection points of the parking spaces; and acquiring the detection points of the parking spaces output by the neural network.
For specific limitations of the parking space recognition device, reference may be made to the above limitations of the parking space recognition method, which is not described herein again. All modules in the parking space identification device can be completely or partially realized through software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent of a processor in the vehicle-mounted terminal, and can also be stored in a memory in the vehicle-mounted terminal in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a vehicle-mounted terminal is provided, and the internal structure thereof may be as shown in fig. 9. The vehicle-mounted terminal comprises a processor, a memory and a network interface which are connected through a system bus. Wherein, the processor of the vehicle-mounted terminal is used for providing calculation and control capability. The memory of the vehicle-mounted terminal comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the vehicle-mounted terminal is used for storing parking space identification data. The network interface of the vehicle-mounted terminal is used for connecting and communicating with an external terminal through a network. The computer program is executed by a processor to implement a parking space recognition method.
Those skilled in the art will appreciate that the structure shown in fig. 9 is only a block diagram of a part of the structure related to the present application, and does not constitute a limitation to the in-vehicle terminal to which the present application is applied, and a specific in-vehicle terminal may include more or less components than those shown in the figure, or combine some components, or have a different arrangement of components.
In one embodiment, a vehicle-mounted terminal is provided, which includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method embodiments when executing the computer program.
In an embodiment, a computer-readable storage medium is provided, on which a computer program is stored, which computer program, when being executed by a processor, carries out the steps of the respective method embodiment as described above.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above examples only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A parking space identification method is characterized by comprising the following steps:
acquiring detection points of all parking spaces obtained by carrying out parking space detection on the parking space map;
taking each parking space angular point in each parking space detection point as a check point of density clustering, and determining respective adjacent check points of each check point based on density to obtain respective adjacent parking space angular points of each parking space angular point;
and determining the parking spaces in the parking space map based on the parking space angular points and the parking space angular points adjacent to the parking space angular points.
2. The method of claim 1, wherein the reachable termination condition for the density of each core point is finding a first core point.
3. The method of claim 2,
the determining of the respective adjacent core points of each core point based on the density can be achieved to obtain the respective adjacent parking space angle points of each parking space angle point, and the method comprises the following steps:
acquiring a pre-constructed density clustering model; the density clustering model selects the corner points of each parking space as the core points in the process of performing density clustering on the detection points of each parking space, and finds the first core point according to the termination condition that the density of each core point can reach;
inputting the parking space detection points into the density clustering model, so that the density clustering model performs density clustering on the parking space detection points according to the kernel point selection condition and the termination condition that the density of each kernel point can reach, and outputs a point set of each kernel point;
and taking other core points in the point set of each core point as core points adjacent to each core point to obtain parking space angle points adjacent to each parking space angle point.
4. The method of claim 1, wherein the parking map comprises a set of handicapped parking spots; each parking space detection point comprises a plurality of parking space detection points which form the incomplete parking space group; each parking space angular point comprises a plurality of parking space angular points forming the incomplete parking space group; determining the parking spaces in the parking space map based on the parking space angular points and the respective adjacent parking space angular points of the parking space angular points, including:
determining each parking space angular point and a convex hull formed by the adjacent parking space angular points of the parking space angular points based on the parking space angular points and the adjacent parking space angular points of the parking space angular points;
determining parking space angular points in a convex hull and straight lines where the parking space angular points adjacent to the parking space angular points in the convex hull are located; the parking space angular points in the convex hull are parking space detection points which are positioned in the convex hull and form the incomplete parking space group;
taking the intersection point of the straight line and the convex hull as a parking space auxiliary point;
determining a finished vehicle position group corresponding to the incomplete vehicle position group in the vehicle bitmap based on the parking space auxiliary points and available parking space angular points; the available parking space angular points are parking space angular points in the plurality of parking space detection points forming the incomplete parking space group except the parking space angular points in the convex hull.
5. The method of claim 4, wherein determining a complete vehicle position group corresponding to the incomplete vehicle position group in the vehicle position map based on the parking space auxiliary point and an available parking space angular point comprises:
determining a first target parking space angular point in the available parking space angular points; the first target parking space angular point is a parking space angular point adjacent to the parking space angular point in the convex hull;
taking the first target parking space angular point as a parking space angular point adjacent to the parking space auxiliary point;
determining a second target parking space angular point in the available parking space angular points; the second target parking space angular point is a parking space angular point with the number of adjacent parking space angular points in the available parking space angular points being only one;
taking the second target parking space angular point as a parking space angular point adjacent to the parking space auxiliary point;
and obtaining a finished vehicle position group corresponding to the incomplete vehicle position group in the vehicle position diagram based on the adjacent vehicle position angular points of the vehicle position auxiliary points and the available vehicle position angular points.
6. The method of claim 4, wherein the set of stub slots includes at least two stub slots sharing the same stub portion; and/or the incomplete parking space group at least comprises one incomplete parking space and a complete parking space adjacent to the incomplete parking space.
7. The method according to any one of claims 1 to 6, wherein the acquiring the detection points of the parking spaces obtained by detecting the parking spaces of the parking space map comprises:
acquiring a ring view of the periphery of a vehicle;
intercepting an image of a parking space detection effective area in the ring view to obtain the parking space map;
inputting the parking space map into a pre-constructed neural network so that the neural network performs parking space detection on the parking space map and outputs the detection points of the parking spaces;
and acquiring the detection points of the parking spaces output by the neural network.
8. The utility model provides a parking stall recognition device which characterized in that, the device includes:
the parking space detection point acquisition module is used for acquiring each parking space detection point obtained by carrying out parking space detection on the parking space map;
the density clustering module is used for taking each parking space angular point in each parking space detection point as a check point of density clustering, determining respective adjacent check points of each check point based on density, and obtaining respective adjacent parking space angular points of each parking space angular point;
and the parking space determining module is used for determining the parking spaces in the parking space map based on the parking space angular points and the parking space angular points adjacent to the parking space angular points.
9. An in-vehicle terminal comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the method of any of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110021253.7A CN112348817B (en) | 2021-01-08 | 2021-01-08 | Parking space identification method and device, vehicle-mounted terminal and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110021253.7A CN112348817B (en) | 2021-01-08 | 2021-01-08 | Parking space identification method and device, vehicle-mounted terminal and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112348817A true CN112348817A (en) | 2021-02-09 |
| CN112348817B CN112348817B (en) | 2021-05-11 |
Family
ID=74427949
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110021253.7A Active CN112348817B (en) | 2021-01-08 | 2021-01-08 | Parking space identification method and device, vehicle-mounted terminal and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112348817B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113627277A (en) * | 2021-07-20 | 2021-11-09 | 的卢技术有限公司 | Method and device for identifying parking space |
| CN113674199A (en) * | 2021-07-06 | 2021-11-19 | 浙江大华技术股份有限公司 | Parking space detection method, electronic device and storage medium |
| CN113822156A (en) * | 2021-08-13 | 2021-12-21 | 北京易航远智科技有限公司 | Parking space detection processing method and device, electronic equipment and storage medium |
| CN115083203A (en) * | 2022-08-19 | 2022-09-20 | 深圳云游四海信息科技有限公司 | Method and system for inspecting parking in road based on image recognition berth |
| CN115206130A (en) * | 2022-07-12 | 2022-10-18 | 合众新能源汽车有限公司 | Parking space detection method, system, terminal and storage medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103600707A (en) * | 2013-11-06 | 2014-02-26 | 同济大学 | Parking position detecting device and method of intelligent parking system |
| US20140114534A1 (en) * | 2012-10-19 | 2014-04-24 | GM Global Technology Operations LLC | Dynamic rearview mirror display features |
| CN108090435A (en) * | 2017-12-13 | 2018-05-29 | 深圳市航盛电子股份有限公司 | One kind can parking area recognition methods, system and medium |
| CN109067725A (en) * | 2018-07-24 | 2018-12-21 | 成都亚信网络安全产业技术研究院有限公司 | Network flow abnormal detecting method and device |
| CN109685000A (en) * | 2018-12-21 | 2019-04-26 | 广州小鹏汽车科技有限公司 | A kind of method for detecting parking stalls and device of view-based access control model |
| CN109712427A (en) * | 2019-01-03 | 2019-05-03 | 广州小鹏汽车科技有限公司 | A kind of method for detecting parking stalls and device |
| CN110390306A (en) * | 2019-07-25 | 2019-10-29 | 湖州宏威新能源汽车有限公司 | Detection method, vehicle and the computer readable storage medium of right angle parking stall |
| US20200294310A1 (en) * | 2019-03-16 | 2020-09-17 | Nvidia Corporation | Object Detection Using Skewed Polygons Suitable For Parking Space Detection |
| CN112052782A (en) * | 2020-08-31 | 2020-12-08 | 安徽江淮汽车集团股份有限公司 | Around-looking-based parking space identification method, device, equipment and storage medium |
-
2021
- 2021-01-08 CN CN202110021253.7A patent/CN112348817B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140114534A1 (en) * | 2012-10-19 | 2014-04-24 | GM Global Technology Operations LLC | Dynamic rearview mirror display features |
| CN103600707A (en) * | 2013-11-06 | 2014-02-26 | 同济大学 | Parking position detecting device and method of intelligent parking system |
| CN108090435A (en) * | 2017-12-13 | 2018-05-29 | 深圳市航盛电子股份有限公司 | One kind can parking area recognition methods, system and medium |
| CN109067725A (en) * | 2018-07-24 | 2018-12-21 | 成都亚信网络安全产业技术研究院有限公司 | Network flow abnormal detecting method and device |
| CN109685000A (en) * | 2018-12-21 | 2019-04-26 | 广州小鹏汽车科技有限公司 | A kind of method for detecting parking stalls and device of view-based access control model |
| CN109712427A (en) * | 2019-01-03 | 2019-05-03 | 广州小鹏汽车科技有限公司 | A kind of method for detecting parking stalls and device |
| US20200294310A1 (en) * | 2019-03-16 | 2020-09-17 | Nvidia Corporation | Object Detection Using Skewed Polygons Suitable For Parking Space Detection |
| CN110390306A (en) * | 2019-07-25 | 2019-10-29 | 湖州宏威新能源汽车有限公司 | Detection method, vehicle and the computer readable storage medium of right angle parking stall |
| CN112052782A (en) * | 2020-08-31 | 2020-12-08 | 安徽江淮汽车集团股份有限公司 | Around-looking-based parking space identification method, device, equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 何俏君等: "基于YOLOv2-Tiny的环视实时车位线识别算法", 《汽车电器》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113674199A (en) * | 2021-07-06 | 2021-11-19 | 浙江大华技术股份有限公司 | Parking space detection method, electronic device and storage medium |
| CN113674199B (en) * | 2021-07-06 | 2024-10-01 | 浙江大华技术股份有限公司 | Parking space detection method, electronic device and storage medium |
| CN113627277A (en) * | 2021-07-20 | 2021-11-09 | 的卢技术有限公司 | Method and device for identifying parking space |
| CN113822156A (en) * | 2021-08-13 | 2021-12-21 | 北京易航远智科技有限公司 | Parking space detection processing method and device, electronic equipment and storage medium |
| CN113822156B (en) * | 2021-08-13 | 2022-05-24 | 北京易航远智科技有限公司 | Parking space detection processing method and device, electronic equipment and storage medium |
| CN115206130A (en) * | 2022-07-12 | 2022-10-18 | 合众新能源汽车有限公司 | Parking space detection method, system, terminal and storage medium |
| CN115206130B (en) * | 2022-07-12 | 2023-07-18 | 合众新能源汽车股份有限公司 | Parking space detection method, system, terminal and storage medium |
| CN115083203A (en) * | 2022-08-19 | 2022-09-20 | 深圳云游四海信息科技有限公司 | Method and system for inspecting parking in road based on image recognition berth |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112348817B (en) | 2021-05-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112348817B (en) | Parking space identification method and device, vehicle-mounted terminal and storage medium | |
| CN113761976B (en) | Scene semantic analysis method based on global guidance selective context network | |
| CN110348297B (en) | Detection method, system, terminal and storage medium for identifying stereo garage | |
| EP3620964A1 (en) | Lane line processing method and device | |
| CN113409382A (en) | Method and device for measuring damaged area of vehicle | |
| CN110211200B (en) | Dental arch wire generating method and system based on neural network technology | |
| CN110991215B (en) | Lane line detection method and device, storage medium and electronic equipment | |
| WO2024197815A1 (en) | Engineering machinery mapping method and device, and readable storage medium | |
| JP2008020225A (en) | Self position estimation program, self position estimation method and self position estimation apparatus | |
| CN116449392B (en) | Map construction method, device, computer equipment and storage medium | |
| CN118115762A (en) | Binocular stereo matching model training method, device, equipment and storage medium | |
| CN117612128B (en) | Lane line generation method, device, computer equipment and storage medium | |
| KR20230009151A (en) | Method and apparatus for building learning data for learning of object recognition neural network for vehicles | |
| US20240070979A1 (en) | Method and apparatus for generating 3d spatial information | |
| CN112212851B (en) | Pose determination method and device, storage medium and mobile robot | |
| CN114926536B (en) | Semantic-based positioning and mapping method and system and intelligent robot | |
| CN116309628A (en) | Lane line recognition method and device, electronic equipment and computer readable storage medium | |
| CN114536326B (en) | Road sign data processing method, device and storage medium | |
| CN112433193B (en) | Multi-sensor-based mold position positioning method and system | |
| CN110686687B (en) | Method for constructing map by visual robot, robot and chip | |
| CN111060127B (en) | Vehicle starting point positioning method and device, computer equipment and storage medium | |
| CN112950621A (en) | Image processing method, apparatus, device and medium | |
| CN113878570A (en) | Wall-following path planning method and device and computer-readable storage medium | |
| CN117419690B (en) | Pose estimation method, device and medium of unmanned ship | |
| CN118518093B (en) | Laser SLAM method, equipment and medium based on multi-frame space occupancy rate |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address |
Address after: Floor 25, Block A, Zhongzhou Binhai Commercial Center Phase II, No. 9285, Binhe Boulevard, Shangsha Community, Shatou Street, Futian District, Shenzhen, Guangdong 518000 Patentee after: Shenzhen Youjia Innovation Technology Co.,Ltd. Address before: 518051 1101, west block, Skyworth semiconductor design building, 18 Gaoxin South 4th Road, Gaoxin community, Yuehai street, Nanshan District, Shenzhen City, Guangdong Province Patentee before: SHENZHEN MINIEYE INNOVATION TECHNOLOGY Co.,Ltd. |
|
| CP03 | Change of name, title or address |