CN118430333A

CN118430333A - Parking space state determining method, controller, vehicle, storage medium and program product

Info

Publication number: CN118430333A
Application number: CN202410808690.7A
Authority: CN
Inventors: 皇甫统帅; 许凌云; 胡德文; 李金涛; 连旭升
Original assignee: BYD Co Ltd
Current assignee: BYD Co Ltd
Priority date: 2024-06-21
Filing date: 2024-06-21
Publication date: 2024-08-02
Anticipated expiration: 2044-06-21
Also published as: CN118430333B

Abstract

The present disclosure relates to the field of vehicle technologies, and in particular, to a parking space state determining method, a controller, a vehicle, a storage medium, and a program product, where the method includes: according to an image acquired by a vehicle-mounted shooting device of a vehicle, determining grounding information of an obstacle in the image in an all-around image and parking space information of a parking space in the image in the all-around image; according to the grounding information and the parking space information, determining first target grounding information of the obstacle relative to the parking space direction; and determining the parking space state of the parking space according to the first target grounding information and the parking space information. Therefore, the automatic identification of the parking space state can be realized according to the grounding information and the parking space information. In addition, the target grounding information is the grounding information of the obstacle relative to the parking space direction, so that the occupation condition of the obstacle in the parking space can be accurately reflected by the target grounding information, and further, when the parking space state is determined according to the target grounding information, the recognition accuracy of the parking space state can be improved, and the parking accident is reduced.

Description

Parking space state determining method, controller, vehicle, storage medium and program product

Technical Field

The disclosure relates to the technical field of vehicles, in particular to a parking space state determining method, a controller, a vehicle, a storage medium and a program product.

Background

With the development of vehicle technology, the application of automatic parking functions has been developed. In the related art, when an automatic parking function is utilized to park, an ultrasonic sensor is generally utilized to transmit ultrasonic waves and receive signals reflected by the ultrasonic sensor so as to judge the parking space state of a parking space.

However, for various reasons, when the parking space state determination is performed by the ultrasonic sensor, there is a case where the parking space state determination is inaccurate.

Disclosure of Invention

The disclosure aims to provide a parking space state determining method, a controller, a vehicle, a storage medium and a program product, so as to solve the technical problems.

To achieve the above object, a first aspect of the present disclosure provides a parking space state determining method, including:

According to an image acquired by a vehicle-mounted shooting device of a vehicle, determining grounding information of an obstacle in the image in an all-around image and parking space information of a parking space in the image in the all-around image;

According to the grounding information and the parking space information, determining first target grounding information of the obstacle relative to the parking space direction;

And determining the parking space state of the parking space according to the first target grounding information and the parking space information, wherein the parking space state is used for representing whether the parking space is occupied by the obstacle or not.

Optionally, the parking space includes a plurality of, according to the grounding information and the parking space information, determining first target grounding information of the obstacle relative to the parking space direction includes:

Determining a target parking space in a plurality of parking spaces according to the grounding information and the parking space information;

according to the grounding information and the parking space information of the target parking space, determining first target grounding information of the obstacle relative to the parking space direction;

The determining the parking space state of the parking space according to the first target grounding information and the parking space information comprises the following steps:

And determining the parking space state of the target parking space according to the first target grounding information and the parking space information of the target parking space.

Optionally, the grounding information includes a grounding point of the obstacle, and determining, according to the grounding information and the parking space information, a target parking space among a plurality of parking spaces includes:

For each parking space, determining the number of grounding points in the parking space according to the grounding points of the obstacle and the parking space information;

And determining the parking space with the largest number of grounding points among the plurality of parking spaces as a target parking space.

Optionally, determining a target parking space in a plurality of parking spaces according to the grounding information and the parking space information includes:

fitting the grounding frame of the obstacle according to the grounding information to obtain a first fitted grounding frame of the obstacle;

for each parking space, determining the overlapping degree between the first fitting grounding frame and the parking space according to the parking space information of the parking space and the first fitting grounding frame;

and determining the parking space with the largest overlapping degree with the first fitting grounding frame among the plurality of parking spaces as a target parking space.

Optionally, the grounding information includes a plurality of grounding points of the obstacle and grounding coordinates of the grounding points, and the fitting is performed on the grounding frame of the obstacle according to the grounding information to obtain a first fitted grounding frame of the obstacle, including:

clustering the grounding points according to the grounding coordinates of the grounding points to obtain a plurality of first clusters;

And performing grounding frame fitting on each first cluster according to the grounding points in the first clusters and the grounding coordinates of the grounding points to obtain a first fitting grounding frame of the obstacle.

Optionally, the grounding information includes a grounding point of the obstacle and a grounding coordinate of the grounding point, and determining, according to the grounding information and the parking space information of the target parking space, first target grounding information of the obstacle relative to the parking space direction includes:

Determining a target grounding point which is farthest from each parking space edge in the grounding points according to the parking space information of the target parking space aiming at each parking space edge of the target parking space;

and determining first target grounding information of the obstacle relative to the parking space direction according to all the target grounding points.

Optionally, the determining, according to all the target grounding points, the first target grounding information of the obstacle relative to the parking space direction includes:

for each target grounding point, determining a straight line which passes through the target grounding point and is parallel to the parking space edge farthest from the target grounding point in the target parking spaces;

and determining first target grounding information of the obstacle relative to the parking space direction according to all the straight lines.

Optionally, the determining the first target grounding information of the obstacle relative to the parking space direction according to all the straight lines includes:

determining grounding frame information of the obstacle according to all the straight lines, and determining the grounding frame information as first target grounding information of the obstacle relative to the parking space direction; or alternatively

And determining intersection point information of all the straight lines, and determining the intersection point information as first target grounding information of the obstacle relative to the parking space direction.

Optionally, the determining, according to the first target grounding information and the parking space information, a parking space state of the parking space includes:

Determining the confidence that the parking space is occupied by the obstacle according to the first target grounding information and the parking space information;

And determining the parking space state of the parking space according to the confidence coefficient and a preset confidence coefficient threshold value.

Optionally, the first target grounding information includes a grounding frame and a grounding area of the obstacle relative to a parking space direction, the parking space information includes a parking space frame and a parking space area of the parking space, and determining, according to the first target grounding information and the parking space information, a confidence level of the parking space being occupied by the obstacle includes:

Determining the coverage areas of the grounding frame and the parking space frame;

Determining a first ratio of the coverage area to the parking space area and a second ratio of the coverage area to the grounding area, determining the first ratio as the confidence of the parking space being encroached by the obstacle when the first ratio is larger than the second ratio, and determining the second ratio as the confidence of the parking space being encroached by the obstacle when the second ratio is larger than the first ratio.

Optionally, the first target grounding information includes corner coordinates of a grounding frame of the obstacle relative to a parking space direction, the parking space information includes corner coordinates of the parking space, and determining, according to the first target grounding information and the parking space information, a parking space state of the parking space includes:

and inputting the corner coordinates of the grounding frame of the obstacle relative to the parking space direction and the corner coordinates of the parking space into a parking space state determining model to obtain the parking space state of the parking space, wherein the parking space state determining model is used for outputting the parking space state according to the input corner coordinates of the grounding frame and the input corner coordinates of the parking space.

Optionally, the first target grounding information includes information of a grounding frame of the obstacle relative to a parking space direction, and the parking space state determining method further includes:

Determining the number of the obstacles in the grounding frame according to the information of the grounding frame of the obstacle relative to the parking space direction and the parking space information of the target parking space;

and when the number of the barriers in the grounding frame is smaller than or equal to the preset number of the barriers, determining the parking space state of the target parking space according to the first target grounding information and the parking space information of the target parking space.

Optionally, the method further comprises:

When the number of the barriers in the grounding frame is larger than or equal to the preset number of the barriers, clustering the grounding points in the first fitting grounding frame according to the grounding coordinates of the grounding points in the first fitting grounding frame to obtain a plurality of second clusters, wherein the number of the second clusters is the same as the number of the barriers in the grounding frame;

For each second cluster, the following steps are performed:

Determining second target grounding information of the obstacle relative to the parking space direction according to the grounding point in the second cluster and the target parking space;

And determining the parking space state of the target parking space according to the second target grounding information and the parking space information of the target parking space.

Optionally, the determining the number of the obstacles in the grounding frame according to the information of the grounding frame of the obstacle relative to the parking space direction and the parking space information of the target parking space includes:

Determining a first width value of the target parking space according to the parking space information of the target parking space, and determining a second width value of the grounding frame according to the information of the grounding frame of the obstacle relative to the parking space direction;

And determining the number of the obstacles in the grounding frame according to the ratio between the first width value and the second width value and a preset ratio.

A second aspect of the present disclosure provides a controller comprising:

A processor;

A memory for storing processor-executable instructions;

wherein the processor is configured to: performing the steps of the method of any of the first aspects.

A third aspect of the present disclosure provides a vehicle comprising the controller of the second aspect, or

The vehicle includes a processor and a memory for storing processor-executable instructions, wherein the processor is configured to: performing the steps of the method of any of the first aspects.

A fourth aspect of the present disclosure provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method of any of the first aspects.

A fifth aspect of the present disclosure provides a computer program product comprising a computer program which, when executed by a processor, implements the steps of the method of any of the first aspects.

Through the technical scheme, the target grounding information of the obstacle relative to the parking space direction can be determined based on the grounding information of the obstacle and the parking space information of the parking space, and then the parking space state of the parking space is determined according to the target grounding information and the parking space information. Therefore, the automatic identification of the parking space state can be realized according to the grounding information and the parking space information. In addition, the target grounding information is the grounding information of the obstacle relative to the parking space direction, so that the occupation condition of the obstacle in the parking space can be accurately reflected by the target grounding information, and further, when the parking space state is determined according to the target grounding information, the recognition accuracy of the parking space state can be improved, and the parking accident is reduced.

Additional features and advantages of the present disclosure will be set forth in the detailed description which follows.

Drawings

The accompanying drawings are included to provide a further understanding of the disclosure, and are incorporated in and constitute a part of this specification, illustrate the disclosure and together with the description serve to explain, but do not limit the disclosure. In the drawings:

FIG. 1 is a flow chart of a method for detecting a parking space in the related art;

FIG. 2 is a flow chart of a parking space detection in the related art;

FIG. 3 is a flow chart of a parking space and obstacle detection method based on deep learning in the related art;

FIG. 4 is a schematic view of an obstacle fitting locating box in the related art;

FIG. 5 is a flow chart illustrating a method of determining a parking spot status according to an exemplary embodiment of the present disclosure;

FIG. 6 is a schematic diagram showing two fisheye images when they are stitched at an overlap, showing the disappearance of the cone;

FIG. 7 is a schematic illustration of a vehicle occluding an adjacent free space in accordance with an exemplary embodiment of the present disclosure;

FIG. 8 is a schematic illustration of a pillar shielding adjacent free spaces according to an exemplary embodiment of the present disclosure;

FIG. 9 is a schematic diagram of ground information identified by a ground point identification model, according to an exemplary embodiment of the present disclosure;

FIG. 10 is a schematic diagram illustrating a conversion relationship of ground information between an image and a look-around image according to an exemplary embodiment of the present disclosure;

FIG. 11 is a diagram illustrating ground contact information of an obstacle in a look-around image according to an exemplary embodiment of the present disclosure;

FIG. 12 is a schematic diagram illustrating a composition of a look-around image from images according to an exemplary embodiment of the present disclosure;

FIG. 13 is a schematic view showing an identification effect of parking space information according to an exemplary embodiment of the present disclosure;

FIG. 14 is a schematic view of the relative positions of 4 corner points of a parking spot, according to an exemplary embodiment of the present disclosure;

FIG. 15 is a first target ground information schematic diagram of an obstacle relative to a parking spot direction, according to an exemplary embodiment of the present disclosure;

FIG. 16 is a diagram illustrating a parking spot status recognition effect according to an exemplary embodiment of the present disclosure;

FIG. 17 is a flowchart illustrating another method of determining a parking spot status according to an exemplary embodiment of the present disclosure;

FIG. 18 is a block diagram illustrating a parking spot status determination device according to an exemplary embodiment of the present disclosure;

Fig. 19 is a functional block diagram schematic of a vehicle according to an exemplary embodiment of the present disclosure.

Detailed Description

Specific embodiments of the present disclosure are described in detail below with reference to the accompanying drawings. It should be understood that the detailed description and specific examples, while indicating and illustrating the disclosure, are not intended to limit the disclosure.

It should be understood that the various steps recited in the method embodiments of the present disclosure may be performed in a different order and/or performed in parallel. Furthermore, method embodiments may include additional steps and/or omit performing the illustrated steps. The scope of the present disclosure is not limited in this respect.

As background art advances, automatic parking function applications have evolved with the development of vehicle technology. In the related art, when an automatic parking function is utilized to park, an ultrasonic sensor is generally utilized to transmit ultrasonic waves and receive signals reflected by the ultrasonic sensor so as to judge the parking space state of a parking space.

For example, patent application 201811559894.2 discloses a method for detecting a parking space, as shown in fig. 1. The method comprises the steps of detecting according to a specified frequency by using an ultrasonic sensor of a vehicle to obtain the distance between the vehicle and an obstacle, determining a plurality of position coordinates of the obstacle according to the position coordinates of the ultrasonic sensor of the vehicle and the distance between the vehicle and the obstacle during multiple detection, fitting the plurality of position coordinates to obtain the boundary of the obstacle, detecting the jump edge of the boundary of the obstacle, adjusting the length of the jump edge according to the calibration relation between measured position data measured by the ultrasonic sensor and position data of the actual obstacle, and finally determining the position of a parking place according to the jump edge after the adjustment of the length.

Then, since the ultrasonic sensor lacks semantic information when judging the parking space state, there is a case where the parking space state judgment is inaccurate. For example, when parking space state determination is performed based on an ultrasonic sensor due to lack of semantic information, there is a problem that the state of a ground lock (ground lock open or ground lock close) in a parking space cannot be distinguished, and thus a parking space (ground lock close) that is actually parked is determined as a parking space in which parking is impossible.

In order to overcome the technical problems, the related art proposes a method for identifying and judging the parking space state based on visual perception.

For example, patent number 202011128953.8 discloses a parking space detection method, as shown in fig. 2, by means of looking around images of a panoramic monitoring image system (Around View Monitor, AVM) to introduce visual semantic information, the problem that the state of a ground lock in a parking space cannot be distinguished is solved.

For example, the invention patent with application number 202211248240.4 discloses a parking space and obstacle detection method based on deep learning, as shown in fig. 3, rich semantic information is introduced through looking around an image to obtain rich information such as a parking space position and an obstacle position at the same time, so that states such as whether the parking space is occupied or not are obtained.

However, although abundant semantic information can be introduced in a manner based on visual perception, when the parking space state of the parking space is determined in a manner based on visual perception, the positioning frame of the obstacle is generally fitted based on the grounding information of the obstacle, so that the fitting positioning frame of the obstacle is obtained, and then the parking space state is determined based on the fitting positioning frame of the obstacle. However, when the positioning frame of the obstacle is fitted, the problem that the fitted positioning frame of the obstacle encroaches on the adjacent parking space exists in the related art, so that judgment of the state of the adjacent parking space is interfered, and the recognition result of the state of the parking space is inaccurate, as shown in fig. 4.

In view of the above, the present disclosure provides a parking space state determining method, a controller, a vehicle, a storage medium and a program product, so as to solve the above technical problems.

Embodiments of the present disclosure are further explained below with reference to the drawings.

Fig. 5 is a flowchart illustrating a parking space state determining method according to an exemplary embodiment of the present disclosure, and referring to fig. 5, the parking space state determining method may include the steps of:

s501: according to an image acquired by a vehicle-mounted shooting device of a vehicle, determining grounding information of an obstacle in the image in an all-around image and parking space information of a parking space in the image in the all-around image;

s502: according to the grounding information and the parking space information, determining first target grounding information of the obstacle relative to the parking space direction;

S503: and determining the parking space state of the parking space according to the first target grounding information and the parking space information, wherein the parking space state is used for representing whether the parking space is occupied by an obstacle or not.

The following describes the parking space state determining method in the present disclosure in detail with reference to each step.

In step S501, the image acquired by the vehicle-mounted photographing device may be one or a plurality of images, which is not limited in any way in the embodiment of the present disclosure. In order to more accurately determine the grounding information and the parking space information, the images acquired by the vehicle-mounted shooting device can be multiple, and the image content corresponding to each image is different. For example, a plurality of images having different image contents may be acquired by a fisheye camera disposed in a direction of front, rear, left front, left rear, right front, right rear, or the like of the vehicle.

After the images are acquired, the grounding information of the obstacle in the looking-around image and the parking space information of the parking space in the looking-around image can be determined according to the images.

It should be understood that, in the related art, when the grounding information of the obstacle in the looking-around image is acquired, the image acquired by the vehicle-mounted photographing device is typically spliced into the looking-around image, and then the grounding information of the obstacle in the looking-around image is obtained by identifying the grounding information of the obstacle in the looking-around image.

However, since the looking-around image is obtained by subjecting a plurality of images to distortion correction, projection transformation, image stitching and the like, human errors or systematic errors (such as blurring or distortion generated in the image stitching process) are inevitably introduced, thereby affecting the recognition accuracy. As shown in fig. 6, the problem of the disappearance of the obstacle (cone) occurs when two fisheye images are spliced at the overlapping portion. In addition, after the stereo obstacle (such as a pillar, a vehicle, a pedestrian and the like) in the fisheye image is mapped to the looking-around image through the steps of distortion correction, projection transformation and the like, obvious deformation can occur. For example, in the looking-around image, the deformed three-dimensional obstacles such as pillars, vehicles, pedestrians and the like will block the adjacent parking space, so that the state of the adjacent parking space cannot be correctly identified. As shown in fig. 7 and 8, the vehicles and the pillars block the adjacent idle parking spaces, so that the corresponding idle parking spaces cannot be released only according to the looking-around image, and the subsequent parking process is affected.

Therefore, in order to acquire accurate grounding information, the grounding signal of the obstacle can be determined according to the image acquired by the vehicle-mounted shooting device, and then the grounding signal of the obstacle is mapped into the looking-around image, so that the grounding information of the obstacle in the looking-around image is acquired.

For example, for each acquired image, the image may be input into a ground point recognition model to recognize ground information of each obstacle in the image by the ground point recognition model, so as to obtain ground information of each obstacle in the image, as shown in fig. 9. And then, based on the mapping relation between the looking-around images and the images, the grounding information of each obstacle in the images is converted into the looking-around images, so that the grounding information of each obstacle in the looking-around images is obtained, as shown in fig. 10 or 11. The grounding information may be set according to practical situations, and the embodiment of the disclosure does not limit the grounding information. In a possible manner, the ground information may include a ground point, a ground coordinate of the ground point, and/or a ground category, etc., wherein the ground category may be used to characterize a category of an obstacle, e.g., a ground lock, a cone, a vehicle, a pedestrian, etc.

The ground point recognition model may be an existing model in the related art, or may be a model improved from the existing model in the related art, which is not limited in any way in the embodiment of the present disclosure.

Illustratively, the YOLO (You Only Look Once, you only see once) model may be trained to arrive at the ground point identification model by:

Acquiring a sample image and a tag, wherein the tag is used for indicating sample grounding information of the sample image; inputting the sample image into a YOLO model to obtain the predicted grounding information of the sample image; and determining a loss function value according to the predicted grounding information and the tag, and updating parameters of the YOLO model according to the loss function value until the preset iteration times are reached or the accuracy of the YOLO model reaches a preset value.

After the ground point recognition model is trained by the method, the ground information can be recognized based on the input image.

When determining the parking space information of the parking space in the looking-around image, the method can firstly perform distortion removal correction and perspective transformation on each image and then splice the images into the looking-around image, and then optimize the looking-around image in an overlapping region eclosion mode to obtain an optimized looking-around image, as shown in fig. 12. And then inputting the optimized looking-around image into a parking space detection model to obtain parking space information in the looking-around image. Or inputting the image into a parking space detection model for each acquired image to obtain parking space information of the parking space in the image, and then converting the parking space information in the image into the looking-around image based on the mapping relation between the looking-around image and the image, so as to obtain the parking space information in the looking-around image. The parking space information can be set according to actual conditions, and the embodiment of the disclosure does not limit the parking space information in any way, and the parking space information can comprise corner points of the parking space, corner point coordinates of the corner points and the like.

The parking space detection model may be an existing model in the related art, or may be a model improved from the existing model in the related art, which is not limited in any way by the embodiment of the present disclosure.

Illustratively, the YOLO (You Only Look Once, you see only once) model may be trained to get the parking space detection model by:

Acquiring a sample looking-around image and a label, wherein the label is used for indicating parking space information of a parking space in the sample looking-around image; inputting the sample looking-around image into a YOLO model to obtain predicted parking space information of the parking spaces in the sample looking-around image; and determining a loss function value according to the predicted parking space information and the label, and updating parameters of the YOLO model according to the loss function value until the preset iteration times are reached or the accuracy of the YOLO model reaches a preset value.

After the parking space detection model is trained by the method, the identification of the parking space information can be performed based on the input looking-around image. In a possible manner, the effect of identifying the parking space information may be as shown in fig. 13, where the relative positions of 4 corner points of the parking space may be as shown in fig. 14.

In step S502, the ground information may be ground information of one obstacle or ground information of a plurality of obstacles. Similarly, the parking space information may be information of one parking space or information of a plurality of parking spaces.

When the grounding information is the grounding information of an obstacle and the parking space information is the parking space information of a parking space, the first target grounding information of the obstacle relative to the parking space direction can be determined directly based on the grounding information and the parking space information.

When the grounding information is the grounding information of one obstacle and the parking space information is the parking space information of a plurality of parking spaces, the obstacle is generally only present in one of the parking spaces, so that when the first target grounding information of the obstacle relative to the parking space direction is determined, the target parking space where the obstacle is located can be determined first, and then the first target grounding information of the obstacle relative to the parking space direction is determined based on the grounding information and the parking space information of the target parking space. That is, in a possible manner, the parking space may include a plurality of parking spaces, and accordingly, determining the first target ground information of the obstacle with respect to the parking space direction according to the ground information and the parking space information may include: determining a target parking space in the plurality of parking spaces according to the grounding information and the parking space information; determining first target grounding information of the obstacle relative to the parking space direction according to the grounding information and the parking space information of the target parking space; accordingly, in step S503, determining the parking space state of the parking space according to the first target grounding information and the parking space information may include: and determining the parking space state of the target parking space according to the first target grounding information and the parking space information of the target parking space.

The method comprises the steps of determining the number of grounding points contained in each parking space, and determining the parking space with the largest number of grounding points as the target parking space. That is, in a possible manner, the grounding information may include a grounding point of the obstacle, and accordingly, determining the target parking space among the plurality of parking spaces according to the grounding information and the parking space information may include: determining the number of grounding points in each parking space according to the grounding points of the obstacle and the parking space information; and determining the parking space with the largest number of grounding points among the plurality of parking spaces as a target parking space.

The method comprises the steps of firstly carrying out ground frame fitting of the obstacle based on ground information to obtain a first fitted ground frame, then determining the overlapping degree of the first fitted ground frame and each parking space, and determining the parking space corresponding to the maximum overlapping degree as a target parking space. That is, in a possible manner, determining the first target ground information of the obstacle with respect to the parking space direction according to the ground information and the parking space information may include: fitting the grounding frame of the obstacle according to the grounding information to obtain a first fitted grounding frame of the obstacle; for each parking space, determining the overlapping degree between the first fitting grounding frame and the parking space according to the parking space information of the parking space and the first fitting grounding frame; and determining the parking space with the largest overlapping degree with the first fitting grounding frame among the plurality of parking spaces as a target parking space.

When the grounding information is the grounding information of a plurality of obstacles and the parking space information is the parking space information of one parking space, the target obstacle positioned in the parking space can be determined first, and then the first target grounding information of the target obstacle relative to the parking space direction is determined based on the grounding information of the target obstacle and the parking space information. Accordingly, in step S503, determining the parking space state of the parking space according to the first target grounding information and the parking space information may include: and determining the parking space state according to the first target grounding information and the parking space information of the target obstacle.

The method for determining the target obstacle in the parking space can refer to the method for determining the target parking space in which the obstacle is located. The ground information includes the ground information of a plurality of obstacles, so that the ground information in the looking-around image is inconvenient to distinguish the obstacles, and the target obstacle in the parking space cannot be accurately determined. In a possible manner, the grounding points of the obstacles can be clustered according to the grounding coordinates of the obstacles, so that the grounding points of the same obstacle are clustered, and the target obstacle can be determined according to the grounding points, the grounding coordinates and the parking space information in a plurality of clusters. For example, for each cluster, a ground frame fitting may be performed according to the ground points and the ground coordinates in the cluster, so as to obtain a first fitted ground frame of the obstacle, and determine the overlapping degree between the first fitted ground frame and the parking space, and then determine the target obstacle according to the overlapping degree between all the first fitted ground frames and the parking space. That is, in a possible manner, the grounding information may include grounding points of a plurality of obstacles and grounding coordinates of the grounding points, and accordingly, fitting the grounding frame of the obstacle according to the grounding information to obtain a first fitted grounding frame of the obstacle may include: clustering the grounding points according to the grounding coordinates of the grounding points to obtain a plurality of first clusters; and fitting the grounding frame according to the grounding points in the first clusters and the grounding coordinates of the grounding points aiming at each first cluster to obtain a first fitting grounding frame of the obstacle.

The first cluster may be obtained by clustering the ground points according to ground coordinates by a clustering algorithm, for example by a k-means clustering algorithm, for example. The first fitted ground box in each cluster is then determined for that cluster by the function cv: MINAREARECT ().

When the ground information is ground information of a plurality of obstacles and the parking space information is parking space information of a plurality of parking spaces, the following steps may be performed for each obstacle: the method comprises the steps of determining a target parking space where an obstacle is located, and then determining first target grounding information of the obstacle relative to the parking space direction of the target parking space based on grounding information of the obstacle and the parking space information of the target parking space. Accordingly, in step S503, determining the parking space state of the parking space according to the first target grounding information and the parking space information may include: and aiming at each obstacle, determining the parking space state of the target parking space according to the first target grounding information of the obstacle and the parking space information of the target parking space.

The specific implementation in this scenario may refer to the foregoing related description, and will not be repeated here.

In a possible manner, determining the first target ground information of the obstacle relative to the parking space direction according to the ground information and the parking space information of the target parking space may include:

Aiming at each parking space edge of the target parking space, determining a target grounding point farthest from the parking space edge in the first fitting grounding frame according to the parking space information of the target parking space; and determining first target grounding information of the obstacle relative to the parking space direction according to all the target grounding points.

For example, the upper left point of the looking-around image can be taken as the origin of coordinates, the right direction is the positive direction of the x axis, the downward direction is the positive direction of the y axis, and a rectangular coordinate system is established according to the angular points (marked as the angular point P1, the angular point P2, the angular point P3 and the angular point P4) and the angular point coordinates (marked as P1 #，)，P2(，)，P3(，) P4%，) Determining the linear equation of 4 parking space edges of the target parking space、、AndThe distance between each grounding point in the first fitting grounding frame and the linear equation can be determinedEquation of straight line、Equation of straight lineThereby obtaining the distance linear equation in the first fitting grounding frameEquation of straight line、Equation of straight lineAnd finally, according to the furthest target grounding points P5, P6, P7 and P8, determining first target grounding information of the obstacle relative to the parking space direction.

The distance linear equation in the first fitted ground frame is determined as followsThe furthest target ground point P5 is illustrated as an example:

By setting straight line equations The expression of (2) is: . Since the coordinates of the corner point P1 and the coordinates of the corner point P2 are known, the slope of the straight line equation passing through the corner point P1 and the corner point P2 Can be expressed as:

Thereby according to the slope 、P1(，) And P2%，) The determined linear equation can be expressed as:

And then can obtain:

And due to the straight line equation The expression of (2) is: This can give:

combining straight line equations

The method can obtain the following steps:

Thus, it can be based on 、、Determining the equation of a straight lineIs a specific expression of (2). Thereafter, the respective ground point-to-straight line equation in the first fitted ground frame can be determined byDistance d of (2):

Wherein, the method comprises the following steps of ，) Representing the first fitting ground frameThe ground coordinates of the individual ground points.

Thereby, each grounding point and straight line in the first fitting grounding frame can be obtainedFurther, the distance between the first fitting grounding frame and the straight line can be obtained through distance comparisonThe target ground point P5 furthest from.

Based on the same way, the distance straight line in the first fitting grounding frame can be obtainedThe target grounding point P6 furthest from the straight lineTarget grounding point P7 farthest from first fitting grounding frame and straight lineThe target ground point P8 furthest from.

After the target grounding point P5, the target grounding point P6, the target grounding point P7 and the target grounding point P8 are obtained, the first target grounding information of the obstacle relative to the parking space direction can be determined.

In a possible manner, determining the first target grounding information of the obstacle relative to the parking space direction according to all the target grounding points may include:

for each target grounding point, determining a straight line which passes through the target grounding point and is parallel to the parking space edge which is farthest from the target grounding point in the target parking spaces; and determining first target grounding information of the obstacle relative to the parking space direction according to all the straight lines.

Illustratively, to determine the target ground point P5 and to match the linear equationThe parallel straight lines are illustrated as examples:

P5 is set up at the target grounding point ，) And with the straight line equationStraight line equations parallel to each otherThe method comprises the following steps: Then:

based on the same manner as above, the target grounding point P6 can be obtained and the equation of straight line Straight line equations parallel to each otherPassing through the target grounding point P7 and connecting with the linear equationStraight line equations parallel to each otherPassing through the target grounding point P8 and being in line with the equation of straight lineStraight line equations parallel to each other. Then, the first target grounding information of the obstacle relative to the parking space direction can be determined according to the four straight lines.

In a possible manner, determining the first target grounding information of the obstacle relative to the parking space direction according to all the straight lines may include:

Determining grounding frame information of the obstacle according to all the straight lines, and determining the grounding frame information as first target grounding information of the obstacle relative to the parking space direction; or determining intersection information of all straight lines, and determining the intersection information as first target grounding information of the obstacle relative to the parking space direction.

The ground frame information may include ground frame coordinates and/or ground frame area, and the intersection information may include intersection points and intersection point coordinates.

Illustratively, in solving the linear equationEquation of lineEquation of straight lineEquation of straight lineAnd then, the intersection point and the intersection point coordinates among the four direct lines are solved, so that intersection point information is obtained.

Illustratively, in solving the linear equationEquation of lineEquation of straight lineEquation of straight lineThen, by solving the intersection coordinates among the four direct points, the grounding frame coordinates and/or the grounding frame area and the like can be determined based on the intersection coordinates, and then the grounding frame information of the obstacle, namely the first target grounding information of the obstacle relative to the parking space direction, is obtained, as shown in fig. 15, wherein the obstacle fitting positioning frame, namely the grounding frame in fig. 15.

In step S503, determining the parking space state of the parking space according to the first target grounding information and the parking space information may include:

Determining the confidence that the parking space is occupied by the obstacle according to the first target grounding information and the parking space information; and determining the parking space state of the parking space according to the confidence coefficient and a preset confidence coefficient threshold value.

The preset confidence threshold may be determined according to practical situations, which is not limited in the embodiments of the present disclosure. For example, the preset confidence threshold may be set to 5, thereby determining that the parking space state of the parking space is an occupied state, that is, the parking space is occupied by the obstacle, when the confidence is greater than 5; and when the confidence coefficient is less than or equal to 5, determining that the parking space state of the parking space is an unoccupied state, namely that the parking space is unoccupied by the obstacle.

In a possible manner, the first target grounding information includes a grounding frame and a grounding area of the obstacle relative to the parking space direction, the parking space information includes a parking space frame and a parking space area of the parking space, and accordingly, determining, according to the first target grounding information and the parking space information, a confidence that the parking space is occupied by the obstacle may include:

Determining a coverage area between the grounding frame and the parking space; and determining a first ratio of the coverage area to the parking space area and a second ratio of the coverage area to the grounding area, determining the first ratio as the confidence that the parking space is occupied by the obstacle when the first ratio is larger than the second ratio, and determining the second ratio as the confidence that the parking space is occupied by the obstacle when the second ratio is larger than the first ratio.

For example, the ground area and parking space area may be brought into the following equation, resulting in a confidence level:

,)

wherein, The degree of confidence is indicated and,Indicating the ground area of the ground frame,The parking space area is represented,Representing the coverage area.

In a possible manner, the first target grounding information includes a grounding frame of the obstacle relative to the parking space direction, the parking space information includes corner coordinates of the parking space, and accordingly, determining the parking space state of the parking space according to the first target grounding information and the parking space information may include:

And inputting the grounding frame and the corner coordinates of the obstacle relative to the parking space direction into a parking space state determining model to obtain the parking space state of the parking space, wherein the parking space state determining model is used for outputting the parking space state according to the input grounding frame and corner coordinates of the obstacle relative to the parking space direction.

The parking space state determining model may be an existing model in the related art, or may be a model improved from the existing model in the related art, which is not limited in any way in the embodiment of the present disclosure.

Illustratively, the CNN (Convolutional Neural Network ) model may be trained to obtain the parking space state determination model by:

Acquiring sample corner coordinates, a sample grounding frame and a tag, wherein the tag is used for indicating the parking space state of a parking space corresponding to the sample corner coordinates; inputting the sample angular point coordinates and the sample grounding frame into a CNN model to obtain a predicted parking space state; and determining a loss function value according to the predicted parking space state and the label, and updating parameters of the CNN model according to the loss function value until the preset iteration times are reached or the precision of the CNN model reaches a preset value.

After the parking space state determination model is trained by the method, the parking space state can be identified based on the input first target grounding information and the parking space information, so that the parking space state is output. In a possible manner, the output parking space status may be as shown in fig. 16.

In a possible manner, the first target grounding information includes information of a grounding frame of the obstacle with respect to the parking space direction, and the parking space state determining method may further include:

Determining the number of the obstacles in the grounding frame according to the information of the grounding frame in the direction of the obstacle relative to the parking space and the parking space information of the target parking space;

Correspondingly, determining the parking space state of the parking space according to the first target grounding information and the parking space information may include:

When the number of the barriers in the grounding frame is smaller than or equal to the preset number of the barriers, the parking space state of the target parking space is determined according to the first target grounding information and the parking space information of the target parking space.

The number of the preset obstacles may be determined according to practical situations, and the embodiment of the present disclosure does not limit this. In a possible manner, the preset number of obstacles is set to 1.

It should be appreciated that, since the ground frame of the obstacle relative to the parking spot direction is determined by the ground information of the obstacle in the look-around image, when the distance between adjacent obstacles is closer, the ground points between adjacent obstacles are closer. Therefore, when the grounding frame is fitted according to the grounding information, the grounding information of a part of adjacent obstacles can be used as the grounding information of the current obstacle, so that the grounding frame fitted based on the grounding information is larger than the real grounding frame, and the parking space state is inaccurately and wrongly identified, so that a safety accident is caused. Therefore, before the parking space state of the target parking space is determined based on the first target grounding information and the parking space information of the target parking space, the number of the obstacles in the grounding frame is judged to avoid the situation, and therefore the recognition accuracy of the parking space state is improved.

In a possible manner, the method further comprises:

For each second cluster, the following steps are performed:

Performing grounding frame fitting according to the grounding points in the second clusters and the grounding coordinates of the grounding points to obtain second fitting grounding frames of the obstacles; determining second target grounding information of the obstacle relative to the parking space direction according to the second fitting grounding frame and the target parking space; and determining the parking space state of the parking space corresponding to the maximum overlapping degree according to the second target grounding information and the target parking space.

The implementation manner of the above steps is just described with reference to the foregoing related description, and will not be repeated here.

In this embodiment, when the number of the obstacles in the grounding frame is greater than or equal to the preset number of the obstacles, the grounding points in the first fitting grounding frame are clustered again, so that the number of the obstacles in each grounding frame is smaller than the preset number of the obstacles, the grounding frame fitted based on the grounding information is prevented from being larger than the real grounding frame, and the recognition accuracy of the parking space state is improved.

In a possible manner, determining the number of the obstacles in the grounding frame according to the information of the grounding frame of the obstacle relative to the parking space direction and the parking space information of the target parking space may include:

determining a first width value of the target parking space according to the parking space information of the target parking space, and determining a second width value of the grounding frame according to the information of the grounding frame of the obstacle relative to the parking space direction; and determining the number of the barriers in the grounding frame according to the ratio between the first width value and the second width value and the preset ratio.

For example, the preset ratio may include a first preset ratio and a second preset ratio, and when the ratio between the second width value and the first width value is smaller than the first preset ratio, the number of obstacles in the ground frame may be determined to be 1; when the ratio between the second width value and the first width value is smaller than the second preset ratio and is larger than or equal to the first preset ratio, the number of the barriers in the grounding frame can be determined to be 2; when the ratio between the second width value and the first width value is greater than or equal to the second preset ratio, the number of obstacles in the ground frame may be determined to be 3.

Wherein the first preset ratio and the second preset ratio may be determined according to actual conditions, and the embodiments of the present disclosure do not limit this at all, and the first preset ratio may be set asThe second preset ratio may be set to. Thus, after obtaining the ratio between the second width value and the first width value, the number of obstacles in the ground frame can be determined by the following formula:

wherein, The number of obstacles is indicated and the number of obstacles,A value of the first width is indicated,A value of the second width is indicated,A first predetermined ratio is indicated and a first predetermined ratio,Representing a second preset ratio.

In order to facilitate understanding of the parking space state determining method of the present disclosure, an embodiment of the present disclosure is described below.

As illustrated in fig. 17, for example, the fisheye image 1 may be acquired by a fisheye camera provided in front of the vehicle, the fisheye image 2 may be acquired by a fisheye camera provided in rear of the vehicle, the fisheye image 3 may be acquired by a fisheye camera provided in left side of the vehicle, and the fisheye image 4 may be acquired by a fisheye camera provided in right side of the vehicle. Then, on the one hand, the fisheye image 1, the fisheye image 2, the fisheye image 3 and the fisheye image 4 are respectively input into the ground point recognition model to obtain ground coordinates COOR1 (P1, P2..Pn) of the obstacle in the 4 fisheye images under the image coordinate system, and then the ground coordinates COOR1 (P1, P2..Pn) are converted into ground coordinates COOR2 (P1, P2..Pn) under the looking-around image based on the coordinate conversion relation between the looking-around image and the fisheye image. On the other hand, the fisheye image 1, the fisheye image 2, the fisheye image 3 and the fisheye image 4 are spliced to obtain a looking-around image, and the looking-around image is input into a parking space detection model to obtain corner coordinates SlotPt (P1, P2, P3, P4) of the parking space under the coordinate system of the looking-around image. Then, the ground coordinates COOR2 (P1, P2..pn) and the corner coordinates SlotPt (P1, P2, P3, P4) may be input to the obstacle fitting module to perform the ground frame fitting of the obstacle according to the ground coordinates COOR2 (P1, P2..pn) by the obstacle fitting module, to obtain a plurality of obstacle fitting frames, and for each obstacle fitting frame, the following steps are performed: determining a target parking space with the largest overlapping degree with the obstacle fitting frame, determining the vertex coordinates of the obstacle fitting frame corresponding to the obstacle relative to the target parking space based on the target parking space and the obstacle fitting frame, and finally inputting the vertex coordinates of the target obstacle relative to the target parking space and the corner coordinates of the target parking space into a parking space state determining model so as to obtain the parking space state of the target parking space based on the vertex coordinates of the target obstacle relative to the target parking space and the corner coordinates of the target parking space through the parking space state determining model, namely, whether the target parking space is occupied by the target obstacle.

Based on the same concept, the embodiment of the present disclosure further provides a parking space state determining apparatus, as shown in fig. 18, where the parking space state determining apparatus 1800 may include:

The first determining module 1801 is configured to determine, according to an image acquired by the vehicle-mounted capturing device of the vehicle, ground connection information of an obstacle in the image in the looking-around image and parking space information of a parking space in the image in the looking-around image;

the second determining module 1802 is configured to determine, according to the grounding information and the parking space information, first target grounding information of the obstacle relative to the parking space direction;

The third determining module 1803 is configured to determine, according to the first target grounding information and the parking space information, a parking space state of the parking space, where the parking space state is used to characterize whether the parking space is occupied by an obstacle.

Through above-mentioned parking stall state determining device 1800, can confirm the target ground connection information of obstacle relative parking stall direction based on the ground connection information of obstacle and the parking stall information of parking stall, and then according to target ground connection information and parking stall information, confirm the parking stall state of parking stall. Therefore, the automatic identification of the parking space state can be realized according to the grounding information and the parking space information. In addition, the target grounding information is the grounding information of the obstacle relative to the parking space direction, so that the occupation condition of the obstacle in the parking space can be accurately reflected by the target grounding information, and further, when the parking space state is determined according to the target grounding information, the recognition accuracy of the parking space state can be improved, and the parking accident is reduced.

In a possible manner, the parking space includes a plurality of parking spaces, and accordingly, the second determining module 1802 may include:

The first determining submodule is used for determining a target parking space in a plurality of parking spaces according to the grounding information and the parking space information;

the second determining submodule is used for determining first target grounding information of the obstacle relative to the parking space direction according to the grounding information and the parking space information of the target parking space;

Correspondingly, the third determining module may be configured to determine a parking space state of the target parking space according to the first target grounding information and the parking space information of the target parking space.

In a possible manner, the ground information may include a ground point of the obstacle, and accordingly, the first determining sub-module may include:

The first determining unit is used for determining the number of grounding points in each parking space according to the grounding points of the obstacle and the parking space information;

The first determining unit is used for determining the parking space with the largest number of grounding points among the plurality of parking spaces as a target parking space.

In a possible manner, the first determining sub-module may comprise:

The fitting unit is used for fitting the grounding frame of the obstacle according to the grounding information to obtain a first fitting grounding frame of the obstacle;

The third determining unit is used for determining the overlapping degree between the first fitting grounding frame and the parking spaces according to the parking space information of each parking space and the first fitting grounding frame;

and the fourth determining unit is used for determining the parking space with the largest overlapping degree with the first fitting grounding frame among the plurality of parking spaces as the target parking space.

In a possible manner, the grounding information includes the grounding points of the plurality of obstacles and the grounding coordinates of the grounding points, and accordingly, the fitting unit may include:

A clustering subunit, configured to cluster the grounding points according to the grounding coordinates of the grounding points, to obtain a plurality of first clusters;

and the fitting subunit is used for fitting the grounding frame according to the grounding point in each first cluster and the grounding coordinates of the grounding point to obtain a first fitting grounding frame of the obstacle.

In a possible manner, the grounding information includes a grounding point of the obstacle and a grounding coordinate of the grounding point, and accordingly, the second determination submodule may include:

the fifth determining unit is used for determining a target grounding point farthest from the parking space edge in the first fitting grounding frame according to the parking space information of the target parking space aiming at each parking space edge of the target parking space;

and the sixth determining unit is used for determining first target grounding information of the obstacle relative to the parking space direction according to all the target grounding points.

In a possible manner, the sixth determining unit may include:

The first determining subunit is used for determining, for each target grounding point, a straight line which passes through the target grounding point and is parallel to the parking space edge farthest from the target grounding point in the target parking spaces;

and the second determining subunit is used for determining the first target grounding information of the obstacle relative to the parking space direction according to all the straight lines.

In a possible manner, the second determining subunit may include:

the first determining component is used for determining grounding frame information of the obstacle according to all straight lines, and determining the grounding frame information as first target grounding information of the obstacle relative to the parking space direction; or alternatively

And the second determining component is used for determining intersection point information of all straight lines and determining the intersection point information as first target grounding information of the obstacle relative to the parking space direction.

In a possible manner, the third determining module 1803 may include:

The third determining submodule is used for determining the confidence that the parking space is occupied by the obstacle according to the first target grounding information and the parking space information;

And the fourth determining submodule is used for determining the parking space state of the parking space according to the confidence coefficient and a preset confidence coefficient threshold value.

In a possible manner, the first target grounding information includes a grounding frame and a grounding area of the obstacle with respect to the parking space direction, the parking space information includes a parking space frame and a parking space area of the parking space, and accordingly, the third determining submodule may include:

A seventh determining unit, configured to determine coverage areas of the ground frame and the parking space frame;

The eighth determining unit is configured to determine a first ratio of the coverage area to the parking space area, and a second ratio of the coverage area to the ground area, determine the first ratio as a confidence level that the parking space is occupied by the obstacle when the first ratio is greater than the second ratio, and determine the second ratio as a confidence level that the parking space is occupied by the obstacle when the second ratio is greater than the first ratio.

In a possible manner, the first target grounding information includes corner coordinates of the ground frame of the obstacle relative to the parking space direction, the parking space information includes corner coordinates of the parking space, and accordingly, the third determining module 1803 may be configured to input the corner coordinates of the ground frame of the obstacle relative to the parking space direction and the corner coordinates of the parking space into the parking space state determining model, to obtain the parking space state of the parking space, where the parking space state determining model is configured to output the parking space state according to the input corner coordinates of the ground frame and the corner coordinates of the parking space.

In a possible manner, the first target grounding information includes information of a grounding frame of the obstacle with respect to the parking space direction, and accordingly, the parking space state determining apparatus 1800 may further include:

The fourth determining module is used for determining the number of the obstacles in the grounding frame according to the information of the grounding frame in the direction of the obstacle relative to the parking space and the parking space information of the target parking space;

Accordingly, the third determining module 1803 may be configured to determine, when the number of obstacles in the grounding frame is less than or equal to the preset number of obstacles, a parking space state of the target parking space according to the first target grounding information and the parking space information of the target parking space.

In a possible manner, the third determining module 1803 may be configured to: when the number of the barriers in the grounding frame is larger than or equal to the preset number of the barriers, clustering the grounding points in the first fitting grounding frame according to the grounding coordinates of the grounding points in the first fitting grounding frame to obtain a plurality of second clusters, wherein the number of the second clusters is the same as the number of the barriers in the grounding frame; for each second cluster, the following steps are performed: performing grounding frame fitting according to the grounding points in the second clusters and the grounding coordinates of the grounding points to obtain second fitting grounding frames of the obstacles; determining second target grounding information of the obstacle relative to the parking space direction according to the second fitting grounding frame and the target parking space; and determining the parking space state of the parking space according to the second target grounding information and the parking space information of the target parking space.

In a possible manner, the fourth determining module may include:

A ninth determining unit, configured to determine a first width value of the target parking space according to the parking space information of the target parking space, and determine a second width value of the grounding frame according to the information of the grounding frame in the direction of the obstacle relative to the parking space;

And a tenth determining unit, configured to determine the number of obstacles in the grounding frame according to the ratio between the first width value and the second width value and the preset ratio.

Based on the same concept, the embodiments of the present disclosure also provide a controller, which may include:

A processor;

A memory for storing processor-executable instructions;

Wherein the processor may be configured to: executing the steps of any one of the vehicle position state determining methods.

Based on the same conception, the embodiment of the disclosure also provides a vehicle comprising the controller, or

The vehicle includes a processor and a memory for storing processor-executable instructions, wherein the processor may be configured to: executing the steps of any one of the vehicle position state determining methods.

Based on the same conception, the disclosed embodiments also provide a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of any of the above-described vehicle position state determination methods.

Based on the same conception, the disclosed embodiment also provides a computer program product comprising a computer program which realizes the steps of any of the above-mentioned vehicle position state determining methods when being executed by a processor.

Referring to fig. 19, fig. 19 is a functional block diagram of a vehicle according to an exemplary embodiment. Vehicle 1900 may include various subsystems, such as an infotainment system 1910, a perception system 1920, a decision control system 1930, a drive system 1940, and a computing platform 1950. Vehicle 1900 may also include more or fewer subsystems, and each subsystem may include multiple components. In addition, interconnections between each subsystem and between each component of vehicle 1900 may be made by wire or wirelessly.

In some embodiments, the infotainment system 1910 can include a communication system, an entertainment system, a navigation system, and the like.

The sensing system 1920 may include several types of sensors for sensing information of the environment surrounding the vehicle 1900. For example, sensing system 1920 may include a global positioning system (which may be a GPS system, or may be a beidou system or other positioning system), an inertial measurement unit (inertial measurement unit, IMU), a lidar, millimeter wave radar, an ultrasonic radar, and a camera device.

Decision control system 1930 can include a computing system, a vehicle controller, a steering system, a throttle, and a braking system.

Drive system 1940 may include components that provide powered movement of vehicle 1900. In one embodiment, the drive system 1940 may include an engine, an energy source, a transmission, and wheels. The engine may be one or a combination of an internal combustion engine, an electric motor, an air compression engine. The engine is capable of converting energy provided by the energy source into mechanical energy.

Some or all of the functions of vehicle 1900 are controlled by computing platform 1950. Computing platform 1950 may include at least one processor 1951 and memory 1952, and processor 1951 may execute instructions 1953 stored in memory 1952.

The processor 1951 may be any conventional processor, such as a commercially available CPU. The processor may also include, for example, an image processor (Graphic Process Unit, GPU), a field programmable gate array (Field Programmable GATE ARRAY, FPGA), a System On Chip (SOC), an Application SPECIFIC INTEGRATED Circuit (ASIC), or a combination thereof.

Memory 1952 may be implemented by any type of volatile or non-volatile memory device or combination thereof, such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disk.

In addition to instructions 1953, the memory 1952 may store data such as road maps, route information, vehicle location, direction, speed, etc. The data stored by memory 1952 may be used by computing platform 1950.

In the disclosed embodiment, the processor 1951 can execute instructions 1953 to complete all or part of the steps of the parking spot status determination method described above.

The preferred embodiments of the present disclosure have been described in detail above with reference to the accompanying drawings, but the present disclosure is not limited to the specific details of the embodiments described above, and various simple modifications may be made to the technical solutions of the present disclosure within the scope of the technical concept of the present disclosure, and all the simple modifications belong to the protection scope of the present disclosure.

In addition, the specific features described in the foregoing embodiments may be combined in any suitable manner, and in order to avoid unnecessary repetition, the present disclosure does not further describe various possible combinations.

Claims

1. The parking space state determining method is characterized by comprising the following steps of:

2. The parking space state determining method according to claim 1, wherein the parking space includes a plurality of the first target ground information of the obstacle with respect to the parking space direction based on the ground information and the parking space information, comprising:

3. The parking space state determination method according to claim 2, wherein the ground information includes a ground point of the obstacle, and the determining a target parking space among a plurality of the parking spaces based on the ground information and the parking space information includes:

4. The parking space state determining method according to claim 2, wherein determining a target parking space among a plurality of the parking spaces based on the ground connection information and the parking space information comprises:

5. The method for determining a parking space state according to claim 4, wherein the grounding information includes grounding points of the plurality of obstacles and grounding coordinates of the grounding points, the fitting the grounding frame of the obstacle according to the grounding information to obtain a first fitted grounding frame of the obstacle includes:

6. The parking space state determining method according to claim 5, wherein determining the first target ground information of the obstacle with respect to the parking space direction based on the ground information and the parking space information of the target parking space comprises:

determining a target grounding point farthest from each parking space edge in the first fitting grounding frame according to the parking space information of the target parking space aiming at each parking space edge of the target parking space;

7. The method of claim 6, wherein determining the first target ground information of the obstacle relative to the parking space direction based on all the target ground points comprises:

8. The parking space state determination method according to claim 7, wherein the determining the first target ground information of the obstacle with respect to the parking space direction based on all the straight lines includes:

9. The parking space state determining method according to any one of claims 1 to 5, wherein the determining the parking space state of the parking space based on the first target ground information and the parking space information includes:

10. The parking space state determination method according to claim 9, wherein the first target ground information includes a ground frame and a ground area of the obstacle with respect to a parking space direction, the parking space information includes a parking space frame and a parking space area of the parking space, and the determining the degree of confidence that the parking space is occupied by the obstacle based on the first target ground information and the parking space information includes:

11. The parking space state determination method according to any one of claims 1 to 5, wherein the first target ground information includes corner coordinates of a ground frame of the obstacle with respect to a parking space direction, the parking space information includes corner coordinates of the parking space, and the determining the parking space state of the parking space based on the first target ground information and the parking space information includes:

12. The parking space state determination method according to claim 5, wherein the first target ground information includes information of a ground frame of the obstacle with respect to a parking space direction, the parking space state determination method further comprising:

13. The parking spot status determination method according to claim 12, further comprising:

For each second cluster, the following steps are performed:

Performing grounding frame fitting according to the grounding points in the second clusters and the grounding coordinates of the grounding points to obtain second fitting grounding frames of the obstacles;

Determining second target grounding information of the obstacle relative to the parking space direction according to the second fitting grounding frame and the target parking space;

14. The parking space state determining method according to claim 12, wherein the determining the number of the obstacles in the ground frame based on the information of the ground frame of the obstacle with respect to the parking space direction and the parking space information of the target parking space includes:

15. A controller, the controller comprising:

A processor;

A memory for storing processor-executable instructions;

Wherein the processor is configured to: performing the steps of the method of any one of claims 1-14.

16. A vehicle comprising a controller as claimed in claim 15, or

The vehicle includes a processor and a memory for storing processor-executable instructions, wherein the processor is configured to: performing the steps of the method of any one of claims 1-14.

17. A non-transitory computer readable storage medium having stored thereon a computer program, which when executed by a processor performs the steps of the method according to any of claims 1-14.

18. A computer program product comprising a computer program, characterized in that the computer program, when executed by a processor, implements the steps of the method according to any one of claims 1-14.