Parking space detection method and parking space detection system

Abstract

The invention provides a parking space detection method. The parking space detection method comprises the steps that target parking space image information collected by an image sensor and a target parking space radar data set collected by a radar sensor are acquired; image parking space point coordinates are determined according to the target parking space image information; according to the target parking space radar data set, radar parking space point coordinates are determined; matching the image parking space point coordinates with the radar parking space point coordinates to obtain a parking space point coordinate set in which the image parking space point coordinates and the radar parking space point coordinates are in one-to-one correspondence; obtaining fusion parameters detected by a current sunlight and rainfall sensor; and according to the fusion parameters, fusing the image parking space point coordinates and the radar parking space point coordinates in one-to-one correspondence in the parking space point coordinate set to obtain corresponding parking space point fusion coordinates, and according to the parking space point fusion coordinates, determining target parking space information. The problem that the parking space cannot be identified or the detection precision is low when the parking space line is lost or worn in severe weather and night conditions can be solved; the precision is higher, and the cost is lower.

Description

Technical Field

[0001] The present invention relates to the field of automobile sensing, and in particular to a parking space detection method and a parking space detection system. Background Art

[0002] With the continuous advancement of automobile technology and the increasing demand of consumers for intelligence and safety, the automatic parking system has gradually changed from a concept to a reality and has become an important part of the field of autonomous driving. The automatic parking system can significantly improve parking efficiency, greatly shorten parking time, and reduce the difficulty and fatigue of the driver. At the same time, the automatic parking system can reduce parking accidents caused by human factors, improve parking safety, and protect the safety of passengers and vehicles. Therefore, the automatic parking system has broad application prospects and huge market potential in the autonomous driving industry. Parking space detection is one of the important components of the automatic parking system technology and plays an indispensable role in realizing the fully automatic parking function. Faced with complex and changeable real scenes, a variety of sensors provide more comprehensive environmental information for the integration of the automatic parking system.

[0003] The most critical issue for automatic parking systems is finding vacant parking spaces.

[0004] The typical approach is to detect existing parking space lines on the ground based on computer vision methods on camera images, and then identify whether they are occupied. However, some parking lots do not draw parking space lines for users, but simply declare vacant spaces as available parking spaces. At the same time, in old parking lots, some parking space lines are severely worn.

[0005] Another typical method is the parking space detection method that combines ultrasonic radar and camera. In this method, ultrasonic radar is used to detect the distance information between obstacles and the vehicle, providing information assistance for the camera-based parking space detection algorithm, but the ranging accuracy of ultrasonic radar is limited. In contrast, the parking space detection method based on millimeter-wave radar can alleviate the above pain points to a certain extent, and more accurate parking space detection results can be obtained by fusing the detection results of the camera and millimeter-wave radar. The weights obtained by deep learning are usually used to determine the parking space, which has low accuracy and high investment costs for data collection, storage, and model training.

[0006] Therefore, how to develop a parking space detection method that can provide more accurate parking space coordinate information and provide more accurate parking space detection results for automatic parking systems in bad weather, low light intensity, worn or missing parking space lines, etc., has become a technical problem that technical personnel in this field urgently need to solve. Summary of the invention

[0007] In order to overcome the above technical defects, the purpose of the present invention is to provide a parking space detection method and a parking space detection system. By fusing the brightness and humidity detected by the sunlight and rain sensor and other fusion parameters to fuse the parking space coordinates detected by the camera and the millimeter wave radar, the target parking space detection can be performed, which can solve the problem of parking space being unable to be identified or low detection accuracy in bad weather, night conditions, missing or worn parking space lines; higher accuracy and lower cost.

[0008] The present invention discloses a parking space detection method, comprising:

[0009] Obtaining target parking space image information collected by the image sensor and target parking space radar data set collected by the radar sensor;

[0010] Determine the image parking space coordinates according to the target parking space image information; determine the radar parking space coordinates according to the target parking space radar data set;

[0011] Matching the image parking spot coordinates with the radar parking spot coordinates to obtain a parking spot coordinate set in which the image parking spot coordinates and the radar parking spot coordinates correspond one to one;

[0012] Get the fusion parameters detected by the current sunlight and rain sensor;

[0013] According to the fusion parameters, the image parking spot coordinates and the radar parking spot coordinates in one-to-one correspondence in the parking spot coordinate set are fused to obtain corresponding parking spot fusion coordinates, and the target parking spot information is determined according to the parking spot fusion coordinates.

[0014] Optionally, the image parking spot coordinates and the radar parking spot coordinates are both parking spot corner point coordinates.

[0015] Optionally, the fusion parameters include ambient humidity and ambient brightness.

[0016] Optionally, the step of fusing the image parking spot coordinates and the radar parking spot coordinates that correspond one to one in the parking spot coordinate set according to the fusion parameters to obtain corresponding parking spot fusion coordinates includes:

[0017] According to the ambient humidity s and the ambient brightness l, the image parking position coordinates corresponding to each other are calculated by equation (1): and the radar parking location coordinates Fusion is performed to obtain the fusion coordinates of the parking spot

[0018]

[0019] Among them, {i=1,2,3,4…}.

[0020] Optionally, four groups of image parking spot coordinates and corresponding four groups of radar parking spot coordinates are obtained to form the parking spot coordinate set;

[0021] The step of fusing the image parking spot coordinates and the radar parking spot coordinates corresponding to each other in the parking spot coordinate set according to the fusion parameters to obtain corresponding parking spot fusion coordinates includes:

[0022] According to the environmental humidity s and the environmental brightness l, the image parking spot coordinates in the parking spot coordinate set And the one-to-one corresponding radar parking position coordinates According to formula (1), i is set to 1, 2, 3, and 4 respectively, and four sets of parking spot fusion coordinates are calculated.

[0023] The present invention also discloses a parking space detection system, comprising:

[0024] Sensor module, parking spot coordinate calculation module, parking spot matching module, sunlight and rainfall sensor and parking spot fusion module;

[0025] The sensor module includes: a radar sensor and an image sensor; the radar sensor is used to collect image information of the target parking space, and the image sensor is used to collect radar data sets of the target parking space;

[0026] The parking spot coordinate calculation module is used to obtain the target parking spot radar data set and determine the radar parking spot coordinates according to the target parking spot radar data set; and to obtain the target parking spot image information and determine the image parking spot coordinates according to the target parking spot image information;

[0027] The parking spot matching module is used to obtain and match the image parking spot coordinates and the radar parking spot coordinates to obtain a parking spot coordinate set in which the image parking spot coordinates and the radar parking spot coordinates correspond one to one;

[0028] The sunlight and rainfall sensor is used to detect and obtain current fusion parameters;

[0029] The parking spot fusion module is used to obtain the fusion parameters and the parking spot coordinate set, and according to the fusion parameters, fuse the image parking spot coordinates and the radar parking spot coordinates in the parking spot coordinate set that correspond one to one to obtain the corresponding parking spot fusion coordinates, and determine the target parking spot information according to the parking spot fusion coordinates.

[0030] Optionally, the parking spot coordinate calculation module includes:

[0031] A millimeter wave radar parking spot coordinate calculation module is used for the target parking spot radar data set, and determines the radar parking spot coordinates according to the target parking spot radar data set;

[0032] The image parking spot coordinate calculation module is used to obtain the target parking spot image information and determine the image parking spot coordinates based on the target parking spot image information.

[0033] Optionally, the image parking space coordinates and the radar parking space coordinates are both parking space corner point coordinates.

[0034] Optionally, the sunlight and rainfall sensor is used to obtain fusion parameters: ambient humidity s and ambient brightness l;

[0035] The parking spot fusion module is used to calculate the one-to-one corresponding parking spot coordinates of the image according to the ambient humidity s and the ambient brightness l through formula (1). and the radar parking location coordinates Fusion is performed to obtain the fusion coordinates of the parking spot

[0036]

[0037] Among them, {i=1,2,3,4…}.

[0038] Optionally, a millimeter wave radar parking spot coordinate calculation module is used to detect and determine the four sets of radar parking spot coordinates;

[0039] An image parking spot coordinate calculation module is used to detect and determine the four groups of image parking spot coordinates;

[0040] The parking spot matching module is used to obtain and match the four groups of image parking spot coordinates and the four groups of radar parking spot coordinates to obtain a parking spot coordinate set in which the image parking spot coordinates and the radar parking spot coordinates correspond to each other.

[0041] The parking spot fusion module is used to obtain the fusion parameters and the parking spot coordinate set, and according to the fusion parameters, fuse the image parking spot coordinates and the radar parking spot coordinates in the parking spot coordinate set one by one to obtain the corresponding parking spot fusion coordinates, including:

[0042] According to the environmental humidity s and the environmental brightness l, the image parking spot coordinates in the parking spot coordinate set And the one-to-one corresponding radar parking position coordinates According to formula (1), i is set to 1, 2, 3, and 4 respectively, and four sets of parking spot fusion coordinates are calculated.

[0043] Compared with the prior art, the above technical solution has the following beneficial effects:

[0044] 1. The present invention fuses the parking space coordinates detected by the camera and the millimeter-wave radar by fusing the fusion parameters such as brightness and humidity detected by the sunlight and rain sensor to perform target parking space detection, which can solve the problem of unrecognizable parking spaces or low detection accuracy in bad weather, night conditions, or when the parking space lines are missing or worn.

[0045] 2. Compared with existing machine learning methods, it has higher accuracy and lower cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 The figure is a flow chart of a parking space detection method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The advantages of the present invention are further described below in conjunction with the accompanying drawings and specific embodiments.

[0048] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0049] The terms used in this disclosure are for the purpose of describing specific embodiments only and are not intended to limit the disclosure. The singular forms of "a", "said" and "the" used in this disclosure and the appended claims are also intended to include plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0050] In the description of the present invention, it is necessary to understand that the terms "inside", "outside", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0051] In the description of the present invention, unless otherwise specified and limited, it should be noted that the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a mechanical connection or an electrical connection, or it can be the internal connection between two components. It can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to the specific circumstances.

[0052] In the following description, the suffixes such as "module", "component" or "unit" used to represent elements are only used to facilitate the description of the present invention, and have no specific meanings. Therefore, "module" and "component" can be used interchangeably.

[0053] Figure 1 A parking space detection method according to an embodiment of the present invention is shown. The parking space detection method comprises:

[0054] S1: Acquire the target parking space image information collected by the image sensor and the target parking space radar data set collected by the radar sensor;

[0055] S2: determining the image parking space coordinates according to the target parking space image information; determining the radar parking space coordinates according to the target parking space radar data set;

[0056] S3: Matching the image parking spot coordinates with the radar parking spot coordinates to obtain a parking spot coordinate set where the image parking spot coordinates and the radar parking spot coordinates correspond one to one;

[0057] S4: Obtain the fusion parameters detected by the current sunlight and rainfall sensor;

[0058] S5: According to the fusion parameters, the image parking spot coordinates and the radar parking spot coordinates in one-to-one correspondence in the parking spot coordinate set are fused to obtain corresponding parking spot fusion coordinates, and the target parking spot information is determined according to the parking spot fusion coordinates.

[0059] The radar sensor is a millimeter-wave radar sensor, and the target parking space radar data set includes distance and angle information of the point cloud of the parking space acquired by the millimeter-wave radar sensor; the image sensor is a camera, and the target parking space image information includes the stationary vehicle position detected on the camera image and the ground contact points of the front and rear wheels.

[0060] The present invention provides a parking space detection method for use in an automatic parking system. The parking space corner coordinates are calculated by using the distance and angle information of the point cloud of the millimeter wave radar sensor. At the same time, the position of the stagnant vehicle and the grounding points of the front and rear wheels are detected on the camera image, and then the parking space corner coordinates from the camera are calculated. Finally, the target parking space detection is realized by using the parking space corner matching algorithm and the fusion algorithm based on the sunlight and rain sensor information. The method detects the target parking space by fusing the fusion parameters detected by the sunlight and rain sensor, which can solve the problem of the inability to identify the parking space or the low detection accuracy in bad weather, night conditions, missing or worn parking space lines. Compared with the existing machine learning methods, it has higher accuracy and lower cost.

[0061] The specific steps of a parking space detection method are described as follows:

[0062] Execute step S1: obtain the target parking space image information collected by the image sensor and the target parking space radar data set collected by the radar sensor.

[0063] Millimeter-wave radar sensors and cameras are required to detect parking spaces simultaneously.

[0064] For the millimeter-wave radar end, first obtain the raw data after millimeter-wave radar ADC sampling, and use signal processing algorithms such as Fourier transform, constant false alarm detection, and super-resolution angle measurement to obtain the millimeter-wave radar point cloud, where each point cloud contains three-dimensional information of distance, angle, and speed; then for each frame of millimeter-wave radar point cloud, filter out the point cloud with a speed greater than 0, retain the point cloud with a speed of 0 (stationary obstacles), and find the point cloud with the shortest distance in the point cloud with a speed of 0, that is, the shortest distance filtering; finally, use the shortest distance point clouds of multiple frames to form a distance-time graph, so as to obtain the target parking space radar data set.

[0065] For the camera side, the camera image is first obtained, and the instance segmentation technology is used to complete the position detection of the parked vehicle for each frame of the image. Then, the detection result is transformed from the image coordinate system to the world coordinate system, namely the BEV (Birdseye of view) perspective, so as to obtain the image information of the target parking space.

[0066] Execute step S2: determine the image parking spot coordinates based on the target parking space image information; determine the radar parking spot coordinates based on the target parking space radar data set; in a typical test method, the image parking spot coordinates and the radar parking spot coordinates are the vehicle's parking space corner point coordinates, preferably the vehicle's four parking space corner point coordinates.

[0067] (I) The millimeter wave radar parking spot coordinate calculation module determines the image parking spot coordinates based on the target parking spot image information. The specific process is described as follows:

[0068] The distance-time graph is constructed based on the shortest distance point cloud. Assume that the point cloud at time t obtained by the shortest distance filtering is P t =(R t ,A t ), in order to reduce the false alarm rate, the present invention uses the point cloud {P t-1 ,P t-2 ,P t-3}, the speed v from the CAN signal of the vehicle is used to estimate the parking space and obtain the coordinates of the parking space. The specific implementation method is as follows:

[0069] Step (1): the number of frames required to find a parking space is estimated based on the estimated driving time during the actual parking search and the frame rates of the corresponding camera and millimeter-wave radar sensor, which is recorded as T.

[0070] Step (2), traverse all T frames, denoted as t∈{1,2,…,T}. The distance moved by the vehicle is D t =v×Δt+D t-1 .

[0071] If R t <θ and (where θ is a predefined depth threshold used to determine the possibility of an empty parking space), then there is a stalled vehicle occupying the parking space, and the width of the stalled vehicle is calculated as W object =W object +D t , parking lot collection

[0072] If R t ≥θ and Then W object =0,

[0073] The width of the parking space is W space =W spac +D t ,P space =P space ∪P t .

[0074] If R t <θ, then W object =0,

[0075] The width of the parking space is W space =W space +D t ,P space =P space ∪P t .

[0076] If R t ≥θ, then W object =W object +D t , parking lot collection

[0077] Step (3): for the stationary vehicle width W output in step (2), objevt 、Parking space width W space and parking lot set D space , if D spac is an empty set, then the parking spot coordinate calculation module outputs that there is no parking space; if W space <W car +offset(W car is the width of the vehicle, and offset is half of the distance between two adjacent vehicles during parking), then the parking spot coordinate calculation module outputs that there is no parking space; W spac >W car+offset, then output That is, we get point P m1 =(R m ,S m1 ), P m2 =(R m ,A m2 ), P n1 =(R n ,A n1 ) and P n2 =(R n ,A n2 ).

[0078] Step (4), point P m1 , P m2 , P n1 and P n2 The radar coordinate system (polar coordinate system) is converted to the coordinate P under the BEV perspective m1 ′=(x m1 ,y m1 ), P m2 ′=(x m2 ,y m2 ), P n1 ′=(x n1 ,y n1 ) and P n2 ′==(x n2 ,y n2 ), the conversion process is:

[0079]

[0080]

[0081] Among them, t x is the installation distance difference between the millimeter wave radar and the camera in the horizontal direction, t y is the installation distance difference between the millimeter-wave radar and the camera in the vertical direction.

[0082] Step (5): For parking safety considerations, the four parking spots need to be fine-tuned. The final image parking spot coordinates are:

[0083]

[0084] Finally, the coordinate set of the four parking spots detected by the millimeter-wave radar sensor is expressed as:

[0085]

[0086] (ii) The camera parking spot coordinate calculation module determines the camera parking spot coordinates based on the target parking spot image information. The specific process is described as follows:

[0087] First, the camera image is input into the instance segmentation module, which outputs the mask of the detected vehicle and removes the detected vehicles outside the parking range, leaving only the stationary vehicles. Next, the mask is divided into the upper half (corresponding to the front wheel at a farther distance) and the lower half (corresponding to the front wheel at a closer distance). For the two parts, the two points closest to the edge of the mask of the two stationary vehicles detected are calculated respectively, and the ground contact points of the four wheels in the image coordinate system can be obtained. Then, in order to obtain the distance between the two stationary vehicles, the calibrated intrinsic and extrinsic matrix is ​​used to transform the detected mask and the four wheel ground contact points in the mask to the world coordinate system, that is, the BEV perspective. Finally, the wheel ground contact points under the BEV perspective are translated by a certain offset offset to obtain the final four parking spot coordinates. The final set of four parking spot coordinates detected by the camera is:

[0088]

[0089] S3: Match the image parking spot coordinates and the radar parking spot coordinates to obtain a parking spot coordinate set in which the image parking spot coordinates and the radar parking spot coordinates correspond one to one.

[0090] The parking spot matching module matches the image parking spot coordinates with the radar parking spot coordinates in the following steps: Assume that the four parking spot coordinate sets detected by the millimeter wave radar are:

[0091]

[0092] The coordinate sets of the four parking spots detected by the camera are:

[0093]

[0094] For the four parking spots of the millimeter-wave radar, traverse each point from the set C, calculate the Euclidean distance between the two points, and select the point with the smallest distance for matching. The specific algorithm is as follows:

[0095] For i = {1,2,3,4}

[0096] Forj = {1,2,3,4}

[0097] Calculate the Euclidean distance d between the i-th point in set R and the j-th point in set C ij :

[0098]

[0099] calculate Find the point that matches the i-th point in the set R, recorded as

[0100]

[0101] End

[0102] Finally, the parking lot matching results of the millimeter-wave radar and the camera are obtained:

[0103]

[0104] The specific implementation algorithms in steps S1, S2, and S3 are all implemented using the respective algorithm functions of the vehicle-mounted millimeter-wave radar parking spot coordinate calculation module, the camera parking spot coordinate calculation module, and the parking spot matching module. The above specific algorithm is only an optional example, and the present invention is not limited thereto. Any similar algorithm known to those skilled in the art can be used as an optional implementation. The preferred embodiment of the present invention will be described later. It is specifically embodied in steps S4-S5.

[0105] Execute step S4: obtain the fusion parameters detected by the current sunlight and rainfall sensor; in this embodiment, the fusion parameters are preferably ambient humidity and ambient brightness.

[0106] Execute step S5: according to the fusion parameters, fuse the image parking spot coordinates and the radar parking spot coordinates in the parking spot coordinate set in one-to-one correspondence to obtain corresponding parking spot fusion coordinates, and determine the target parking spot information according to the parking spot fusion coordinates.

[0107] First, four groups of image parking spot coordinates and corresponding four groups of radar parking spot coordinates are obtained to form the parking spot coordinate set as follows:

[0108]

[0109] Secondly, according to the environmental humidity s and the environmental brightness l, the image parking spot coordinates in the parking spot coordinate set And the one-to-one corresponding radar parking position coordinates According to formula (1), i is set to 1, 2, 3, and 4 respectively, and the fusion coordinates of four parking spots are calculated: and

[0110]

[0111] Among them, {i=1,2,3,4}.

[0112] The target parking space information is determined based on the four sets of parking space fusion coordinates. Therefore, if the camera and the millimeter-wave radar detect the parking space coordinates at the same time, the parking space fusion coordinates obtained by the above fusion method are used to determine the final target parking space information. In the process of fusion using formula (1), different weights are assigned to the camera and the millimeter-wave radar for different temperature and humidity environments; due to the low detection accuracy of the camera under bad weather and night conditions, the detection performance of the millimeter-wave radar will not be affected by bad weather and lighting. Therefore, the fusion weight assigned to the camera will become lower, and the fusion weight assigned to the millimeter-wave radar will become higher. The fusion result can better reflect the accurate information of the target parking space, thereby solving the problem of the inability to identify the parking space or the low detection accuracy in bad weather, night conditions, missing or worn parking space lines.

[0113] If the camera cannot detect the parking space line well, the detection result of the millimeter wave radar can be used; if the millimeter wave also cannot detect the four coordinates of the parking space, it is considered that there is no empty parking space and parking is not possible.

[0114] It can be seen that the present invention aims at the pain point that parking spaces cannot be identified or the detection accuracy is low when the parking space lines are missing or worn. By fusing the brightness and humidity detected by the sunlight and rain sensors and other fusion parameters, the coordinates of the parking space detected by the camera and the millimeter wave radar are fused to achieve the weight adjustment of the two sensors under different weather and environmental conditions, and perform target parking space detection, which can solve the problem of parking spaces cannot be identified or the detection accuracy is low in bad weather, night conditions, missing or worn parking space lines. And compared with the existing machine learning methods, it has higher accuracy and lower cost.

[0115] The present invention also provides a parking space detection system, comprising:

[0116] Sensor module, parking spot coordinate calculation module, parking spot matching module, sunlight and rainfall sensor and parking spot fusion module;

[0117] The sensing module includes: a radar sensor and an image sensor; the radar sensor is used to collect image information of the target parking space, and the image sensor is used to collect radar data sets of the target parking space;

[0118] The parking spot coordinate calculation module is used to obtain the target parking spot radar data set and determine the radar parking spot coordinates based on the target parking spot radar data set; and is used to obtain the target parking spot image information and determine the image parking spot coordinates based on the target parking spot image information;

[0119] The parking spot matching module is used to obtain and match the image parking spot coordinates and the radar parking spot coordinates to obtain a parking spot coordinate set in which the image parking spot coordinates and the radar parking spot coordinates correspond one to one;

[0120] The sunlight and rainfall sensor is used to detect and obtain current fusion parameters;

[0121] The parking spot fusion module is used to obtain the fusion parameters and the parking spot coordinate set, and fuse the image parking spot coordinates and the radar parking spot coordinates in the parking spot coordinate set one by one according to the fusion parameters to obtain the corresponding parking spot fusion coordinates, and determine the target parking spot information according to the parking spot fusion coordinates. Further, the parking spot coordinate calculation module includes:

[0122] A millimeter wave radar parking spot coordinate calculation module is used for the target parking spot radar data set, and determines the radar parking spot coordinates according to the target parking spot radar data set;

[0123] Image parking spot coordinate calculation module: used to obtain the target parking spot image information, and determine the image parking spot coordinates according to the target parking spot image information. Preferably, the image parking spot coordinates and the radar parking spot coordinates are both parking spot corner point coordinates.

[0124] Furthermore, the sunlight and rainfall sensor is used to obtain fusion parameters: ambient humidity s and ambient brightness l;

[0125] The parking spot fusion module is used to calculate the one-to-one corresponding parking spot coordinates of the image according to the ambient humidity s and the ambient brightness l through formula (1). and the radar parking location coordinates Fusion is performed to obtain the fusion coordinates of the parking spot

[0126]

[0127] Among them, {i=1,2,3,4…}.

[0128] Furthermore, the millimeter wave radar parking spot coordinate calculation module is used to detect and determine the four sets of radar parking spot coordinates;

[0129] An image parking spot coordinate calculation module is used to detect and determine the four groups of image parking spot coordinates;

[0130] The parking spot matching module is used to obtain and match the four groups of image parking spot coordinates and the four groups of radar parking spot coordinates to obtain a parking spot coordinate set in which the image parking spot coordinates and the radar parking spot coordinates correspond to each other.

[0131] The parking spot fusion module is used to obtain the fusion parameters and the parking spot coordinate set, and according to the fusion parameters, fuse the one-to-one corresponding image parking spot coordinates and the radar parking spot coordinates in the parking spot coordinate set to obtain the corresponding parking spot fusion coordinates, including: according to the environmental humidity s and the environmental brightness l, the image parking spot coordinates in the parking spot coordinate set And the one-to-one corresponding radar parking position coordinates According to formula (1), i is set to 1, 2, 3, and 4 respectively, and four sets of parking spot fusion coordinates are calculated.

[0132] It should be noted that the embodiments of the present invention have better practicability and do not impose any form of limitation on the present invention. Any technician familiar with the field may use the technical content disclosed above to change or modify it into an equivalent effective embodiment. However, any modification or equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A parking space detection method, characterized in that: include: Obtaining target parking space image information collected by the image sensor and target parking space radar data set collected by the radar sensor; Determine the image parking space coordinates according to the target parking space image information; Determine the radar parking space coordinates according to the target parking space radar data set; Matching the image parking spot coordinates with the radar parking spot coordinates to obtain a parking spot coordinate set in which the image parking spot coordinates and the radar parking spot coordinates correspond one to one; Get the fusion parameters detected by the current sunlight and rain sensor; According to the fusion parameters, the image parking spot coordinates and the radar parking spot coordinates in one-to-one correspondence in the parking spot coordinate set are fused to obtain corresponding parking spot fusion coordinates, and the target parking spot information is determined according to the parking spot fusion coordinates.

2. The parking space detection method according to claim 1, characterized in that: The image parking spot coordinates and the radar parking spot coordinates are both parking spot corner point coordinates.

3. The parking space detection method according to claim 2, characterized in that: The fusion parameters include ambient humidity and ambient brightness.

4. The parking space detection method according to claim 3, characterized in that: The step of fusing the image parking spot coordinates and the radar parking spot coordinates corresponding to each other in the parking spot coordinate set according to the fusion parameters to obtain corresponding parking spot fusion coordinates includes: According to the ambient humidity s and the ambient brightness l, the image parking position coordinates corresponding to each other are calculated by equation (1): and the radar parking location coordinates Fusion is performed to obtain the fusion coordinates of the parking spot Among them, {i=1,2,3,4…}.

5. The parking space detection method according to claim 4, characterized in that: Acquire four groups of image parking spot coordinates and corresponding four groups of radar parking spot coordinates to form the parking spot coordinate set; The step of fusing the image parking spot coordinates and the radar parking spot coordinates corresponding to each other in the parking spot coordinate set according to the fusion parameters to obtain corresponding parking spot fusion coordinates includes: According to the environmental humidity s and the environmental brightness l, the image parking spot coordinates in the parking spot coordinate set And the one-to-one corresponding radar parking position coordinates According to formula (1), i is set to 1, 2, 3, and 4 respectively, and four sets of parking spot fusion coordinates are calculated.

6. A parking space detection system, characterized in that: include: Sensor module, parking spot coordinate calculation module, parking spot matching module, sunlight and rainfall sensor and parking spot fusion module; The sensor module includes: a radar sensor and an image sensor; the radar sensor is used to collect image information of the target parking space, and the image sensor is used to collect radar data sets of the target parking space; The parking spot coordinate calculation module is used to obtain the target parking spot radar data set and determine the radar parking spot coordinates according to the target parking spot radar data set; and to obtain the target parking spot image information and determine the image parking spot coordinates according to the target parking spot image information; The parking spot matching module is used to obtain and match the image parking spot coordinates and the radar parking spot coordinates to obtain a parking spot coordinate set in which the image parking spot coordinates and the radar parking spot coordinates correspond one to one; The sunlight and rainfall sensor is used to detect and obtain current fusion parameters; The parking spot fusion module is used to obtain the fusion parameters and the parking spot coordinate set, and fuse the image parking spot coordinates and the radar parking spot coordinates in the parking spot coordinate set that correspond to each other one by one according to the fusion parameters to obtain corresponding parking spot fusion coordinates, and determine the target parking spot information according to the parking spot fusion coordinates.

7. The parking space detection system according to claim 6, characterized in that: The parking spot coordinate calculation module includes: A millimeter wave radar parking spot coordinate calculation module is used for the target parking spot radar data set, and determines the radar parking spot coordinates according to the target parking spot radar data set; Image parking spot coordinate calculation module; Used to obtain the target parking space image information, and determine the image parking space point coordinates based on the target parking space image information.

8. The parking space detection system according to claim 7, characterized in that: The image parking spot coordinates and the radar parking spot coordinates are both parking spot corner point coordinates.

9. The parking space detection system according to claim 8, characterized in that: The sunlight and rainfall sensor is used to obtain fusion parameters: ambient humidity s and ambient brightness l; The parking spot fusion module is used to calculate the one-to-one corresponding parking spot coordinates of the image according to the ambient humidity s and the ambient brightness l through formula (1). and the radar parking location coordinates Fusion is performed to obtain the fusion coordinates of the parking spot Among them, {i=1,2,3,4…}.

10. The parking space detection system according to claim 9, characterized in that: A millimeter wave radar parking spot coordinate calculation module is used to detect and determine the four sets of radar parking spot coordinates; An image parking spot coordinate calculation module is used to detect and determine the four groups of image parking spot coordinates; The parking spot matching module is used to obtain and match the four groups of image parking spot coordinates and the four groups of radar parking spot coordinates to obtain a parking spot coordinate set in which the image parking spot coordinates and the radar parking spot coordinates correspond to each other. The parking spot fusion module is used to obtain the fusion parameters and the parking spot coordinate set, and according to the fusion parameters, fuse the image parking spot coordinates and the radar parking spot coordinates in the parking spot coordinate set one by one to obtain the corresponding parking spot fusion coordinates, including: According to the environmental humidity s and the environmental brightness l, the image parking spot coordinates in the parking spot coordinate set And the one-to-one corresponding radar parking position coordinates According to formula (1), i is set to 1, 2, 3, and 4 respectively, and four sets of parking spot fusion coordinates are calculated.

Images