[go: up one dir, main page]

CN119229415A - Automatic parking space detection method, device, computer equipment and medium - Google Patents

Automatic parking space detection method, device, computer equipment and medium Download PDF

Info

Publication number
CN119229415A
CN119229415A CN202411263828.6A CN202411263828A CN119229415A CN 119229415 A CN119229415 A CN 119229415A CN 202411263828 A CN202411263828 A CN 202411263828A CN 119229415 A CN119229415 A CN 119229415A
Authority
CN
China
Prior art keywords
parking space
line
surround
detection method
area
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202411263828.6A
Other languages
Chinese (zh)
Inventor
程建伟
马村
王雅儒
王述良
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan Jimu Intelligent Technology Co ltd
Original Assignee
Wuhan Jimu Intelligent Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Wuhan Jimu Intelligent Technology Co ltd filed Critical Wuhan Jimu Intelligent Technology Co ltd
Priority to CN202411263828.6A priority Critical patent/CN119229415A/en
Publication of CN119229415A publication Critical patent/CN119229415A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
    • G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
    • G06V20/586—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of parking space
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T3/00—Geometric image transformations in the plane of the image
    • G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
    • G06T3/4038—Image mosaicing, e.g. composing plane images from plane sub-images
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T3/00—Geometric image transformations in the plane of the image
    • G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
    • G06T3/4046—Scaling of whole images or parts thereof, e.g. expanding or contracting using neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/80—Geometric correction
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30244—Camera pose
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30248—Vehicle exterior or interior
    • G06T2207/30252—Vehicle exterior; Vicinity of vehicle
    • G06T2207/30264—Parking
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Databases & Information Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Molecular Biology (AREA)
  • Biomedical Technology (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Traffic Control Systems (AREA)

Abstract

本发明实施例提供了一种自动泊车停车位检测方法，涉及自动泊车技术领域，所述检测方法包括：获取多视角图像；基于标定的多个图像采集设备的内外参，对多视角图像进行拼接，得到环视拼接图；将所述环视拼接图输入多任务网络，预测出停车位的位置参数；根据停车位的位置参数，确定出停车位类型；基于环视拼接图的畸变拉伸特性，将环视拼接图划分为多个区域，确定出每个区域中的空停车位。该检测方法设计了多任务网络模型，涵盖语义分割、关键点检测和方向角预测，使得预测停车位的信息更全面，并提出了区域划分和多策略融合的方法，对于环视图畸变引起的车位遮挡、非连续的间隔车位以及非垂直的斜向式车位等复杂场景的空停车位检测更稳定有效。

The embodiment of the present invention provides a method for detecting parking spaces for automatic parking, which relates to the field of automatic parking technology. The detection method includes: acquiring multi-view images; stitching the multi-view images based on the internal and external parameters of multiple calibrated image acquisition devices to obtain a surround stitching map; inputting the surround stitching map into a multi-task network to predict the location parameters of the parking space; determining the type of the parking space according to the location parameters of the parking space; dividing the surround stitching map into multiple regions based on the distortion and stretching characteristics of the surround stitching map, and determining the empty parking spaces in each region. The detection method designs a multi-task network model, covering semantic segmentation, key point detection and direction angle prediction, so that the information of the predicted parking space is more comprehensive, and proposes a method of regional division and multi-strategy fusion, which is more stable and effective for detecting empty parking spaces in complex scenes such as parking space occlusion caused by surround view distortion, non-continuous interval parking spaces, and non-vertical oblique parking spaces.

Description

Automatic parking space detection method and device, computer equipment and medium
Technical Field
The present invention relates to the field of automatic parking technologies, and in particular, to a method and apparatus for detecting an automatic parking space, a computer device, and a medium.
Background
With the rapid development of automatic driving technology, automatic parking has become an important component of intelligent automobiles. In an automatic parking system, accurate detection of an empty parking space is a precondition for achieving automatic parking. However, the existing automatic parking system has a certain limitation in the detection of the parking space, and most of the existing automatic parking space detection methods are to obtain the corner points of the parking space or obtain the parking space detection frame through a detection network, so that the position of the parking space is determined. The method is effective for parking spaces in simple scenes, but is often poor in effect for complex scenes such as parking space shielding, discontinuous interval parking spaces, non-vertical oblique parking spaces and the like caused by distortion of the looking-around view.
Patent CN202310670523.6 proposes a single-stage parking space detection method based on a ring view, and the information of left and right corner points of a parking space entry line, the type of a parking space, the occupation condition of the parking space and the like are obtained by constructing a detection network. The network structure designed by the patent belongs to a multi-task network, more information of the parking spaces can be obtained, not only the corner information of the parking spaces can be directly predicted, but also the types of the parking spaces, whether vehicles occupy the parking spaces or not and the like can be predicted, so that the detected empty parking spaces can be directly obtained. However, the end-to-end method for directly predicting the empty parking space through a single stage does not solve the problems of parking space shielding caused by distortion of an looking-around view and detection of non-vertical inclined parking spaces.
Therefore, there is a need to provide a parking space detection method that is also used in non-standard parking spaces or complex environments.
Disclosure of Invention
In view of the above, an embodiment of the present invention provides an automatic parking space detection method based on a ring view, so as to solve the problem that the parking space detection method in the prior art cannot be applied to a complex environment, where the detection method includes:
Acquiring multi-view images according to a plurality of image acquisition devices arranged at different positions of the vehicle;
based on the internal and external parameters of the calibrated multiple image acquisition devices, splicing the multi-view images by an image splicing technology to obtain a circular view splicing diagram with a overlooking view;
Inputting the look-around mosaic into a multi-task network, and predicting the position parameters of a parking space, wherein the position parameters comprise corner information, semantic segmentation information of a borderline, a borderline direction angle and occupation condition information;
determining the width of an entering line of the parking space according to the position parameters of the parking space, and determining the type of the parking space according to the width of the entering line or the direction angle of the borderline;
Based on distortion stretching characteristics of the looking-around spliced graph, dividing the looking-around spliced graph into a plurality of areas, determining a cutoff line of a parking space boundary according to semantic dividing line information of two corner points and the boundary of an entering line in each area, and determining an empty parking space in each area according to different relations of the cutoff line of the parking space boundary.
The automatic parking space detection method provided by the invention is also characterized in that the multi-view images are acquired according to a plurality of image acquisition devices arranged at different positions of the vehicle, and four fisheye cameras are arranged at the front, rear, left and right directions of the vehicle and used for acquiring surrounding images.
The automatic parking space detection method provided by the invention is also characterized in that the looking-around splice diagram is input into a multi-task network, and the position parameter multi-task network of the parking space is predicted, and the method comprises the following steps:
inputting the look-around mosaic image into a multi-task network, obtaining prediction information of each branch through a multi-layer convolutional neural network for the input look-around mosaic image, outputting the corner information through a key point branch of the multi-task network, outputting semantic segmentation information of the boundary line through a segmentation branch of the multi-task network, outputting the boundary line direction angle through a direction angle branch of the multi-task network, and outputting the occupation condition information through an occupation classification branch of the multi-task network.
The automatic parking space detection method provided by the invention is also characterized in that the method for detecting the automatic parking space is used for determining the width of an entering line of the parking space according to the position parameter of the parking space and determining the type of the parking space according to the width of the entering line or the direction angle of the borderline, and comprises the following steps:
calculating the width of the entering line according to the coordinates of the left corner point and the right corner point of the entering line;
Comparing the width of the access line with the preset width and the preset length of the parking space respectively to determine whether the current parking space belongs to a vertical parking space or a parallel parking space or
And judging whether the current parking space belongs to an inclined parking space according to the difference value between the borderline direction angle and 90 degrees, wherein the borderline direction angle is an included angle between an entering line of the parking space and a left boundary line.
The automatic parking space detection method provided by the invention also has the characteristics that the method further comprises the following steps:
If no line exists between the left corner point and the right corner point of the entering line according to the semantic segmentation information of the edge line of the parking space, and the width of the entering line calculated according to the left corner point and the right corner point of the entering line is smaller than the preset width of the parking space, two parking spaces to which the left corner point and the right corner point of the entering line respectively belong are determined to be interval parking spaces.
The automatic parking space detection method provided by the invention is also characterized in that the method divides the looking-around splice diagram into a plurality of areas based on the distortion stretching characteristic of the looking-around splice diagram, and comprises the following steps:
The method comprises the steps that the vehicle is always located in the center of the all-around spliced image, the all-around spliced image is divided into four areas, a first area is located in front of the left side of the vehicle, a second area is located in front of the right side of the vehicle, a third area is located in rear of the right side of the vehicle, and a fourth area is located in rear of the left side of the vehicle.
The automatic parking space detection method provided by the invention is also characterized in that in each region, according to the semantic parting line information of two corner points and the edge line of the entering line, the broken line of the edge line of the parking space is determined, and according to the judging modes of different relations of the broken line of the edge line of the parking space, the empty parking space in each region is determined, and the method comprises the following steps:
judging the area falling into more than w/2 of the width of the entering line of the current parking space as the area to which the current parking space belongs;
in the first region and the third region, a parking space in which a broken line of a parking space borderline satisfies the following formula is determined as an empty parking space:
l0>max(l1,c)
wherein c is a preset length of the parking space, l 0 is a cutoff line of a left boundary line of the parking space, and l 1 is a cutoff line of a right boundary line of the parking space;
In the second region and the fourth region, a parking space in which a broken line of a parking space borderline satisfies the following formula is determined as an empty parking space:
l1>max(l0,c)。
a second object of the present invention is to provide an automatic parking space detection apparatus including:
the image acquisition module is used for acquiring multi-view images of different positions of the vehicle;
The image processing module is used for splicing the multi-view images through an image splicing technology based on the internal and external parameters of the calibrated multiple image acquisition devices to obtain a look-around splicing image with a overlooking view;
The computing module is used for outputting the look-around splice graph to a multi-task network, predicting the position parameters of the parking space and determining the type of the parking space according to the position parameter information;
the judging module is used for dividing the looking-around splice graph into a plurality of areas based on the distortion stretching characteristics of the looking-around splice graph, and determining the empty parking spaces of each area.
A third object of the present invention is to provide a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the method for detecting an automatic parking space according to any one of the preceding claims when executing the computer program.
A fourth object of the present invention is to provide a computer-readable storage medium storing a program for executing the automatic parking space detection method according to any one of the foregoing.
Compared with the prior art, the technical scheme provided by the embodiment of the invention has the advantages that a multi-task network model is designed, semantic segmentation, key point detection and direction angle prediction are covered, so that information of a predicted parking space is more comprehensive, and a method for regional division and multi-strategy fusion is provided, so that parking space shielding, discontinuous interval parking spaces, non-vertical oblique parking spaces and other complex scenes caused by distortion of an looking-around view are detected more stably and effectively.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a parking space detection method provided by an embodiment of the present invention;
FIG. 2 is a block diagram of empty parking space detection based on a ring view in an embodiment of the present invention;
FIG. 3 is a block diagram of a multi-tasking network provided by an embodiment of the present invention;
FIG. 4 is a schematic view of three common parking space types according to an embodiment of the present invention;
FIG. 5 is a schematic view of the structure of two special parking space types in an embodiment of the invention;
FIG. 6 is a schematic diagram of a space region error detection bit parking space of a space parking space in an embodiment of the invention;
FIG. 7 is a schematic view of a region division of a ring view in an embodiment of the invention;
FIG. 8 is a schematic illustration of empty space occlusion for different areas of FIG. 7;
figure 9 is a schematic illustration of the semantic segmentation borderline of the empty parking space of region I in figure 7,
FIG. 10 is a block diagram of a computer device according to an embodiment of the present invention;
fig. 11 is a block diagram of an automatic parking space detection device according to an embodiment of the present invention.
Detailed Description
Embodiments of the present application will be described in detail below with reference to the accompanying drawings.
Other advantages and effects of the present application will become apparent to those skilled in the art from the following disclosure, which describes the embodiments of the present application with reference to specific examples. It will be apparent that the described embodiments are only some, but not all, embodiments of the application. The application may be practiced or carried out in other embodiments that depart from the specific details, and the details of the present description may be modified or varied from the spirit and scope of the present application. It should be noted that the following embodiments and features in the embodiments may be combined with each other without conflict. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
In an embodiment of the present invention, there is provided an automatic parking space detection method based on a ring view, as shown in fig. 1, the detection method includes:
s1, acquiring multi-view images according to a plurality of image acquisition devices arranged at different positions of a vehicle;
S2, based on the internal and external parameters of the calibrated multiple image acquisition devices, splicing the multi-view images through an image splicing technology to obtain a look-around splicing image with a overlooking view;
S3, inputting the look-around mosaic into a multi-task network, and predicting position parameters of a parking space, wherein the position parameters comprise corner information, semantic segmentation information of a boundary, a boundary direction angle and occupation condition information;
S4, determining the width of an entering line of the parking space according to the position parameters of the parking space, and determining the type of the parking space according to the width of the entering line or the direction angle of the borderline;
S5, dividing the look-around splice diagram into a plurality of areas based on distortion stretching characteristics of the look-around splice diagram, determining a broken line of a parking space boundary according to semantic dividing line information of two corner points and the boundary of an entering line in each area, and determining an empty parking space in each area according to different relations of the broken line of the parking space boundary.
In the above embodiment, each information is listed as follows:
(1) And the corner information of the parking space, namely the coordinates of four corners of the parking space and the confidence scores corresponding to the corners, and reflecting that the corners are shielded when the corner confidence scores are lower. In general, the confidence score is higher for two corner points near the entry line of the parking space, and lower for two corner points behind the parking space.
(2) The occupancy information of the parking space, namely a probability value output by occupancy classification branches in the multi-task network, needs to be compared with a threshold value to reflect whether the parking space is occupied by a vehicle or not, and is used for obtaining a preliminary empty parking space.
(3) The segmentation information of the parking spaces is semantic segmentation lines corresponding to four edges of the parking spaces, and is used for assisting in judging continuous parking spaces and interval parking spaces and also used for assisting in judging empty space detection of parking space shielding caused by distortion of a circular view.
(4) The direction angle information of the parking space, namely the included angle between the entering line and the boundary line of the parking space, is defined as the included angle at the left side of the parking space.
The technical scheme provided by the embodiment obtains more information of the parking spaces based on the multi-task network, combines the characteristics of the look-around map to construct region division, and fuses various logic strategies so that the empty parking spaces can be detected more accurately and the application range is wide.
In some embodiments, the acquiring multi-view images according to the plurality of image acquisition devices arranged at different positions of the vehicle includes installing four fisheye cameras at four directions of front, rear, left and right of the vehicle for acquiring images of four sides.
In some embodiments, inputting the look-around mosaic into a multi-task network predicts a location parameter of a parking space, including:
inputting the look-around mosaic image into a multi-task network, obtaining prediction information of each branch through a multi-layer convolutional neural network for the input look-around mosaic image, outputting the corner information through a key point branch of the multi-task network, outputting semantic segmentation information of the boundary line through a segmentation branch of the multi-task network, outputting the boundary line direction angle through a direction angle branch of the multi-task network, and outputting the occupation condition information through an occupation classification branch of the multi-task network.
In the above embodiment, as shown in fig. 2, a multi-task network is constructed, and for the input looking-around spliced image, the prediction information of each branch is obtained through a multi-layer convolutional neural network (including a backhaul network structure and a Head network structure). Wherein:
(1) The key point branches consist of a parking space corner output head and a parking space corner position branch output head. The parking space corner output head is used for obtaining coordinates of four corners of a parking space, the parking space corner confidence score output head is used for obtaining confidence scores of the corners, the reliability of the corners can be measured, and the visibility information of the corners can be obtained by whether the threshold is met or not.
(2) The segmentation branches are 'edge semantic segmentation output heads'. The method is used for detecting borderlines (including entry lines and boundary lines) of the parking space and obtaining semantic segmentation information of the borderlines.
(3) The direction angle branch is a "sideline direction angle output head". The angle between the boundary line and the entrance line is measured, and the angle can be used for assisting in judging the type of the parking space.
(4) The occupation classification branch is a parking space occupation classification output head. The method is used for obtaining preliminary empty parking space information, and a more accurate empty parking space can be obtained by combining the regional division strategy.
For the key point branches, 4 corner points of the parking space are defined, the order of the 4 corner points is, starting from the left end point of the parking space entry line in the clockwise direction, an upper left corner point, an upper right corner point, a lower right corner point and a lower left corner point, each point is represented as ((x0,y0,v0),(x1,y1,v1),(x2,y2,v2),(x3,y3,v3)),, x i、yi is the horizontal and vertical coordinates of an image of the ith key point respectively, v i is a visible attribute, 0 represents that the key point is invisible, and 1 represents that the key point is visible.
Let the predicted value of the "parking space corner output head" be z a, train by MSE loss function, the specific formula is as follows:
Where N is the training sample size and z gt is the truth label of the parking space corner points in the training sample, and its form is composed of the truth coordinates (x 0,y0,x1,y1,x2,y2,x3,y3) of 4 corner points. In practical training, z gt can be converted into a form of offset with the difference value of the anchor point through the coding process, then the predicted z a is also an offset, and the offset and the anchor point are added through the subsequent decoding process to obtain the final predicted coordinate value.
Assuming that the predicted value of the parking space angular point position component output head is s a, training by adopting a Smooth L1 loss function, and the specific formula is as follows:
S gt is a confidence score value tag of the parking space corner point in the training sample, and the confidence score value tag is formed by (v 0,v1,v2,v3). In actual training, the confidence score v i (i=1, 2,3, 4) has a value of 0 or 1, and the confidence score in prediction is a probability value between 0 and 1, and needs to be compared with a threshold value, and then the confidence score v i is used for judging the visibility of the corner points.
For a segmentation branch, a predicted value of an edge semantic segmentation output head is set as X, and the prediction value is trained by using a Dice Loss training, wherein the specific formula is as follows:
Where X is the set of pixel classes of the predicted segmented image, Y is the set of pixel classes of the true segmented image, |X n Y| represents the common element between sets X and Y, |X| represents the number of elements in set X, and|Y| represents the number of elements in set Y. The above formula can be used to measure the overlap between the predicted semantic segmentation pixel X and the true semantic segmentation pixel Y, which takes a value between 0 and 1. When the predicted semantic segmentation result is closer to the real semantic segmentation case, the Dice Loss is closer to 0.
For the direction angle branch, the predicted value of the 'edge direction angle output head' is set as p a, the Smooth L1 loss function is also adopted for training, and the formula is the same as the loss function formula of the confidence score, and the formula is expressed as:
wherein p gt is the included angle between the parking space entry line and the boundary line, which is defined as the included angle with the left end point of the entry line as the vertex and the two edges of the entry line and the left boundary line, and the true value label is generally calculated by the labeled parking space corner point. When the corner points of the parking spaces are not complete and the true value of the included angles cannot be obtained, the true value label is obtained by calculating the semantic dividing line of the borderline.
For the occupation classification branch, a predicted value of a parking space occupation classification output head is set as y a, and a cross entropy loss function is adopted for training, wherein the specific formula is as follows:
H(ya)＝-(ygtlog(ya)+(1-ygt)log(1-ya))
Wherein y gt is the true value of the parking space occupation, the occupation classification value is 0 or 1,0 indicates that the parking space is unoccupied (i.e. an empty parking space), and 1 indicates that the parking space is occupied. The predicted value y a is a probability value between 0 and 1, and then needs to be compared with a threshold value to determine whether the predicted current parking space is occupied by the vehicle.
The total loss function is defined as:
Loss＝α1MSE(za)+α2SLL1(sa)+α3DL(X)+α4SLL1(pa)+α5H(ya)
where α 1、α2、α3、α4、α5 represents the weight ratio of each lost task, and the weight value is set by an empirical value at the time of training of the actual model.
In some embodiments, as shown in figure 4,
Determining the width of an entering line of the parking space according to the position parameter of the parking space, and determining the type of the parking space according to the width of the entering line or the direction angle of the borderline, wherein the method comprises the following steps:
calculating the width of the entering line according to the coordinates of the left corner point and the right corner point of the entering line;
Comparing the width of the access line with the preset width and the preset length of the parking space respectively to determine whether the current parking space belongs to a vertical parking space or a parallel parking space or
And judging whether the current parking space belongs to an inclined parking space according to the difference value between the borderline direction angle and 90 degrees, wherein the borderline direction angle is an included angle between an entering line of the parking space and a left boundary line.
In some embodiments, four corner points of the parking space are ((x0,y0,v0),(x1,y1,v1),(x2,y2,v2),(x3,y3,v3)),, where (x i,yi) represents the corner coordinates and v i represents the corner confidence score. According to the corner sequence defined above, the coordinates of the left and right end points of the entry line of the parking space are (x 0,y0) and (x 1,y1), so that the entry line width w of the parking space can be obtained, and the calculation formula is:
The prior parking space width w 0 and the length d 0 are preset, and according to the calculated specific value of the width w of the entering line, the specific value is compared with the difference values of w 0 and d 0 respectively, so that whether the current parking space belongs to a vertical parking space or a parallel parking space can be determined.
Note that, the borderline direction angle obtained above is θ, and the predicted value of θ is around 90 degrees in the case of a vertical parking space and a parallel parking space, and the predicted value of θ is greatly different from 90 degrees in the case of an oblique parking space. It is then possible to determine whether the current parking space is an oblique parking space by means of the borderline direction angle.
Therefore, the type of the parking space can be judged according to the width w of the parking space entering line and the boundary included angle theta. In order to increase the robustness of judgment, according to the parking space similarity principle, W and theta of a plurality of parking spaces near the current parking space are counted and respectively stored in a set W and theta, outliers and outliers are removed, and an average value is calculated after filteringAndAnd judging the type of the parking space by the method.
The detection method further comprises the following steps:
If no line exists between the left corner point and the right corner point of the entering line according to the semantic segmentation information of the edge line of the parking space, and the width of the entering line calculated according to the left corner point and the right corner point of the entering line is smaller than the preset width of the parking space, two parking spaces to which the left corner point and the right corner point of the entering line respectively belong are determined to be interval parking spaces.
For example, in a scene such as a warehouse, a space parking space is seen in addition to a conventional contiguous parking space. Such spaced parking spaces often occur in garages due to the isolation of the garage support posts. As shown in fig. 5:
In order to prevent the middle space region of the space parking space from being erroneously recognized as the parking space, as shown in fig. 6, judgment and filtering are required in combination with the prior width w 0 of the parking space and the semantic division line of the parking space. First, according to the left and right end points (x 0,y0) and (x 1,y1) of the parking space entry line, in combination with the parking space semantic division line predicted by the multi-task network, it is possible to obtain that no line exists between the point (x 0,y0) and the point (x 1,y1). Second, the width w of the incoming line is calculated from the point (x 0,y0) and the point (x 1,y1) to be significantly smaller than the a priori width w 0. By combining the two conditions, the false detection situation of the interval area of the interval parking space can be filtered.
In the above embodiment, the coordinates of the parking space, the width of the entry line, the type of the parking space, and the like may be obtained through the parking space corner information, the division information, and the direction corner information obtained by the multitasking network. In addition, the multitasking network predicts the occupation situation of the parking spaces, namely whether each parking space in the circular view is occupied by a vehicle or not can be obtained, and the multitasking network can be used for preliminarily judging whether the parking space is an empty parking space or not.
In some embodiments, the ring view is obtained by converting the images of the fish-eye cameras around the vehicle into a top view angle for stitching, so that objects with a high height in the original image are distorted and stretched in the ring view, and the phenomenon that the stretched vehicle shields a far parking space in the ring view can be caused, and the missed detection of an empty parking space can be caused. According to the characteristic of the ring view, namely that the vehicle is always positioned in the center of the ring view, the ring view can be divided into areas, and the ring view is used for treating the empty vehicle position missing detection phenomenon caused by the parking space shielding caused by the distortion of the ring view.
In some embodiments, as shown in fig. 3, the dividing the view-around mosaic into a plurality of regions based on the distortion stretching characteristics of the view-around mosaic includes:
The method comprises the steps that the vehicle is always located in the center of the all-around spliced image, the all-around spliced image is divided into four areas, a first area is located in front of the left side of the vehicle, a second area is located in front of the right side of the vehicle, a third area is located in rear of the right side of the vehicle, and a fourth area is located in rear of the left side of the vehicle.
In some embodiments, in each area, determining a cutoff line of a parking space boundary according to semantic division line information of two corner points and the boundary of an entry line, and determining an empty parking space in each area according to a judgment mode of different relationships of the cutoff lines of the parking space boundary, including:
Judging that the area falling into more than w/2 of the width of the entering line of the current parking space is the area to which the current parking space belongs;
in the first region and the third region, a parking space in which a broken line of a parking space borderline satisfies the following formula is determined as an empty parking space:
l0>max(l1,c)
wherein c is a preset length of the parking space, l 0 is a cutoff line of a left boundary line of the parking space, and l 1 is a cutoff line of a right boundary line of the parking space;
In the second region and the fourth region, a parking space in which a broken line of a parking space borderline satisfies the following formula is determined as an empty parking space:
l1>max(l0,c)。
As in the previous embodiment, the region division of the ring view is shown in FIG. 7:
And judging the empty parking spaces by using different logics for the divided areas I, II, III and IV, and judging by mainly combining the coordinates of the entry line end points of the parking spaces and the semantic dividing lines of the borderlines. The situation of parking space shielding caused by vehicle distortion in four areas is sketched as shown in fig. 8:
The area attribution of the parking spaces is defined, for any parking space, the width w of the entering line is calculated according to two end points (x 0,y0) and points (x 1,y1) of the entering line, and when the entering line has a length exceeding w/2 line segments and falls into a certain area, the parking space is defined as belonging to the area. That is, it is determined to which region R i (i=i, II, III, IV) the parking space belongs, and one of the following two conditions needs to be satisfied:
①(x0,y0)∈Ri And (2) and
②(x1,y1)∈Ri And (2) and
For parking spaces in region R i (i=i, II, III, IV), a cut-off line of the parking space borderline can be further obtained by combining the two end points of the entry line and the borderline semantic division line. Taking area I as an example, the broken line of the parking space borderline is shown in figure 9,
Judging the empty parking spaces in the area I, the following formula needs to be satisfied:
l0>max(l1,c)
The constant c is generally taken as 1/3 of the prior length of the parking space, is used for representing the threshold lower limit of the longest borderline of the current parking space, and can be adjusted and set according to experiments, wherein c=d 0/3, and d 0 is the prior length of the parking space.
Similarly, for region II, according to the definition of the left and right end sequences of the entry line, at this time, l 1 is the long side and l 0 is the short side, and the formula for determining that the empty parking space needs to be satisfied is:
l1>max(l0,c)
for region III, the formula for judging the empty parking space is the same as that of region I, and for region IV, the formula for judging the empty parking space is the same as that of region II.
And combining the logic judgment of the regional division and the judgment result of the parking space occupation information output by the multitasking network, so as to obtain all the empty parking spaces in the current ring view.
In the foregoing embodiments, accurate and efficient detection of empty parking spaces is critical to constructing a map of parking spaces, which is a critical aspect in automated parking. According to the method, more parking space information is obtained through constructing the multi-task network, then the position, the entrance line width, the edge orientation angle, the type of the parking space and the like of the parking space are obtained based on the logic strategy, and then the comprehensive position result of the empty parking space is obtained through combining region division, so that the detection result is more robust and accurate, and the method is suitable for parking space detection of complex scenes.
In some embodiments, a computer device is provided, as shown in fig. 10, including a memory 201, a processor 202, and a computer program stored in the memory 201 and executable on the processor 202, where the processor 202 implements the automatic parking space detection method of any of the above embodiments when executing the computer program.
In particular, the computer device may be a computer terminal, a server or similar computing means.
In some embodiments, a computer readable storage medium is provided, where the computer readable storage medium stores instructions for performing the automatic parking space detection method according to any of the above embodiments.
In particular, computer-readable storage media, including both permanent and non-permanent, removable and non-removable media, may be used to implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer-readable storage media include, but are not limited to, phase-change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Disks (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable storage media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated data signals and carrier waves.
Based on the same inventive concept, the embodiment of the invention also provides an automatic parking space detection device, as described in the following embodiment. Because the principle of solving the problem of the automatic parking space detection device is similar to that of an automatic parking space detection method, the implementation of the automatic parking space detection device can refer to the implementation of the automatic parking space detection method, and the repetition is omitted. As used below, the term "unit" or "module" may be a combination of software and/or hardware that implements the intended function. While the means described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware, is also possible and contemplated.
An automatic parking space detection device is shown in fig. 11, and comprises an image acquisition module, an image processing module, a calculation module and a judgment module.
An image acquisition module 301, configured to acquire multi-view images of different positions of the vehicle;
The image processing module 302 is configured to stitch the multi-view images through an image stitching technology based on the internal and external parameters of the calibrated multiple image acquisition devices, so as to obtain a look-around stitching graph with a top view;
the computing module 303 is configured to output the look-around mosaic to a multi-task network, predict a position parameter of a parking space, and determine a type of the parking space according to the position parameter information;
The judging module 304 is configured to divide the looking-around mosaic into a plurality of regions based on the distortion stretching characteristic of the looking-around mosaic, and determine an empty parking space in each region.
It will be apparent to those skilled in the art that the modules or steps of the embodiments of the invention described above may be implemented in a general purpose computing device, they may be concentrated on a single computing device, or distributed across a network of computing devices, they may alternatively be implemented in program code executable by computing devices, so that they may be stored in a storage device for execution by computing devices, and in some cases, the steps shown or described may be performed in a different order than what is shown or described, or they may be separately fabricated into individual integrated circuit modules, or a plurality of modules or steps in them may be fabricated into a single integrated circuit module. Thus, embodiments of the invention are not limited to any specific combination of hardware and software.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, and various modifications and variations can be made to the embodiments of the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (10)

1.一种自动泊车停车位检测方法，其特征在于，所述检测方法包括：1. A method for detecting an automatic parking space, characterized in that the detection method comprises: 根据设置在自车不同位置的多个图像采集设备获取多视角图像；Acquire multi-view images using multiple image acquisition devices arranged at different positions of the vehicle; 基于标定的多个图像采集设备的内外参，通过图像拼接技术对多视角图像进行拼接，得到俯视视角的环视拼接图；Based on the calibrated internal and external parameters of multiple image acquisition devices, the multi-view images are stitched together using image stitching technology to obtain a surround stitching image from a bird's-eye view. 将所述环视拼接图输入多任务网络，预测出停车位的位置参数，所述位置参数包括角点信息、边线的语义分割信息、边线方向角以及占用情况信息；Inputting the surround stitching image into a multi-task network to predict the position parameters of the parking space, wherein the position parameters include corner point information, semantic segmentation information of the edge line, edge line direction angle, and occupancy information; 根据停车位的位置参数，确定停车位的进入线宽度，根据所述进入线宽度或边线方向角确定出停车位类型；Determine the entry line width of the parking space according to the location parameters of the parking space, and determine the parking space type according to the entry line width or the sideline direction angle; 基于环视拼接图的畸变拉伸特性，将环视拼接图划分为多个区域，在每个区域中，根据进入线的两个角点和边线的语义分割线信息，确定出停车位边线的截断线，并根据停车位边线的截断线的不同关系，确定出每个区域中的空停车位。Based on the distortion and stretching characteristics of the surround view mosaic map, the surround view mosaic map is divided into multiple regions. In each region, the truncation line of the parking space edge line is determined according to the semantic segmentation line information of the two corner points and edge line of the entry line. According to the different relationships of the truncation lines of the parking space edge line, the empty parking spaces in each region are determined. 2.根据权利要求1所述的自动泊车停车位检测方法，其特征在于，所述根据设置在自车不同位置的多个图像采集设备获取多视角图像，包括在自车的前后左右四个方位安装四个鱼眼相机用于获取四周的图像。2. The automatic parking space detection method according to claim 1 is characterized in that the multi-view images are obtained based on multiple image acquisition devices arranged at different positions of the vehicle, including installing four fisheye cameras at four positions of the vehicle to obtain images of the surrounding areas. 3.根据权利要求1所述的自动泊车停车位检测方法，其特征在于，所述将所述环视拼接图输入多任务网络，预测出停车位的位置参数多任务网络，包括：3. The automatic parking space detection method according to claim 1, characterized in that the step of inputting the surround view mosaic image into a multi-task network to predict the location parameters of the parking space comprises: 将所述环视拼接图输入多任务网络，对于输入的环视拼接图像，经过多层卷积神经网络得到各分支的预测信息通过所述多任务网络的关键点分支输出所述角点信息，通过所述多任务网络的分割分支输出所述边线的语义分割信息，通过所述多任务网络的方向角分支输出所述边线方向角，通过所述多任务网络的占用分类分支输出所述占用情况信息。The surround view stitching image is input into a multi-task network. For the input surround view stitching image, prediction information of each branch is obtained through a multi-layer convolutional neural network, the corner point information is output through the key point branch of the multi-task network, the semantic segmentation information of the edge line is output through the segmentation branch of the multi-task network, the edge line direction angle is output through the direction angle branch of the multi-task network, and the occupancy situation information is output through the occupancy classification branch of the multi-task network. 4.根据权利要求1所述的自动泊车停车位检测方法，其特征在于，所述根据停车位的位置参数，确定停车位的进入线宽度，根据所述进入线宽度或边线方向角确定出停车位类型，包括：4. The automatic parking space detection method according to claim 1, characterized in that the step of determining the entry line width of the parking space according to the position parameters of the parking space, and determining the parking space type according to the entry line width or the sideline direction angle comprises: 根据进入线的左右两个角点的坐标计算所述进入线宽度；Calculate the width of the entry line according to the coordinates of the left and right corner points of the entry line; 将所述进入线宽度分别与停车位的预设宽度和预设长度进行对比，以确定出当前停车位属于垂直式停车位还是平行式停车位；或者，Comparing the entry line width with a preset width and a preset length of the parking space respectively to determine whether the current parking space is a vertical parking space or a parallel parking space; or, 根据所述边线方向角与90度的差值大小，判断当前停车位是否属于斜向式停车位，其中，所述边线方向角为车位的进入线与左边界线的夹角。According to the difference between the sideline direction angle and 90 degrees, it is determined whether the current parking space is an oblique parking space, wherein the sideline direction angle is the angle between the entrance line and the left boundary line of the parking space. 5.根据权利要求1所述的自动泊车停车位检测方法，其特征在于，还包括：5. The automatic parking space detection method according to claim 1, further comprising: 若根据停车位的所述边线的语义分割信息确定出进入线的左右两个角点之间不存在线条，且，根据所述进入线的左右两个角点计算出的进入线的宽度小于停车位的预设宽度，则确定出所述进入线的左右两个角点分别所属的两个停车位为间隔停车位。If it is determined based on the semantic segmentation information of the edge line of the parking space that there is no line between the left and right corner points of the entry line, and the width of the entry line calculated based on the left and right corner points of the entry line is smaller than the preset width of the parking space, then it is determined that the two parking spaces to which the left and right corner points of the entry line respectively belong are interval parking spaces. 6.根据权利要求1-5任一项所述的自动泊车停车位检测方法，其特征在于，所述基于环视拼接图的畸变拉伸特性，将环视拼接图划分为多个区域，包括：6. The automatic parking space detection method according to any one of claims 1 to 5, characterized in that the surround view splicing image is divided into a plurality of regions based on the distortion stretching characteristics of the surround view splicing image, including: 以所述自车始终位于所述环视拼接图的中央，将所述环视拼接图划分为四个区域，位于所述自车的左前方的为第一区域、位于所述自车的右前方的为第二区域、位于所述自车的右后方的为第三区域以及位于所述自车的左后方的为第四区域。Assuming that the vehicle is always located in the center of the surround view mosaic map, the surround view mosaic map is divided into four areas: the first area is located in the left front of the vehicle, the second area is located in the right front of the vehicle, the third area is located in the right rear of the vehicle, and the fourth area is located in the left rear of the vehicle. 7.根据权利要求5所述的自动泊车停车位检测方法，其特征在于，所述在每个区域中，根据进入线的两个角点和边线的语义分割线信息，确定出停车位边线的截断线，并根据停车位边线的截断线的不同关系的判断方式，确定出每个区域中的空停车位，包括：7. The automatic parking space detection method according to claim 5, characterized in that, in each area, according to the semantic segmentation line information of the two corner points of the entry line and the edge line, the truncation line of the parking space edge line is determined, and according to the judgment method of different relationships of the truncation lines of the parking space edge line, the empty parking space in each area is determined, including: 判断当前停车位的进入线宽度的w/2以上落入的区域为当前停车位属于的区域；Determine that the area where the width w /2 or more of the entry line of the current parking space falls is the area to which the current parking space belongs; 在所述第一区域和所述第三区域中，将停车位边线的截断线满足以下公式的停车位确定为空停车位：In the first area and the third area, a parking space whose truncation line of the parking space boundary line satisfies the following formula is determined as an empty parking space: l0>max(l1,)l 0 >max(l 1 ,) 其中，c为停车位的预设长度，l0为停车位的左边界线的截断线，l1为停车位的右边界线的截断线；Wherein, c is the preset length of the parking space, l 0 is the cutoff line of the left boundary line of the parking space, and l 1 is the cutoff line of the right boundary line of the parking space; 在所述第二区域和所述第四区域中，将停车位边线的截断线满足以下公式的停车位确定为空停车位：In the second area and the fourth area, a parking space whose truncation line of the parking space boundary line satisfies the following formula is determined as an empty parking space: l1>max(l0,)。l 1 >max(l 0 ,). 8.一种自动泊车停车位检测装置，其特征在于，包括：8. An automatic parking space detection device, characterized by comprising: 图像获取模块，用于获取自车不同位置的多视角图像；An image acquisition module is used to acquire multi-view images of different positions of the vehicle; 图像处理模块，用于基于标定的多个图像采集设备的内外参，通过图像拼接技术对多视角图像进行拼接，得到俯视视角的环视拼接图；An image processing module is used to stitch the multi-view images by image stitching technology based on the internal and external parameters of the calibrated multiple image acquisition devices to obtain a surround stitching image of a bird's-eye view; 计算模块，用于将所述环视拼接图输出多任务网络，预测出停车位的位置参数，根据位置参数信息确定出停车位类型；A calculation module, used for outputting the surround stitching graph to a multi-task network, predicting the position parameters of the parking space, and determining the type of the parking space according to the position parameter information; 判断模块，用于基于环视拼接图的畸变拉伸特性，将环视拼接图划分为多个区域，确定出每个区域的空停车位。The judgment module is used to divide the surround view splicing image into multiple areas based on the distortion and stretching characteristics of the surround view splicing image, and determine the empty parking spaces in each area. 9.一种计算机设备，包括存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序，其特征在于，所述处理器执行所述计算机程序时实现权利要求1至7中任一项所述的自动泊车停车位检测方法。9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the automatic parking space detection method according to any one of claims 1 to 7 when executing the computer program. 10.一种计算机可读存储介质，其特征在于，所述计算机可读存储介质存储有执行权利要求1至7中任一项所述的自动泊车停车位检测方法。10 . A computer-readable storage medium, characterized in that the computer-readable storage medium stores instructions for executing the automatic parking space detection method according to any one of claims 1 to 7.
CN202411263828.6A 2024-09-10 2024-09-10 Automatic parking space detection method, device, computer equipment and medium Pending CN119229415A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202411263828.6A CN119229415A (en) 2024-09-10 2024-09-10 Automatic parking space detection method, device, computer equipment and medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202411263828.6A CN119229415A (en) 2024-09-10 2024-09-10 Automatic parking space detection method, device, computer equipment and medium

Publications (1)

Publication Number Publication Date
CN119229415A true CN119229415A (en) 2024-12-31

Family

ID=93942764

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202411263828.6A Pending CN119229415A (en) 2024-09-10 2024-09-10 Automatic parking space detection method, device, computer equipment and medium

Country Status (1)

Country Link
CN (1) CN119229415A (en)

Similar Documents

Publication Publication Date Title
US11380104B2 (en) Method and device for detecting illegal parking, and electronic device
CN111754394B (en) Method and device for detecting object in fisheye image and storage medium
US9311533B2 (en) Device and method for detecting the presence of a logo in a picture
CN111709343A (en) Point cloud detection method and device, computer equipment and storage medium
CN110516514B (en) Modeling method and device of target detection model
CN113901961B (en) Parking space detection method, device, equipment and storage medium
CN111967396A (en) Processing method, device and equipment for obstacle detection and storage medium
CN112906471A (en) Traffic signal lamp identification method and device
CN113191221B (en) Vehicle detection method and device based on panoramic camera and computer storage medium
CN111724607B (en) Steering lamp use detection method and device, computer equipment and storage medium
CN111292331B (en) Image processing method and device
CN113312949A (en) Video data processing method, video data processing device and electronic equipment
CN115346193A (en) Parking space detection method and tracking method thereof, parking space detection device, parking space detection equipment and computer readable storage medium
CN118823071A (en) River video flow velocity identification method, device, storage medium and program product
CN119229415A (en) Automatic parking space detection method, device, computer equipment and medium
CN110796129A (en) Text line region detection method and device
CN114820748B (en) Full-view-angle characteristic intelligent evidence obtaining method based on panoramic data
CN114494968B (en) A method and device for identifying abnormal intrusion behavior in a restricted area
CN116433893A (en) Target detection method, system, medium and equipment based on arbitrary direction rotating frame
CN115578669A (en) Method and device for identifying camera lens foreign matters
CN104809438A (en) Method and device for detecting electronic eyes
CN114399733A (en) Image detection method, device, equipment and storage medium
CN103729619B (en) The automatic identifying method of car plate frame in preventing road monitoring system
CN114639084A (en) Road side end vehicle sensing method based on SSD (solid State disk) improved algorithm
CN113147746A (en) Method and device for detecting ramp parking space

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination