CN114120259A - Empty parking space identification method and system, computer equipment and storage medium - Google Patents
Empty parking space identification method and system, computer equipment and storage medium Download PDFInfo
- Publication number
- CN114120259A CN114120259A CN202010900989.7A CN202010900989A CN114120259A CN 114120259 A CN114120259 A CN 114120259A CN 202010900989 A CN202010900989 A CN 202010900989A CN 114120259 A CN114120259 A CN 114120259A
- Authority
- CN
- China
- Prior art keywords
- points
- boundary
- parking space
- empty
- point
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 48
- 230000011218 segmentation Effects 0.000 claims abstract description 49
- 238000012216 screening Methods 0.000 claims abstract description 19
- 238000012545 processing Methods 0.000 claims abstract description 12
- 238000004590 computer program Methods 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 4
- 230000001105 regulatory effect Effects 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 6
- 238000001514 detection method Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 3
- 230000004927 fusion Effects 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G06T3/047—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformation in the plane of the image
- G06T3/40—Scaling the whole image or part thereof
- G06T3/4038—Scaling the whole image or part thereof for image mosaicing, i.e. plane images composed of plane sub-images
Abstract
The invention relates to an empty parking space identification method and a system, computer equipment and a storage medium thereof, wherein the method comprises the following steps: responding to the received starting identification signal, and acquiring a panoramic image around the current vehicle; performing semantic segmentation recognition on the panoramic image by using a pre-trained semantic segmentation model to obtain a drivable area; carrying out regularization processing on the boundary of the travelable area to obtain boundary information, wherein the boundary information comprises a plurality of boundary points; screening the plurality of boundary points according to a preset screening rule to obtain a plurality of empty parking space angular point preselected points; randomly selecting any 4 points as a group of preselected points to enumerate the multiple empty parking space angular point preselected points to obtain multiple groups of preselected points, respectively judging whether multiple parking spaces formed by the multiple groups of preselected points meet preset parking space conditions, generating corresponding candidate empty parking spaces according to one or more groups of preselected points meeting the parking space conditions, and determining the finally output target empty parking spaces according to the candidate empty parking spaces. The invention can solve the technical problems of poor robustness and accuracy in the existing empty parking space identification.
Description
Technical Field
The invention relates to the technical field of automatic parking, in particular to an empty parking space identification method and system, computer equipment and a storage medium.
Background
The empty parking space identification is the basis of the automatic parking technology, and most of parking space detection methods adopted in automatic parking systems can be divided into three methods, namely a camera-based parking space detection method, an ultrasonic radar-based parking space detection method, a camera and ultrasonic radar fusion-based parking space detection method and the like. The parking space detection method based on the camera can only detect the line parking space, namely the parking space is a regular marking parking space; the method for detecting the empty parking space based on the ultrasonic radar can sense the non-linear parking space (space parking space), but the method is difficult to obtain accurate information such as the size, the posture and the like of the parking space, and has poor robustness and accuracy in application; the method based on the fusion of the camera and the ultrasonic radar integrates the advantages of the former two methods, can detect the line parking space and the space parking space at the same time, but the identification of the space parking space still has the defects of poor robustness and accuracy.
Disclosure of Invention
The invention aims to provide an empty parking space identification method and system, computer equipment and a computer readable storage medium thereof for automatic parking, so as to solve the technical problems of poor robustness and accuracy in the existing empty parking space identification.
In order to achieve the above object, according to a first aspect, an embodiment of the present invention provides an empty space identification method, including:
step S1, responding to the received start identification signal, and acquiring a panoramic image of the surrounding environment of the current vehicle;
step S2, performing semantic segmentation and identification on the panoramic image by using a pre-trained semantic segmentation model to obtain a drivable area;
step S3, carrying out regularization processing on the boundary of the travelable area to obtain boundary information, wherein the boundary information comprises a plurality of boundary points;
s4, screening the boundary points according to a preset screening rule to obtain a plurality of empty parking space corner point preselected points;
step S5, enumerating the multiple vacant parking space angular point preselected points by randomly selecting any 4 points as a group of preselected points to obtain multiple groups of preselected points, respectively judging whether multiple parking spaces formed by the multiple groups of preselected points meet preset parking space conditions, generating corresponding candidate vacant parking spaces according to one or more groups of preselected points meeting the parking space conditions, and determining a target vacant parking space to be finally output according to the candidate vacant parking spaces.
Optionally, the extracting boundary information of the travelable region includes:
extracting the boundary of the travelable area to form a closed curve;
expanding 1 pixel point into the closed curve to form a boundary search area;
selecting any 1 adjacent pixel point on the boundary search area as an initial search point of boundary search;
performing boundary search in the boundary search area according to the initial search point; in the searching process, calculating and searching the cost distance of other pixel points adjacent to the current searching point according to a preset cost function, and selecting the pixel point with the minimum cost distance as the next searching point to continue searching until the initial searching point is searched;
and carrying out regularization processing on a plurality of search points searched by the boundary to obtain the boundary information.
Optionally, the calculating, according to a preset cost function, a cost distance for searching for another pixel point adjacent to the current search point includes:
d＝d1×ɑ+d2×(1-ɑ)
d1＝dismin
d2＝abs(thetai-2,i-1-thetai-1,i)
wherein d is a cost distance; d1The closest distance between the pixel point and the non-empty parking space is obtained; d2The included angle between the direction of the preorder search path and the current search path direction; theta (theta)i-1,iRepresents a vector (x)i-xi-1，yi-yi-1) The camber value of an included angle between the camber value and the x-axis direction in a vehicle coordinate system; alpha is a preset scale factor used for d1And d2The ratio in the cost calculation.
Optionally, the regularizing the plurality of search points obtained by the boundary search to obtain the boundary information of the travelable area includes:
performing straight line fitting between three or more adjacent search points; if the fitting error is less than or equal to a preset error threshold value, using end points at two ends of the fitted straight line as boundary points; if the fitting error is larger than a preset error threshold value, taking three or more adjacent search points as boundary points; and finally obtaining a plurality of boundary points of the regulated empty parking spaces.
Optionally, the step S4 includes:
step S41, calculating the path direction and the path corner of each boundary point; wherein the path direction a of the ith boundary pointiAn included angle formed by a directed line segment formed between the ith boundary point and the (i + 1) th boundary point and clockwise rotation of the positive direction of the x axis in the vehicle coordinate system is formed; path rotation angle theta of the ith boundary pointiIs the ith boundary point path direction aiAnd the (i-1) th boundary point path direction ai-1The absolute value of the difference;
step S42, connecting a plurality of boundary pointsIntermediate path angle thetaiAnd outputting boundary points meeting the preset numerical range as the empty parking space angular point preselected points.
Optionally, the step S5 includes:
step S51, randomly selecting any 4 points as a group of preselected points to enumerate the preselected points of the angular points of the plurality of vacant parking spaces to obtain a plurality of groups of preselected points;
step S52, respectively judging whether a plurality of parking spaces formed by the plurality of groups of pre-selection points meet preset parking space conditions, generating corresponding candidate empty parking spaces according to one or more groups of pre-selection points meeting the parking space conditions, and calculating the credibility of the candidate empty parking spaces;
wherein for any one set of preselected points (P)k,1，Pk,2，Pk,3，Pk,4) Calculating a point Pk,1To point Pk,4Accumulated path length of lkEuclidean distance l between two adjacent pointsk-12、lk-23、lk-34、lk-41Judging the calculated lk、lk-12、lk-23、lk-34、lk-41If the preset parking space condition is met, if the group of pre-selected points (P)k,1，Pk,2，Pk,3，Pk,4) If the parking space condition is met, the group of pre-selected points (P) is selectedk,1，Pk,2，Pk,3，Pk,4) Marking as a pre-selected vacant parking space angular point sequence, and predicting parking spaces according to a closed quadrangle formed by 4 points in each sequence to obtain a candidate vacant parking space;
and step S53, determining the finally output target empty space according to the credibility of the candidate empty space.
Optionally, the parking space condition specifically includes:
(carw×2+carh)×kcmin<lk<(carw+carh×2)×kcmax
kmin<lk-12/lk-34<kmax
kmin<lk-23/lk-41<kmax
wherein, carwAnd carhRespectively width and height of the current vehicle, kcmin、kcmax、kminAnd kmaxIs a preset value.
Optionally, wherein:
for any pre-selected empty parking space angle point sequence (P)k,1，Pk,2，Pk,3，Pk,4)，Pk,1Pk,2Pk,3Pk,4Is an original closed quadrangle; let Pk,1’Pk,2’Pk,3’Pk,4' is a predicted empty space, C1、C2Are respectively line segment Pk,1Pk,2、Pk,3Pk,4Mid point of (A), O2O4Is C1C2Perpendicular bisector of, O2O4And O1O3Are mutually perpendicular and bisected; carrying out parking space prediction according to preset constraint conditions to obtain a plurality of candidate empty parking spaces; the constraint conditions are that the length and the width of the candidate empty parking space are respectively m times of the length and the width of the current vehicle;
credibility score of candidate empty parking spacekThe calculation method is as follows:
scorek＝w1×S1/S+w2×S2/S+w3×S3/S
wherein S is Pk,1’Pk,2’Pk,3’Pk,4Area of `, S1Is Pk,1’Pk,2’Pk,3’Pk,4' and Pk,1Pk,2Pk,3Pk,4Area of intersection region, S2Is Pk,1’Pk,2’Pk,3’Pk,4' and Pk,1Pk,2Pk,3Pk,4Cross-over ratio of (S)3Is Pk,1’Pk,2’Pk,3’Pk,4' is the area of the travelable region, w1、w2、w3M is a preset value, and is a preset weight value.
Optionally, the step S53 includes:
judging whether the credibility of the candidate empty parking spaces is greater than the credibility threshold value or not;
if the credibility of only one candidate empty parking space is greater than the credibility threshold, determining the candidate empty parking space as an empty parking space to be output;
if the credibility of at least two candidate empty parking spaces is greater than the credibility threshold, selecting the candidate empty parking space with the highest credibility as the empty parking space to be output;
and if the credibility of the candidate empty parking spaces is less than the credibility threshold, outputting a result that the empty parking spaces are not detected.
Optionally, the method further comprises:
acquiring current time information and vehicle position information, and judging whether to perform image semantic segmentation according to the current time information and the vehicle position information; if the time interval between the current time and the time when the image semantic segmentation is carried out last time is larger than a preset time threshold, or if the distance between the current vehicle position and the vehicle position when the image semantic segmentation is carried out last time is larger than a preset distance threshold, the image semantic segmentation is judged to be carried out, and the starting identification signal is generated.
According to a second aspect, an embodiment of the present invention provides an empty space recognition system, configured to execute the empty space recognition method according to the first aspect, where the system includes:
the image acquisition unit is used for responding to the received starting identification signal and acquiring the panoramic image around the current vehicle;
the semantic segmentation unit is used for performing semantic segmentation and identification on the panoramic image through a pre-trained semantic segmentation model to obtain a drivable area;
the boundary extraction unit is used for carrying out regularization processing on the boundary of the travelable area to obtain boundary information, and the boundary information comprises a plurality of boundary points;
the preselected point screening unit is used for screening the boundary points according to a preset screening rule to obtain a plurality of vacant parking space angular point preselected points; and
and the empty parking space prediction unit is used for enumerating the multiple empty parking space angular point preselected points by randomly selecting any 4 points as a group of preselected points to obtain multiple groups of preselected points, respectively judging whether multiple parking spaces formed by the multiple groups of preselected points meet preset parking space conditions, and if any one group of preselected points meet the parking space conditions, generating corresponding empty parking spaces according to the group of preselected points.
Optionally, the system further comprises:
the triggering unit is used for acquiring current time information and vehicle position information and judging whether to perform image semantic segmentation according to the current time information and the vehicle position information; if the time interval between the current time and the time when the image is subjected to semantic segmentation last time is larger than a preset time threshold value, or if the distance between the current vehicle position and the vehicle position when the image is subjected to semantic segmentation last time is larger than a preset distance threshold value, the image is judged to be subjected to semantic segmentation, a starting identification signal is generated, and the starting identification signal is sent to the image acquisition unit.
According to a third aspect, an embodiment of the present invention provides a computer apparatus, including: an empty space identification system according to the first aspect; alternatively, a memory and a processor, the memory having stored therein computer readable instructions, which, when executed by the processor, cause the processor to carry out the steps of the empty space identification method according to the second aspect.
According to a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, which, when executed by a processor, implements the empty space identification method according to the first aspect.
The embodiment of the invention provides an empty parking space identification method and system, computer equipment and a computer readable storage medium thereof, which are applied to automatic parking, wherein when the embodiment is implemented, an automatic parking system of a vehicle is started to obtain a panoramic image of the surrounding environment of the current vehicle; performing semantic segmentation recognition on the panoramic image by using a pre-trained semantic segmentation model to obtain a drivable area; carrying out regularization processing on the boundary of the travelable area to obtain boundary information, wherein the boundary information comprises a plurality of boundary points; screening the plurality of boundary points according to a preset screening rule to obtain a plurality of empty parking space angular point preselected points; randomly selecting any 4 points as a group of preselected points to enumerate the multiple vacant parking space angular point preselected points to obtain multiple groups of preselected points, respectively judging whether multiple parking spaces formed by the multiple groups of preselected points meet preset parking space conditions, and if any one group of preselected points meet the parking space conditions, generating corresponding vacant parking spaces according to the group of preselected points. The method and the device predict the empty parking spaces based on the boundary points of the drivable areas, can effectively inhibit the influence of illumination change and weather change on the accuracy of the empty parking space positioning result in various parking scenes, are not limited by parking space types, and can simultaneously identify the line parking spaces and the space parking spaces (parking spaces without parking spaces).
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart illustrating an empty space identification method according to an embodiment of the present invention.
Fig. 2 is a schematic view of a travelable area according to an embodiment of the present invention.
FIG. 3 is a schematic diagram of a closed curve according to an embodiment of the present invention.
FIG. 4 is a diagram illustrating a boundary search area according to an embodiment of the present invention.
FIG. 5 is a diagram illustrating a boundary search result according to an embodiment of the present invention.
Fig. 6 is a schematic diagram illustrating a candidate empty space prediction according to an embodiment of the present invention.
Fig. 7 is a schematic diagram of a frame of an empty space recognition system according to another embodiment of the present invention.
Detailed Description
Various exemplary embodiments, features and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. In addition, in the following detailed description, numerous specific details are set forth in order to provide a better understanding of the present invention. It will be understood by those skilled in the art that the present invention may be practiced without some of these specific details. In some instances, well known means have not been described in detail so as not to obscure the present invention.
Referring to fig. 1, an embodiment of the present invention provides a method for identifying an empty parking space, including the following steps S1-S5:
step S1, responding to the received start identification signal, and acquiring a panoramic image of the surrounding environment of the current vehicle;
specifically, the method of the embodiment is applied to an automatic parking system to realize recognition of a drivable area, and when the automatic parking is started, initialization of the automatic parking system is finished, and one-time image semantic segmentation is performed to obtain the drivable area.
Illustratively, the panoramic image for looking around may be obtained by: firstly, collecting four fisheye images, and carrying out distortion correction on four fisheye videos; then, automatically selecting four outermost inner corner points of the checkerboard as perspective transformation source points, selecting four corresponding target points of the target image to obtain four pairs of corresponding points, and respectively carrying out perspective transformation on the four fish-eye images through a perspective transformation matrix obtained by calculation to obtain a four-way aerial view; and finally, splicing the four fisheye aerial view images. Optionally, the panoramic image is an overhead image covering a rectangular area with the center of the rear axle of the vehicle as a midpoint, the length of the rectangular area is 18.6 meters, and the width of the rectangular area is 15.4 meters.
It should be noted that the above-mentioned manner is merely an example, and the method of the present embodiment is not limited to a certain means for obtaining the panoramic image.
The splicing positions of two adjacent image areas are provided with obvious splicing seams, and the positions of parts of the splicing seams are inaccurate, so that the problem of camera installation is considered; the four paths of video have large differences in brightness, and have uneven brightness. In order to deal with the seam problem, optionally, a gradually changing weight value may be added to the overlapping area of the two mutually spliced images, and the overlapping area is subjected to weighted fusion to achieve the effect of smooth transition.
The parking system initialization section in the present embodiment involves determination of the world coordinate system and initialization of the initial posture of the vehicle. Once the world coordinate system is determined, all coordinate conversions are uniformly converted into the world coordinate system for calculation in the whole automatic parking process until parking is terminated, the determination of the world coordinate system is favorable for simplifying the complexity of an algorithm in a subsequent drivable area map updating stage, the initialization information of the vehicle posture and the determination of the world coordinate system are synchronously performed, the directions of an x axis and a y axis in the world coordinate system are determined by the vehicle posture at the current moment initialized by a system, the expression mode of the vehicle posture is probably synchronously converted into the world coordinate system, and the coordinates of the center of a rear axis of the vehicle and the vehicle posture are updated in real time in the whole operating stage of the automatic parking system.
Step S2, performing semantic segmentation and identification on the panoramic image by using a pre-trained semantic segmentation model to obtain a drivable area;
specifically, semantic segmentation is an important field in computer vision, and refers to identifying an image at a pixel level, that is, marking an object class to which each pixel in the image belongs, and assigning different pixel values to different object classes in the image, specifically, in this embodiment, a drivable region and a non-drivable region are segmented. Through semantic segmentation, a travelable region under the current vehicle camera view can be obtained.
Step S3, carrying out regularization processing on the boundary of the travelable area to obtain boundary information, wherein the boundary information comprises a plurality of boundary points;
s4, screening the boundary points according to a preset screening rule to obtain a plurality of empty parking space corner point preselected points;
step S5, enumerating the multiple vacant parking space angular point preselected points by randomly selecting any 4 points as a group of preselected points to obtain multiple groups of preselected points, respectively judging whether multiple parking spaces formed by the multiple groups of preselected points meet preset parking space conditions, generating corresponding candidate vacant parking spaces according to one or more groups of preselected points meeting the parking space conditions, and determining a target vacant parking space to be finally output according to the candidate vacant parking spaces.
As an optional scheme, the step S3 in this embodiment includes the following steps S31 to 236:
step S31, extracting the boundary of the travelable area to form a closed curve;
for example, the drivable region is shown in fig. 2 and the closed curve is shown in fig. 3.
Step S32, expanding 1 pixel point in the closed curve to form a boundary search area;
illustratively, based on the closed curve of fig. 3, the corresponding boundary search area is shown in fig. 4.
Step S34, selecting any 1 adjacent pixel point on the boundary searching area as an initial searching point of boundary searching;
step S35, carrying out boundary search in the boundary search area according to the initial search point; in the searching process, calculating and searching the cost distance of other pixel points adjacent to the current searching point according to a preset cost function, and selecting the pixel point with the minimum cost distance as the next searching point to continue searching until the initial searching point is searched;
illustratively, based on the border search area of fig. 4, the corresponding border search result is shown in fig. 5, where the border search result in fig. 5 includes a plurality of search points.
And step S36, performing regularization processing on the plurality of search points obtained by the boundary search to obtain boundary information of the travelable area.
Optionally, the calculating, according to a preset cost function, a cost distance for searching for another pixel point adjacent to the current search point includes:
d＝d1×ɑ+d2×(1-ɑ)
d1＝dismin
d2＝abs(thetai-2,i-1-thetai-1,i)
wherein d is a cost distance; d1The shortest distance between the pixel point and the non-driving area; d2For example, if the 2 search points obtained by the previous sequential search are D1 and D2, and D3 is another pixel point adjacent to the current search point D2, the direction of the preamble search path is D1 → D2, and the direction of the current search path is D2 → D3; theta (theta)i-1,iRepresents a vector (x)i-xi-1，yi-yi-1) The camber value of an included angle between the camber value and the x-axis direction in a vehicle coordinate system; alpha is a preset scale factor used for d1And d2The ratio in the cost calculation is preferably, but not limited to, 0.3 in the present embodiment.
Optionally, the regularizing the plurality of search points obtained by the boundary search to obtain the boundary information of the travelable area includes:
performing straight line fitting between three or more adjacent search points; if the fitting error is less than or equal to a preset error threshold value, using end points at two ends of the fitted straight line as boundary points; if the fitting error is larger than a preset error threshold value, taking three or more adjacent search points as boundary points; and finally obtaining a plurality of boundary points of the regulated empty parking spaces.
Specifically, the fitting error refers to a difference between a fitted straight line and a line segment formed by connecting original search points, and the difference can be determined by, for example, calculating a similarity between two line segments. Specifically, the difference in coordinate position information of two line segments is another example.
It should be noted that, based on the method of this embodiment, according to a large number of sample tests, the boundary of the travelable region before the boundary point is regularized is composed of 200 to 300 pixel points on average, the boundary of the travelable region after the boundary regularized extraction can be represented by 20 to 50 pixel points, and the error after the regularization is controlled within the error range allowed by the precision. The significance of the regularization of the boundary points of the drivable region is to simplify the complexity of a subsequent algorithm in the automatic parking system and improve the overall operation efficiency of the system.
Optionally, the step S4 includes steps S41 to S42:
step S41, calculating the path direction and the path corner of each boundary point; wherein the path direction a of the ith boundary pointiAn included angle formed by a directed line segment formed between the ith boundary point and the (i + 1) th boundary point and clockwise rotation of the positive direction of the x axis in the vehicle coordinate system is formed; path rotation angle theta of the ith boundary pointiIs the ith boundary point path direction aiAnd the (i-1) th boundary point path direction ai-1The absolute value of the difference;
specifically, the path direction in the step is specifically the search direction in the boundary search, and is defined by the current boundary point (x)i,yi) Point to the next boundary point (x)i+1,yi+1) (ii) a Wherein, thetai＝abs(ai-ai-1)。
Step S42, turning the path in the plurality of boundary points by the angle thetaiAnd outputting boundary points meeting the preset numerical range as the empty parking space angular point preselected points.
For example, in this embodiment, the boundary point where the path corner is greater than or equal to 30 and less than or equal to 150, or the path corner is greater than or equal to 210 and less than or equal to 330 is output as the empty corner point preselected point. Of course, the above numerical ranges can be adjusted properly in practical application.
Optionally, the step S5 includes:
step S51, randomly selecting any 4 points as a group of preselected points to enumerate the preselected points of the angular points of the plurality of vacant parking spaces to obtain a plurality of groups of preselected points;
step S52, respectively judging whether a plurality of parking spaces formed by the plurality of groups of pre-selection points meet preset parking space conditions, generating corresponding candidate empty parking spaces according to one or more groups of pre-selection points meeting the parking space conditions, and calculating the credibility of the candidate empty parking spaces;
wherein for any one set of preselected points (P)k,1，Pk,2，Pk,3，Pk,4) Calculating a point Pk,1To point Pk,4Accumulated path length of lkEuclidean distance l between two adjacent pointsk-12、lk-23、lk-34、lk-41Judging the calculated lk、lk-12、lk-23、lk-34、lk-41If the preset parking space condition is met, if the group of pre-selected points (P)k,1，Pk,2，Pk,3，Pk,4) If the parking space condition is met, the group of pre-selected points (P) is selectedk,1，Pk,2，Pk,3，Pk,4) Marking as a pre-selected vacant parking space angular point sequence, and predicting parking spaces according to a closed quadrangle formed by 4 points in each sequence to obtain a candidate vacant parking space;
and step S53, determining the finally output target empty space according to the credibility of the candidate empty space.
Optionally, the parking space condition specifically includes:
(carw×2+carh)×kcmin<lk<(carw+carh×2)×kcmax
kmin<lk-12/lk-34<kmax
kmin<lk-23/lk-41<kmax
wherein, carwAnd carhRespectively width and height of the current vehicle, kcmin、kcmax、kminAnd kmaxIs a preset numerical value;
preferably, k in this embodimentcminAnd kcmaxAre taken to be 1.1 and 1.5, kminAnd kmaxValues of (a) are 0.8 and 1.3; and marking the enumeration sequence meeting the condition as a pre-selected empty parking space angle point sequence.
Optionally, wherein:
FIG. 6 is a schematic diagram of the candidate empty space prediction, referring to FIG. 6, for any one of the pre-selected empty space angle point sequences (P)k,1，Pk,2，Pk,3，Pk,4)，Pk,1Pk,2Pk,3Pk,4Is an original closed quadrangle; let Pk,1’Pk,2’Pk,3’Pk,4' is a predicted empty space, C1、C2Are respectively line segment Pk,1Pk,2、Pk,3Pk,4Mid point of (A), O2O4Is C1C2Perpendicular bisector of, O2O4And O1O3Are mutually perpendicular and bisected; carrying out parking space prediction according to preset constraint conditions to obtain a plurality of candidate empty parking spaces; the constraint conditions are that the length and the width of the candidate empty parking space are respectively m times of the length and the width of the current vehicle;
specifically, one empty space P corresponding to the original closed quadrangle can be predicted based on the set length and width of the candidate empty spacesk,1’Pk,2’Pk,3’Pk,4’。
Further, the credibility score of the candidate empty parking spacekThe calculation method is as follows:
scorek＝w1×S1/S+w2×S2/S+w3×S3/S
wherein S is Pk,1’Pk,2’Pk,3’Pk,4Area of `, S1Is Pk,1’Pk,2’Pk,3’Pk,4' and Pk,1Pk,2Pk,3Pk,4Area of intersection region, S2Is Pk,1’Pk,2’Pk,3’Pk,4' and Pk,1Pk,2Pk,3Pk,4Cross-over ratio of (S)3Is Pk,1’Pk,2’Pk,3’Pk,4' is the area of the travelable region, w1、w2、w3Is a preset weight value, the weight w1、w2、w3In the examples, 0.3, 0.2, 0.5 are preferred; m is a predetermined number, and m is preferably 1.2 in this embodiment.
Optionally, the step S53 includes:
judging whether the credibility of the candidate empty parking spaces is greater than the credibility threshold value or not;
wherein:
(1) if the credibility of only one candidate empty parking space is greater than the credibility threshold, determining the candidate empty parking space as an empty parking space to be output;
(2) if the credibility of at least two candidate empty parking spaces is greater than the credibility threshold, selecting the candidate empty parking space with the highest credibility as the empty parking space to be output;
(3) and if the credibility of the candidate empty parking spaces is less than the credibility threshold, outputting a result that the empty parking spaces are not detected.
Specifically, in the present embodiment, the reliability threshold is preferably 0.9.
Optionally, the method of this embodiment further includes:
acquiring current time information and vehicle position information, and judging whether to perform image semantic segmentation according to the current time information and the vehicle position information; if the time interval between the current time and the time when the image semantic segmentation is carried out last time is larger than a preset time threshold, or if the distance between the current vehicle position and the vehicle position when the image semantic segmentation is carried out last time is larger than a preset distance threshold, the image semantic segmentation is judged to be carried out, and the starting identification signal is generated.
Specifically, in the automatic parking process, the method of the embodiment always updates the position information of the vehicle in real time, and only when the semantic segmentation condition is triggered, the semantic segmentation model is used for identifying the drivable area in the current visual field. The semantic division triggering condition is set in the embodiment to ensure the high efficiency and effectiveness of the travelable region identification. Because the same image data is input, the semantic segmentation model based on the deep learning with strong robustness can feed back the same or similar recognition output, if the surrounding environmental characteristics of the vehicle at the same position are not changed much, the recognition of the travelable region is repeated, a large amount of unnecessary computing resources are occupied, and the overall computing efficiency of the algorithm is affected, however, the recognition of the travelable region is performed redundantly, and no significant improvement in performance is brought.
Another embodiment of the present invention provides an empty space recognition system, configured to execute the empty space recognition method according to the above embodiment, where the system includes:
the image acquisition unit 1 is used for responding to the received starting identification signal and acquiring a panoramic image around the current vehicle;
the semantic segmentation unit 2 is used for performing semantic segmentation and identification on the panoramic image through a pre-trained semantic segmentation model to obtain a drivable area;
a boundary extraction unit 3, configured to perform regularization processing on a boundary of the travelable region to obtain boundary information, where the boundary information includes a plurality of boundary points;
the preselected point screening unit 4 is used for screening the boundary points according to a preset screening rule to obtain a plurality of empty parking space angular point preselected points; and
and the empty parking space prediction unit 5 is used for enumerating the multiple empty parking space angular point preselected points by randomly selecting any 4 points as a group of preselected points to obtain multiple groups of preselected points, respectively judging whether multiple parking spaces formed by the multiple groups of preselected points meet preset parking space conditions, generating corresponding candidate empty parking spaces according to one or more groups of preselected points meeting the parking space conditions, and determining a target empty parking space to be finally output according to the candidate empty parking spaces.
Optionally, the system further comprises:
the triggering unit is used for acquiring current time information and vehicle position information and judging whether to perform image semantic segmentation according to the current time information and the vehicle position information; if the time interval between the current time and the time when the image is subjected to semantic segmentation last time is larger than a preset time threshold value, or if the distance between the current vehicle position and the vehicle position when the image is subjected to semantic segmentation last time is larger than a preset distance threshold value, the image is judged to be subjected to semantic segmentation, a starting identification signal is generated, and the starting identification signal is sent to the image acquisition unit.
The above-described system embodiments are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.
It should be noted that the system described in the foregoing embodiment corresponds to the method described in the foregoing embodiment, and therefore, portions of the system described in the foregoing embodiment that are not described in detail can be obtained by referring to the content of the method described in the foregoing embodiment, and details are not described here.
Moreover, if the empty space recognition system described in the above embodiment is implemented in the form of a software functional unit and sold or used as an independent product, it may be stored in a computer-readable storage medium.
Illustratively, the computer-readable storage medium may include: any entity or device capable of carrying the computer program code, recording medium, usb disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM), Random Access Memory (RAM), electrical carrier wave signals, telecommunications signals, software distribution medium, and the like.
In addition, another embodiment of the present invention further provides another computer device, which includes a memory and a processor, where the memory stores computer readable instructions, and the computer readable instructions, when executed by the processor, cause the processor to execute the steps of the method for empty space identification according to the above-mentioned embodiment.
Of course, the computer device may also have components such as a wired or wireless network interface, a keyboard, and an input/output interface, so as to perform input/output, and the computer device may also include other components for implementing the functions of the device, which are not described herein again.
Illustratively, the computer program may be divided into one or more units, which are stored in the memory and executed by the processor to accomplish the present invention. The one or more units may be a series of computer program instruction segments capable of performing certain functions, which are used to describe the execution of the computer program in the computer device.
The Processor may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component, etc. The general purpose processor may be a microprocessor or the processor may be any conventional processor or the like, which is the control center for the computer device and connects the various parts of the overall computer device using various interfaces and lines.
The memory may be used for storing the computer program and/or unit, and the processor may implement various functions of the computer device by executing or executing the computer program and/or unit stored in the memory and calling data stored in the memory. In addition, the memory may include high speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), at least one magnetic disk storage device, a Flash memory device, or other volatile solid state storage device.
Another embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the steps of the empty space identification method according to the above embodiment.
Specifically, the computer-readable storage medium may include: any entity or device capable of carrying the computer program code, recording medium, usb disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM), Random Access Memory (RAM), electrical carrier wave signals, telecommunications signals, software distribution medium, and the like.
Having described embodiments of the present invention, the foregoing description is intended to be exemplary, not exhaustive, and not limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen in order to best explain the principles of the embodiments, the practical application, or improvements made to the technology in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.
Claims (14)
1. The utility model provides an empty parking stall identification method which characterized in that includes:
step S1, responding to the received start identification signal, and acquiring a panoramic image of the surrounding environment of the current vehicle;
step S2, performing semantic segmentation and identification on the panoramic image by using a pre-trained semantic segmentation model to obtain a drivable area;
step S3, carrying out regularization processing on the boundary of the travelable area to obtain boundary information, wherein the boundary information comprises a plurality of boundary points;
s4, screening the boundary points according to a preset screening rule to obtain a plurality of empty parking space corner point preselected points;
step S5, enumerating the multiple vacant parking space angular point preselected points by randomly selecting any 4 points as a group of preselected points to obtain multiple groups of preselected points, respectively judging whether multiple parking spaces formed by the multiple groups of preselected points meet preset parking space conditions, generating corresponding candidate vacant parking spaces according to one or more groups of preselected points meeting the parking space conditions, and determining a target vacant parking space to be finally output according to the candidate vacant parking spaces.
2. The empty space recognition method according to claim 1, wherein the extracting boundary information of the driving area includes:
extracting the boundary of the travelable area to form a closed curve;
expanding 1 pixel point into the closed curve to form a boundary search area;
selecting any 1 adjacent pixel point on the boundary search area as an initial search point of boundary search;
performing boundary search in the boundary search area according to the initial search point; in the searching process, calculating and searching the cost distance of other pixel points adjacent to the current searching point according to a preset cost function, and selecting the pixel point with the minimum cost distance as the next searching point to continue searching until the initial searching point is searched;
and carrying out regularization processing on a plurality of search points searched by the boundary to obtain the boundary information.
3. The empty space recognition method according to claim 2, wherein the step of calculating the cost distance for searching other pixel points adjacent to the current search point according to a preset cost function comprises:
d＝d1×ɑ+d2×(1-ɑ)
d1＝dismin
d2＝abs(thetai-2,i-1-thetai-1,i)
wherein d is a cost distance; d1The closest distance between the pixel point and the non-empty parking space is obtained; d2The included angle between the direction of the preorder search path and the current search path direction; theta (theta)i-1,iRepresents a vector (x)i-xi-1，yi-yi-1) The camber value of an included angle between the camber value and the x-axis direction in a vehicle coordinate system; alpha is a preset scale factor used for d1And d2The ratio in the cost calculation.
4. The empty space recognition method according to claim 2, wherein the obtaining of the boundary information of the drivable area by regularizing the plurality of search points obtained by the boundary search comprises:
performing straight line fitting between three or more adjacent search points; if the fitting error is less than or equal to a preset error threshold value, using end points at two ends of the fitted straight line as boundary points; if the fitting error is larger than a preset error threshold value, taking three or more adjacent search points as boundary points; and finally obtaining a plurality of boundary points of the regulated empty parking spaces.
5. The empty space identification method according to any one of claims 1 to 4, wherein the step S4 includes:
calculating the path direction and the path corner of each boundary point; wherein the path direction a of the ith boundary pointiAn included angle formed by a directed line segment formed between the ith boundary point and the (i + 1) th boundary point and clockwise rotation of the positive direction of the x axis in the vehicle coordinate system is formed; path rotation angle theta of the ith boundary pointiIs the ith boundary point path direction aiAnd the (i-1) th boundary point path direction ai-1The absolute value of the difference;
rotating the path in a plurality of boundary points by a angle thetaiAnd outputting boundary points meeting the preset numerical range as the empty parking space angular point preselected points.
6. The empty space recognition method according to claim 1, wherein the step S5 includes:
step S51, randomly selecting any 4 points as a group of preselected points to enumerate the preselected points of the angular points of the plurality of vacant parking spaces to obtain a plurality of groups of preselected points;
step S52, respectively judging whether a plurality of parking spaces formed by the plurality of groups of pre-selection points meet preset parking space conditions, generating corresponding candidate empty parking spaces according to one or more groups of pre-selection points meeting the parking space conditions, and calculating the credibility of the candidate empty parking spaces;
wherein for any one set of preselected points (P)k,1，Pk,2，Pk,3，Pk,4) Calculating a point Pk,1To point Pk,4Accumulated path length of lkEuclidean distance l between two adjacent pointsk-12、lk-23、lk-34、lk-41Judging the calculated lk、lk-12、lk-23、lk-34、lk-41If the preset parking space condition is met, if the group of pre-selected points (P)k,1，Pk,2，Pk,3，Pk,4) If the parking space condition is met, the group of pre-selected points (P) is selectedk,1，Pk,2，Pk,3，Pk,4) Marking as a pre-selected vacant parking space angular point sequence, and predicting parking spaces according to a closed quadrangle formed by 4 points in each sequence to obtain a candidate vacant parking space;
and step S53, determining the finally output target empty space according to the credibility of the candidate empty space.
7. The empty parking space recognition method according to claim 6, wherein the parking space conditions specifically are:
(carw×2+carh)×kcmin<lk<(carw+carh×2)×kcmax
kmin<lk-12/lk-34<kmax
kmin<lk-23/lk-41<kmax
wherein, carwAnd carhRespectively width and height of the current vehicle, kcmin、kcmax、kminAnd kmaxIs a preset value.
8. The empty space recognition method according to claim 6, wherein:
for any pre-selected empty parking space angle point sequence (P)k,1，Pk,2，Pk,3，Pk,4)，Pk,1Pk,2Pk,3Pk,4Is an original closed quadrangle; let Pk,1’Pk,2’Pk,3’Pk,4' is a predicted empty space, C1、C2Are respectively line segment Pk,1Pk,2、Pk,3Pk,4Mid point of (A), O2O4Is C1C2Perpendicular bisector of, O2O4And O1O3Are mutually perpendicular and bisected; carrying out parking space prediction according to preset constraint conditions to obtain a plurality of candidate empty parking spaces; the constraint conditions are that the length and the width of the candidate empty parking space are respectively m times of the length and the width of the current vehicle;
credibility score of candidate empty parking spacekThe calculation method is as follows:
scorek＝w1×S1/S+w2×S2/S+w3×S3/S
wherein S is Pk,1’Pk,2’Pk,3’Pk,4Area of `, S1Is Pk,1’Pk,2’Pk,3’Pk,4' and Pk,1Pk,2Pk,3Pk,4Area of intersection region, S2Is Pk,1’Pk,2’Pk,3’Pk,4' and Pk,1Pk,2Pk,3Pk,4Cross-over ratio of (S)3Is Pk,1’Pk,2’Pk,3’Pk,4' is the area of the travelable region, w1、w2、w3M is a preset value, and is a preset weight value.
9. The empty space recognition method according to claim 8, wherein the step S53 includes:
judging whether the credibility of the candidate empty parking spaces is greater than the credibility threshold value or not;
if the credibility of only one candidate empty parking space is greater than the credibility threshold, determining the candidate empty parking space as an empty parking space to be output;
if the credibility of at least two candidate empty parking spaces is greater than the credibility threshold, selecting the candidate empty parking space with the highest credibility as the empty parking space to be output;
and if the credibility of the candidate empty parking spaces is less than the credibility threshold, outputting a result that the empty parking spaces are not detected.
10. The empty space recognition method according to claim 1, further comprising:
acquiring current time information and vehicle position information, and judging whether to perform image semantic segmentation according to the current time information and the vehicle position information; if the time interval between the current time and the time when the image semantic segmentation is carried out last time is larger than a preset time threshold, or if the distance between the current vehicle position and the vehicle position when the image semantic segmentation is carried out last time is larger than a preset distance threshold, the image semantic segmentation is judged to be carried out, and the starting identification signal is generated.
11. An empty space recognition system for performing the empty space recognition method according to any one of claims 1 to 10, the system comprising:
the image acquisition unit is used for responding to the received starting identification signal and acquiring the panoramic image around the current vehicle;
the semantic segmentation unit is used for performing semantic segmentation and identification on the panoramic image through a pre-trained semantic segmentation model to obtain a drivable area;
the boundary extraction unit is used for carrying out regularization processing on the boundary of the travelable area to obtain boundary information, and the boundary information comprises a plurality of boundary points;
the preselected point screening unit is used for screening the boundary points according to a preset screening rule to obtain a plurality of vacant parking space angular point preselected points; and
and the empty parking space prediction unit is used for enumerating the multiple empty parking space angular point preselected points by randomly selecting any 4 points as a group of preselected points to obtain multiple groups of preselected points, respectively judging whether multiple parking spaces formed by the multiple groups of preselected points meet preset parking space conditions, generating corresponding candidate empty parking spaces according to one or more groups of preselected points meeting the parking space conditions, and determining a finally output target empty parking space according to the candidate empty parking spaces.
12. The empty space recognition system of claim 11, further comprising:
the triggering unit is used for acquiring current time information and vehicle position information and judging whether to perform image semantic segmentation according to the current time information and the vehicle position information; if the time interval between the current time and the time when the image is subjected to semantic segmentation last time is larger than a preset time threshold value, or if the distance between the current vehicle position and the vehicle position when the image is subjected to semantic segmentation last time is larger than a preset distance threshold value, the image is judged to be subjected to semantic segmentation, a starting identification signal is generated, and the starting identification signal is sent to the image acquisition unit.
13. A computer device, comprising: an empty space identification system according to any one of claims 11-12; or a memory and a processor, the memory having stored therein computer readable instructions which, when executed by the processor, cause the processor to perform the steps of the empty space identification method according to any one of claims 1 to 10.
14. A computer-readable storage medium having stored thereon a computer program, characterized in that: the computer program, when executed by a processor, implements the empty space identification method of any one of claims 1-10.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010900989.7A CN114120259A (en) | 2020-09-01 | 2020-09-01 | Empty parking space identification method and system, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010900989.7A CN114120259A (en) | 2020-09-01 | 2020-09-01 | Empty parking space identification method and system, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114120259A true CN114120259A (en) | 2022-03-01 |
Family
ID=80360435
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010900989.7A Pending CN114120259A (en) | 2020-09-01 | 2020-09-01 | Empty parking space identification method and system, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114120259A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114782928A (en) * | 2022-06-22 | 2022-07-22 | 苏州魔视智能科技有限公司 | Parking space detection method and device, electronic equipment and storage medium |
| CN116206483A (en) * | 2023-04-21 | 2023-06-02 | 深圳市速腾聚创科技有限公司 | Parking position determining method, electronic device and computer readable storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105574485A (en) * | 2015-11-11 | 2016-05-11 | 北京四象网讯科技有限公司 | Vehicle information identification method and system |
| CN109871745A (en) * | 2018-12-29 | 2019-06-11 | 百度在线网络技术（北京）有限公司 | Identify method, system and the vehicle of empty parking space |
| WO2020019930A1 (en) * | 2018-07-25 | 2020-01-30 | 广州小鹏汽车科技有限公司 | Automatic parking method and device |
-
2020
- 2020-09-01 CN CN202010900989.7A patent/CN114120259A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105574485A (en) * | 2015-11-11 | 2016-05-11 | 北京四象网讯科技有限公司 | Vehicle information identification method and system |
| WO2020019930A1 (en) * | 2018-07-25 | 2020-01-30 | 广州小鹏汽车科技有限公司 | Automatic parking method and device |
| CN109871745A (en) * | 2018-12-29 | 2019-06-11 | 百度在线网络技术（北京）有限公司 | Identify method, system and the vehicle of empty parking space |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114782928A (en) * | 2022-06-22 | 2022-07-22 | 苏州魔视智能科技有限公司 | Parking space detection method and device, electronic equipment and storage medium |
| CN116206483A (en) * | 2023-04-21 | 2023-06-02 | 深圳市速腾聚创科技有限公司 | Parking position determining method, electronic device and computer readable storage medium |
| CN116206483B (en) * | 2023-04-21 | 2023-08-04 | 深圳市速腾聚创科技有限公司 | Parking position determining method, electronic device and computer readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110378297B (en) | Remote sensing image target detection method and device based on deep learning and storage medium | |
| Lee et al. | Accurate traffic light detection using deep neural network with focal regression loss | |
| CN110766002B (en) | Ship name character region detection method based on deep learning | |
| CN110084230B (en) | Image-based vehicle body direction detection method and device | |
| CN116188999B (en) | Small target detection method based on visible light and infrared image data fusion | |
| CN114120259A (en) | Empty parking space identification method and system, computer equipment and storage medium | |
| CN111382625A (en) | Road sign identification method and device and electronic equipment | |
| CN111626295A (en) | Training method and device for license plate detection model | |
| CN114708437A (en) | Training method of target detection model, target detection method, device and medium | |
| CN114820679A (en) | Image annotation method and device, electronic equipment and storage medium | |
| CN111401143A (en) | Pedestrian tracking system and method | |
| CN111428566A (en) | Deformation target tracking system and method | |
| CN112198878B (en) | Instant map construction method and device, robot and storage medium | |
| Esfahani et al. | DeepDSAIR: Deep 6-DOF camera relocalization using deblurred semantic-aware image representation for large-scale outdoor environments | |
| CN111553242B (en) | Training method for generating countermeasure network for predicting driving behavior and electronic device | |
| CN113591543B (en) | Traffic sign recognition method, device, electronic equipment and computer storage medium | |
| CN114067598B (en) | Method and device for determining idle parking spaces | |
| CN111709377B (en) | Feature extraction method, target re-identification method and device and electronic equipment | |
| CN113807354B (en) | Image semantic segmentation method, device, equipment and storage medium | |
| CN114332814A (en) | Parking frame identification method and device, electronic equipment and storage medium | |
| CN114120260A (en) | Method and system for identifying travelable area, computer device, and storage medium | |
| CN112150538B (en) | Method and device for determining vehicle pose in three-dimensional map construction process | |
| CN116884003B (en) | Picture automatic labeling method and device, electronic equipment and storage medium | |
| CN113362372B (en) | Single target tracking method and computer readable medium | |
| US20240046601A1 (en) | Deep recognition model training method, electronic device and readable storage medium |
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 |