CN114241395A - Roadside parking berth abnormity identification method and device based on berth number - Google Patents
Roadside parking berth abnormity identification method and device based on berth number Download PDFInfo
- Publication number
- CN114241395A CN114241395A CN202111613262.1A CN202111613262A CN114241395A CN 114241395 A CN114241395 A CN 114241395A CN 202111613262 A CN202111613262 A CN 202111613262A CN 114241395 A CN114241395 A CN 114241395A
- Authority
- CN
- China
- Prior art keywords
- parking
- berth
- video frames
- video frame
- roadside
- 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
Images
Classifications
-
- 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
- G06N3/045—Combinations of networks
-
- 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
- G06N3/08—Learning methods
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/14—Traffic control systems for road vehicles indicating individual free spaces in parking areas
- G08G1/145—Traffic control systems for road vehicles indicating individual free spaces in parking areas where the indication depends on the parking areas
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
- H04N7/181—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a plurality of remote sources
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Image Analysis (AREA)
- Traffic Control Systems (AREA)
Abstract
The application discloses a roadside parking berth abnormity identification method and device based on berth numbers. The method comprises the following steps: acquiring a plurality of video frames, and acquiring a berth number position and berth number information corresponding to each video frame according to the plurality of video frames; calculating the area intersection ratio of the berth numbers corresponding to the berth number information in two adjacent video frames according to the berth number position; if the area intersection ratio of the parking numbers is smaller than a preset threshold value, the roadside parking positions corresponding to the parking number information are abnormal, otherwise, the roadside parking positions are not abnormal; if the roadside berthing position is abnormal, judging whether an obstacle exists at the berthing position corresponding to each video frame in the two adjacent video frames; and if the position of the parking number has the obstacle, calculating the intersection and comparison of the areas of the parking numbers corresponding to the parking number information in the video frame and the next video frame, judging whether the roadside parking corresponding to the parking number information is abnormal or not, and if the position of the parking number has no obstacle, judging that the roadside parking is abnormal.
Description
Technical Field
The application relates to the technical field of target detection, in particular to a roadside parking berth abnormity identification method and device based on berth numbers.
Background
In recent years, difficulty and disorder in parking become one of the main problems to be solved urgently in urban traffic development in China. Traffic management departments in various regions actively promote and solve the problems of difficult and disordered roadside parking. In order to effectively solve the problems of difficult and disordered parking at the roadside, various traffic management departments try and implement various intelligent parking management technologies. The roadside parking management technology based on the high-level video is one of the mainstream schemes at present, data acquisition is carried out by installing the high-level video camera at the roadside, license plate detection, license plate recognition and the like are carried out by utilizing a visual algorithm to realize parking charging and management, the method is not easy to suffer from artificial interference and damage after installation, the intelligent processing does not need artificial operation, a plurality of adjacent parking positions can be processed simultaneously, and the roadside parking management technology has outstanding advantages compared with the schemes such as video piles and geomagnetic sensors.
In the traditional parking space abnormity identification method, a marked parking space position in video data acquired by an installed high-level camera is used as a criterion for judging whether a vehicle entering or exiting a parking space has parking behaviors. However, when the vehicle is influenced by weather such as strong wind and heavy rain, obstacle shielding such as green plants, vehicle collision, personnel construction and other factors, the high-level video camera is displaced, and the original marked berth position is wrong, so that the calculation and judgment of the vehicle entering and leaving fields are influenced.
Content of application
The method aims to solve the technical problem that in a traditional parking position abnormity identification method, an original marked parking position is changed mistakenly due to the fact that a high-position video camera is shifted. In order to achieve the purpose, the application provides a roadside parking berth abnormity identification method and device based on the berth number.
The application provides a roadside parking berth abnormity identification method based on a berth number, which comprises the following steps:
acquiring a plurality of video frames, and acquiring a berth number position and berth number information corresponding to each video frame according to the plurality of video frames;
calculating the area intersection ratio of the berth numbers corresponding to the berth number information in two adjacent video frames according to the berth number position;
if the area intersection ratio of the parking numbers is smaller than a preset threshold value, the roadside parking positions corresponding to the parking number information are abnormal, otherwise, the roadside parking positions are not abnormal;
if the roadside berthing position is abnormal, judging whether an obstacle exists at the berthing position corresponding to each video frame in the two adjacent video frames;
if the position of the parking number has an obstacle, calculating the area intersection and comparison of the parking number corresponding to the parking number information in the video frame and the next video frame, and judging whether the roadside parking position corresponding to the parking number information is abnormal or not, if the position of the parking number has no obstacle, judging that the roadside parking position is abnormal; wherein the next video frame comprises video frames other than the two adjacent video frames.
In one embodiment, the calculating, according to the position of the berth number, a berth number area intersection ratio corresponding to the berth number information in two adjacent video frames includes:
obtaining the coordinate positions of the upper left corner and the lower right corner of the position of the parking number in each video frame according to the position of the parking number corresponding to each video frame in the two adjacent video frames;
calculating the area of the parking number position according to the coordinate positions of the upper left corner and the lower right corner of the parking number position;
and calculating the area intersection ratio of the berth numbers corresponding to the berth number information in the two adjacent video frames according to the area of the berth number position.
In one embodiment, the area intersection ratio of the berth numbers corresponding to the berth number information in the two adjacent video frames is:
where M represents the area of the position of the parking number in the first video frame and N represents the area of the position of the parking number in the second video frame.
In one embodiment, the obtaining, according to the plurality of video frames, a parking number position and parking number information corresponding to each video frame includes:
inputting the video frames into a target detection network, and outputting a plurality of berth number positions corresponding to each video frame;
according to the plurality of berth number positions, cutting each video frame to obtain a plurality of berth number target images;
and inputting the plurality of berth number target images into a full convolution neural network, and outputting berth number information corresponding to each berth number target image.
In one embodiment, the obtaining of the plurality of video frames and the obtaining of the parking number position and the parking number information corresponding to each video frame according to the plurality of video frames are performed, where the plurality of video frames are a plurality of video frames of different time periods collected by the same high-order video camera.
In one embodiment, the present application provides a roadside parking lot abnormality recognition device based on a parking lot number, including:
the data acquisition module is used for acquiring a plurality of video frames and acquiring the corresponding berth number position and berth number information of each video frame according to the plurality of video frames;
the area intersection ratio acquisition module is used for calculating the intersection ratio of the areas of the berth numbers corresponding to the berth number information in two adjacent video frames according to the berth number positions;
the first judgment module is used for judging that the roadside berthing corresponding to the berthing number information is abnormal if the merging ratio of the berthing number areas is smaller than a preset threshold value, and otherwise, judging that the roadside berthing is not abnormal;
the second judgment module is used for judging whether an obstacle exists at the position of the parking position number if the roadside parking position is abnormal;
and the third judging module is used for calculating the intersection and comparison of the areas of the berth numbers corresponding to the berth number information in the video frame and the next video frame if the berth number position has an obstacle, judging whether the roadside berth corresponding to the berth number information is abnormal or not, and if the berth number position has no obstacle, judging that the roadside berth is abnormal.
In one embodiment, the area intersection ratio obtaining module includes:
the coordinate position acquisition module is used for acquiring the coordinate positions of the upper left corner and the lower right corner of the parking number position in each video frame according to the parking number position corresponding to each video frame in the two adjacent video frames;
the area acquisition module is used for calculating the area of the position of the parking number according to the coordinate positions of the upper left corner and the lower right corner of the position of the parking number;
and the calculating module is used for calculating the area intersection ratio of the berth numbers corresponding to the berth number information in the two adjacent video frames according to the area of the berth number position.
In one embodiment, the area intersection ratio of the berth numbers corresponding to the berth number information in two adjacent video frames in the computing module is:
where M represents the area of the position of the parking number in the first video frame and N represents the area of the position of the parking number in the second video frame.
In one embodiment, the data acquisition module comprises:
the parking position acquiring module is used for inputting the video frames to a target detection network and outputting a plurality of parking position corresponding to each video frame;
the berth number target image acquisition module is used for cutting each video frame according to the berth number positions to obtain a plurality of berth number target images;
and the berth number information acquisition module is used for inputting the multiple berth number target images into the full convolution neural network and outputting berth number information corresponding to each berth number target image.
In one embodiment, the plurality of video frames in the data acquisition module are a plurality of video frames of different time periods collected by the same high-order video camera.
In the method for recognizing the roadside parking lot abnormality based on the parking lot number, the area intersection and comparison calculation can be carried out on the parking lot number positions in the adjacent video frames, and the roadside parking lot abnormality is judged based on the area intersection and comparison. The roadside parking lot abnormality identification method based on the parking lot number can further perform obstacle identification judgment according to the preliminary judgment of the abnormal state of the roadside parking lot, and perform secondary area intersection and comparison calculation based on the obstacle identification judgment result to realize double judgment, thereby realizing the abnormality identification of the roadside parking lot. According to the roadside parking lot abnormality identification method based on the parking lot number, the change conditions of the same parking lot number information in different video frames can be compared, and whether roadside parking lots are abnormal or not can be identified more clearly according to the judgment of the area intersection ratio and the obstacles. By the roadside parking berth abnormity identification method based on the berth number, identification accuracy and robustness are improved, and calculation and judgment accuracy of vehicle entrance and exit is improved.
Drawings
Fig. 1 is a schematic overall flow chart of the roadside parking lot abnormality identification method based on the parking lot number provided by the present application.
Fig. 2 is a schematic view of the overall structure of the roadside parking lot abnormality recognition device for a parking lot number according to the present application.
Detailed Description
The technical solution of the present application is further described in detail by the accompanying drawings and examples.
Referring to fig. 1, the present application provides a roadside parking lot abnormality identification method based on a parking lot number, including:
s10, acquiring a plurality of video frames, and acquiring the corresponding berth number position and berth number information of each video frame according to the plurality of video frames;
s20, calculating the area intersection ratio of the berth numbers corresponding to the berth number information in two adjacent video frames according to the berth number position;
s30, if the area intersection ratio of the parking numbers is smaller than a preset threshold value, the roadside parking positions corresponding to the parking number information are abnormal, otherwise, the roadside parking positions are not abnormal;
s40, if the roadside berthing is abnormal, judging whether an obstacle exists at the berthing number position corresponding to each video frame in the two adjacent video frames;
s50, if the position of the berth number has an obstacle, calculating the intersection and comparison of the areas of the berth numbers corresponding to the berth number information in the video frame and the next video frame, and judging whether the roadside berth corresponding to the berth number information is abnormal or not, if the position of the berth number has no obstacle, judging that the roadside berth is abnormal; wherein the next video frame comprises video frames other than the two adjacent video frames.
In S10, a plurality of parking number positions may be obtained from each video frame. One parking number position corresponds to one parking number information. The same parking number may exist between multiple video frames, for example, the same parking number may exist between a first video frame and a second video frame. The position of the parking number may represent the position occupied by the sequence of parking number, which may also be understood as a region of the parking number. The parking number information may represent character sequence information of the parking number.
In S20, the area of the parking number can be understood as the area of the position occupied by the parking number. One video frame in two adjacent video frames corresponds to one position area, and the other video frame corresponds to the other position area. Alternatively, it can also be understood that two adjacent video frames are represented as a first video frame and a second video frame, the first video frame corresponds to the position area of one parking number, and the second video frame corresponds to the position area of one parking number. And respectively solving intersection or union between the areas of the two positions, and comparing the intersection with the union to obtain an area union ratio. The coincidence degree of the positions of the berth numbers in the two adjacent video frames can be represented by calculating the intersection ratio of the two position areas corresponding to the two adjacent video frames. Therefore, whether the corresponding roadside berthage is abnormal or not can be judged through the area intersection ratio.
In S30, the comparison relationship between the areas occupied by the same parking number in two adjacent video frames can be determined by area intersection and comparison, and the degree of coincidence of the same parking number in two adjacent video frames can be determined, so as to determine whether the parking number is abnormal. And further judging whether the roadside berth is abnormal or not according to whether the berth number position is abnormal or not. The abnormal position of the parking number comprises the shielding condition of obstacles such as green plants, vehicle collision, personnel construction and the like.
In S40, the obstacle may include a green plant, a vehicle collision, a person construction, or the like. The abnormal position of the parking number can be understood as that an obstacle exists in the parking number position, and the parking number area is shielded and covered. When the position of the parking position number is abnormal, whether the position is abnormal caused by the obstacle is further judged. On the contrary, if the position of the parking number is not abnormal, the roadside parking position corresponding to the parking number information is not abnormal, so that the vehicle can be parked, and the vehicle can be judged to enter or exit.
In S50, when the position of the parking number is abnormal and there is an obstacle blocking the position of the parking number, area intersection and comparison calculation is performed on the areas of the parking numbers corresponding to the same parking number information in the video frame and the next video frame, and S30 to S50 are continuously and circularly performed to perform secondary judgment, and whether there is an abnormal condition in the roadside parking position is determined again. Otherwise, if the position of the parking position number is abnormal and no obstacle exists in the position of the parking position number, the roadside parking position is abnormal. If the roadside berth is abnormal, the vehicle cannot be parked, and the vehicle cannot be judged to enter or exit.
The roadside parking lot abnormality identification method based on the parking lot number can perform area intersection comparison calculation on the positions of the parking lot numbers in adjacent video frames, and judge the abnormal state of roadside parking lots based on the area intersection comparison. The roadside parking lot abnormality identification method based on the parking lot number can further perform obstacle identification judgment according to the preliminary judgment of the abnormal state of the roadside parking lot, and perform secondary area intersection and comparison calculation based on the obstacle identification judgment result to realize double judgment, thereby realizing the abnormality identification of the roadside parking lot. According to the roadside parking lot abnormality identification method based on the parking lot number, the change conditions of the same parking lot number information in different video frames can be compared, and whether roadside parking lots are abnormal or not can be identified more clearly according to the judgment of the area intersection ratio and the obstacles. By the roadside parking berth abnormity identification method based on the berth number, identification accuracy and robustness are improved, and calculation and judgment accuracy of vehicle entrance and exit is improved.
In one embodiment, S10, multiple video frames are obtained, and according to the multiple video frames, the multiple video frames are obtained in different time periods collected by the same high-level video camera, where the parking number position and the parking number information correspond to each video frame.
In this embodiment, a plurality of video frames of different time periods collected by a high-order video camera at the same position are obtained. In one embodiment, data acquisition is performed every minute, obtaining a plurality of video frames. And judging whether the berth abnormity occurs in one minute or not according to the plurality of video frames.
In one embodiment, S10, obtain a plurality of video frames, and obtain, according to the plurality of video frames, a corresponding parking position and parking number information of each video frame, and obtain, through the parking number detection and identification model, the parking position and parking number information of each video frame. A plurality of parking numbers exist in one video frame, and one parking number corresponds to one parking number information one by one. The parking number position may also be understood as a parking number region. The position of the parking number and the information of the parking number are in unique corresponding relation.
And acquiring a plurality of parking numbers in a plurality of video frames according to the parking number detection and identification model, and acquiring parking number information and parking number position information corresponding to the plurality of parking numbers respectively. The parking number position information includes coordinate positions of the upper left corner and the lower right corner of the parking, and width and height information of the parking number. The parking number information includes character sequence information of the parking number.
In one embodiment, the berthage number detection and identification model can adopt algorithms such as a deep learning algorithm, a machine learning algorithm and the like for detection and identification.
According to the berth number detection and identification model, acquiring a plurality of berth numbers in a plurality of video frames, and acquiring berth number information and berth number positions corresponding to the plurality of berth numbers respectively, the method comprises the following steps:
s110, inputting a plurality of video frames into a target detection network, and outputting a plurality of berth number positions corresponding to each video frame;
s120, cutting each video frame according to the plurality of berth number positions to obtain a plurality of berth number target images;
and S130, inputting the plurality of berth number target images into a full convolution neural network, and outputting berth number information corresponding to each berth number target image.
In this embodiment, the target detection network includes, but is not limited to, two-stage or single-stage target detection methods such as fast-RCNN, YOLO, SSD, and the like. The detection frame type information includes two types, one type is a parking number area detection frame, and the other type is a non-parking number detection frame. The detection frame position information may be represented by coordinate information of the detection frame, such as the center point coordinate and the width and height information of the minimum rectangular frame. Normalization layers for full convolution neural networks include, but are not limited to, example normalization layers, adaptive example normalization layers, and the like. The nonlinear activation layer of the full convolution neural network includes, but is not limited to, nonlinear activation functions such as ReLu, Leaky ReLu, and the like. The full convolutional neural network can recover the class to which each pixel belongs from the abstract features. I.e. from image-level classification to pixel-level classification, the berth number corresponding to each berth can be accurately identified.
In one embodiment, the step S20 of calculating an area intersection ratio corresponding to the berth number information in two adjacent video frames according to the berth number position includes:
s210, obtaining the coordinate positions of the upper left corner and the lower right corner of the position of the parking number in each video frame according to the position of the parking number corresponding to each video frame in two adjacent video frames;
s220, calculating the area of the position of the parking number according to the coordinate positions of the upper left corner and the lower right corner of the position of the parking number;
and S230, calculating the area intersection ratio corresponding to the berth number information in two adjacent video frames according to the area of the berth number position.
In S210, two adjacent video frames may be represented as a first video frame and a second video frame. According to the corresponding parking position number positions of the same corresponding parking position number in the first video frame and the second video frame, the coordinate positions of the upper left corner and the lower right corner corresponding to each parking position number position can be obtained respectively. The same parking number corresponds to an upper left corner coordinate position and a lower right corner coordinate position in the first video frame respectively, and corresponds to an upper left corner coordinate position and a lower right corner coordinate position in the second video frame respectively.
In S220, the same parking number is in the first video frame, and according to the coordinate position of the upper left corner and the coordinate position of the lower right corner, the area of the parking number position in the first video frame can be calculated and obtained, and can also be understood as the area of the parking number region. And the same parking number is in the second video frame, and the area of the parking number position in the second video frame can be calculated and obtained according to the coordinate position of the upper left corner and the coordinate position of the lower right corner.
In S230, according to the area of the position of the parking position in the first video frame and the area of the position of the parking position in the second video frame, an area intersection ratio corresponding to the same parking position number in two adjacent video frames is calculated. The area intersection ratio can be understood as the ratio of the intersection and union of the area of the position of the parking number in the first video frame and the area of the position of the parking number in the second video frame.
The coincidence degree of the positions of the berth numbers in the two adjacent video frames can be represented by calculating the intersection ratio of the two position areas corresponding to the two adjacent video frames. Therefore, whether the corresponding roadside berthage is abnormal or not can be judged through the area intersection ratio. The first video frame and the second video frame form a reference contrast with respect to each other. The first video frame may serve as a reference for the second video frame, which may serve as a reference for the first video frame. By comparing two adjacent video frames, whether the parking position corresponding to the same parking number is covered and occupied by the barrier can be judged. Therefore, a plurality of video frames can be mutually used as reference, whether the parking position is abnormal or not is judged within a certain time period, and the accuracy of abnormal parking position identification is improved.
In one embodiment, in S230, the area intersection ratio corresponding to the berthage number information in two adjacent video frames is:
where M represents the area of the position of the parking number in the first video frame and N represents the area of the position of the parking number in the second video frame.
In this embodiment, the area of the position of the parking position in the first video frame and the area of the position of the parking position in the second video frame are calculated and obtained through the coordinate positions of the upper left corner and the lower right corner corresponding to the same parking position number in the first video frame and the second video frame. And calculating the ratio of intersection and union of the area of the position of the berth number in the first video frame and the area of the position of the berth number in the second video frame by using the intersection ratio IoU to obtain the intersection ratio of the areas corresponding to the berth number information in two adjacent video frames, namely the intersection ratio of the areas corresponding to the berth number in the two adjacent video frames. By the cross-over ratio IoU, the degree of coincidence of the position of the parking number in the first video frame with the position of the parking number in the second video frame can be achieved.
Therefore, through the area intersection comparison formula, the change conditions of the same parking number information in different video frames can be compared, whether roadside parking berths are abnormal or not can be identified more clearly, the identification accuracy and robustness are improved, and the vehicle entrance and exit can be calculated and judged accurately.
In one embodiment, S30, if the merging ratio of the parking number areas is smaller than the preset threshold, the roadside parking position corresponding to the parking number information is abnormal, otherwise, the roadside parking position has no abnormality, including:
s310, if the area intersection ratio is not greater than a preset threshold value, the roadside berth corresponding to the berth number information is not abnormal;
and S320, if the area intersection ratio is smaller than the preset threshold value, the roadside berthing corresponding to the berthing number information is abnormal.
In this embodiment, the preset threshold may be in the range of 0.8 to 1. In one embodiment, the specific value of the preset threshold may be defined according to actual conditions. And establishing a judgment basis for judging whether the roadside berth is abnormal or not by setting a preset threshold value. Comparing the area intersection ratio with a preset threshold value can judge whether the position of the berth number corresponding to the same berth number information in the adjacent video frames is abnormal or not. The presence or absence of an anomaly in the position of the parking number may be understood as whether the position of the parking number is blocked by an obstacle. If the parking position is blocked by the barrier, the parking position is abnormal, and further the roadside parking position corresponding to the parking position information is abnormal, so that the calculation and judgment of the vehicle entering and exiting time cannot be carried out. If the parking position is not shielded by the barrier, the parking position is not abnormal, and the roadside parking position corresponding to the parking position information is not abnormal, so that the vehicle entrance and exit time can be calculated and judged.
Therefore, the abnormal state of the roadside berth is judged based on the area merging comparison, the obstacle identification judgment is further carried out, the secondary area merging comparison calculation is carried out based on the obstacle identification judgment result to realize double judgment, and the roadside berth is subjected to abnormal identification. By comparing the change conditions of the same berth number information in different video frames and judging according to the area intersection ratio, whether the roadside berth is abnormal or not is more definitely identified, and the identification accuracy rate and the robustness are improved.
In one embodiment, the step S40, if the roadside parking position is abnormal, determining whether there is an obstacle in the position of the parking position number in the video frame includes:
and when the area intersection ratio is smaller than a preset threshold value, the roadside berth is abnormal. When the roadside berth is abnormal, whether obstacles such as vehicles, pedestrians, non-motor vehicles and the like exist in the video frame is judged according to the obstacle detection model. In one embodiment, whether obstacles such as vehicles, pedestrians, non-motor vehicles and the like exist in any one of the adjacent video frames is judged. Or judging whether the first video frame has obstacles such as vehicles, pedestrians, non-motor vehicles and the like. Or judging whether the second video frame has obstacles such as vehicles, pedestrians, non-motor vehicles and the like.
In one embodiment, the obstacle detection model may include a deep learning algorithm, a machine learning algorithm, or the like. The obstacle detection model can be a YOLO neural network, an obstacle image data set is created, model training is carried out on the YOLO neural network, and the trained YOLO neural network is obtained. And detecting the obstacle according to the trained YOLO neural network, and outputting a regression frame of the obstacle and the obstacle category.
In one embodiment, S50, if there is an obstacle in the position of the parking number, calculating an intersection and comparison of areas of the parking numbers corresponding to the parking number information in the video frame and the next video frame, and determining whether the roadside parking position corresponding to the parking number information is abnormal, and if there is no obstacle in the position of the parking number, the roadside parking position is abnormal.
In this embodiment, when the position of the parking number in the video frame does not have obstacles such as vehicles, pedestrians, non-motor vehicles, and the like, it is determined that the roadside parking position is abnormal. When obstacles such as vehicles, pedestrians, non-motor vehicles and the like exist in the video frames, the area merging ratio of the parking numbers is calculated again by using other video frames, the calculation of the double area merging ratio is realized, and whether the roadside parking is abnormal or not is confirmed again. A reference contrast is formed by the first video frame, the second video frame and the next video frame except the first video frame and the second video frame. By comparing a plurality of video frames, whether the roadside berths corresponding to the same berth number information are abnormal or not can be judged. Therefore, a plurality of video frames can be mutually used as reference, whether the parking position is abnormal or not is judged within a certain time period, and the accuracy of abnormal parking position identification is improved. Therefore, according to the roadside parking lot abnormality identification method based on the parking lot number, the area intersection ratio of the parking lot numbers corresponding to the parking lot number information in the video frames can be calculated for multiple times, and abnormality caused by errors or false detection is avoided. By the roadside parking lot abnormality identification method based on the parking lot number, multiple times of verification can be achieved, and accuracy of roadside parking lot abnormality identification is improved.
Referring to fig. 2, the present application provides a roadside parking lot abnormality recognition apparatus 100 based on a parking lot number. The roadside parking lot abnormality recognition device 100 based on the parking lot number includes a data acquisition module 10, an area intersection ratio acquisition module 20, a first judgment module 30, a second judgment module 40, and a third judgment module 50. The data obtaining module 10 is configured to obtain a plurality of video frames, and obtain a parking number position and parking number information corresponding to each video frame according to the plurality of video frames. The area cross-over ratio obtaining module 20 is configured to calculate, according to the position of the parking number, a parking number area cross-over ratio corresponding to the parking number information in two adjacent video frames. The first determining module 30 is configured to determine that the roadside berthing position corresponding to the berthing number information is abnormal if the merging ratio of the berthing number areas is smaller than a preset threshold, and otherwise, determine that the roadside berthing position is not abnormal. The second determining module 40 is configured to determine whether an obstacle exists at a position of a parking number corresponding to each of two adjacent video frames if the roadside parking position is abnormal. The third judging module 50 is configured to calculate a merging ratio of the area of the parking number corresponding to the parking number information in the video frame and the next video frame if the position of the parking number has an obstacle, and judge whether the roadside parking position corresponding to the parking number information is abnormal, and if the position of the parking number does not have an obstacle, the roadside parking position is abnormal; wherein the next video frame comprises video frames other than the adjacent two video frames.
In this embodiment, the relevant description of the data obtaining module 10 may refer to the relevant description of S10 in the above embodiment. The description of the area intersection ratio obtaining module 20 may refer to the description of S20 in the above embodiment. The related description of the first judging module 30 can refer to the related description of S30 in the above embodiment. The related description of the second judging module 40 can refer to the related description of S40 in the above embodiment. The related description of the third judging module 50 can refer to the related description of S50 in the above embodiment.
The roadside parking lot abnormality recognition device based on the parking lot number can perform area intersection and comparison calculation on the positions of the parking lot numbers in adjacent video frames, and judge the abnormal state of roadside parking lots based on the area intersection and comparison. The roadside parking lot abnormality recognition device based on the parking lot number can further perform obstacle recognition and judgment according to preliminary judgment of the abnormal state of the roadside parking lot, and perform secondary area intersection and comparison calculation based on the obstacle recognition and judgment result to realize multiple judgment, so that the roadside parking lot is recognized abnormally. By the roadside parking lot abnormality recognition device based on the parking lot number, the change conditions of the same parking lot number information in different video frames can be compared, and whether roadside parking lots are abnormal or not can be recognized more clearly according to the judgment of the area intersection ratio and the obstacles. By the roadside parking berth abnormity identification device based on the berth number, the identification accuracy and robustness are improved, and the calculation and judgment accuracy of the vehicle entering and leaving is improved.
In one embodiment, the area intersection ratio acquiring module 20 includes a coordinate position acquiring module (not shown), an area acquiring module (not shown), and a calculating module (not shown). The coordinate position obtaining module is used for obtaining the coordinate positions of the upper left corner and the lower right corner of the position of the parking number in each video frame according to the position of the parking number corresponding to each video frame in the two adjacent video frames. The area obtaining module is used for calculating the area of the position of the parking number according to the coordinate positions of the upper left corner and the lower right corner of the position of the parking number. The calculation module is used for calculating the area intersection ratio of the berth numbers corresponding to the berth number information in two adjacent video frames according to the area of the berth number position.
In this embodiment, the relevant description of the coordinate position obtaining module may refer to the relevant description of S210 in the above embodiment. The relevant description of the area acquisition module may refer to the relevant description of S220 in the above embodiment. The relevant description of the calculation module can refer to the relevant description of S230 in the above embodiment.
In one embodiment, the area intersection ratio of the berth numbers corresponding to the berth number information in two adjacent video frames in the calculation module is:
where M represents the area of the position of the parking number in the first video frame and N represents the area of the position of the parking number in the second video frame.
In this embodiment, the area intersection ratio of the berth numbers corresponding to the berth number information in two adjacent video frames in the calculation module may refer to the related description in S230 in the above embodiment.
In one embodiment, the data obtaining module 10 includes a parking number position obtaining module (not shown), a parking number target image obtaining module (not shown), and a parking number information obtaining module (not shown). The parking position acquiring module is used for inputting a plurality of video frames into the target detection network and outputting a plurality of parking position corresponding to each video frame. The berth number target image acquisition module is used for cutting each video frame according to the berth number positions to obtain a plurality of berth number target images. The berth number information acquisition module is used for inputting a plurality of berth number target images into the full convolution neural network and outputting berth number information corresponding to each berth number target image.
In this embodiment, the relevant description of the parking position number obtaining module may refer to the relevant description of S110 in the above embodiment. The relevant description of the berth number target image acquisition module may refer to the relevant description of S120 in the above embodiment. The relevant description of the parking number information acquiring module may refer to the relevant description of S130 in the above embodiment.
In one embodiment, the multiple video frames in the data acquisition module 10 are multiple video frames of different time periods collected by the same high-order video camera.
In this embodiment, reference may be made to the description about S10 in the above embodiment for the related description of the plurality of video frames in the data obtaining module 10.
Those of skill in the art will further appreciate that the various illustrative logical blocks, units, and steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate the interchangeability of hardware and software, various illustrative components, elements, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design requirements of the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the embodiments of the present application.
The various illustrative logical blocks, or elements described in this application may be implemented or operated by a general purpose processor, a digital signal processor, an Application Specific Integrated Circuit (ASIC), a field programmable gate array or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described. A general-purpose processor may be a microprocessor, but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine. A processor may also be implemented as a combination of computing devices, e.g., a digital signal processor and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a digital signal processor core, or any other similar configuration.
The steps of a method or algorithm described in the embodiments herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module may be stored in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. For example, a storage medium may be coupled to the processor such the processor can read information from, and write information to, the storage medium. In the alternative, the storage medium may be integral to the processor. The processor and the storage medium may reside in an ASIC, which may be located in a user terminal. In the alternative, the processor and the storage medium may reside in different components in a user terminal.
The above-mentioned embodiments, objects, technical solutions and advantages of the present application are described in further detail, it should be understood that the above-mentioned embodiments are merely exemplary embodiments of the present application, and are not intended to limit the scope of the present application, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present application should be included in the scope of the present application.
Claims (10)
1. A roadside parking berth abnormality identification method based on berth numbers is characterized by comprising the following steps:
acquiring a plurality of video frames, and acquiring a berth number position and berth number information corresponding to each video frame according to the plurality of video frames;
calculating the area intersection ratio of the berth numbers corresponding to the berth number information in two adjacent video frames according to the berth number position;
if the area intersection ratio of the parking numbers is smaller than a preset threshold value, the roadside parking positions corresponding to the parking number information are abnormal, otherwise, the roadside parking positions are not abnormal;
if the roadside berthing position is abnormal, judging whether an obstacle exists at the berthing position corresponding to each video frame in the two adjacent video frames;
if the position of the parking number has an obstacle, calculating the area intersection and comparison of the parking number corresponding to the parking number information in the video frame and the next video frame, and judging whether the roadside parking position corresponding to the parking number information is abnormal or not, if the position of the parking number has no obstacle, judging that the roadside parking position is abnormal; wherein the next video frame comprises video frames other than the two adjacent video frames.
2. The method for recognizing the roadside parking lot abnormality based on the parking lot number according to claim 1, wherein the calculating the intersection and combination ratio of the parking lot number areas corresponding to the parking lot number information in two adjacent video frames according to the parking lot number position comprises:
obtaining the coordinate positions of the upper left corner and the lower right corner of the position of the parking number in each video frame according to the position of the parking number corresponding to each video frame in the two adjacent video frames;
calculating the area of the parking number position according to the coordinate positions of the upper left corner and the lower right corner of the parking number position;
and calculating the area intersection ratio of the berth numbers corresponding to the berth number information in the two adjacent video frames according to the area of the berth number position.
3. The roadside parking lot abnormality identification method based on the parking lot number as recited in claim 2, wherein the area intersection ratio of the parking lot numbers corresponding to the parking lot number information in the two adjacent video frames is:
wherein M represents the area of the position of the berth number in the first video frame of the two adjacent video frames, and N represents the area of the position of the berth number in the second video frame.
4. The method for recognizing the roadside parking lot abnormality based on the parking lot number as recited in claim 1, wherein the obtaining the parking lot number position and the parking lot number information corresponding to each video frame according to the plurality of video frames comprises:
inputting the video frames into a target detection network, and outputting a plurality of berth number positions corresponding to each video frame;
according to the plurality of berth number positions, cutting each video frame to obtain a plurality of berth number target images;
and inputting the plurality of berth number target images into a full convolution neural network, and outputting berth number information corresponding to each berth number target image.
5. The roadside parking lot abnormality identification method based on the parking lot number as recited in claim 1, wherein the obtaining a plurality of video frames and obtaining the parking lot number position and the parking lot number information corresponding to each video frame according to the plurality of video frames, wherein the plurality of video frames are a plurality of video frames collected by a same high-order video camera in different time periods.
6. A roadside parking berth abnormality recognition device based on a berth number, characterized by comprising:
the data acquisition module is used for acquiring a plurality of video frames and acquiring the corresponding berth number position and berth number information of each video frame according to the plurality of video frames;
the area intersection ratio acquisition module is used for calculating the intersection ratio of the areas of the berth numbers corresponding to the berth number information in two adjacent video frames according to the berth number positions;
the first judgment module is used for judging that the roadside berthing corresponding to the berthing number information is abnormal if the merging ratio of the berthing number areas is smaller than a preset threshold value, and otherwise, judging that the roadside berthing is not abnormal;
the second judgment module is used for judging whether an obstacle exists at the position of the berth number corresponding to each video frame in the two adjacent video frames or not if the roadside berth is abnormal;
a third determining module, configured to calculate a cross-over and comparison between areas of the parking numbers corresponding to the parking number information in the video frame and a next video frame if an obstacle exists at the parking number position, and determine whether a roadside parking position corresponding to the parking number information is abnormal, and if an obstacle does not exist at the parking number position, the roadside parking position is abnormal; wherein the next video frame comprises video frames other than the two adjacent video frames.
7. The roadside parking lot abnormality recognition device based on the parking lot number as recited in claim 6, wherein the area intersection ratio obtaining module comprises:
the coordinate position acquisition module is used for acquiring the coordinate positions of the upper left corner and the lower right corner of the parking number position in each video frame according to the parking number position corresponding to each video frame in the two adjacent video frames;
the area acquisition module is used for calculating the area of the position of the parking number according to the coordinate positions of the upper left corner and the lower right corner of the position of the parking number;
and the calculating module is used for calculating the area intersection ratio of the berth numbers corresponding to the berth number information in the two adjacent video frames according to the area of the berth number position.
8. The roadside parking lot abnormality recognition device based on the parking lot number as recited in claim 7, wherein the area intersection ratio of the parking lot numbers corresponding to the parking lot number information in two adjacent video frames in the calculation module is:
where M represents the area of the position of the parking number in the first video frame and N represents the area of the position of the parking number in the second video frame.
9. The roadside parking lot abnormality recognition device based on the parking lot number as recited in claim 6, wherein the data acquisition module comprises:
the parking position acquiring module is used for inputting the video frames to a target detection network and outputting a plurality of parking position corresponding to each video frame;
the berth number target image acquisition module is used for cutting each video frame according to the berth number positions to obtain a plurality of berth number target images;
and the berth number information acquisition module is used for inputting the multiple berth number target images into the full convolution neural network and outputting berth number information corresponding to each berth number target image.
10. The roadside parking lot abnormality recognition device according to claim 6, wherein the plurality of video frames in the data acquisition module are a plurality of video frames of different time periods collected by the same high-order video camera.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111613262.1A CN114241395A (en) | 2021-12-27 | 2021-12-27 | Roadside parking berth abnormity identification method and device based on berth number |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111613262.1A CN114241395A (en) | 2021-12-27 | 2021-12-27 | Roadside parking berth abnormity identification method and device based on berth number |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114241395A true CN114241395A (en) | 2022-03-25 |
Family
ID=80763404
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111613262.1A Pending CN114241395A (en) | 2021-12-27 | 2021-12-27 | Roadside parking berth abnormity identification method and device based on berth number |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114241395A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115082571A (en) * | 2022-07-20 | 2022-09-20 | 深圳云游四海信息科技有限公司 | Anomaly detection method and system for in-road parking camera |
| CN116863712A (en) * | 2023-09-01 | 2023-10-10 | 成都宜泊信息科技有限公司 | Method and system for accurately judging vehicle parking position of road side inspection vehicle |
-
2021
- 2021-12-27 CN CN202111613262.1A patent/CN114241395A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115082571A (en) * | 2022-07-20 | 2022-09-20 | 深圳云游四海信息科技有限公司 | Anomaly detection method and system for in-road parking camera |
| CN116863712A (en) * | 2023-09-01 | 2023-10-10 | 成都宜泊信息科技有限公司 | Method and system for accurately judging vehicle parking position of road side inspection vehicle |
| CN116863712B (en) * | 2023-09-01 | 2023-11-28 | 成都宜泊信息科技有限公司 | Method and system for accurately judging vehicle parking position of road side inspection vehicle |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR102054061B1 (en) | Recognition method of Vehicle Licence Plate | |
| KR102197946B1 (en) | object recognition and counting method using deep learning artificial intelligence technology | |
| CN106652465B (en) | Method and system for identifying abnormal driving behaviors on road | |
| Rahman et al. | A real time vehicle's license plate recognition system | |
| CN114241395A (en) | Roadside parking berth abnormity identification method and device based on berth number | |
| CN111382704B (en) | Vehicle line pressing violation judging method and device based on deep learning and storage medium | |
| CN111523464A (en) | Method and device for detecting illegal lane change of vehicle | |
| KR101908611B1 (en) | Parking control method for vehicle | |
| CN108335489A (en) | Vehicle on highway behavior semantic analysis and abnormal behaviour monitoring system and method | |
| CN106373426A (en) | Computer vision-based parking space and illegal lane occupying parking monitoring method | |
| CN110298300B (en) | Method for detecting vehicle illegal line pressing | |
| CN102722704A (en) | Method and system for recognizing vehicle license plate by integrating video dynamic tracking | |
| CN100454355C (en) | Video method for collecting information of vehicle flowrate on road in real time | |
| CN113205692B (en) | Automatic identification method for road side parking position abnormal change | |
| CN112185098B (en) | Shared bicycle monitoring method and system based on city monitoring video | |
| CN104392239A (en) | License plate identification method and system | |
| CN108154146A (en) | A kind of car tracing method based on image identification | |
| CN115116012A (en) | Method and system for detecting parking state of vehicle parking space based on target detection algorithm | |
| CN101222618A (en) | Methods and systems for identifying events for a vehicle | |
| CN114141022B (en) | Emergency lane occupation behavior detection method and device, electronic equipment and storage medium | |
| CN112560814A (en) | Method for identifying vehicles entering and exiting parking spaces | |
| CN113206995A (en) | Method for judging deflection angle of parking line based on video frame | |
| CN112906428A (en) | Image detection area acquisition method and space use condition judgment method | |
| CN113128264B (en) | Vehicle region determining method and device and electronic equipment | |
| Bachtiar et al. | Parking management by means of computer vision |
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 |