CN108694387B - False license plate filtering method and device - Google Patents
False license plate filtering method and device Download PDFInfo
- Publication number
- CN108694387B CN108694387B CN201810459996.0A CN201810459996A CN108694387B CN 108694387 B CN108694387 B CN 108694387B CN 201810459996 A CN201810459996 A CN 201810459996A CN 108694387 B CN108694387 B CN 108694387B
- Authority
- CN
- China
- Prior art keywords
- license plate
- vehicle
- module
- false
- region
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
- G06V20/625—License plates
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
- Traffic Control Systems (AREA)
- Image Processing (AREA)
Abstract
The invention provides a false license plate filtering method, which comprises the following steps: the method comprises the steps of scene image acquisition and input, license plate queue acquisition, license plate vertical displacement criterion, vehicle body vertical displacement criterion, vehicle detection criterion, vehicle and license plate position relation criterion, vertical edge detection criterion, real license plate processing and false license plate processing. Compared with the existing license plate recognition technology, the method can effectively filter false license plates.
Description
Technical Field
The invention relates to image processing, video monitoring and intelligent traffic, in particular to a license plate recognition method and device.
Background
With the increasingly modern city, the holding capacity of motor vehicles is continuously increased, and the intelligent traffic system plays an increasingly important role. The license plate recognition technology is an important component of a modern intelligent traffic system and has very wide application.
However, in an actual scene, a lawless person uses a printed license plate or a mobile phone electronic license plate to perform fraud brake opening. The existing license plate recognition technology cannot distinguish false license plates.
In summary, there is a need to provide a license plate recognition method capable of filtering false license plates.
Disclosure of Invention
In view of the above, the main purpose of the present invention is to quickly recognize a license plate and to filter false license plates.
To achieve the above object, according to a first aspect of the present invention, there is provided a false license plate filtering method, comprising:
a scene image acquisition input step, wherein a video scene image is acquired or input;
a license plate queue acquisition step, namely performing license plate identification on each frame of scene image, and storing a license plate identification result into a license plate queue;
a license plate vertical displacement criterion step, namely calculating the displacement of a license plate area in the vertical direction, and if the displacement is smaller than a first threshold value, switching to a false license plate processing step;
a step of judging the vertical displacement of the vehicle body, namely setting a vehicle body area, calculating the displacement of the vehicle body area in the vertical direction, and turning to a step of processing a false license plate if the displacement is less than a first threshold value;
a vehicle detection criterion step, in which vehicle detection is carried out on the scene image, and if no vehicle region exists, a false license plate processing step is carried out;
judging the position relation between the vehicle and the license plate, and switching to a false license plate processing step according to the position relation between the vehicle region and the license plate region;
a vertical edge detection criterion step, namely calculating a difference image of two adjacent frames of scene images, carrying out edge detection, and if the number of vertical edges is not less than a second threshold value, switching to a false license plate processing step;
a step of processing a real license plate, namely, considering that the current license plate is the real license plate, outputting a license plate identification result and emptying a license plate queue;
and a false license plate processing step, namely considering that the current license plate is a false license plate, and filtering the license plate.
The vehicle detection method comprises a vehicle plate vertical displacement judging step, a vehicle body vertical displacement judging step, a vehicle detection judging step, a vehicle and plate position relation judging step and a vertical edge detection judging step, wherein the sequences of the vehicle plate vertical displacement judging step, the vehicle body vertical displacement judging step, the vehicle detection judging step, the vehicle and plate position relation judging step and the vertical edge detection judging step can be interchanged, and the vehicle detection judging step is performed before the vehicle and plate position relation judging step.
Further, the license plate recognition result in the license plate queue obtaining step includes: license plate region, license plate characters; further, the license plate recognition result also comprises confidence of license plate character recognition.
Further, the license plate vertical displacement criterion step comprises: and calculating the displacement of the current license plate region and the first license plate region in the vertical direction, if the displacement of the license plate region in the vertical direction is smaller than a first threshold value, switching to a false license plate processing step, and otherwise, switching to the next step.
Further, the vehicle body vertical displacement criterion step comprises:
the method comprises the following steps of setting a vehicle body area, namely calculating the width and the height of a license plate area in a current frame scene image, setting the vehicle body area of the current frame according to the width, the height and coordinates of the license plate area, and storing the vehicle body area into a vehicle body queue;
and a step of judging the vertical displacement of the vehicle body area, namely calculating the displacement of the current vehicle body area and the first vehicle body area in the vertical direction, if the displacement of the vehicle body area in the vertical direction is less than a first threshold value, switching to the step of processing the false license plate, and otherwise, switching to the next step.
Further, the step of judging the position relation between the vehicle and the license plate comprises the following steps: a horizontal center line judging step, a vertical distance judging step, a left and right boundary judging step and a width judging step. The horizontal center line judging step, the vertical distance judging step, the left and right boundary judging step and the width judging step can be interchanged in sequence.
Further, the horizontal center line judging step includes: and calculating the horizontal center line of the current frame vehicle area, if the license plate area is positioned above the horizontal center line, switching to a false license plate processing step, and otherwise, switching to the next step.
Further, the step of determining the vertical distance includes: and calculating the distance between the lower boundary of the current frame license plate region and the lower boundary of the vehicle region, if the distance is greater than a third threshold value, switching to a false license plate processing step, and otherwise, switching to the next step.
Further, the left and right boundary determining step includes: and respectively counting the left boundary and the right boundary of the license plate region and the vehicle region of the current frame, if the left boundary of the license plate region is positioned on the left side of the left boundary of the vehicle region or the right boundary of the license plate region is positioned on the right side of the right boundary of the vehicle region, switching to a false license plate processing step, and otherwise, switching to the next step.
Further, the width judging step includes: and counting the width of the vehicle region of the current frame, if the width of the vehicle region is smaller than a fourth threshold value, switching to a false license plate processing step, and otherwise, switching to the next step.
The vertical edge detection criterion step comprises: calculating a difference image of two adjacent frames of scene images, performing edge detection on the difference image by adopting an edge detection method, counting the number of vertical edges, switching to a false license plate processing step if the number of the vertical edges is not less than a second threshold value, and otherwise, switching to the next step.
According to another aspect of the present invention, there is provided a false license plate filtering device, the device comprising:
the scene image acquisition input module is used for acquiring or inputting a video scene image;
the license plate queue acquisition module is used for carrying out license plate identification on each frame of scene image and storing a license plate identification result into a license plate queue;
the license plate vertical displacement criterion module is used for calculating the displacement of the license plate area in the vertical direction, and if the displacement is smaller than a first threshold value, the license plate area is switched to the false license plate processing module;
the vehicle body vertical displacement criterion module is used for setting a vehicle body area, calculating the displacement of the vehicle body area in the vertical direction, and switching to the false license plate processing module if the displacement is smaller than a first threshold value;
the vehicle detection criterion module is used for detecting vehicles in the scene image, and if no vehicle region exists, the false license plate processing module is switched to;
the vehicle and license plate position relation criterion module is used for switching to the false license plate processing module according to the position relation between the vehicle region and the license plate region;
the vertical edge detection criterion module is used for calculating a difference image of two adjacent frames of scene images, carrying out edge detection, and switching to the false license plate processing module if the number of vertical edges is not less than a second threshold value;
the real license plate processing module is used for considering that the current license plate is a real license plate, outputting a license plate identification result and emptying a license plate queue;
and the false license plate processing module is used for considering that the current license plate is a false license plate and filtering the license plate.
The vehicle detection criterion module is arranged in front of the vehicle and license plate position relation criterion module.
Further, the license plate recognition result in the license plate queue obtaining module includes: license plate region, license plate characters; further, the license plate recognition result also comprises confidence of license plate character recognition.
Further, the license plate vertical displacement criterion module comprises: the license plate processing module is used for calculating the displacement of the current license plate region and the first license plate region in the vertical direction, if the displacement of the license plate region in the vertical direction is smaller than a first threshold value, the license plate processing module is switched to the false license plate processing module, and if not, the license plate processing module enters the next module.
Further, the vehicle body vertical displacement criterion module comprises:
the vehicle body area setting module is used for calculating the width and the height of a license plate area in a current frame scene image, setting the vehicle body area of the current frame according to the width, the height and the coordinates of the license plate area, and storing the vehicle body area into a vehicle body queue;
and the vehicle body area vertical displacement judgment module is used for calculating the displacement of the current vehicle body area and the first vehicle body area in the vertical direction, if the displacement of the vehicle body area in the vertical direction is less than a first threshold value, the vehicle body area is switched to the false license plate processing module, and if not, the vehicle body area enters the next module.
Further, the vehicle detection criterion module comprises: the false license plate processing module is used for adopting vehicle detection equipment to detect vehicles in the scene image of the current frame, if no vehicle region exists, the false license plate processing module is switched to, and if the vehicle region exists, the vehicle region is output and the next module is started.
Further, the vehicle and license plate position relation criterion module comprises: the device comprises a horizontal center line judging module, a vertical distance judging module, a left and right boundary judging module and a width judging module. The horizontal center line judging module, the vertical distance judging module, the left and right boundary judging module and the width judging module can be interchanged in sequence.
Further, the horizontal center line judgment module includes: and the false license plate processing module is used for calculating the horizontal center line of the current frame vehicle area, if the license plate area is positioned above the horizontal center line, the false license plate processing module is switched to, and if not, the next module is switched to.
Further, the vertical distance determination module includes: and the false license plate processing module is used for calculating the distance between the lower boundary of the current frame license plate region and the lower boundary of the vehicle region, if the distance is greater than a third threshold value, the false license plate processing module is switched to, and if not, the false license plate processing module enters the next module.
Further, the left and right boundary determining module includes: the license plate area counting module is used for counting the left boundary and the right boundary of the current frame license plate area and the vehicle area respectively, if the left boundary of the license plate area is positioned on the left side of the left boundary of the vehicle area or the right boundary of the license plate area is positioned on the right side of the right boundary of the vehicle area, the license plate area counting module is switched to the false license plate processing module, and if the left boundary of the license plate area is not positioned on the right side of the right boundary of the vehicle area, the license plate area counting module enters the next module.
Further, the width determination module includes: and the width of the vehicle region of the current frame is counted, if the width of the vehicle region is smaller than a fourth threshold value, the false license plate processing module is switched to, and if not, the next module is switched to.
Compared with the existing license plate recognition technology, the false license plate filtering method and the false license plate filtering device can quickly recognize the license plate and filter the false license plate.
Drawings
Fig. 1 shows a flowchart of a first embodiment of a false license plate filtering method according to the present invention.
Fig. 2 is a flowchart illustrating a false license plate filtering method according to a second embodiment of the present invention.
FIG. 3 shows a block diagram of a first embodiment of a false license plate filter device according to the present invention.
FIG. 4 is a block diagram of a second embodiment of a false license plate filter device according to the present invention.
Detailed Description
To further clarify the structure, characteristics and other objects of the present invention, those skilled in the art will now describe in detail the preferred embodiments of the present invention with reference to the attached drawings, which are provided for the purpose of describing the technical solutions of the present invention only and are not intended to limit the present invention.
The false license plate filtering method comprises the following steps:
a scene image acquisition input step, wherein a video scene image is acquired or input;
a license plate queue acquisition step, namely performing license plate identification on each frame of scene image, and storing a license plate identification result into a license plate queue;
a license plate vertical displacement criterion step, namely calculating the displacement of a license plate area in the vertical direction, and if the displacement is smaller than a first threshold value, switching to a false license plate processing step;
a step of judging the vertical displacement of the vehicle body, namely setting a vehicle body area, calculating the displacement of the vehicle body area in the vertical direction, and turning to a step of processing a false license plate if the displacement is less than a first threshold value;
a vehicle detection criterion step, in which vehicle detection is carried out on the scene image, and if no vehicle region exists, a false license plate processing step is carried out;
judging the position relation between the vehicle and the license plate, and switching to a false license plate processing step according to the position relation between the vehicle region and the license plate region;
a vertical edge detection criterion step, namely calculating a difference image of two adjacent frames of scene images, carrying out edge detection, and if the number of vertical edges is not less than a second threshold value, switching to a false license plate processing step;
a step of processing a real license plate, namely, considering that the current license plate is the real license plate, outputting a license plate identification result and emptying a license plate queue;
and a false license plate processing step, namely considering that the current license plate is a false license plate, and filtering the license plate.
The vehicle detection method comprises a vehicle plate vertical displacement judging step, a vehicle body vertical displacement judging step, a vehicle detection judging step, a vehicle and plate position relation judging step and a vertical edge detection judging step, wherein the sequences of the vehicle plate vertical displacement judging step, the vehicle body vertical displacement judging step, the vehicle detection judging step, the vehicle and plate position relation judging step and the vertical edge detection judging step can be interchanged, and the vehicle detection judging step is performed before the vehicle and plate position relation judging step.
Fig. 1 is a flowchart illustrating a method for filtering false license plates according to a first embodiment of the present invention. As shown in fig. 1, a method for filtering false license plates according to the present invention includes:
a scene image acquisition input step S1 of acquiring or inputting a video scene image;
a license plate queue acquisition step S2, wherein license plate recognition is carried out on each frame of scene image, and a license plate recognition result is stored in a license plate queue;
a license plate vertical displacement criterion step S3, calculating the displacement of the license plate area in the vertical direction, and if the displacement is smaller than a first threshold value, switching to a false license plate processing step S9;
a step S4 of judging the vertical displacement of the vehicle body, which is to set a vehicle body area, calculate the displacement of the vehicle body area in the vertical direction, and if the displacement is smaller than a first threshold value, shift to a step S9 of processing a false license plate;
a vehicle detection criterion step S5, vehicle detection is carried out on the scene image, and if no vehicle region exists, a false license plate processing step S9 is carried out;
a vehicle and license plate position relation criterion step S6, wherein a false license plate processing step S9 is carried out according to the position relation between a vehicle region and a license plate region;
a vertical edge detection criterion step S7, calculating a difference image of two adjacent frames of scene images, carrying out edge detection, and if the number of vertical edges is not less than a second threshold value, switching to a false license plate processing step S9;
a real license plate processing step S8, wherein the current license plate is a real license plate, and a license plate recognition result is output and a license plate queue is emptied;
and a false license plate processing step S9, wherein the current license plate is considered as a false license plate, and the license plate is filtered.
The scene image capturing input step S1 may capture a video scene image through an image capturing device, or may input the video scene image through an input device. In the embodiment, the video scene images of the entrance and the exit of the parking lot are acquired through the camera installed at the entrance and the exit.
Further, the license plate queue obtaining step S2 includes: and (3) performing license plate recognition on each frame of scene image by adopting a license plate recognition method based on a classifier or a neural network, acquiring license plate recognition results, and sequentially storing the license plate recognition results into a license plate queue. The license plate recognition result comprises: license plate region, license plate characters. Further, the license plate recognition result also comprises confidence of license plate character recognition.
Further, the step S3 of determining the vertical displacement of the license plate includes: and calculating the displacement of the current license plate region and the first license plate region in the vertical direction, if the displacement of the license plate region in the vertical direction is smaller than a first threshold value, switching to a false license plate processing step S9, and otherwise, switching to the next step.
Further, the value range of the first threshold is 0.5-2 times of the height of the license plate area of the current frame.
In the embodiment, the license plate vertical displacement criterion step S3 is: calculating the longitudinal coordinate y of the central point of the first license plate area in the license plate queue1Calculating the longitudinal coordinate y of the central point of the current frame license plate area in the license plate queueiIf yi-y1If the | is less than the height of the license plate area of the current frame, the false license plate processing step S9 is carried out, otherwise, the vehicle body vertical displacement criterion step S4 is carried out.
Further, the vehicle body vertical displacement criterion step S4 includes:
a vehicle body area setting step S41, wherein the width and the height of a license plate area in a current frame scene image are calculated, and the vehicle body area of the current frame is set according to the width, the height and the coordinates of the license plate area and stored in a vehicle body queue;
and a step S42 of judging the vertical displacement of the vehicle body area, which is to calculate the displacement of the current vehicle body area and the first vehicle body area in the vertical direction, if the displacement of the vehicle body area in the vertical direction is less than a first threshold value, the step S9 of processing a false license plate is carried out, and if not, the next step is carried out.
Further, the vehicle body region setting step S41 is: counting the upper boundary y of the license plate area in the current frame scene image as y1And the lower boundary y is y2X is left boundary1And the right boundary x ═ x2Calculating the width W of the license plate regionpAnd height HpSetting the upper boundary of the current frame vehicle body area as y ═ y1+λ1×HpThe lower boundary is y ═y1+λ2×HpX is the left boundary1-λ3×WpX is the right boundary2+λ4×WpAnd stored in the vehicle queue. Wherein λ is1The value range of (A) is 3.5-4.5, lambda2The value range of (A) is 1.5-2.5, lambda3The value range of (A) is 0-0.1, lambda4The value range of (A) is 0-0.1.
In an embodiment, the vehicle body region setting step S41 is: counting the upper boundary y of the license plate area in the current frame scene image as y1And the lower boundary y is y2X is left boundary1And the right boundary x ═ x2Calculating the width W of the license plate regionpAnd height HpSetting the upper boundary of the current frame vehicle body area as y ═ y1+4HpThe lower boundary is y ═ y1+2HpX is the left boundary1X is the right boundary2And stored in the vehicle queue.
In an embodiment, the step S42 of determining the vertical displacement of the vehicle body region is: calculating center point ordinate y 'of first vehicle body area in vehicle body queue'1Calculating the center point ordinate y 'of the current frame body area'jIf y'j-y1If the | is less than the height of the license plate area of the current frame, the false license plate processing step S9 is carried out, otherwise, the vehicle detection criterion step S5 is carried out.
Further, the vehicle detection criterion step S5 is to perform vehicle detection on the scene image of the current frame by using a vehicle detection method, and if no vehicle region exists, the false license plate processing step S9 is performed, and if a vehicle region exists, the vehicle region is output and the next step is performed. The vehicle detection method can be realized by the existing image-based vehicle detection method. The embodiment adopts a method in the document of 'model-based vehicle detection and tracking', Hu in, Yang Jingyu, China Picture and graphics, 2008,13(3): 450-.
Further, the step S6 of determining a relationship between the vehicle and the license plate position includes: a horizontal center line judgment step S61, a vertical pitch judgment step S62, a left-right boundary judgment step S63, and a width judgment step S64. The horizontal center line determining step S61, the vertical space determining step S62, the left and right boundary determining step S63, and the width determining step S64 may be interchanged.
Further, the horizontal center line judging step S61 includes: and calculating the horizontal center line of the current frame vehicle area, if the license plate area is positioned above the horizontal center line, switching to a false license plate processing step S9, and otherwise, switching to the next step.
Further, the vertical interval determination step S62 includes: and calculating the distance between the lower boundary of the current frame license plate region and the lower boundary of the vehicle region, if the distance is greater than a third threshold value, switching to a false license plate processing step S9, and otherwise, switching to the next step.
Further, the left-right boundary determination step S63 includes: and respectively counting the left boundary and the right boundary of the current license plate region and the vehicle region, if the left boundary of the license plate region is positioned on the left side of the left boundary of the vehicle region, or the right boundary of the license plate region is positioned on the right side of the right boundary of the vehicle region, switching to a false license plate processing step S9, and otherwise, switching to the next step.
Further, the width determination step S64 includes: and counting the width of the vehicle region of the current frame, if the width of the vehicle region is smaller than a fourth threshold value, switching to a false license plate processing step S9, and otherwise, switching to the next step.
Further, the value range of the third threshold is 0.4-0.6 times of the height of the current frame vehicle area; the value range of the fourth threshold is 1.5-2.5 times of the width of the license plate area of the current frame.
In the first embodiment, the horizontal center line determination step S61, the vertical pitch determination step S62, the left-right boundary determination step S63, and the width determination step S64 are performed in this order. The step S6 of determining the relationship between the vehicle and the license plate position includes: horizontal center line determination step S61: calculating the horizontal center line of the current frame vehicle area, if the license plate area is positioned above the horizontal center line, switching to a false license plate processing step S9, otherwise, switching to a vertical distance judgment step S62; the vertical pitch determination step S62 includes: calculating the distance between the lower boundary of the current frame license plate region and the lower boundary of the vehicle region, if the distance is greater than 0.5 times of the height of the current frame vehicle region, switching to a false license plate processing step S9, otherwise, switching to a left and right boundary judgment step S63; the left-right boundary determination step S63 includes: respectively counting the left boundary and the right boundary of the license plate region and the vehicle region of the current frame, if the left boundary of the license plate region is positioned on the left side of the left boundary of the vehicle region or the right boundary of the license plate region is positioned on the right side of the right boundary of the vehicle region, switching to a false license plate processing step S9, and otherwise, switching to a width judgment step S64; the width determination step S64 includes: otherwise, counting the width of the current frame vehicle region, if the width of the vehicle region is less than 2 times of the width of the current frame vehicle license plate region, switching to a false vehicle license plate processing step S9, and otherwise, switching to a vertical edge detection criterion step S7.
In the second embodiment, the step of determining the position relationship between the vehicle and the license plate may be performed in the order of the left-right boundary determining step S63, the width determining step S64, the vertical distance determining step S62, and the horizontal center line determining step S61.
The vertical edge detection criterion step S7 includes: calculating a difference image of two adjacent frames of scene images, performing edge detection on the difference image by adopting an edge detection method, counting the number of vertical edges, and if the number of the vertical edges is not less than a second threshold value, switching to a false license plate processing step S9, otherwise, switching to the next step.
The edge detection method can be implemented by existing edge detection operators, including but not limited to: sobel edge detection operator, Laplace edge detection operator, Canny edge detection operator, etc.
Further, the value range of the second threshold is 2-4. In the embodiment, a difference image of two adjacent frames of scene images is calculated, a Sobel edge detection operator is adopted to perform vertical edge detection on the difference image, the number of vertical edges is counted, if the number of the vertical edges is not less than 2, a false license plate processing step S9 is carried out, otherwise, a real license plate processing step S8 is carried out.
Further, the real license plate processing step S8 includes: and (4) considering the license plate in the current frame scene image as a real license plate, outputting a license plate recognition result, and emptying a license plate queue.
Further, the false license plate processing step S9 includes: and (4) considering the license plate in the current frame scene image as a false license plate, and filtering the license plate recognition result of the current frame.
Fig. 2 is a flowchart illustrating a second embodiment of a false license plate filtering method according to the present invention. As shown in fig. 2, a method for filtering false license plates according to the present invention includes:
a scene image acquisition input step S1 of acquiring or inputting a video scene image;
a license plate queue acquisition step S2, wherein license plate recognition is carried out on each frame of scene image, and a license plate recognition result is stored in a license plate queue;
a license plate vertical displacement criterion step S3, calculating the displacement of the license plate area in the vertical direction, and if the displacement is smaller than a first threshold value, switching to a false license plate processing step S9;
a vehicle detection criterion step S5, vehicle detection is carried out on the scene image, and if no vehicle region exists, a false license plate processing step S9 is carried out;
a vehicle and license plate position relation criterion step S6, wherein a false license plate processing step S9 is carried out according to the position relation between a vehicle region and a license plate region;
a step S4 of judging the vertical displacement of the vehicle body, which is to set a vehicle body area, calculate the displacement of the vehicle body area in the vertical direction, and if the displacement is smaller than a first threshold value, shift to a step S9 of processing a false license plate;
a vertical edge detection criterion step S7, calculating a difference image of two adjacent frames of scene images, carrying out edge detection, and if the number of vertical edges is not less than a second threshold value, switching to a false license plate processing step S9;
a real license plate processing step S8, wherein the current license plate is a real license plate, and a license plate recognition result is output and a license plate queue is emptied;
and a false license plate processing step S9, wherein the current license plate is considered as a false license plate, and the license plate is filtered.
A false license plate filtering device according to the present invention comprises:
the scene image acquisition input module is used for acquiring or inputting a video scene image;
the license plate queue acquisition module is used for carrying out license plate identification on each frame of scene image and storing a license plate identification result into a license plate queue;
the license plate vertical displacement criterion module is used for calculating the displacement of the license plate area in the vertical direction, and if the displacement is smaller than a first threshold value, the license plate area is switched to the false license plate processing module;
the vehicle body vertical displacement criterion module is used for setting a vehicle body area, calculating the displacement of the vehicle body area in the vertical direction, and switching to the false license plate processing module if the displacement is smaller than a first threshold value;
the vehicle detection criterion module is used for detecting vehicles in the scene image, and if no vehicle region exists, the false license plate processing module is switched to;
the vehicle and license plate position relation criterion module is used for switching to the false license plate processing module according to the position relation between the vehicle region and the license plate region;
the vertical edge detection criterion module is used for calculating a difference image of two adjacent frames of scene images, carrying out edge detection, and switching to the false license plate processing module if the number of vertical edges is not less than a second threshold value;
the real license plate processing module is used for considering that the current license plate is a real license plate, outputting a license plate identification result and emptying a license plate queue;
and the false license plate processing module is used for considering that the current license plate is a false license plate and filtering the license plate.
The vehicle detection criterion module is arranged in front of the vehicle and license plate position relation criterion module.
FIG. 3 is a block diagram of a first embodiment of a false license plate filter arrangement according to the present invention. As shown in fig. 3, a false license plate filtering device according to the present invention includes:
the scene image acquisition input module 1 is used for acquiring or inputting a video scene image;
the license plate queue acquisition module 2 is used for carrying out license plate identification on each frame of scene image and storing a license plate identification result into a license plate queue;
the license plate vertical displacement criterion module 3 is used for calculating the displacement of the license plate area in the vertical direction, and if the displacement is smaller than a first threshold value, the license plate area is switched to a false license plate processing module 9;
the vehicle body vertical displacement criterion module 4 is used for setting a vehicle body area, calculating the displacement of the vehicle body area in the vertical direction, and if the displacement is smaller than a first threshold value, switching to the false license plate processing module 9;
the vehicle detection criterion module 5 is used for detecting vehicles in the scene image, and if no vehicle region exists, the false license plate processing module 9 is switched to;
the vehicle and license plate position relation criterion module 6 is used for switching to a false license plate processing module 9 according to the position relation between a vehicle region and a license plate region;
the vertical edge detection criterion module 7 is used for calculating a difference image of two adjacent frames of scene images, carrying out edge detection, and switching to the false license plate processing module 9 if the number of vertical edges is not less than a second threshold value;
the real license plate processing module 8 is used for considering the current license plate as a real license plate, outputting a license plate identification result and emptying a license plate queue;
and the false license plate processing module 9 is used for considering that the current license plate is a false license plate and filtering the license plate.
The scene image acquisition input module 1 may acquire a video scene image through an image acquisition device, and may also input the video scene image through an input device.
Further, the license plate queue obtaining module 2 includes: the method is used for carrying out license plate recognition on each frame of scene image by adopting a classifier or neural network-based license plate recognition method, obtaining license plate recognition results and sequentially storing the license plate recognition results into a license plate queue. The license plate recognition result comprises: license plate region, license plate characters. Further, the license plate recognition result also comprises confidence of license plate character recognition.
Further, the license plate vertical displacement criterion module 3 includes: the vehicle body vertical displacement judgment module is used for calculating the vertical displacement of the current vehicle license plate region and the first vehicle license plate region, if the vertical displacement of the vehicle license plate region is smaller than a first threshold value, the false vehicle license plate processing module 9 is switched to, and if not, the vehicle body vertical displacement judgment module 4 is switched to.
Further, the value range of the first threshold is half of the height of the license plate area to 2 times of the height of the license plate area.
Further, the body vertical displacement criterion module 4 includes:
the vehicle body area setting module 41 is configured to calculate the width and height of a license plate area in a current frame scene image, set a vehicle body area of the current frame according to the width, height and coordinates of the license plate area, and store the vehicle body area in a vehicle body queue;
and the vehicle body area vertical displacement judgment module 42 is used for calculating the displacement of the current vehicle body area and the first vehicle body area in the vertical direction, if the displacement of the vehicle body area in the vertical direction is smaller than a first threshold value, switching to the false license plate processing module 9, and otherwise, entering the vehicle detection criterion module 5.
Further, the vehicle detection criterion module 5 includes: the false license plate processing module 9 is used for detecting vehicles by adopting vehicle detection equipment to the scene image of the current frame, if no vehicle region exists, the false license plate processing module is switched to, and if the vehicle region exists, the vehicle region is output and enters the vehicle and license plate position relation judging module 6.
Further, the vehicle and license plate position relation criterion module 6 includes: a horizontal center line judging module 61, a vertical distance judging module 62, a left and right boundary judging module 63 and a width judging module 64. The horizontal center line judging module 61, the vertical distance judging module 62, the left and right boundary judging module 63 and the width judging module 64 may be interchanged in sequence.
Further, the horizontal center line determining module 61 includes: and the false license plate processing module 9 is used for calculating the horizontal center line of the current frame vehicle area, if the license plate area is positioned above the horizontal center line, the false license plate area is switched to, and if not, the next module is switched to.
Further, the vertical distance determination module 62 includes: and the false license plate processing module 9 is used for calculating the distance between the lower boundary of the current frame license plate region and the lower boundary of the vehicle region, and if the distance is greater than a third threshold value, the false license plate processing module is switched to, otherwise, the false license plate processing module enters the next module.
Further, the left-right boundary determining module 63 includes: and the false license plate processing module 9 is used for respectively counting the left boundary and the right boundary of the license plate region and the vehicle region of the current frame, and if the left boundary of the license plate region is positioned on the left side of the left boundary of the vehicle region or the right boundary of the license plate region is positioned on the right side of the right boundary of the vehicle region, the false license plate processing module is switched to, or the false license plate processing module enters the next module.
Further, the width determination module 64 includes: and the width of the vehicle region of the current frame is counted, if the width of the vehicle region is smaller than a fourth threshold value, the false license plate processing module 9 is switched to, and if not, the next module is switched to.
FIG. 4 is a block diagram of a second embodiment of a false license plate filter arrangement according to the present invention. As shown in fig. 4, a false license plate filtering device according to the present invention includes:
the scene image acquisition input module 1 is used for acquiring or inputting a video scene image;
the license plate queue acquisition module 2 is used for carrying out license plate identification on each frame of scene image and storing a license plate identification result into a license plate queue;
the license plate vertical displacement criterion module 3 is used for calculating the displacement of the license plate area in the vertical direction, and if the displacement is smaller than a first threshold value, the license plate area is switched to a false license plate processing module 9;
the vehicle detection criterion module 5 is used for detecting vehicles in the scene image, and if no vehicle region exists, the false license plate processing module 9 is switched to;
the vehicle and license plate position relation criterion module 6 is used for switching to a false license plate processing module 9 according to the position relation between a vehicle region and a license plate region;
the vehicle body vertical displacement criterion module 4 is used for setting a vehicle body area, calculating the displacement of the vehicle body area in the vertical direction, and if the displacement is smaller than a first threshold value, switching to the false license plate processing module 9;
the vertical edge detection criterion module 7 is used for calculating a difference image of two adjacent frames of scene images, carrying out edge detection, and switching to the false license plate processing module 9 if the number of vertical edges is not less than a second threshold value;
the real license plate processing module 8 is used for considering the current license plate as a real license plate, outputting a license plate identification result and emptying a license plate queue;
and the false license plate processing module 9 is used for considering that the current license plate is a false license plate and filtering the license plate.
Compared with the existing license plate recognition technology, the false license plate filtering method and the false license plate filtering device can quickly recognize the license plate and filter the false license plate.
While the foregoing is directed to the preferred embodiment of the present invention, and is not intended to limit the scope of the invention, it will be understood that the invention is not limited to the embodiments described herein, which are described to assist those skilled in the art in practicing the invention. Further modifications and improvements may readily occur to those skilled in the art without departing from the spirit and scope of the invention, and it is intended that the invention be limited only by the terms and scope of the appended claims, as including all alternatives and equivalents which may be included within the spirit and scope of the invention as defined by the appended claims.
Claims (14)
1. A false license plate filtering method is characterized by comprising the following steps:
a scene image acquisition input step, wherein a video scene image is acquired or input;
a license plate queue acquisition step, namely performing license plate identification on each frame of scene image, and storing a license plate identification result into a license plate queue;
a license plate vertical displacement criterion step, namely calculating the displacement of a license plate area in the vertical direction, and if the displacement is smaller than a first threshold value, switching to a false license plate processing step;
a step of judging the vertical displacement of the vehicle body, namely setting a vehicle body area, calculating the displacement of the vehicle body area in the vertical direction, and turning to a step of processing a false license plate if the displacement is less than a first threshold value;
a vehicle detection criterion step, in which vehicle detection is carried out on the scene image, and if no vehicle region exists, a false license plate processing step is carried out;
judging the position relation between the vehicle and the license plate, and switching to a false license plate processing step according to the position relation between the vehicle region and the license plate region;
a vertical edge detection criterion step, namely calculating a difference image of two adjacent frames of scene images, carrying out edge detection, and if the number of vertical edges is not less than a second threshold value, switching to a false license plate processing step;
a step of processing a real license plate, namely, considering that the current license plate is the real license plate, outputting a license plate identification result and emptying a license plate queue; and
a false license plate processing step, namely considering that the current license plate is a false license plate, and filtering the license plate;
the vehicle detection criterion step is carried out before the vehicle and license plate position relation criterion step;
further, the value range of the first threshold is 0.5-2 times of the height of the license plate area of the current frame; the value range of the second threshold is 2-4.
2. The method of claim 1, wherein the license plate recognition result in the license plate queue obtaining step comprises: license plate region, license plate characters.
3. The method of claim 2, further comprising the license plate recognition result further comprising a confidence level of the license plate character recognition.
4. The method of claim 1, wherein the license plate vertical displacement criterion step comprises: and calculating the displacement of the current license plate region and the first license plate region in the vertical direction, if the displacement of the license plate region in the vertical direction is smaller than a first threshold value, switching to a false license plate processing step, and otherwise, switching to the next step.
5. The method of claim 1, wherein the body vertical displacement criterion step comprises: the method comprises the following steps of setting a vehicle body area, namely calculating the width and the height of a license plate area in a current frame scene image, setting the vehicle body area of the current frame according to the width, the height and coordinates of the license plate area, and storing the vehicle body area into a vehicle body queue;
and a step of judging the vertical displacement of the vehicle body area, namely calculating the displacement of the current vehicle body area and the first vehicle body area in the vertical direction, if the displacement of the vehicle body area in the vertical direction is less than a first threshold value, switching to the step of processing the false license plate, and otherwise, switching to the next step.
6. The method of claim 5, further the body region setting step is: counting the upper boundary y of the license plate area in the current frame scene image as y1And the lower boundary y is y2X is left boundary1And the right boundary x ═ x2Calculating the width W of the license plate regionpAnd height HpSetting the upper boundary of the current frame vehicle body area as y ═ y1+λ1×HpThe lower boundary is y ═ y1+λ2×HpX is the left boundary1-λ3×WpX is the right boundary2+λ4×WpAnd storing the data in a vehicle body queue;
wherein λ is1The value range of (A) is 3.5-4.5, lambda2The value range of (A) is 1.5-2.5, lambda3The value range of (A) is 0-0.1, lambda4The value range of (A) is 0-0.1.
7. The method of claim 1, wherein the vehicle to license plate location relationship criterion step comprises: a horizontal center line judgment step, a vertical distance judgment step, a left and right boundary judgment step and a width judgment step; the sequence of the horizontal center line judging step, the vertical distance judging step, the left and right boundary judging step and the width judging step can be interchanged;
the horizontal center line judging step includes: calculating the horizontal center line of the current frame vehicle area, if the license plate area is positioned above the horizontal center line, switching to a false license plate processing step, and otherwise, switching to the next step;
the vertical distance judging step includes: calculating the distance between the lower boundary of the current frame license plate region and the lower boundary of the vehicle region, if the distance is greater than a third threshold value, switching to a false license plate processing step, otherwise, switching to the next step;
the left and right boundary determining step includes: respectively counting the left boundary and the right boundary of the license plate region and the vehicle region of the current frame, if the left boundary of the license plate region is positioned on the left side of the left boundary of the vehicle region or the right boundary of the license plate region is positioned on the right side of the right boundary of the vehicle region, switching to a false license plate processing step, and otherwise, switching to the next step;
the width judging step includes: and counting the width of the vehicle region of the current frame, if the width of the vehicle region is smaller than a fourth threshold value, switching to a false license plate processing step, and otherwise, switching to the next step.
8. The method of claim 7, further comprising the step of setting the third threshold value to be 0.4-0.6 times of the height of the vehicle region of the current frame; the value range of the fourth threshold is 1.5-2.5 times of the width of the license plate area of the current frame.
9. A false license plate filtering device, the device comprising:
the scene image acquisition input module is used for acquiring or inputting a video scene image;
the license plate queue acquisition module is used for carrying out license plate identification on each frame of scene image and storing a license plate identification result into a license plate queue;
the license plate vertical displacement criterion module is used for calculating the displacement of the license plate area in the vertical direction, and if the displacement is smaller than a first threshold value, the license plate area is switched to the false license plate processing module;
the vehicle body vertical displacement criterion module is used for setting a vehicle body area, calculating the displacement of the vehicle body area in the vertical direction, and switching to the false license plate processing module if the displacement is smaller than a first threshold value;
the vehicle detection criterion module is used for detecting vehicles in the scene image, and if no vehicle region exists, the false license plate processing module is switched to;
the vehicle and license plate position relation criterion module is used for switching to the false license plate processing module according to the position relation between the vehicle region and the license plate region;
the vertical edge detection criterion module is used for calculating a difference image of two adjacent frames of scene images, carrying out edge detection, and switching to the false license plate processing module if the number of vertical edges is not less than a second threshold value;
the real license plate processing module is used for considering that the current license plate is a real license plate, outputting a license plate identification result and emptying a license plate queue;
the false license plate processing module is used for considering that the current license plate is a false license plate and filtering the license plate;
the vehicle detection criterion module is arranged in front of the vehicle and license plate position relation criterion module.
10. The apparatus of claim 9, wherein the license plate recognition result in the license plate queue obtaining module comprises: license plate region, license plate characters.
11. The apparatus of claim 10, further wherein the license plate recognition result further comprises a confidence level of the license plate character recognition.
12. The apparatus of claim 9, wherein the license plate vertical displacement criteria module comprises: the license plate processing module is used for calculating the displacement of the current license plate region and the first license plate region in the vertical direction, if the displacement of the license plate region in the vertical direction is smaller than a first threshold value, the license plate processing module is switched to the false license plate processing module, and if not, the license plate processing module enters the next module.
13. The apparatus of claim 9, wherein the body vertical displacement criteria module comprises: the vehicle body area setting module is used for calculating the width and the height of a license plate area in a current frame scene image, setting the vehicle body area of the current frame according to the width, the height and the coordinates of the license plate area, and storing the vehicle body area into a vehicle body queue; and the vehicle body area vertical displacement judgment module is used for calculating the displacement of the current vehicle body area and the first vehicle body area in the vertical direction, if the displacement of the vehicle body area in the vertical direction is less than a first threshold value, the vehicle body area is switched to the false license plate processing module, and if not, the vehicle body area enters the next module.
14. The apparatus of claim 9, wherein the vehicle to license plate positional relationship criterion module comprises: the device comprises a horizontal central line judging module, a vertical distance judging module, a left and right boundary judging module and a width judging module; the horizontal center line judging module, the vertical distance judging module, the left and right boundary judging module and the width judging module can be interchanged in sequence;
the horizontal center line judging module comprises: the license plate processing module is used for calculating the horizontal center line of the current frame vehicle area, if the license plate area is positioned above the horizontal center line, the false license plate processing module is switched to, and if not, the next module is switched to; the vertical distance judgment module comprises: the false license plate processing module is used for calculating the distance between the lower boundary of the current frame license plate region and the lower boundary of the vehicle region, if the distance is greater than a third threshold value, the false license plate processing module is switched to, and if not, the next module is switched to;
the left and right boundary judgment module includes: the system comprises a false license plate processing module, a current frame license plate area, a vehicle area and a vehicle area, wherein the false license plate processing module is used for counting the left boundary and the right boundary of the current frame license plate area and the vehicle area respectively, and if the left boundary of the license plate area is positioned on the left side of the left boundary of the vehicle area or the right boundary of the license plate area is positioned on the right side of the right boundary of the vehicle area, the false license plate processing module is switched to, or the false license plate processing module enters the next module;
the width judging module includes: and the width of the vehicle region of the current frame is counted, if the width of the vehicle region is smaller than a fourth threshold value, the false license plate processing module is switched to, and if not, the next module is switched to.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810459996.0A CN108694387B (en) | 2018-05-15 | 2018-05-15 | False license plate filtering method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810459996.0A CN108694387B (en) | 2018-05-15 | 2018-05-15 | False license plate filtering method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108694387A CN108694387A (en) | 2018-10-23 |
| CN108694387B true CN108694387B (en) | 2021-10-26 |
Family
ID=63846454
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810459996.0A Active CN108694387B (en) | 2018-05-15 | 2018-05-15 | False license plate filtering method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108694387B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109583442B (en) * | 2018-11-07 | 2020-12-11 | 北京智芯原动科技有限公司 | False license plate detection method and device based on line segment detection |
| CN110533039B (en) * | 2019-09-04 | 2022-05-24 | 深圳市捷顺科技实业股份有限公司 | Method, device and equipment for detecting authenticity of license plate |
| CN113408364B (en) * | 2021-05-26 | 2022-11-11 | 深圳市捷顺科技实业股份有限公司 | Temporary license plate recognition method, system, device and storage medium |
| CN114842462B (en) * | 2022-04-07 | 2024-04-09 | 深圳市捷顺科技实业股份有限公司 | License plate anti-fake method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101021973A (en) * | 2006-08-04 | 2007-08-22 | 艾伯资讯（深圳）有限公司 | Method for fraudelent electronic vehicle number plate preventing identifying system |
| CN101931792A (en) * | 2010-08-04 | 2010-12-29 | 丁天 | Multi-plate number locating method in high definition video steaming |
| CN102054174A (en) * | 2009-10-30 | 2011-05-11 | 佳能株式会社 | Method and system for positioning licence plate |
| CN103235938A (en) * | 2013-05-03 | 2013-08-07 | 北京国铁华晨通信信息技术有限公司 | Method and system for detecting and identifying license plate |
| CN107103314A (en) * | 2017-06-15 | 2017-08-29 | 浙江南自智能科技股份有限公司 | A kind of fake license plate vehicle retrieval system based on machine vision |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2006020337A2 (en) * | 2004-07-20 | 2006-02-23 | E-Views Safety Systems, Inc. | Distributed, roadside-based real-time id recognition system and method |
-
2018
- 2018-05-15 CN CN201810459996.0A patent/CN108694387B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101021973A (en) * | 2006-08-04 | 2007-08-22 | 艾伯资讯（深圳）有限公司 | Method for fraudelent electronic vehicle number plate preventing identifying system |
| CN102054174A (en) * | 2009-10-30 | 2011-05-11 | 佳能株式会社 | Method and system for positioning licence plate |
| CN101931792A (en) * | 2010-08-04 | 2010-12-29 | 丁天 | Multi-plate number locating method in high definition video steaming |
| CN103235938A (en) * | 2013-05-03 | 2013-08-07 | 北京国铁华晨通信信息技术有限公司 | Method and system for detecting and identifying license plate |
| CN107103314A (en) * | 2017-06-15 | 2017-08-29 | 浙江南自智能科技股份有限公司 | A kind of fake license plate vehicle retrieval system based on machine vision |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108694387A (en) | 2018-10-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108694387B (en) | False license plate filtering method and device | |
| CN106600977B (en) | Multi-feature recognition-based illegal parking detection method and system | |
| EP1796043B1 (en) | Object detection | |
| CN103824066B (en) | A kind of licence plate recognition method based on video flowing | |
| US7218207B2 (en) | Method for detecting position of lane marker, apparatus for detection position of lane marker and alarm apparatus for lane deviation | |
| CN105741559B (en) | A kind of illegal occupancy Emergency Vehicle Lane detection method based on track line model | |
| CN111382704B (en) | Vehicle line pressing violation judging method and device based on deep learning and storage medium | |
| CN105740836B (en) | A kind of illegal detection method for occupying Emergency Vehicle Lane | |
| CN101369312B (en) | Method and equipment for detecting intersection in image | |
| CN112861797A (en) | Method and device for identifying authenticity of license plate and related equipment | |
| US20120189161A1 (en) | Visual attention apparatus and control method based on mind awareness and display apparatus using the visual attention apparatus | |
| JP5306124B2 (en) | Number plate reading apparatus and number plate reading method | |
| CN105206060B (en) | A kind of vehicle type recognition device and its method based on SIFT feature | |
| JP4739699B2 (en) | License plate recognition apparatus and method | |
| JP3364771B2 (en) | Moving object detection device | |
| US11157755B2 (en) | Image processing apparatus | |
| CN109886120B (en) | Zebra crossing detection method and system | |
| JPH1031797A (en) | Method for specifying vehicle in parking lot, system therefor, parking charge calculation method and system therefor | |
| EP3329419A1 (en) | Method for capturing an object on a road in the environment of a motor vehicle, camera system and motor vehicle using the same | |
| DE102020213799A1 (en) | Obstacle detection device and obstacle detection method | |
| KR20080049472A (en) | Information detecting system using photographing apparatus load in vehicle and artificial neural network | |
| CN112329724A (en) | Real-time detection and snapshot method for lane change of motor vehicle | |
| Kuo et al. | An autonomous license plate detection method | |
| CN114842462B (en) | License plate anti-fake method | |
| CN108470137B (en) | Video-based automobile sun visor detection method |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |