CN117218681A - Height estimation method of monocular lens, child passing gate device and judging method - Google Patents
Height estimation method of monocular lens, child passing gate device and judging method Download PDFInfo
- Publication number
- CN117218681A CN117218681A CN202311486914.9A CN202311486914A CN117218681A CN 117218681 A CN117218681 A CN 117218681A CN 202311486914 A CN202311486914 A CN 202311486914A CN 117218681 A CN117218681 A CN 117218681A
- Authority
- CN
- China
- Prior art keywords
- lens
- target pedestrian
- target
- gate
- height
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000001514 detection method Methods 0.000 claims abstract description 24
- 230000003287 optical effect Effects 0.000 claims description 32
- 238000003384 imaging method Methods 0.000 claims description 9
- 238000005096 rolling process Methods 0.000 claims description 9
- 230000001960 triggered effect Effects 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 3
- 238000009434 installation Methods 0.000 abstract description 5
- 230000000694 effects Effects 0.000 abstract description 4
- 238000004364 calculation method Methods 0.000 abstract description 3
- 238000013461 design Methods 0.000 description 7
- 230000007547 defect Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 239000011521 glass Substances 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 230000001680 brushing effect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000010408 sweeping Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Traffic Control Systems (AREA)
Abstract
The invention discloses a height estimation method of a monocular lens, a child passing device and a judging method, wherein the height of a target pedestrian can be directly calculated by utilizing a geometric relationship through carrying out certain specific limitations and requirements on the installation of the monocular RGB lens, and the precision is very high and is no worse than that of a module with a depth lens. According to the invention, the monocular RGB lens is designed at the entrance and exit of the gate, and the monocular RGB lens is utilized to operate the target detection and multi-target tracking algorithm, so that the required calculation resources are fewer compared with that of binocular lenses, the requirement on hardware is lower, on the basis of the original gate, the intelligent recognition of the gate passing child is realized with a lower cost, a simpler scheme and more accurate and practical effect, the safe and convenient passing of the child is ensured, and the passing efficiency and user experience of the whole gate are improved.
Description
Technical Field
The invention relates to the technical field of self-service passing gate machines, in particular to a height estimation method of a monocular lens, a child passing gate device based on the monocular lens and a judging method.
Background
With the rapid development of technology, self-service traffic is becoming more and more common in various public transportation places such as subway stations, high-speed rail stations, airports, ferries, customs and the like, and even the entrance gates of places such as communities, factories, construction sites, tourist attractions and the like are covered with various self-service traffic gates. People can realize self-help traffic through various methods such as ticket brushing, face brushing, code sweeping and the like, so that the daily life of people is greatly facilitated, and traffic efficiency is improved. However, many public places have special consideration or limitation on children (such as height lower than 1 meter, 1.2 meter, 1.4 meter, etc.), such as high-speed rail, subway, airport, tourist attractions, etc., and special treatment is needed on the aspects of charging, safety, etc. for the children, only people meeting the specific height can pass through, or special channels are used for passing through. Therefore, how to actively and intelligently identify whether the passing person is a child or not in the gate passing or at the end of the passing becomes significant. If the result of the active identification can be further reminiscent to the service personnel on site, better passing experience can be generated, and better passing safety can be realized.
If only the passing of the gate is realized, if the gate is used for passing by children, various methods can be adopted, such as: the scheme based on RGB-D module, such as the module with structured light and TOF can output depth information, can directly output depth information when outputting RGB image, combine target detection algorithm and depth information, can directly give whether there is a person's gate and how many person's gates of height, but the cost of this kind of module is extremely expensive, generally needs more than hundreds or even thousand yuan, and this kind of module can't embed into current floodgate machine perfectly, accomplish hidden design moreover, because the glass board outside the floodgate machine has very strong shielding and reflection effect to the signal that the module launched, and then seriously influences the precision of depth information. If better depth information is to be obtained, the module needs to be fully exposed, which is an unacceptable design for the gate.
The scheme of binocular (double RGB) can achieve hidden design, but the requirement is higher when the scheme is used, two lenses are required to be calibrated, good frame synchronization is required, relatively accurate depth information can be calculated, the height of a passman is further obtained, and in addition, the requirement for hardware is also very high due to the fact that the depth is calculated through the binocular, the consumption of resources is very high;
the monocular RGB scheme has the advantages of lower cost, hidden design and low requirement on computing resources, but has the defect that a depth map cannot be output, so that the height of a passer-by cannot be calculated through the depth map, and further whether a passer-by is a child cannot be judged.
In view of the above, the present inventors have intensively conceived against the drawbacks and inconveniences of the conventional monocular RGB scheme, and have actively studied and tried improvement efforts to develop the present invention.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide the height estimation method of the monocular lens, which has low cost and can accurately estimate the height of a target pedestrian.
In order to achieve the above object, the solution of the present invention is:
a height estimation method of a monocular lens comprises the following steps:
the mounting height of the monocular RGB lens is obtained and is recorded asThe lens has no pitch angle and no rolling angle;
the monocular RGB lens shoots a target pedestrian, and an image plane is perpendicular to the ground and parallel to the plane where the target pedestrian is located;
the optical center of the monocular RGB lens is set asIts coordinates in the image plane are +.>The vertex of the target pedestrian is +.>Which is +.>The coordinates are +.>The method comprises the steps of carrying out a first treatment on the surface of the The bottom vertex of the target pedestrian is +.>Which is +.>The coordinates are +.>;
Monocular RGB lens at heightThe light beams horizontally injected form a plane parallel to the ground, the optical axis is also arranged on the plane, and the intersection point of the plane and the target pedestrian is +.>Obviously->The point on the image plane is marked +.>The ordinate is the same as the optical axis, is +.>The height of the target pedestrian is +.>;
The relationship according to the principle of pinhole imaging and similar triangles is:
the method comprises the following steps:
finally, it can be calculated:
。
another object of the present invention is to overcome the drawbacks of the prior art and to provide a child passing device based on a monocular lens, which is easy to install, low in cost and accurate in height estimation.
In order to achieve the above object, the solution of the present invention is:
the target pedestrian passing gate device based on the monocular lens comprises a monocular RGB lens, a target detection module, a multi-target tracking module and a passing gate judging module;
the monocular RGB lens is horizontally arranged at the gate inlet and/or the gate outlet, the lens faces the gate channel to shoot, and the optical axis direction of the monocular RGB lens is parallel to the ground;
the target detection module is connected with the monocular RGB lens, and detects a target pedestrian in the monocular RGB lens in real time or in each frame to give a coordinate position of the target pedestrian in the image;
the multi-target tracking module is connected with the monocular RGB lens, and in the whole process of crossing the gate of the target pedestrian, the multi-target tracking module distributes a unique track ID and represents the unique track of the target pedestrian, so that the tracks of crossing the gate of different target pedestrians are counted separately, and a single frame is used as a result;
the gate crossing judging module is connected with the monocular RGB lens, when the track of the target pedestrian leaves the picture, the gate crossing judging module is triggered, the gate crossing judging module can automatically call the height estimating module of the monocular RGB lens, the height condition of the gate crossing target pedestrian is counted and automatically learned, if the height of the target pedestrian crossing the gate is found to be lower than the set threshold value, the current target pedestrian is judged to be a child, and an alarm is actively sent to the service desk and a scene picture and/or video is sent;
the height estimation method of the monocular RGB lens comprises the following steps:
the mounting height of the monocular RGB lens is obtained and is recorded asThe lens has no pitch angle and no rolling angle;
the monocular RGB lens shoots a target pedestrian, and an image plane is perpendicular to the ground and parallel to the plane where the target pedestrian is located;
the optical center of the monocular RGB lens is set asIts coordinates in the image plane are +.>The vertex of the target pedestrian is +.>Which is +.>The coordinates are +.>The method comprises the steps of carrying out a first treatment on the surface of the The bottom vertex of the target pedestrian is +.>Which is +.>The coordinates are +.>;
Monocular RGB lens at heightThe light beams horizontally incident form a plane parallel to the ground, the optical axis is also on the plane, and the plane is parallel to the targetThe intersection of pedestrians is->Obviously->The point on the image plane is marked +.>The ordinate is the same as the optical axis, is +.>The height of the target pedestrian is +.>;
The relationship according to the principle of pinhole imaging and similar triangles is:
the method comprises the following steps:
finally, it can be calculated:
。
furthermore, the crossing gate judging module caches all heights of the track ID, and adopts a certain learning strategy to estimate an average height.
The invention further aims to overcome the defects of the prior art and provide a child passing gate judging method based on a monocular lens, which can accurately estimate the height of a child, realize the safety of gate passing and improve user experience.
In order to achieve the above object, the solution of the present invention is:
a child passing gate judging method comprises the following steps:
step S1: reading monocular RGB lens data;
step S2: the target detection module acquires one frame of image data of a monocular RGB lens, a coordinate position of a target pedestrian in an image is acquired through a target detection algorithm in the target detection module, and height estimation is carried out on the target pedestrian;
the height estimation method of the monocular RGB lens comprises the following steps:
the mounting height of the monocular RGB lens is obtained and is recorded asThe lens has no pitch angle and no rolling angle;
the monocular RGB lens shoots a target pedestrian, and an image plane is perpendicular to the ground and parallel to the plane where the target pedestrian is located;
the optical center of the monocular RGB lens is set asIts coordinates in the image plane are +.>The vertex of the target pedestrian is +.>Which is +.>The coordinates are +.>The method comprises the steps of carrying out a first treatment on the surface of the The bottom vertex of the target pedestrian is +.>Which is +.>The coordinates are +.>;
Monocular RGB lens at heightThe light beams horizontally injected form a plane parallel to the ground, the optical axis is also arranged on the plane, and the intersection point of the plane and the target pedestrian is +.>Obviously->The point on the image plane is marked +.>The ordinate is the same as the optical axis, is +.>The height of the target pedestrian is +.>;
The relationship according to the principle of pinhole imaging and similar triangles is:
the method comprises the following steps:
finally, it can be calculated:
;
step S3: the multi-target tracking module distributes unique track IDs for each target pedestrian by utilizing a multi-target tracking algorithm;
step S4: judging whether the track of the target pedestrian disappears, if yes, entering a step S5, otherwise, entering a step S1;
step S5: the passing gate judging module counts and automatically learns the height condition of a passing gate target pedestrian;
step S6: judging whether the height of the target pedestrian is smaller than a threshold value, if so, entering a step S7, otherwise, ending;
step S7: confirming a target behavior child, actively sending an alarm to a service desk and sending a scene picture and/or video;
and (5) ending.
Further, in step S5, the passing judgment module caches all heights of the track ID, and adopts a certain learning strategy to estimate an average height.
After the scheme is adopted, the child height estimation and passing judgment method and device based on the monocular lens can be used for actively and intelligently identifying whether the target pedestrian passing through the self-service gate is a child or not, can realize two basic tasks of target detection and child identification, is easy to conceal and design in the gate, and is not easy to be interfered by glass shielding and the like. Compared with the existing RGB-D, binocular and other module schemes, the monocular RGB lens is easier to install, has lower cost and needs less operation resources, and more importantly, after meeting the installation position limit and requirement of the gate, the height of a target can be accurately estimated through a geometric relationship, so that the height estimation effect comparable with that of the binocular, RGBD and other expensive modules is achieved.
The invention carries out improved design on the basis of a monocular RGB scheme in the background technology, and can directly calculate the height of a target pedestrian by utilizing a geometric relationship by carrying out certain specific limitations and requirements on the installation of the monocular lens, and the invention has very high precision and is not inferior to a module with a depth lens. According to the invention, through designing and improving the functions and characteristics of the passing gate channel, a monocular RGB lens device is designed at the entrance and exit of the gate, based on the monocular RGB lens device, a pedestrian detection and height estimation method is provided, on the basis of the original gate, intelligent recognition of passing gate children is realized with a lower cost, a simpler scheme, a more accurate and more practical effect, safety and convenience of passing gate children are ensured, and the efficiency and user experience of the whole gate passing system are improved.
Compared with the prior art, the child height estimation and gate passing judgment method and device based on the monocular lens have the following advantages:
the invention adopts the single-purpose RGB lens, has simple and convenient installation, is not easy to be shielded by glass, is convenient for hidden design and has low cost;
compared with binocular, the method does not need to perform stereo matching, parallax calculation and the like, and has fewer required calculation resources and lower requirements on hardware;
according to the invention, the height of the target can be rapidly and accurately estimated through the similarity relation of the solid geometry, so that the universality is strong;
according to the invention, the multi-target tracking module and the gate-crossing alarming strategy module are additionally arranged, and the accuracy rate of identifying children and the final use experience of the gate can be improved through comprehensive analysis of the whole track.
Drawings
FIG. 1 is a schematic diagram of the invention for estimating height based on monocular RGB shots.
Fig. 2 is a diagram of an actual image taken by the monocular RGB lens of the present invention.
FIG. 3 is a schematic diagram of the overall functional framework of the present invention.
FIG. 4 is a flow chart of the child passing gate judgment in the present invention.
Detailed Description
In order to further explain the technical scheme of the invention, the invention is explained in detail by specific examples.
As shown in fig. 1 and 2, the invention discloses a height estimation method of a monocular lens, which comprises the following steps:
the mounting height of the monocular RGB lens (the mounting height of the monocular RGB lens is known or can be obtained by measurement) is obtained and is recorded asThe monocular RGB lens is horizontally placed, the optical axis direction of the monocular RGB lens is parallel to the ground, no pitch angle and no rolling angle are formed, an included angle in the horizontal direction can be formed between the monocular RGB lens and the side face of the gate, the adjusted included angle ensures that the whole channel can be completely shot, an actual shot picture is shown in fig. 2, a target pedestrian passes the gate from the front, and from the perspective of the monocular RGB lens,i.e. to the inside of the picture, a schematic perspective view is shown in fig. 1;
the monocular RGB lens horizontally shoots a target pedestrian, and the image plane is perpendicular to the ground and parallel to the plane where the target pedestrian is located;
the optical center of the monocular RGB lens is set asIts coordinates in the image plane are +.>The vertex of the target pedestrian is +.>Which is +.>The coordinates are +.>The method comprises the steps of carrying out a first treatment on the surface of the The bottom vertex of the target pedestrian is +.>Which is +.>The coordinates are +.>;
Monocular RGB lens at heightThe light beams are horizontally injected into a plurality of light beams to form a plane parallel to the ground, and because the lens has no pitch angle, the optical axis is also arranged on the plane, and the intersection point of the plane and the target pedestrian is +.>Obviously->The point on the image plane is marked +.>The ordinate is the same as the optical axis, is +.>The height of the target pedestrian is +.>;
The relationship according to the principle of pinhole imaging and similar triangles is:
the method comprises the following steps:
finally, it can be calculated:
。
as shown in fig. 3 and fig. 4, the general idea of the present invention is to install a single-purpose RGB lens on the basis of the existing gate, where the single-purpose RGB lens may be connected to a central control unit of the gate or a single hardware module, and when the gate is opened to perform work, the data of the single-purpose RGB lens may be continuously read, and a target detection algorithm is invoked to obtain the coordinate position of the target pedestrian in the image in real time, where the detection algorithm is not limited, and the detectors of one stage and two stages may be, for example, a conventional Yolo series algorithm, as long as the coordinate position can be obtained. The detection frame information of the target pedestrians can be used for estimating the heights, meanwhile, the detection frame information can be used as input of multi-target tracking, the multi-target tracking algorithm is not limited in the invention, classic algorithms such as a common sort and a deep sort can be used, after the multi-target tracking algorithm is adopted, each target can be distributed with a unique ID to represent the unique track of the target pedestrians, after the track ID disappears, a passing judgment module can be triggered, otherwise, the tracking is continued until the target pedestrians disappear. After the target pedestrians disappear, the passing judgment module works, although the target of each frame of detection result can estimate the height, in consideration of the fluctuation of a single frame of result, in order to make the result more accurate, the invention adopts all heights of the cache track, adopts a certain learning strategy to estimate an average height, one of the feasible learning strategies is to directly calculate the average value, or remove the maximum and minimum average value, or learn a distribution, and the center point of the distribution is the required method, and the invention is not limited, and if the height is found to be lower than a certain threshold (can be adjusted according to the actual application scene, such as 1.2 meters, and the like), the conclusion of the passing of the child is automatically generated, and the result, the video picture on site, and the like are saved and sent to ground service personnel for on-site confirmation guidance so as to finally realize better passing service and higher passing security guarantee.
Specifically, as shown in fig. 3, the invention also discloses a child passing gate device based on a monocular lens, which comprises a monocular RGB lens, a target detection module, a multi-target tracking module and a passing gate judging module;
the monocular RGB lens is horizontally arranged at the gate inlet and/or the gate outlet, the lens faces the gate channel to shoot, and the optical axis direction of the monocular RGB lens is parallel to the ground; in order to realize accurate estimation of the height of a target pedestrian, the invention has a special requirement for the installation of a monocular RGB lens: the lens needs to be ensured to be horizontally placed, the shooting direction is parallel to the ground (namely, the optical axis direction of the lens is parallel to the ground), the pitch angle and the roll angle cannot be provided, and the yaw angle can be provided in the horizontal direction. The monocular RGB lens can be connected with a central control machine of a gate or a single hardware module.
The target detection module is connected with the monocular RGB lens, and detects a target pedestrian in the monocular RGB lens in real time or in each frame to give a coordinate position of the target pedestrian in the image;
the multi-target tracking module is connected with the monocular RGB lens, and in the whole process of crossing the gate of the target pedestrian, the multi-target tracking module distributes a unique track ID and represents the unique track of the target pedestrian, so that the tracks of crossing the gate of different target pedestrians are counted separately, and a single frame is used as a result;
the gate crossing judging module is connected with the monocular RGB lens, when the track of the target pedestrian leaves the picture, the gate crossing judging module is triggered, the gate crossing judging module can automatically call the height estimating module of the monocular RGB lens, the height condition of the gate crossing target pedestrian is counted and automatically learned, if the height of the target pedestrian crossing the gate is found to be lower than the set threshold value, the current target pedestrian is judged to be a child, and an alarm is actively sent to the service desk and a scene picture and/or video is sent;
the height estimation method of the monocular RGB lens comprises the following steps:
the mounting height of the monocular RGB lens is obtained and is recorded asThe lens has no pitch angle and no rolling angle;
the monocular RGB lens shoots a target pedestrian, and an image plane is perpendicular to the ground and parallel to the plane where the target pedestrian is located;
the optical center of the monocular RGB lens is set asIts coordinates in the image plane are +.>The vertex of the target pedestrian is +.>Which is +.>The coordinates are +.>The method comprises the steps of carrying out a first treatment on the surface of the The bottom vertex of the target pedestrian is +.>Which is +.>The coordinates are +.>;
Monocular RGB lens at heightA plurality of light rays are horizontally injected to form a plane parallel to the ground, the optical axis is also arranged on the plane, and the intersection point of the plane and the target pedestrian is +.>Obviously->The point on the image plane is marked +.>The ordinate is the same as the optical axis, is +.>The height of the target pedestrian is +.>;
The relationship according to the principle of pinhole imaging and similar triangles is:
the method comprises the following steps:
finally, it can be calculated:
。
furthermore, the multi-target tracking module caches all heights of the track IDs, and adopts a certain learning strategy to estimate an average height.
The invention also discloses a child passing gate judging method based on the monocular lens, which comprises the following steps:
step S1: reading monocular RGB lens data;
step S2: the target detection module acquires one frame of image data of a monocular RGB lens, a coordinate position of a target pedestrian in an image is acquired through a target detection algorithm in the target detection module, and height estimation is carried out on the target pedestrian;
the height estimation method of the monocular RGB lens comprises the following steps:
the mounting height of the monocular RGB lens is obtained and is recorded asThe lens has no pitch angle and no rolling angle;
the monocular RGB lens shoots a target pedestrian, and an image plane is perpendicular to the ground and parallel to the plane where the target pedestrian is located;
the optical center of the monocular RGB lens is set asIts coordinates in the image plane are +.>The vertex of the target pedestrian is +.>Which is +.>The coordinates are +.>The method comprises the steps of carrying out a first treatment on the surface of the The bottom vertex of the target pedestrian is +.>Which is +.>The coordinates are +.>;
Monocular RGB lens at heightA plurality of light rays are horizontally injected to form a plane parallel to the ground, the optical axis is also arranged on the plane, and the intersection point of the plane and the target pedestrian is +.>Obviously->The point on the image plane is marked +.>The ordinate is the same as the optical axis, is +.>The height of the target pedestrian is +.>;
The relationship according to the principle of pinhole imaging and similar triangles is:
the method comprises the following steps:
finally, it can be calculated:
;
step S3: the multi-target tracking module distributes unique track IDs for each target pedestrian by utilizing a multi-target tracking algorithm;
step S4: judging whether the track of the target pedestrian disappears, if yes, entering a step S5, otherwise, entering a step S1;
step S5: the passing gate judging module counts and automatically learns the height condition of a passing gate target pedestrian;
step S6: judging whether the height of the target pedestrian is smaller than a threshold value, if so, entering a step S7, otherwise, ending;
step S7: confirming a target behavior child, actively sending an alarm to a service desk and sending a scene picture and/or video;
and (5) ending.
In step S5, the passing judgment module caches all heights of the track ID, and adopts a certain learning strategy to estimate an average height.
The above examples and drawings are not intended to limit the form or form of the present invention, and any suitable variations or modifications thereof by those skilled in the art should be construed as not departing from the scope of the present invention.
Claims (5)
1. The height estimation method based on the monocular lens is characterized by comprising the following steps of:
the mounting height of the monocular RGB lens is obtained and is recorded asThe lens has no pitch angle and no rolling angle;
the monocular RGB lens shoots a target pedestrian, and an image plane is perpendicular to the ground and parallel to the plane where the target pedestrian is located;
the optical center of the monocular RGB lens is set asIts coordinates in the image plane are +.>The vertex of the target pedestrian is +.>Which is +.>The coordinates are +.>The method comprises the steps of carrying out a first treatment on the surface of the The bottom vertex of the target pedestrian is +.>Which is +.>The coordinates are +.>;
The lens is at the heightThe light beams horizontally injected form a plane parallel to the ground, the optical axis is also arranged on the plane, and the intersection point of the plane and the target pedestrian is +.>Obviously->The point on the image plane is marked +.>The ordinate is the same as the optical axis, is +.>The height of the target pedestrian is +.>;
The relationship according to the principle of pinhole imaging and similar triangles is:
the method comprises the following steps:
and finally, calculating:
。
2. a child passing brake apparatus, comprising: monocular RGB lens, target detection module, multi-target tracking module and brake judgment module;
the monocular RGB lens is horizontally arranged at the gate inlet and/or the gate outlet, the lens faces the gate channel to shoot, and the optical axis direction of the monocular RGB lens is parallel to the ground;
the target detection module is connected with the monocular RGB lens, and detects a target pedestrian in the monocular RGB lens in real time or in each frame to give a coordinate position of the target pedestrian in the image;
the multi-target tracking module is connected with the monocular RGB lens, and in the whole process of crossing the gate of the target pedestrian, the multi-target tracking module distributes a unique track ID and represents the unique track of the target pedestrian, so that the tracks of crossing the gate of different target pedestrians are counted separately, and a single frame is used as a result;
the gate passing judging module is connected with the monocular RGB lens, when the track of the target pedestrian leaves the picture, the gate passing judging module is triggered, the gate passing judging module can automatically call the height estimating module of the monocular RGB lens, the height situation of the gate passing target pedestrian is counted and automatically learned, and if the gate passing target pedestrian is found to be a child, an alarm is actively sent to the service desk and a scene picture and/or video is sent;
the height estimation method of the monocular RGB lens comprises the following steps:
the mounting height of the monocular RGB lens is obtained and is recorded asThe lens has no pitch angle and no rolling angle;
the monocular RGB lens shoots a target pedestrian, and an image plane is perpendicular to the ground and parallel to the plane where the target pedestrian is located;
the optical center of the monocular RGB lens is set asIts coordinates in the image plane are +.>The vertex of the target pedestrian is +.>Which is +.>The coordinates are +.>The method comprises the steps of carrying out a first treatment on the surface of the The bottom vertex of the target pedestrian is +.>Which is +.>The coordinates are +.>;
The lens is at the heightThe light beams horizontally injected form a plane parallel to the ground, the optical axis is also arranged on the plane, and the intersection point of the plane and the target pedestrian is +.>Obviously->The point on the image plane is marked +.>The ordinate is the same as the optical axis, is +.>The height of the target pedestrian is +.>;
The relationship according to the principle of pinhole imaging and similar triangles is:
the method comprises the following steps:
and finally, calculating:
。
3. a child passing brake apparatus according to claim 2 wherein: the crossing gate judging module caches all heights of the track ID, and adopts a certain learning strategy to estimate an average height.
4. A child passing gate judgment method based on the child passing gate device as claimed in any one of claims 2 or 3, comprising the steps of:
step S1: reading monocular RGB lens data;
step S2: the target detection module acquires one frame of image data of a monocular RGB lens, a coordinate position of a target pedestrian in an image is acquired through a target detection algorithm in the target detection module, and height estimation is carried out on the target pedestrian;
the height estimation method of the monocular RGB lens comprises the following steps:
the mounting height of the monocular RGB lens is obtained and is recorded asThe lens has no pitch angle and no rolling angle;
the monocular RGB lens shoots a target pedestrian, and an image plane is perpendicular to the ground and parallel to the plane where the target pedestrian is located;
the optical center of the monocular RGB lens is set asIts coordinates in the image plane are +.>The vertex of the target pedestrian is +.>Which is +.>The coordinates are +.>The method comprises the steps of carrying out a first treatment on the surface of the The bottom vertex of the target pedestrian is +.>Which is +.>The coordinates are +.>;
The lens is at the heightThe light beams horizontally injected form a plane parallel to the ground, the optical axis is also arranged on the plane, and the intersection point of the plane and the target pedestrian is +.>Obviously->The point on the image plane is marked +.>The ordinate is the same as the optical axis, is +.>The height of the target pedestrian is +.>;
The relationship according to the principle of pinhole imaging and similar triangles is:
the method comprises the following steps:
and finally, calculating:
;
step S3: the multi-target tracking module distributes unique track IDs for each target pedestrian by utilizing a multi-target tracking algorithm;
step S4: judging whether the track of the target pedestrian disappears, if yes, entering a step S5, otherwise, entering a step S1;
step S5: the passing gate judging module counts and automatically learns the height condition of a passing gate target pedestrian;
step S6: judging whether the height of the target pedestrian is smaller than a threshold value, if so, entering a step S7, otherwise, ending;
step S7: passing the gate by children;
and (5) ending.
5. The method for determining a child passing gate according to claim 4, wherein: in step S5, the crossing judgment module caches all heights of the track ID, and adopts a certain learning strategy to estimate an average height.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311486914.9A CN117218681B (en) | 2023-11-09 | 2023-11-09 | Height estimation method of monocular lens, child passing gate device and judging method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311486914.9A CN117218681B (en) | 2023-11-09 | 2023-11-09 | Height estimation method of monocular lens, child passing gate device and judging method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN117218681A true CN117218681A (en) | 2023-12-12 |
CN117218681B CN117218681B (en) | 2024-02-06 |
Family
ID=89046710
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202311486914.9A Active CN117218681B (en) | 2023-11-09 | 2023-11-09 | Height estimation method of monocular lens, child passing gate device and judging method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117218681B (en) |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104899554A (en) * | 2015-05-07 | 2015-09-09 | 东北大学 | Vehicle ranging method based on monocular vision |
CN105286871A (en) * | 2015-11-27 | 2016-02-03 | 西安交通大学 | Video processing-based body height measurement method |
CN106361345A (en) * | 2016-11-29 | 2017-02-01 | 公安部第三研究所 | System and method for measuring height of human body in video image based on camera calibration |
US20180005055A1 (en) * | 2016-07-04 | 2018-01-04 | Postech Academy-Industry Foundation | Moving object detection method in dynamic scene using monocular camera |
CN111080679A (en) * | 2020-01-02 | 2020-04-28 | 东南大学 | Method for dynamically tracking and positioning indoor personnel in large-scale place |
CN112950822A (en) * | 2021-02-05 | 2021-06-11 | 北京轨道交通路网管理有限公司 | Automatic identification system and method for gate passing logic |
WO2021208231A1 (en) * | 2020-04-15 | 2021-10-21 | 上海工程技术大学 | Gap measuring system and measuring method |
-
2023
- 2023-11-09 CN CN202311486914.9A patent/CN117218681B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104899554A (en) * | 2015-05-07 | 2015-09-09 | 东北大学 | Vehicle ranging method based on monocular vision |
CN105286871A (en) * | 2015-11-27 | 2016-02-03 | 西安交通大学 | Video processing-based body height measurement method |
US20180005055A1 (en) * | 2016-07-04 | 2018-01-04 | Postech Academy-Industry Foundation | Moving object detection method in dynamic scene using monocular camera |
CN106361345A (en) * | 2016-11-29 | 2017-02-01 | 公安部第三研究所 | System and method for measuring height of human body in video image based on camera calibration |
CN111080679A (en) * | 2020-01-02 | 2020-04-28 | 东南大学 | Method for dynamically tracking and positioning indoor personnel in large-scale place |
WO2021208231A1 (en) * | 2020-04-15 | 2021-10-21 | 上海工程技术大学 | Gap measuring system and measuring method |
CN112950822A (en) * | 2021-02-05 | 2021-06-11 | 北京轨道交通路网管理有限公司 | Automatic identification system and method for gate passing logic |
Also Published As
Publication number | Publication date |
---|---|
CN117218681B (en) | 2024-02-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106980829B (en) | Abnormal behaviour automatic testing method of fighting based on video analysis | |
CN105373135B (en) | A kind of method and system of aircraft docking guidance and plane type recognition based on machine vision | |
CN106778655B (en) | Human body skeleton-based entrance trailing entry detection method | |
CN103605967A (en) | Subway fare evasion prevention system and working method thereof based on image recognition | |
WO2004042673A2 (en) | Automatic, real time and complete identification of vehicles | |
CN103489199A (en) | Video image target tracking processing method and system | |
CN102609724B (en) | Method for prompting ambient environment information by using two cameras | |
CN101847206A (en) | Pedestrian traffic statistical method and system based on traffic monitoring facilities | |
CN103164711A (en) | Regional people stream density estimation method based on pixels and support vector machine (SVM) | |
CN105608417A (en) | Traffic signal lamp detection method and device | |
CN103268470A (en) | Method for counting video objects in real time based on any scene | |
CN107103300B (en) | Off-duty detection method and system | |
CN105303844B (en) | Night expressway fog automatic detection device and its detection method based on laser | |
CN102915433A (en) | Character combination-based license plate positioning and identifying method | |
CN104966062A (en) | Video monitoring method and device | |
KR20160035121A (en) | Method and Apparatus for Counting Entity by Using Location Information Extracted from Depth Image | |
Akamatsu et al. | Development of a person counting system using a 3D laser scanner | |
CN107274396B (en) | Device for counting number of people | |
Shafie et al. | Smart video surveillance system for vehicle detection and traffic flow control | |
CN117218681B (en) | Height estimation method of monocular lens, child passing gate device and judging method | |
Peláez et al. | Road detection with thermal cameras through 3D information | |
CN102930618B (en) | Real-time people counting method and system based on ranging principle | |
CN114596657B (en) | Gate passing system based on depth data | |
CN113468985B (en) | Locking method for suspicious radiation source carrying personnel | |
Lin et al. | Adaptive inverse perspective mapping transformation method for ballasted railway based on differential edge detection and improved perspective mapping model |
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 |