CN107977601B - Human body recognition system above rail - Google Patents
Human body recognition system above rail Download PDFInfo
- Publication number
- CN107977601B CN107977601B CN201710895001.0A CN201710895001A CN107977601B CN 107977601 B CN107977601 B CN 107977601B CN 201710895001 A CN201710895001 A CN 201710895001A CN 107977601 B CN107977601 B CN 107977601B
- Authority
- CN
- China
- Prior art keywords
- real
- time
- rail
- image
- dynamics
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
- G06V20/42—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items of sport video content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/103—Static body considered as a whole, e.g. static pedestrian or occupant recognition
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Human Computer Interaction (AREA)
- Emergency Alarm Devices (AREA)
Abstract
The present invention relates to human body recognition systems above a kind of rail, including：Dynamics detection device is arranged in the change rail joint of rail, the real-time dynamics born at the change rail connector for detecting and exporting rail；Real-time display equipment, change rail near joints in rail are set, it is connect with dynamics detection device, the real-time dynamics born at change rail connector for receiving rail, and the real-time dynamics born at the change rail connector of rail is when being more than or equal to default dynamics threshold value, real-time display dynamics overload alarm information；Fixed mechanism, including egg type plastic shell, annular connect band, jitter analysis element and several capturing elements, egg type plastic shell is fixedly connected on telescope support, annular connect band is used to be socketed in the outer surface of egg type plastic shell, several capturing elements are non-to be equally spacedly distributed in annular connect band.By means of the invention it is possible to improve the accuracy of identification of human body above rail.
Description
Technical field
The present invention relates to human body recognition systems above field of target recognition more particularly to a kind of rail.
Background technology
Rail is made of potassium steel.It is obtained if the manganese of 2-3% is added in steel for manganese steel
Low manganess steel it is just glass-like crisp, it is broken once striking.However, if being added>13% manganese, is made potassium steel, then he is just
Become not only hard but also rich in toughness.So people manufacture wear-resisting machine parts and rail, bridge etc. with manganese steel.
In addition, on the top of rail, that is, that layer with wheel contact, add another metal：Indium, his adds
Enter, part on rail is made to have more some toughness in the case where ensureing hardness, this is critically important to the wheel of train, he and " wheel
When hoop " contact, collision, certain elasticity is had, ensure that the safe operation of train.
In the prior art, the research of rail is concentrated on rail material itself, and to the detection of rail ambient enviroment, especially
Its detection to rail surrounding body, due to the difficulty of human bioequivalence, and lacks effective testing mechanism.
Invention content
To solve the above-mentioned problems, it the present invention provides human body recognition system above a kind of rail, solves around rail
The problem that human body is difficult to.
The present invention at least has following two important inventive points：
(1) it based on the analysis of image content to several capturing elements on holder, determines and is transported in each capturing element
The motion vector of animal body determines that the holder that stretches shakes arrow accordingly based on several motion vectors of several capturing elements
Amount, consequently facilitating the capturing element of maximum quantity carries out track up to moving target；
(2) by with pre-stored each visual angle still image carry out it is matched in a manner of obtain best background image, meanwhile,
Each foreground point is determined by randomization judgment model, to lay solid data base for the identification of subsequent motion target
Plinth.
According to an aspect of the present invention, a kind of rail top human body recognition system is provided, the system comprises：
Dynamics detection device is arranged in the change rail joint of rail, the change rail connector place for detecting and exporting rail
The real-time dynamics born；
Real-time display equipment is arranged the change rail near joints in rail, is connect with dynamics detection device, for receiving rail
Change rail connector at the real-time dynamics born, and the real-time dynamics born at the change rail connector of rail be more than or equal to it is default
When dynamics threshold value, real-time display dynamics overload alarm information；
Fixed mechanism, including egg type plastic shell, annular connect band, jitter analysis element and several capturing elements,
Egg type plastic shell is fixedly connected on telescope support, and annular connect band is used to be socketed in the outer surface of egg type plastic shell,
Several capturing elements are non-to be equally spacedly distributed in annular connect band, wherein between telescope support, capturing element
Spacing is bigger；
Several capturing elements for shooting scene around to obtain several real-time scene images respectively respectively
Stream, jitter analysis element are connect with several capturing elements respectively, and following processing is executed to each capturing element：Based on real-time
The data analysis of front and back real-time scene image, judges and exports the motion vector of moving object therein in scene image stream；
Jitter analysis element determines that the holder that stretches is shaken accordingly based on several motion vectors of several capturing elements
Vector；
Flexible holder, including telescope support, trapezoidal erecting bed and direct current generator, telescope support are located on trapezoidal erecting bed,
Direct current generator is located in trapezoidal erecting bed, and direct current generator is based on shake vector controlled telescope support for receiving shake vector
Carry out corresponding expanding-contracting action；
Amount of noise detection device is connect with several capturing elements, if for receiving the shooting of several capturing elements
Dry real-time scene image carries out noise type detection to determine in the real-time scene image to each real-time scene image
Main noise number of types, using the minimum real-time scene image of main noise number of types as pending image；Wherein, mainly
Noise refers to that noise in image amplitude is more than or equal to the noise of default amplitude thresholds；
Background obtains equipment, is carried out for receiving pending image, and by pending image and each visual angle still image
It matches, will be exported as real-time background image with the highest visual angle still image of pending images match degree, each is regarded one by one
Angle still image is is shot the image obtained to the static scene under the corresponding visual angle without any Moving Objects in advance；
Target analysis equipment obtains equipment with background and connect, for receiving pending image and real-time background image, determines
The variance of the mean value of the pixel value of real-time background image and the pixel value of real-time background image, by each in pending image
The mean value of the pixel value of pixel and the pixel value of real-time background image does difference, when difference is more than or equal to real-time background image
At five times of the variance of pixel value, which is determined as foreground point, the pixel is otherwise determined as background dot, and will wait for
All foreground points handled in image form real-time foreground image, determine whether there is human body from real-time foreground image, and
It determines there are when human body, sends out emergency alarm.
Description of the drawings
Embodiment of the present invention is described below with reference to attached drawing, wherein：
Fig. 1 is the structural representation according to rail where human body recognition system above the rail shown in embodiment of the present invention
Figure.
Fig. 2 is the structural representation according to the fixed mechanism of human body recognition system above the rail shown in embodiment of the present invention
Figure.
Specific implementation mode
The embodiment of human body recognition system above the rail of the present invention is described in detail below with reference to accompanying drawings.
In order to overcome above-mentioned deficiency, the present invention to build a kind of rail top human body recognition system, implemented targetedly
Image Acquisition and image procossing mechanism processing, determine there are when human body, send out emergency alarm.
Fig. 1 is the structural representation according to rail where human body recognition system above the rail shown in embodiment of the present invention
Figure, Fig. 2 are according to the structural schematic diagram of the fixed mechanism of human body recognition system above the rail shown in embodiment of the present invention, such as
Shown in figure, the system comprises：
Dynamics detection device is arranged in the change rail joint of rail, the change rail connector place for detecting and exporting rail
The real-time dynamics born；
Real-time display equipment is arranged the change rail near joints in rail, is connect with dynamics detection device, for receiving rail
Change rail connector at the real-time dynamics born, and the real-time dynamics born at the change rail connector of rail be more than or equal to it is default
When dynamics threshold value, real-time display dynamics overload alarm information；
Fixed mechanism, including egg type plastic shell, annular connect band, jitter analysis element and several capturing elements,
Egg type plastic shell is fixedly connected on telescope support, and annular connect band is used to be socketed in the outer surface of egg type plastic shell,
Several capturing elements are non-to be equally spacedly distributed in annular connect band, wherein between telescope support, capturing element
Spacing is bigger；
Several capturing elements for shooting scene around to obtain several real-time scene images respectively respectively
Stream, jitter analysis element are connect with several capturing elements respectively, and following processing is executed to each capturing element：Based on real-time
The data analysis of front and back real-time scene image, judges and exports the motion vector of moving object therein in scene image stream；
Jitter analysis element determines that the holder that stretches is shaken accordingly based on several motion vectors of several capturing elements
Vector；
Flexible holder, including telescope support, trapezoidal erecting bed and direct current generator, telescope support are located on trapezoidal erecting bed,
Direct current generator is located in trapezoidal erecting bed, and direct current generator is based on shake vector controlled telescope support for receiving shake vector
Carry out corresponding expanding-contracting action；
Amount of noise detection device is connect with several capturing elements, if for receiving the shooting of several capturing elements
Dry real-time scene image carries out noise type detection to determine in the real-time scene image to each real-time scene image
Main noise number of types, using the minimum real-time scene image of main noise number of types as pending image；Wherein, mainly
Noise refers to that noise in image amplitude is more than or equal to the noise of default amplitude thresholds；
Background obtains equipment, is carried out for receiving pending image, and by pending image and each visual angle still image
It matches, will be exported as real-time background image with the highest visual angle still image of pending images match degree, each is regarded one by one
Angle still image is is shot the image obtained to the static scene under the corresponding visual angle without any Moving Objects in advance；
Target analysis equipment obtains equipment with background and connect, for receiving pending image and real-time background image, determines
The variance of the mean value of the pixel value of real-time background image and the pixel value of real-time background image, by each in pending image
The mean value of the pixel value of pixel and the pixel value of real-time background image does difference, when difference is more than or equal to real-time background image
At five times of the variance of pixel value, which is determined as foreground point, the pixel is otherwise determined as background dot, and will wait for
All foreground points handled in image form real-time foreground image, determine whether there is human body from real-time foreground image, and
It determines there are when human body, sends out emergency alarm.
Then, the concrete structure for continuing human body recognition system above the rail to the present invention is further detailed.
In human body recognition system above the rail：Real-time display equipment is in real-time display dynamics overload alarm information
Meanwhile the real-time dynamics also born at the change rail connector of real-time display rail.
In human body recognition system above the rail：The real-time display equipment is also connect with target analysis equipment, with
It receives and shows the emergency alarm.
In human body recognition system above the rail：The real-time display equipment is embedded in the side of flexible holder,
The real-time display equipment is connect with jitter analysis element.
In human body recognition system above the rail：The real-time display equipment is additionally operable to real-time display, and several shoot
Several motion vectors of element, and vector is shaken accordingly for the real-time display holder that stretches.
In human body recognition system above the rail：The real-time display equipment is LED display or LCD display.
And in human body recognition system above the rail, further include：
TF storage devices obtain equipment with the real-time display equipment and the background respectively and connect；
Wherein, the TF storage devices are additionally operable to store each visual angle static map for storing the default dynamics threshold value
Picture.
In addition, several cmos sensors can be used to realize in several capturing elements.Cmos image sensor have with
Under several advantages：1), random window reading capability.Random window read operation is that cmos image sensor is functionally better than
The one side of CCD, also referred to as area-of-interest are chosen.In addition, the highly integrated characteristic of cmos image sensor makes it hold very much
Easily realize while opening the function of multiple tracking windows.2), capability of resistance to radiation.Generally speaking, cmos image sensor potentially resists
Radiance has important enhancing relative to CCD performances.3), system complexity and reliability.It can using cmos image sensor
To be greatly simplified system hardware structure.4), non-destructive data playback mode.5), the spectrum assignment optimized.It is noticeable
Be, due to be integrated with multiple functional transistors in pixel structure, cmos image sensor there is also several disadvantages,
Mainly noise and filling rate two indices.In view of the relatively superior performance of cmos image sensor so that cmos image senses
Device is widely used in every field.
Human body recognition system above rail using the present invention is difficult to detect for rail environment human body in the prior art
Technical problem is carried out pending image and each visual angle still image by pending image of the acquisition from rail environment
It matches one by one, will export, and determine real-time as real-time background image with the highest visual angle still image of pending images match degree
The variance of the mean value of the pixel value of background image and the pixel value of real-time background image, by each pixel in pending image
The mean value of the pixel value of point and the pixel value of real-time background image does difference, when difference is more than or equal to the pixel of real-time background image
At five times of the variance of value, which is determined as foreground point, the pixel is otherwise determined as background dot, and will be pending
All foreground points in image form real-time foreground image, human body are determined whether there is from real-time foreground image, and in determination
There are when human body, emergency alarm is sent out, to ensure the safe operation of related railroad.
It is understood that although the present invention has been disclosed in the preferred embodiments as above, above-described embodiment not to
Limit the present invention.For any person skilled in the art, without departing from the scope of the technical proposal of the invention,
Many possible changes and modifications all are made to technical solution of the present invention using the technology contents of the disclosure above, or are revised as
With the equivalent embodiment of variation.Therefore, every content without departing from technical solution of the present invention is right according to the technical essence of the invention
Any simple modifications, equivalents, and modifications made for any of the above embodiments still fall within the range of technical solution of the present invention protection
It is interior.
Claims (4)
1. human body recognition system above a kind of rail, the system comprises：
Dynamics detection device is arranged in the change rail joint of rail, is born at the change rail connector for detecting and exporting rail
Real-time dynamics；
Real-time display equipment is arranged the change rail near joints in rail, is connect with dynamics detection device, the change for receiving rail
The real-time dynamics born at rail connector, and the real-time dynamics born at the change rail connector of rail is more than or equal to default dynamics
When threshold value, real-time display dynamics overload alarm information；
Fixed mechanism, including egg type plastic shell, annular connect band, jitter analysis element and several capturing elements, egg type
Plastic shell is fixedly connected on telescope support, and annular connect band is used to be socketed in the outer surface of egg type plastic shell, several
A capturing element is non-to be equally spacedly distributed in annular connect band, wherein the spacing between telescope support, capturing element
It is bigger；
Several capturing elements are trembled for being shot respectively to scene around to obtain several real-time scene image streams respectively
Dynamic analysis element is connect with several capturing elements respectively, and following processing is executed to each capturing element：Based on real-time scene
The data analysis of front and back real-time scene image, judges and exports the motion vector of moving object therein in image stream；
Jitter analysis element determines that the holder that stretches shakes vector accordingly based on several motion vectors of several capturing elements；
Flexible holder, including telescope support, trapezoidal erecting bed and direct current generator, telescope support are located on trapezoidal erecting bed, direct current
Motor is located in trapezoidal erecting bed, and direct current generator is carried out for receiving shake vector, and based on shake vector controlled telescope support
Corresponding expanding-contracting action；
Amount of noise detection device is connect with several capturing elements, for receive several capturing elements shooting several
It is main in the real-time scene image to determine to carry out noise type detection to each real-time scene image for real-time scene image
Noise type quantity, using the minimum real-time scene image of main noise number of types as pending image；Wherein, main noise
Refer to that noise in image amplitude is more than or equal to the noise of default amplitude thresholds；
Background obtains equipment, is carried out one by one for receiving pending image, and by pending image and each visual angle still image
Matching will be exported with the highest visual angle still image of pending images match degree as real-time background image, each visual angle is quiet
State image is is shot the image obtained to the static scene under the corresponding visual angle without any Moving Objects in advance；
Target analysis equipment obtains equipment with background and connect, and for receiving pending image and real-time background image, determines real-time
The variance of the mean value of the pixel value of background image and the pixel value of real-time background image, by each pixel in pending image
The mean value of the pixel value of point and the pixel value of real-time background image does difference, when difference is more than or equal to the pixel of real-time background image
At five times of the variance of value, which is determined as foreground point, the pixel is otherwise determined as background dot, and will be pending
All foreground points in image form real-time foreground image, human body are determined whether there is from real-time foreground image, and in determination
There are when human body, emergency alarm is sent out；
Real-time display equipment goes back the change rail connector place of real-time display rail while real-time display dynamics overload alarm information
The real-time dynamics born；
The real-time display equipment is also connect with target analysis equipment, to receive and show the emergency alarm；
The real-time display equipment is embedded in the side of flexible holder, and the real-time display equipment connects with jitter analysis element
It connects.
2. human body recognition system above rail as described in claim 1, it is characterised in that：
The real-time display equipment is additionally operable to several motion vectors of several capturing elements of real-time display, and for real-time
The flexible holder of display shakes vector accordingly.
3. human body recognition system above rail as claimed in claim 2, it is characterised in that：
The real-time display equipment is LED display or LCD display.
4. human body recognition system above rail as described in any one of claims 1-3, which is characterized in that further include：
TF storage devices obtain equipment with the real-time display equipment and the background respectively and connect；
Wherein, the TF storage devices are additionally operable to store each visual angle still image for storing the default dynamics threshold value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710895001.0A CN107977601B (en) | 2017-09-28 | 2017-09-28 | Human body recognition system above rail |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710895001.0A CN107977601B (en) | 2017-09-28 | 2017-09-28 | Human body recognition system above rail |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107977601A CN107977601A (en) | 2018-05-01 |
| CN107977601B true CN107977601B (en) | 2018-10-30 |
Family
ID=62012333
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710895001.0A Active CN107977601B (en) | 2017-09-28 | 2017-09-28 | Human body recognition system above rail |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107977601B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109725280B (en) * | 2019-01-17 | 2021-05-18 | 乐清市启程电气有限公司 | Watt-hour meter field parameter extracting platform |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TW200710765A (en) * | 2005-05-31 | 2007-03-16 | Objectvideo Inc | Human detection and tracking for security applications |
| CN103186766A (en) * | 2011-12-30 | 2013-07-03 | 中国铁道科学研究院电子计算技术研究所 | Integrated portrait recognition device |
| CN104207787A (en) * | 2014-09-28 | 2014-12-17 | 广州视源电子科技股份有限公司 | Human body identification method |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AU2004280500A1 (en) * | 2003-10-09 | 2005-04-21 | De Beers Consolidated Mines Limited | Enhanced video based surveillance system |
| CN105075234B (en) * | 2013-09-30 | 2019-08-13 | 株式会社尼康 | Electronic equipment, the control method of electronic equipment and control program |
| CN104477185B (en) * | 2014-11-05 | 2016-09-21 | 中国铁路总公司 | Orbit connection device and gapless line transport trainset |
-
2017
- 2017-09-28 CN CN201710895001.0A patent/CN107977601B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TW200710765A (en) * | 2005-05-31 | 2007-03-16 | Objectvideo Inc | Human detection and tracking for security applications |
| CN103186766A (en) * | 2011-12-30 | 2013-07-03 | 中国铁道科学研究院电子计算技术研究所 | Integrated portrait recognition device |
| CN104207787A (en) * | 2014-09-28 | 2014-12-17 | 广州视源电子科技股份有限公司 | Human body identification method |
Non-Patent Citations (1)
| Title |
|---|
| 轨道交通闸机系统中的人体识别技术;牛国柱 等;《城市轨道交通研究》;20121231;第29-32页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107977601A (en) | 2018-05-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105512632B (en) | Biopsy method and device | |
| EP3321844B1 (en) | Action recognition in a video sequence | |
| EP3420530B1 (en) | A device and method for determining a pose of a camera | |
| US7542080B2 (en) | Measuring device using image and measurement method using image | |
| JP5898199B2 (en) | Posture estimation apparatus, posture estimation method, and posture estimation program | |
| KR101898782B1 (en) | Apparatus for tracking object | |
| CN110072673B (en) | Robot for tracking human and method thereof | |
| US20160061582A1 (en) | Scale estimating method using smart device and gravity data | |
| KR20160088896A (en) | Detection of an object's varying features with a non-stationary device | |
| US9842254B1 (en) | Calibrating inertial measurement units using image data | |
| WO2003053241A3 (en) | Device, system and method for image based size analysis | |
| JP2007011815A (en) | Driving recorder | |
| US9256957B1 (en) | Method for moving-object detection tracking identification cueing of videos | |
| US11388343B2 (en) | Photographing control method and controller with target localization based on sound detectors | |
| Bian et al. | Fall detection based on skeleton extraction | |
| CN107977601B (en) | Human body recognition system above rail | |
| JP2013197625A5 (en) | ||
| CN113168224A (en) | Information processing apparatus, information processing method, and program | |
| JP4361913B2 (en) | Motion calculation device | |
| US11399133B2 (en) | Image capture device with an automatic image capture capability | |
| CN107784272B (en) | Human body recognition method | |
| US20180169473A1 (en) | Motion analyzing apparatus, motion analyzing method, and recording medium | |
| Brunner et al. | Combining multiple sensor modalities for a localisation robust to smoke | |
| WO2016063595A1 (en) | Image processing device, image processing method and program | |
| KR101958249B1 (en) | Underwater golf ball regression by small unmanned underwater vehicle |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20180918 Address after: 362000 Xu Xu 151, Xu CuO village, Hou Long Town, Quangang District, Quanzhou, Fujian Applicant after: Zhang Sanmei Address before: 215400 Luoyang Road, Taicang Economic Development Zone, Suzhou, Jiangsu 28 Applicant before: She Yidao |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20200103 Address after: 362000 No. 236, Liang Liang village, Tu Tu Town, Huian County, Fujian, Quanzhou Patentee after: Wang Peijie Address before: 362000 Xu Xu 151, Xu CuO village, Hou Long Town, Quangang District, Quanzhou, Fujian Patentee before: Zhang Sanmei |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20200708 Address after: 362000 Xu Xu 151, Xu CuO village, Hou Long Town, Quangang District, Quanzhou, Fujian Patentee after: Zhang Sanmei Address before: 362000 No. 236, Liang Liang village, Tu Tu Town, Huian County, Fujian, Quanzhou Patentee before: Wang Peijie |
|
| TR01 | Transfer of patent right |