CN209949108U - VLC dynamic positioning system based on mean shift and unscented Kalman filtering - Google Patents

VLC dynamic positioning system based on mean shift and unscented Kalman filtering Download PDF

Info

Publication number
CN209949108U
CN209949108U CN201920840596.4U CN201920840596U CN209949108U CN 209949108 U CN209949108 U CN 209949108U CN 201920840596 U CN201920840596 U CN 201920840596U CN 209949108 U CN209949108 U CN 209949108U
Authority
CN
China
Prior art keywords
vlc
led
module
image
led lamp
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.)
Expired - Fee Related
Application number
CN201920840596.4U
Other languages
Chinese (zh)
Inventor
文尚胜
董芝强
关伟鹏
谢泽堃
陈邦栋
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
South China University of Technology SCUT
Original Assignee
South China University of Technology SCUT
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by South China University of Technology SCUT filed Critical South China University of Technology SCUT
Priority to CN201920840596.4U priority Critical patent/CN209949108U/en
Application granted granted Critical
Publication of CN209949108U publication Critical patent/CN209949108U/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Image Analysis (AREA)
  • Length Measuring Devices By Optical Means (AREA)

Abstract

The utility model discloses a VLC dynamic positioning system based on mean shift and unscented Kalman filtering, including VLC transmitting terminal, VLC receiving terminal and server end, wherein, VLC transmitting terminal include 28V constant voltage source, LED drive circuit, LED lamp and power adapter; the VLC receiving end comprises a positioning terminal and a camera subsystem; the server side comprises an image processing subsystem, an image tracking module and an unscented Kalman filter. The utility model discloses have the ability of tracking high-speed target, improved the positioning accuracy when LED is sheltered from, even half LED is shielded, the precision also can keep. In addition, the method has good robustness and real-time performance, and has wide application prospect in the field of indoor positioning.

Description

VLC dynamic positioning system based on mean shift and unscented Kalman filtering
Technical Field
The utility model relates to a visible light communication technical field, concretely relates to VLC dynamic positioning system based on mean value drift and unscented Kalman filtering.
Background
The indoor positioning technology has wide application prospect in indoor places such as commercial centers, large public buildings (subways, airports, libraries and the like), high-risk industrial parks, hospitals, nursing homes and the like, which have low GPS assistance and urgently need navigation positioning service.
Common indoor location technologies include Infrared (IR), ultrasonic, Radio Frequency Identification (RFID), Wireless Local Area Network (WLAN), bluetooth, and Ultra Wideband (UWB). Based on these techniques, different implementations have been developed. However, these have significant disadvantages and are difficult to popularize in view of positioning accuracy, anti-electromagnetic interference capability, or high cost of hardware devices.
Compared with the indoor positioning technology, the Visible Light Positioning (VLP) technology has the outstanding advantages of abundant bandwidth resources, strong anti-electromagnetic interference capability, high positioning precision, strong illumination capability and the like. Furthermore, the hardware equipment supporting the VLP system is relatively inexpensive, since no complex equipment is required to make accurate measurements. Therefore, the VLP technology has wide application prospect in the field of indoor positioning.
There are two modes of VLP-based indoor positioning systems, Photodiode (PD) -based positioning and Image Sensor (IS) -based positioning. The mobility of the positioning terminal is severely limited due to the sensitivity of the photodiode to the direction of the light beam, and there is another significant disadvantage that the robustness is poor. Even at the same location, repeated measurements can produce fluctuating values of light intensity. Furthermore, PD-based VLP systems are susceptible to interference from ambient light and reflected light. At the same time, accurate measurements of the received angle and signal strength are required. Otherwise the positioning result will have a significant error.
Theoretically, IS performs better than PD in the VLP domain, but existing studies have not achieved satisfactory results in terms of localization accuracy, real-time and robustness. When the target moves at a faster speed, the image of the LED may become blurred, which may lead to a positioning failure. In addition, when the light path between the LED and the positioning terminal is blocked, since most positioning algorithms are based on two or more LEDs, the algorithm will likely fail in the absence of one LED in the image. In addition, the masking effect is a fatal problem in the VLP field, and the use of optical flow to solve the problem in turn makes the calculation amount large.
On the other hand, smart phones are currently equipped with high resolution Complementary Metal Oxide Semiconductor (CMOS) sensor cameras, which can be easily combined with the IS-based VLP approach, and are of great commercial value.
SUMMERY OF THE UTILITY MODEL
The utility model aims at providing a VLC dynamic positioning system based on mean value drift and unscented Kalman filtering in order to the not enough of the performance of current method and VLP system to and cater for current requirement to indoor location technique.
The technical purpose of the utility model can be achieved by adopting the following technical scheme:
a VLC dynamic positioning system based on mean shift and unscented Kalman filtering comprises a VLC transmitting end, a VLC receiving end and a server end;
the VLC transmitting end comprises a 28V constant voltage source, an LED driving circuit, an LED lamp and a power adapter; the LED lamp is powered by a 28V constant voltage source; the LED driving circuit controls the LED lamp to generate high-frequency on-off; the power adapter converts the input voltage into 5V voltage and then supplies power to the LED driving circuit independently;
the VLC receiving end comprises a positioning terminal and a camera subsystem; the camera subsystem comprises a CMOS camera and a CMOS camera parameter setting module; the CMOS camera is used for continuously shooting real-time images of the LED lamp; the CMOS camera is connected with the positioning terminal, and transmits the obtained image to the positioning terminal; the positioning terminal comprises a WIFI module and a liquid crystal display screen, and the WIFI module transmits images to the server side;
the server end comprises an image processing subsystem, an image tracking module and an unscented Kalman filter; the image processing subsystem comprises a color image conversion gray level image module, an image filtering module, an image binarization module and an LED lamp identification module, wherein the modules are sequentially connected to receive and process images to realize LED lamp identification, and the IDs of the images and the LED lamps are transmitted to an unscented Kalman filter and an image tracking module; the unscented Kalman filter comprises an LED lamp position prediction module, an LED lamp position optimization module and a Kalman filtering parameter updating module, and transmits the current LED prediction result to the image tracking module; the image tracking module tracks and positions the target in real time, transmits positioning information to the positioning terminal through the WIFI module, and finally displays the positioning information by applying control voltage to a liquid crystal display screen of the positioning terminal.
Furthermore, the LED driving circuit adopts an STM32 system board as a chip, and simultaneously adopts a DD311 high-frequency switch to control the on-off of the current of the LED lamp, so that the LED lamp is turned on and off at high frequency.
Furthermore, the CMOS camera parameter setting module is used for setting the focal length, the exposure time, the exposure compensation and the light sensitivity of the CMOS camera, so that the CMOS camera can capture the LED image clearly.
Further, the locating method of the VLC dynamic locating system based on the mean shift and the unscented kalman filter includes the following steps;
s1, controlling the high-frequency on-off of the LED lamp through the LED driving circuit, finding the area where the LED exists through LED-ID identification, identifying the ID of the LED, and obtaining the initial position of the positioning terminal;
s2, dynamically tracking the LED lamps in the image sequence by using a mean shift algorithm and unscented Kalman filtering, and calculating the relative position relation of the pixel coordinates of the LED lamps of the current frame and the initial frame;
and S3, combining the initial position and the relative position relation of the positioning terminal to obtain the current position of the positioning terminal in the world coordinate system.
Further, the step S1 is as follows:
s101, controlling the on-off of output current through an LED driving circuit to enable an LED lamp to flicker according to specified frequency;
s102, shooting the LED lamp by using a camera, and converting the on-off frequency of the LED lamp into the number information of the LED lamp stripes on the image shot by the camera;
s103, the camera transmits the image to a positioning terminal through a USB connecting line, and then transmits the image to a server end through a local area network, the server end converts the LED image into a gray-scale image, then Gaussian filtering and binarization processing are carried out to obtain a binarization LED stripe image, and feature extraction and feature detection are carried out on the binarization LED stripe image to obtain the ID of the LED lamp;
and S104, calculating the actual initial position of the positioning terminal, namely the robot by using the world coordinates of the LED lamps and the geometric relation in the world coordinate system.
Further, the step S2 includes the following steps:
s201, describing a model of a tracking target by using an Epanechov kernel function, and assuming that the initial position of the tracking target of the current frame image is firstly positioned at a central point b of the target at the position of the previous frame image0Estimating the probability density of the target area in the current image frame, and calculating the center point b0Has a Bhattacharyya coefficient rho (b)0) Bhattacharyya coefficient rho (b) with other points if rho (b)>ρ(b0) The center of the search area is shifted to point b, if point b is the center point b0If the distance is smaller than any small constant epsilon, the cycle is ended, the point b is the central point of the tracking target area in the current frame image, and if not, the calculation is repeated;
the above is the process of the mean shift algorithm;
s202, introducing an unscented Kalman filtering optimization algorithm, and settingA state vector for the target, where x and y represent coordinates of the center point of the target;
Figure BDA0002085204180000042
and
Figure BDA0002085204180000043
the derivatives of x and y, i.e., the velocity of the target, respectively; h isk|k-1Representing the variation of the bandwidth of the kernel function of the target from time k-1 to time k, YkInitializing the unscented Kalman filtering for the observation variable of the target:
Figure BDA0002085204180000044
Figure BDA0002085204180000045
Figure BDA0002085204180000046
is a state vector XkMean value of PkIs a state vector XkThe variance of (a);
the ith Sigma point corresponding to the state vector of the target at the moment of k-1 is set as
Figure BDA0002085204180000047
The number of Sigma dots is 2n + 1.
Calculate Sigma point:
Figure BDA0002085204180000051
Figure BDA0002085204180000053
in the formula
Figure BDA0002085204180000054
Alpha is a candidate parameter, 0<α≤10-4；
By usingRepresenting the observed value at the target center at the ith Sigma point at time k, substituting the Sigma point into the following state transition equation F and observation equation H:
Figure BDA0002085204180000056
Figure BDA0002085204180000057
calculating the average value of the state vector and the observation variable at the moment before updating:
Figure BDA0002085204180000058
Figure BDA0002085204180000059
Figure BDA00020852041800000510
for observing variable YkIs determined by the average value of (a) of (b),
Figure BDA00020852041800000511
is a weight coefficient of the mean value,
Figure BDA00020852041800000512
Figure BDA00020852041800000513
calculating a Kalman gain K:
Figure BDA00020852041800000514
in the formula
Figure BDA00020852041800000515
Figure BDA00020852041800000516
Beta.gtoreq.0 and here the value is zero.
And the mean value and the variance of the state vector are updated in the following two formulas;
Figure BDA00020852041800000517
Figure BDA00020852041800000518
s203, introducing unscented Kalman filtering into a mean shift algorithm, predicting the most probable position of a current frame target by using the unscented Kalman filtering, taking the unscented Kalman filtering as prior information of the mean shift algorithm, searching the area by using the mean shift algorithm, taking a tracking result of the mean shift algorithm on the target as an observation model of the unscented Kalman filtering when the LED lamp is shielded, and combining a noise matrix to obtain final output.
Further, in step S3, a linear mapping from pixel coordinates to world coordinates is established through coordinate transformation between different coordinate systems by using a relative position relationship between the current frame LED lamp and the initial frame LED lamp, then a relative position relationship between the current frame LED lamp and the initial frame LED lamp is calculated by using a geometric position relationship between the LED lamp and the positioning terminal, and then a position of the positioning terminal in the world coordinate system is calculated by combining the initial position of the positioning terminal.
Further, the process of estimating the probability density in step S201 is as follows:
in the k-th frame image, n is assumed to be presentkA pixel
Figure BDA0002085204180000061
In the target region, the kernel function bandwidth is h, the feature space is uniformly divided into m subintervals, and then the probability density estimate of the target model feature value u being 1, …, m is as follows:
Figure BDA0002085204180000062
wherein
Figure BDA0002085204180000063
For normalization constants, the function k () is a contour function of the kernel function, passing through the distance center point (i.e., a in the formula)0) Measure the weight of each pixel, t (a)i) Is a point aiThe characteristic value of (d), δ (t (a)i) U) is used for judging whether the characteristic value of the pixel belongs to the u-th interval or not, and the initial frame central point is a0The probability density of which is estimated as
Figure BDA0002085204180000064
Further, the calculation equation of the Bhattacharyya coefficient is as follows:
Figure BDA0002085204180000066
wherein
Figure BDA0002085204180000067
Is the weight of the ith point.
Further, the Gaussian white noise matrix W of the state transformation equation F and the observation equation HkAnd VkSatisfies the following statistical properties:
Wk～N(0,Qk)Vk～N(0,Rk)
wherein Q iskAnd RkRespectively covariance matrices of the two noise matrices.
The utility model discloses for prior art have following advantage and effect:
the utility model discloses have the ability of tracking high-speed target, improved the positioning accuracy when LED is sheltered from, even half LED is shielded, the precision also can keep. In addition, the method has good robustness and real-time performance, and has wide application prospect in the field of indoor positioning.
Drawings
Fig. 1 is a schematic diagram of a VLC dynamic positioning method based on mean shift and unscented kalman filtering disclosed in the present invention;
fig. 2 is a schematic diagram of an industrial camera rolling door effect working principle in an embodiment of the present invention;
FIG. 3 is a schematic diagram of the unscented transformation process of the present invention;
FIG. 4 is a schematic diagram of the world coordinate system, the image coordinate system, and the camera coordinate system of the present invention;
fig. 5 is a schematic diagram of the geometric relationship between the LED and the image sensor in the embodiment of the present invention;
fig. 6 is a structural diagram of the device components of the VLC dynamic positioning system based on mean shift and unscented kalman filter disclosed in the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by a person skilled in the art without creative work belong to the protection scope of the present invention.
Example one
As shown in fig. 1, the present embodiment discloses a VLC dynamic positioning method based on mean shift and unscented kalman filter, including the following steps:
and S1, controlling the LED lamps to be on and off at different frequencies in a high frequency mode. When the dynamic visible light positioning algorithm is started, finding the area where the LED lamp is located through an LED-ID recognition algorithm, then recognizing the ID of each LED, and obtaining the initial position of the positioning terminal;
in a specific embodiment, the step S1 includes the following sub-steps:
s101, programming a single chip microcomputer to enable different pins of the single chip microcomputer to output pulse sequences with different frequencies, inputting the pulse sequences with different frequencies into an LED driving circuit, and controlling the on-off of output current by the LED driving circuit so as to enable an LED to flicker according to specified frequency;
s102, as shown in the attached figure 2, shooting the LED lamp by using a camera of a rolling shutter mechanism with a Complementary Metal Oxide Semiconductor (CMOS) image sensor, and converting the on-off frequency of the LED lamp into the number information of the LED lamp stripes shot by the camera by utilizing the characteristics of line-by-line exposure and line-by-line data reading;
s103, the camera transmits the image to a positioning terminal through a USB connecting line, and then transmits the image to a server end through a local area network, the server end converts the LED image into a gray-scale image, then Gaussian filtering and binarization processing are carried out to obtain a binarization LED stripe image, and feature extraction and feature detection are carried out on the binarization image, so that the ID of the LED lamp is obtained;
s104, obtaining the ID of each LED to obtain the world coordinate of each LED, and calculating the actual initial position of the positioning terminal, namely the robot by using the world coordinates of the LED lamps and the geometric relation in a world coordinate system;
s2, dynamically tracking the LED lamps in the image sequence by using a mean shift algorithm and unscented Kalman filtering, and calculating the relative position relation of the pixel coordinates of the LED lamps of the current frame and the initial frame by using the algorithm;
in a specific embodiment, the step S2 includes the following sub-steps:
s201, describing a model of a tracking target by using an Epanechov kernel function, and assuming that n exists in a k frame imagekA pixelIn the target area, it is assumed that the initial position of the tracking target of the current frame image is first positioned at the center point b of the target at the position of the previous frame image0The kernel function has a bandwidth of h, and the feature space is uniformly divided into m sub-regionsIn this case, the probability density of the target model eigenvalue u is 1, …, and m is estimated as:
Figure BDA0002085204180000082
wherein
Figure BDA0002085204180000083
For normalization constants, the function k () is a contour function of the kernel function, passing through the distance center point (i.e., a in the formula)0) Measure the weight of each pixel, t (a)i) Is a point aiThe characteristic value of (d), δ (t (a)i) The function of u) is to determine whether the characteristic value of the pixel belongs to the u-th interval. The central point of the initial frame is a0The probability density of which is estimated as
Figure BDA0002085204180000091
Using kernel density estimation, the weight of each sample point is related to their distance from the center point, which ensures that the mean shift process converges the kernel function to near the point where the estimated density is zero.
B is calculated by using the following equation0Has a Bhattacharyya coefficient rho (b)0)：
Figure BDA0002085204180000092
The Bhattacharyya coefficient ρ (b) for each point is then calculated using the following equation:
wherein
Figure BDA0002085204180000094
The weight for each point.
The Bhattacharyya coefficients have good fitness to the mean shift algorithm and are therefore selected to measure the similarity between the target model and the candidate model.
If ρ (b)>ρ(b0) Then the center of the search area is shifted to point b, if b and b0If the distance is smaller than any small constant epsilon, the circulation is ended, the point b is the central point of the tracking target area in the current frame image, and if not, the calculation is repeated.
S202, introducing an unscented Kalman filtering optimization algorithm, and setting
Figure BDA0002085204180000095
A state vector for the target, where x and y represent coordinates of the center point of the target;
Figure BDA0002085204180000096
and
Figure BDA0002085204180000097
the derivatives of x and y, i.e., the velocity of the target, respectively; h isk|k-1Representing the variation of the bandwidth of the kernel function from time k-1 to time k. Y iskIs the observed variable of the target. Firstly, initializing unscented Kalman filtering:
Figure BDA0002085204180000099
Figure BDA00020852041800000910
is a state vector XkMean value of PkIs a state vector XkThe variance of (a);
the ith Sigma point corresponding to the state vector of the target at the moment of k-1 is set as
Figure BDA00020852041800000911
The number of Sigma dots is 2n + 1. The Sigma points are then calculated as shown in fig. 3 using the following three equations:
Figure BDA0002085204180000101
Figure BDA0002085204180000102
Figure BDA0002085204180000103
in the formula
Figure BDA0002085204180000104
α is a candidate parameter, 0<α≤10-4。
By using
Figure BDA0002085204180000105
Representing the observed value at the center of the object at the ith Sigma point at time k. The Sigma point is substituted into the following state transition equation F and observation equation H:
Figure BDA0002085204180000107
the mean value before this time update of the state vector and the observed variable can then be calculated from the following equation:
Figure BDA0002085204180000109
Figure BDA00020852041800001010
for observing variable YkIs determined by the average value of (a) of (b),
Figure BDA00020852041800001011
is a weight coefficient of the mean value,
Figure BDA00020852041800001012
and calculating the Kalman gain K according to the following formula:
in the formula
Figure BDA00020852041800001015
Figure BDA00020852041800001016
Beta.gtoreq.0 and here the value is zero.
Finally, K is substituted into the following two formulas to update the mean and variance of the state vector:
Figure BDA00020852041800001017
Figure BDA00020852041800001018
the target position calculated by the unscented Kalman filtering algorithm at each time is compared with the last position of the unscented Kalman filtering algorithm, so that the state model of the unscented Kalman filtering algorithm is updated and corrected;
s203, introducing the unscented Kalman filtering into a mean shift algorithm, predicting the most possible position of the current frame target by using the unscented Kalman filtering, and taking the position as prior information of the mean shift algorithm to search the area by using the mean shift algorithm. By the scheme, the number of iterations can be effectively reduced, the positioning instantaneity is improved, and the maximum allowable operation speed of the positioning terminal is improved.
In addition, the Gaussian white noise matrix W of the state transition equation and the observation equationkAnd VkSatisfies the following statistical properties:
Wk～N(0,Qk)Vk～N(0,Rk)
wherein QkAnd RkRespectively covariance matrices of the two noise matrices. When the LED is shielded, the tracking result of the mean shift algorithm on the target is used as an observation model of unscented Kalman filtering, and the final output is obtained by combining a noise matrix, so that the problem that the LED is shielded can be effectively solved.
The relative position relation of the pixel coordinates of the current frame LED lamp and the initial frame LED lamp can be obtained through the mean shift algorithm combined with unscented Kalman filtering.
S3, obtaining the current position of the positioning terminal in the world coordinate system by using the initial position of the positioning terminal obtained in S1 and the relative position relation obtained in S2;
in a specific embodiment, the procedure of step S3 is as follows:
the relative position relationship of the pixel coordinates of the current frame LED lamp and the initial frame LED lamp is utilized, as shown in figure 4, linear mapping from the pixel coordinates to world coordinates is established through coordinate transformation between different coordinate systems, then, as shown in figure 5, the relative position relationship of the pixel coordinates of the current frame LED lamp and the initial frame LED lamp can be calculated through similar triangles by utilizing the geometric position relationship between the LED lamps and the positioning terminals, and then the position of the positioning terminals in the world coordinate system is obtained through calculation by combining the initial positions of the positioning terminals.
Example two
As shown in fig. 6, the present embodiment discloses a VLC dynamic positioning system based on mean shift and unscented kalman filter, including: VLC transmitting end, VLC receiving end and server end.
The VLC transmitting end comprises a 28V constant voltage source, an LED driving circuit, an LED lamp and a power adapter;
the LED lamp is powered by a 28V constant voltage source, so that the LED lamp has constant and proper light intensity; the LED driving circuit takes an STM32 system board as a chip, a DD311 high-frequency switch is used for controlling the on-off of the current of the LED lamp, so that the LED lamp generates high-frequency on-off which cannot be identified by human eyes, and the on-off frequency of each LED lamp is different, so that the LED can be used as a unique ID optical signal transmitter; the power adapter converts the input voltage into about 5V voltage to supply power for the LED drive circuit independently.
The VLC receiving end comprises a positioning terminal and a camera subsystem;
the camera subsystem consists of a CMOS camera and a CMOS camera parameter setting module, and the CMOS camera parameter setting module sets the focal length, the exposure time, the exposure compensation and the light sensitivity of the CMOS camera, so that the CMOS camera can capture an LED image clearly; the CMOS camera continuously shoots real-time images of the LED lamps, and due to the rolling door effect of the camera, namely the characteristics of line-by-line exposure and line-by-line data reading, images of the LED lamps with different numbers of light and shade stripes can be obtained; the camera is directly connected with the positioning terminal, and the obtained image is transmitted to the positioning terminal; the positioning terminal comprises a WIFI module and a liquid crystal display screen, and the image is transmitted to the server side through the WIFI module;
the server side comprises an image processing subsystem, an image tracking module and an unscented Kalman filter;
the image processing subsystem comprises a color image conversion gray level image module, an image filtering module, an image binarization module and an LED lamp identification module, wherein the modules are sequentially connected to receive and process images to realize LED identification, and then the images and the IDs of the LEDs are transmitted to an unscented Kalman filter and an image tracking module; the unscented Kalman filter comprises an LED lamp position prediction module, an LED lamp position optimization module and a Kalman filtering parameter updating module; the unscented Kalman filter transmits the current LED prediction result to an image tracking module; the image tracking module tracks and positions the target in real time, transmits positioning information to the positioning terminal by using the WIFI module, and finally displays the positioning information by applying control voltage to a liquid crystal display screen of the positioning terminal.
The utility model discloses an unscented kalman filter assists image tracking module to fix a position, can effectively improve the real-time and the robustness of location, is applicable to mobile terminal equipment, like smart mobile phone, panel computer, mobile robot etc..
The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments, and any other changes, modifications, substitutions, combinations, and simplifications which do not depart from the spirit and principle of the present invention should be equivalent replacement modes, and all are included in the scope of the present invention.

Claims (3)

1. A VLC dynamic positioning system based on mean shift and unscented Kalman filtering is characterized in that the VLC dynamic positioning system comprises a VLC transmitting end, a VLC receiving end and a server end;
the VLC transmitting end comprises a 28V constant voltage source, an LED driving circuit, an LED lamp and a power adapter; the LED lamp is powered by a 28V constant voltage source; the LED driving circuit controls the LED lamp to generate high-frequency on-off; the power adapter converts the input voltage into 5V voltage and then supplies power to the LED driving circuit independently;
the VLC receiving end comprises a positioning terminal and a camera subsystem; the camera subsystem comprises a CMOS camera and a CMOS camera parameter setting module; the CMOS camera is used for continuously shooting real-time images of the LED lamp; the CMOS camera is connected with the positioning terminal, and transmits the obtained image to the positioning terminal; the positioning terminal comprises a WIFI module and a liquid crystal display screen, and the WIFI module transmits images to the server side;
the server end comprises an image processing subsystem, an image tracking module and an unscented Kalman filter; the image processing subsystem comprises a color image conversion gray level image module, an image filtering module, an image binarization module and an LED lamp identification module, wherein the modules are sequentially connected to receive and process images to realize LED lamp identification, and the images and the IDs of the LED lamps are transmitted to an unscented Kalman filter and an image tracking module; the unscented Kalman filter comprises an LED lamp position prediction module, an LED lamp position optimization module and a Kalman filtering parameter updating module, and transmits the current LED prediction result to the image tracking module; the image tracking module tracks and positions the target in real time, transmits positioning information to the positioning terminal through the WIFI module, and finally displays the positioning information by applying control voltage to a liquid crystal display screen of the positioning terminal.
2. The VLC dynamic positioning system based on mean shift and unscented Kalman filtering of claim 1, characterized in that the LED driving circuit uses an STM32 system board as a chip, and a DD311 high frequency switch is used to control the on-off of the current of the LED lamp, so that the LED lamp generates high frequency on-off.
3. The VLC dynamic positioning system based on mean shift and unscented Kalman filtering of claim 1, wherein the CMOS camera parameter setting module is used to set the focal length, exposure duration, exposure compensation, and sensitivity of the CMOS camera, so that the CMOS camera can capture the LED image clearly.
CN201920840596.4U 2019-06-05 2019-06-05 VLC dynamic positioning system based on mean shift and unscented Kalman filtering Expired - Fee Related CN209949108U (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201920840596.4U CN209949108U (en) 2019-06-05 2019-06-05 VLC dynamic positioning system based on mean shift and unscented Kalman filtering

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201920840596.4U CN209949108U (en) 2019-06-05 2019-06-05 VLC dynamic positioning system based on mean shift and unscented Kalman filtering

Publications (1)

Publication Number Publication Date
CN209949108U true CN209949108U (en) 2020-01-14

Family

ID=69135181

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201920840596.4U Expired - Fee Related CN209949108U (en) 2019-06-05 2019-06-05 VLC dynamic positioning system based on mean shift and unscented Kalman filtering

Country Status (1)

Country Link
CN (1) CN209949108U (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110233667A (en) * 2019-06-05 2019-09-13 华南理工大学 VLC dynamic positioning method and system based on average drifting and Unscented kalman filtering

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110233667A (en) * 2019-06-05 2019-09-13 华南理工大学 VLC dynamic positioning method and system based on average drifting and Unscented kalman filtering

Similar Documents

Publication Publication Date Title
Maheepala et al. Light-based indoor positioning systems: A review
US10256906B2 (en) Position localization using visible light communication
CN107421506B (en) Visual tracking and positioning system and method based on indoor visible light communication
Guan et al. High-speed robust dynamic positioning and tracking method based on visual visible light communication using optical flow detection and Bayesian forecast
US9791543B2 (en) System and method for estimating the position and orientation of a mobile communications device in a beacon-based positioning system
Liu et al. Towards indoor localization using visible light communication for consumer electronic devices
Guan et al. Robust robotic localization using visible light positioning and inertial fusion
CN109116298B (en) Positioning method, storage medium and positioning system
CN108692701B (en) Mobile robot multi-sensor fusion positioning method based on particle filter
Xu et al. Experimental indoor visible light positioning systems with centimeter accuracy based on a commercial smartphone camera
CN108871290B (en) visible light dynamic positioning method based on optical flow detection and Bayesian prediction
Chen et al. A novel visible light positioning system with event-based neuromorphic vision sensor
CN110233667A (en) VLC dynamic positioning method and system based on average drifting and Unscented kalman filtering
Mao et al. High precision indoor positioning method based on visible light communication using improved Camshift tracking algorithm
CN108459597A (en) A kind of mobile electronic device and method for handling the task of mission area
CN108495251B (en) Combined positioning method based on Bluetooth and LED optical positioning
Guan et al. Visible light dynamic positioning method using improved camshift-kalman algorithm
JP2019100817A (en) Position estimation device, position estimation method and terminal device
Wu et al. Visible light positioning system based on CMOS image sensor using particle filter tracking and detecting algorithm
CN209949108U (en) VLC dynamic positioning system based on mean shift and unscented Kalman filtering
US20240074018A1 (en) System and method for controlling a light source for illuminating a scene of interest
CN105790837A (en) LED visual light indoor positioning method and LED visual light indoor positioning system based on image matching and fingerprint database
US20240127576A1 (en) System and method for matching of image features with indistinguishable landmarks
Li et al. High‐Accuracy and Real‐Time Indoor Positioning System Based on Visible Light Communication and Mobile Robot
CN110349196B (en) Depth fusion method and device

Legal Events

Date Code Title Description
GR01 Patent grant
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20200114

CF01 Termination of patent right due to non-payment of annual fee