CN106767822A - Indoor locating system and method based on camera communication with framing technology - Google Patents

Abstract

The invention provides an indoor positioning system and method based on camera communication and image positioning technology. The system includes a transmitting end and a receiving end, the transmitting end is used to transmit the position information of the transmitting end to the space where the optical signal is coded and modulated; the receiving end is used to receive the optical signal, and the After the optical signal is decoded and demodulated, the image positioning calculation is performed to obtain the position information. This application proposes an indoor positioning system and method based on camera communication and image positioning technology. It is designed to solve the problems of high hardware cost, high-speed movement, and low positioning accuracy of existing indoor positioning systems. It has green environmental protection and natural high-density coverage. , The signal is easy to realize the advantages of space separation.

Description

Indoor positioning system and method based on camera communication and image positioning technology

technical field

The present invention relates to the technical field of positioning, and more specifically, to an indoor positioning system and method based on camera communication and image positioning technology.

Background technique

At present, positioning technology develops very rapidly and is widely used. The commonly used positioning methods mainly include GPS positioning technology, RFID indoor positioning technology and Wi-Fi positioning technology.

Among them, the system that uses GPS positioning satellites to perform real-time positioning and navigation around the world is called the global satellite positioning system, or GPS for short. GPS is a comprehensive, all-weather, all-time, high-precision satellite navigation system developed and established by the US Department of Defense. It can provide global users with low-cost, high-precision three-dimensional position, speed and precise timing navigation information. The application model of satellite communication technology in the field of navigation has greatly improved the informatization level of the earth society and strongly promoted the development of the digital economy. However, the ability of GPS signals to penetrate walls is very weak, so in indoor environments, GPS signals are very weak and difficult to locate, so GPS technology is not suitable for indoor positioning.

Among them, the RFID indoor positioning system usually consists of electronic tags, radio frequency readers, middleware and computer databases. The radio frequency tag and the reader exchange data through the transmission channel of the space electromagnetic wave erected by the antenna. In the application of the positioning system, the radio frequency reader is placed on the moving object to be tested, and the radio frequency electronic tag is embedded in the operating environment. The position identification information is stored on the electronic tag, and the reader is connected to the information database through wired or wireless.

The radio frequency identification indoor positioning technology has a very short working distance, but it can obtain centimeter-level positioning accuracy information within a few milliseconds, and due to the advantages of non-line-of-sight electromagnetic fields, the transmission range is large, and the volume of the identification is relatively small, and the cost is relatively low. However, it has no communication capability, poor anti-interference ability, and is not easy to integrate into other systems, and the user's security and privacy protection and international standardization are not perfect.

Among them, there are two types of Wi-Fi positioning technology, one is to use the wireless signal strength of the mobile device and three wireless network access points, and use the difference algorithm to more accurately triangulate the positioning of people and vehicles. The other is to record the signal strength of a huge number of certain location points in advance, and determine the location ("fingerprint" positioning) by comparing the signal strength of the newly added device with the database with huge data. The overall accuracy of Wi-Fi positioning technology for indoor positioning is not high, only 2 meters, and requires a large number of network access points, which is expensive.

Contents of the invention

The present invention provides an indoor positioning system and method based on camera communication and image positioning technology that overcomes the above problems or at least partially solves the above problems.

According to one aspect of the present invention, an indoor positioning system based on camera communication and image positioning technology is provided, including a transmitting end and a receiving end,

The transmitting end is used to transmit the location information of the transmitting end to the space where the optical signal is coded and modulated;

The receiving end is configured to receive the optical signal, perform image positioning calculation to obtain position information after decoding and demodulating the optical signal.

According to another aspect of the present invention, an indoor positioning method based on camera communication and image positioning technology is provided, including:

S1, receiving a modulated optical signal with position coding information to generate image information;

S2. Extract the vertical distance from the receiving end to the light source, the position information of the light source, and the optical signal strength from the image information, and obtain the horizontal distance from the receiving end to the light source;

S3, coordinate positioning and coordinate screening are performed according to the position information of the two light sources in the image information and the horizontal distance from the receiving end to the light source, so as to locate the position coordinates of the receiving end

The indoor positioning system and method based on camera communication and image positioning technology proposed by this application are designed to solve the problems of high hardware cost, high-speed movement, and low positioning accuracy of existing indoor positioning systems. Coverage, signal easy to achieve space separation and other advantages.

Description of drawings

Fig. 1 schematic diagram of BP neural network model of the present invention;

2 is a schematic diagram of an indoor positioning system based on camera communication and image positioning technology according to the present invention;

Fig. 3 is a schematic diagram of the first embodiment of the system of the present invention;

Fig. 4 is the schematic diagram of the second embodiment of the system of the present invention;

Fig. 5 is the waveform schematic diagram of UPSOOK modulation method described in the present invention;

Fig. 6 is a schematic diagram of the height measurement principle of the present invention;

Fig. 7 is a schematic diagram of the coordinate positioning principle of the present invention;

Fig. 8 is a diagram showing the position of the coordinates of the receiving end and the normal display of the coordinates of the receiving end according to the principle of coordinate screening according to the present invention;

Fig. 9 is a schematic diagram of the positions of the receiving end and the LED light when the receiving end of the coordinate screening principle of the present invention rotates;

FIG. 10 is a flowchart of an indoor positioning method based on camera communication and image positioning technology according to the present invention.

detailed description

The specific implementation manners of the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

The implementation of the present invention needs to rely on the BP neural network model, and the basic principles and algorithm flow of the BP neural network are as follows:

BP (Back Propagation) neural network, the full name of the multi-layer forward neural network of error backpropagation, is one of the most widely used neural network models at present, which can be used for classification, clustering and prediction, etc., through historical data During training, the network can learn the knowledge implicit in the data. The main feature of the BP neural network is that the signal is transmitted forward and the error is reversed; it is usually composed of an input layer, a hidden layer and an output layer. The layers are fully interconnected, and the nodes of each layer are not connected. In the process of forward propagation, information is calculated layer by layer from the input layer nodes to the output layer through the hidden layer, and the state of neurons in each layer only affects the state of neurons in the next layer. If the expected value is not obtained at the output layer, the error of the output layer is calculated, and then backpropagation, and the error signal is sent back through the network according to the original connection path to modify its parameters until the desired goal is achieved. Theory proves that the three-layer BP neural network has the ability to simulate any complex nonlinear mapping as long as the number of nodes in the hidden layer is large enough. Combined with the actual needs of this system, the specific three-layer BP neural network structure is shown in Figure 1.

In the figure above, X ₁ X ₂ ... X ₅ is the five-dimensional input value of the BP neural network, Output is the predicted output of the BP neural network, and ω _ij and ω _jk are the weights of the BP neural network. It can be seen from Figure 1 that the BP neural network can be regarded as a nonlinear function, and the input value and predicted value of the network are the independent variable and dependent variable of the function respectively. When the number of input nodes is 5 and the number of output nodes is 1, the BP neural network expresses the functional mapping relationship from 5 independent variables to 1 dependent variable.

Before the prediction of BP neural network, the network must be trained first, and the network has associative memory and prediction ability through training. The training process of BP neural network includes the following steps.

Step 1: Network initialization. The structure of the network is determined according to the input-output sequence (X, Y) of the system. Including the number of nodes (n, m, l) of the input/output layer and the hidden layer, the connection weights ω _ij , ω _jk between the input layer, hidden layer and output neurons, and the hidden layer threshold a, output Initialization of layer threshold b, given learning rate and neuron activation function.

Step 2: Hidden layer output calculation. According to the input vector X, the connection weight ω _ij between the output layer and the hidden layer, and the hidden layer threshold, the hidden layer output H is calculated.

In formula (1), f is the hidden layer activation function. The present invention selects the logarithmic S-shaped transfer activation function:

Step 3: Output layer output calculation. According to the output H of the hidden layer, connect the weight ω _jk and the threshold b, and calculate the prediction output O of the BP neural network.

Step 4: Error calculation. Calculate the network prediction error e according to the network prediction output O and the expected output Y.

e _k = Y _k - O _k k = 1, 2, ..., m (4)

Step 5: Weight update. Update the network connection weights ω _ij , ω _jk according to the pre-value error e of the network.

ω _jk ＝ω _jk +γH _j e _k j＝1,2,...,l; k＝1,2,...,m (6)

In the formula, γ is the activation function of the hidden layer.

Step 6: Threshold update. Update the network node thresholds a and b according to the pre-value error e of the network.

b _k ＝b _k +e _k k＝1,2,...,m (8)

Step 7: Determine whether the algorithm iteration is over, if not, return to step 2.

The above is an introduction to the BP neural network model and its working principle.

As shown in Figure 2, it is an indoor positioning system based on camera communication and image positioning technology of the present invention, including a transmitting end and a receiving end,

The transmitting end is used to transmit the location information of the transmitting end to the space where the optical signal is coded and modulated;

The receiving end is configured to receive the optical signal, perform image positioning calculation to obtain position information after decoding and demodulating the optical signal.

The transmitting end includes a light source, an encoder and a modulator,

The encoder is used to perform optical signal encoding on the position information of the light source;

The modulator is used to modulate the encoded optical signal;

The light source is used to emit the modulated optical signal.

The receiver includes:

An image sensor for receiving light signals and generating image information;

a demodulator, configured to demodulate the image information opposite to the modulation process, so as to obtain the optical signal strength of the transmitting end;

a decoder, used to decode the demodulated image information opposite to the encoding process, so as to obtain the position information of the light source at the transmitting end;

The image processing module is used to obtain the vertical distance and horizontal distance from the receiving end to the light source at the transmitting end according to the optical signal strength and the position information of the light source at the transmitting end, so as to locate the position coordinates of the receiving end.

The image processing module includes:

The height measurement module is used to obtain the vertical distance from the receiving end to the light source according to the distance between the two light sources in the image information and the physical focal length of the image sensor;

The BP neural network model is used to output the horizontal distance that the receiving end deviates from different light sources according to the vertical distance from the receiving end to the light source and the optical signal strength;

The positioning module is used to perform coordinate positioning and coordinate screening according to the position information of the two light sources in the image information and the horizontal distance of the receiving end from the light source, and locate the position coordinates of the receiving end.

The light source used in the system of the present invention is an LED light source, the image sensor is a camera, etc., and an indoor positioning system is formed by using the visible light LED and the receiving end camera. The data sending and receiving of the present invention are based on the camera communication technology (OCC), and the positioning technology utilizes the image and data collected by the camera, fully utilizes the advantages of the camera communication, and is a brand-new positioning means.

The system of the present invention has two implementations, respectively the first embodiment shown in FIG. 3 and the second embodiment shown in FIG. 4 .

The implementation environment of the embodiment of the present invention is:

The transmitter encoder is a GeoHash encoder, the modulator is a UPSOOK modulator, and the light source is an LED light source.

LED lights with coded modulation function are installed on the top of the indoor area of the building that needs to be positioned according to predetermined rules. These LED lights have fixed coordinates and positions, and the distance between the LED lights is known through installation.

For the coordinates of the LED lights, a traditional Cartesian coordinate system is used, that is, the coordinates of each LED are (x, y).

At the same time, if you send the (x, y) coordinates directly, you need to send the horizontal and vertical coordinates separately, which is difficult to send, and the Cartesian coordinate system is not suitable for positioning, so you need to send the (x, y) coordinate information of the LED light Encoded and sent. The invention adopts the GeoHash algorithm to encode the coordinates of the LED lights.

The GeoHash encoding method has the following advantages:

First of all, if you use Cartesian coordinates directly, you need to apply indexes to both horizontal and vertical coordinates at the same time. In some cases, you cannot apply indexes to both columns at the same time (for example, versions before MySQL4, the data layer of Google App Engine, etc.). GeoHash uses a string to represent the horizontal and vertical coordinates. Just apply an index on one column.

Secondly, what GeoHash represents is not a point, but a rectangular area. For example, the code 11110101 represents a rectangular area. The longer the code, the smaller the rectangular area represented, the higher the positioning accuracy, and the higher the communication cost and cost; the shorter the code, the larger the rectangular area represented, and the lower the positioning accuracy, but the precise coordinate information of the user may not be exposed , maintaining user privacy and reducing communication costs. Users can flexibly adjust the GeoHash encoding length according to actual positioning requirements.

Third, coded prefixes can represent larger regions. For example, 11110101, its prefix 1111 indicates a larger range including the code 11110101. This feature can be used for nearby place searches. Calculate GeoHash (for example, 11110101) based on the user's current coordinates and then take its prefix to query, and you can query all nearby locations. At the same time, according to this feature, GeoHash coding has a certain error correction function, because the code word prefixes of adjacent lights are the same. If the prefixes of adjacent lights are different after decoding, it can be judged that there is an error in decoding.

GeoHash encoding the coordinates of LED lights is much more efficient than directly using rectangular coordinates.

Before each LED lamp emits light, after performing GeoHash encoding on each position, it is modulated and then emitted, and the modulation is UPSOOK modulation. UPSOOK is a relatively successful modulation method in camera communication. It is a modulation method based on the state of the LED light on and off.

The schematic diagram of UPSOOK modulated LED light waveform is shown in Figure 5. Although the frame rate of the camera has a precise correspondence with the baud rate of the transmitted signal and the carrier frequency, the sampling phase of the camera and the phase of the transmitted signal cannot be synchronized, so the sampling time of the camera is random. In this way, there are two possibilities for the sampled LED on and off states.

As shown in Figures 5(b) and 5(c), for the LED light + camera receiving system that sends the same signal, the received LED states are completely opposite except for the FH state due to the different sampling positions of the image sensor.

It can be seen from the figure that the LED state detected in Figure 5(b) corresponds to the original data in Figure 5(a) one by one ("on" corresponds to "1"; "off" corresponds to "0"), but in Figure 5(c) The detected LED state is completely opposite to the original data in Figure 5(a) ("on" corresponds to "0"; "off" corresponds to "1"), so a flag bit can be added to check whether there is a phase error at the receiving end . The present invention has stipulated that the flag must be bit 1 in advance, as the state of the LED light of the flag is the same as the original data (i.e. bit 1), it means that the sampling phase is correct; as the state of the LED light of the flag is opposite to the original data (bit 1), It means that the sampling phase is incorrect, and the received bits need to be inverted (that is, 0 becomes 1, and 1 becomes 0).

The image information of the visible light LED is captured by the camera at the receiving end, as shown in Figure 5(d). Use image segmentation and image recognition technology to identify two lights, use UPSOOK demodulation technology to receive the information contained in the LEDs, and each picture can receive one bit of information. Use the GeoHash decoding method corresponding to the GeoHash encoding method to obtain the coordinate position of the corresponding LED light.

As shown in Figure 3, the first embodiment of the indoor positioning system based on camera communication and image positioning technology includes a transmitting end and a receiving end, and the receiving end is a mobile terminal, which can be any intelligent terminal with a camera, including a smart phone , Tablet PC, etc.

The transmitter includes GeoHash encoder, UPSOOK modulator and LED light source. A GeoHash encoder, a UPSOOK modulator and an LED light source are a set of light-emitting components. The transmitting end includes multiple light-emitting components. The number depends on the size of the indoor positioning space and the installation layout of the light-emitting components.

Each light-emitting component performs GeoHash encoding on its own coordinate position through the GeoHash encoder, performs UPSOOK modulation through the UPSOOK modulator, and then transmits an optical signal to the indoor positioning area.

The mobile terminal includes image sensor, UPSOOK demodulator, GeoHash decoder, altitude measurement module, BP neural network model and positioning module.

The image sensor receives the light signal and generates image information, and processes the recognized LED light image through image processing; then performs demodulation corresponding to UPSOOK modulation through the UPSOOK demodulator, and this process can obtain the light signal intensity.

Thereafter, it is divided into two processes. In one process, the GeoHash decoder performs GeoHash decoding corresponding to the GeoHash encoding method on the information after UPSOOK demodulation, and obtains the coordinate position of the corresponding LED light.

In another process, the height measurement module first obtains the vertical distance from the receiving end to the light source according to the distance between the two LEDs in the image information and the physical focal length of the image sensor; then, according to the measured vertical distance and optical signal strength, the BP neural network model processing to obtain the horizontal distance of the mobile terminal away from different light sources.

The coordinate position of the LED light and the horizontal distance of the mobile terminal from different light sources obtained in the last two processes need to be processed by the positioning module, and finally the mobile terminal is positioned.

The specific implementation of height measurement is:

As shown in Figure 6, the two LED lights on the top of the indoor positioning area are A and B, and the distance between A and B is known as AB. f is the physical focal length of the camera of the mobile terminal, which can be obtained by querying the camera parameters; O is the focal point, CD is the distance between A and B on the picture after being captured by the camera of the mobile terminal, and the actual size corresponding to the image pixel is determined by the size of the image sensor of the camera Decision, which can be obtained by querying the camera parameters.

Then AOB and COD form a similar triangle, and we can get:

The vertical height of the mobile terminal from the LED light is:

The specific implementation of the BP neural network model is:

For the LED light source described in the present invention, particularly, for the Lambertian LED, the channel gain between the i-th LED lamp and the mobile terminal is H _i (0), and its expression is:

Among them, m is the Lambert order of the LED lamp, _AR is the area of the detector, d _i is the distance between the LED and the detector, is the LED emission angle, ψ _i is the angle between the incident light of the LED lamp and the normal direction of the detector, Ψ _c is the detector FOV, T _s (ψ _i ) is the transmittance of the receiving filter, g(ψ _i ) is the receiving Concentrator gain.

Assuming that the receiving system does not use filters and concentrators, the formula (10) becomes:

It can be obtained from the above formula that the relationship between the distance and the light intensity at the receiving end is a nonlinear relationship, and it is relatively complicated to fit the functional relationship between the distance and the signal intensity at the receiving end by conventional methods. The present invention solves this problem through a BP neural network.

At the same distance, the height is different, and the brightness of the LED at the receiving end is obviously different. Therefore, relying solely on the signal strength of the LED at the receiving end is not enough to determine the distance between the receiving end and the light-emitting LED. The horizontal distance from the receiving end to the LED at the lower height.

Therefore, it is necessary to measure the horizontal distance between the receiving end and the LED and the corresponding signal strength of the receiving end when the receiving end is at different vertical heights from the light source, and use these data to train a BP neural network model.

In actual situations, the light-emitting LEDs are arranged above the building, so the vertical height between the receiving end and the light-emitting LEDs can be converted into the height coordinates of the receiving end.

In this way, the vertical distance from the receiving end to the light source, the horizontal distance from the receiving end to the light-emitting LED, and the optical signal strength information received by the corresponding receiving end are used as a data set, and these data sets are used to train the BP neural network; the trained BP neural network The model can calculate the horizontal distance of the receiving end away from the LED according to the vertical distance from the receiving end to the light source and the received optical signal strength of the receiving end.

When calculating the intensity of the optical signal, the optical signal received by the image sensor of the mobile terminal is a picture. If the gray value of each pixel of the image at the receiving end is directly added to represent the signal intensity at the receiving end, the neural network input vector dimension If it is too low, it is not conducive to taking advantage of the neural network; and if the common gray histogram is used to represent the intensity information of the image at the receiving end, then in the input vector, the dimension of the intensity information has 256 dimensions, which is much larger than the 1 dimension of the height coordinate. The intensity information will overwhelm the height information, which will greatly affect the accuracy of the neural network output. Therefore, we compress the information of the gray histogram. The specific implementation is: 1. Convert the color picture at the receiving end to a grayscale picture, and the grayscale value of the image is between 0-256. 2. Divide the grayscale value of the grayscale image into four intervals, which are: 0-64 , 64-128, 128-192, 192-256, and count the number of pixels in the four gray-scale intervals respectively to form a four-dimensional vector to represent the strength value of the signal at the receiving end.

The input of the BP neural network is a five-dimensional vector, which is composed of the signal strength of the receiving end and the height coordinate of the receiving end; the output is the horizontal distance between the receiving end and the corresponding light-emitting LED. That is, the five-dimensional input values X ₁ , X ₂ , ... X ₅ of the BP neural network in Figure 1 are composed of the height coordinates and signal strength of the receiving end, and the predicted output of the BP neural network is the horizontal distance between the mobile terminal and the corresponding LED.

The specific implementation of the positioning module is:

The traditional positioning technology based on intensity ranging usually adopts the trilateral positioning method, and the trilateral positioning method requires the information of three LED lights to be received at the same time. Once the user starts to move at high speed, it is extremely difficult for the mobile terminal to continuously receive the information of the three lights. Therefore, trilateral positioning is not enough to support the high-speed movement of the receiving end, which has great limitations.

Therefore, the present invention innovatively proposes a brand-new indoor positioning method based on image processing, which adopts the method of coordinate positioning, and only needs two LED lights to realize positioning. This positioning method is more flexible, makes good use of the advantages of camera communication, and can support high-speed movement of the receiving end under certain circumstances.

The positioning module can be divided into a coordinate positioning unit and a coordinate screening unit.

As shown in Figure 7, the coordinates of two LED lights A and B, and the horizontal distances d _A and d _B between the mobile terminal and A and B have been obtained through the BP neural network model. The coordinate positioning unit calculates coordinates by using a circle with A and B as centers and d _A and d _B as radii. The intersection points of the two circles are P ₁ and P ₂ , and the coordinates of the receiving end are one of P ₁ and P ₂ .

The coordinates of P ₁ and P ₂ can be obtained from the above binary quadratic equation.

Then the coordinate filtering unit needs to filter out the correct coordinates. Taking a smart phone as an example, the pictures taken by the front camera of the mobile phone are reversed left and right (the principle is like looking in a mirror), and the up and down remain unchanged. When arranging the coordinates of the light-emitting LEDs, the present invention stipulates that the vertical coordinate y of the LEDs increases sequentially from south to north, and the horizontal coordinate x of the LEDs increases sequentially from west to east. In one embodiment, the LEDs are arranged in a square, and adjacent LEDs must have the same abscissa and different ordinates; or the same ordinate but different abscissas.

In the process of positioning, it is assumed that the mobile receiving end is placed so that the front end of the mobile receiving end points to the north direction, which is reflected in the picture taken by the front camera as: if the receiving end is on the west side of the two light-emitting LEDs, that is, the receiving end When the abscissa is smaller than the abscissa of the two light-emitting LEDs, the two light-emitting LEDs in the picture taken by the camera are located in the left half of the picture; if the receiving end is on the east side of the two light-emitting LEDs, that is, the abscissa of the receiving end is greater than two When lighting LEDs, the two lighted LEDs in the picture are located in the right half of the picture. When the receiving end is on the north or south side of the two light-emitting LEDs, it has a similar principle.

According to this characteristic, the following positioning algorithm is summarized:

Establish a Cartesian coordinate system with the center of the picture as the origin, and each pixel is a unit, then the Cartesian coordinates of each point on the picture can be easily calculated. Due to the occurrence of rotation and other situations during the picture shooting process, it is more complicated to directly use the Cartesian coordinate system for calculation, and the Cartesian coordinate system can be converted into polar coordinates (ρ, θ) during specific implementation. The present invention uses θ in polar coordinates.

When the receiving end is at a certain position, assuming that the receiving end is on the side of LED light A and LED light B, the captured images may be different due to different rotation angles. Assume that FIG. 8 is a schematic diagram of the position of the receiving end and the LED light when the coordinates are normal, and FIG. 9 is a schematic diagram of the position of the receiving end and the LED light when the angular rotation occurs.

In this embodiment, it is stipulated that the included angle of LEDs with larger actual coordinates on the picture is θ ₁ , and the included angle of LEDs with smaller actual coordinates on the picture is θ ₂ . The coordinates of point P ₁ are (x ₁ , y ₁ ), and the coordinates of point P ₂ are (x ₂ , y ₂ ). Let Δθ=θ ₂ −θ ₁ . The coordinates of the receiving end are M(x,y).

If 0°<Δθ<180° or Δθ<-180°,

x=min(x ₁ , x ₂ )

y=min(y ₁ , y ₂ )

If Δθ>180° or -180°<Δθ<0°,

x=max(x ₁ , x ₂ )

y=max(y ₁ , y ₂ )

Combined with the height coordinates of the receiving end obtained in the second step, the coordinates of the receiving end can be obtained: (x, y, z).

As shown in Figure 4, the second embodiment of the indoor positioning system based on camera communication and image positioning technology, the transmitting end of the second embodiment is the same as the first embodiment; the receiving end includes a mobile terminal and a server, and the mobile terminal includes Image sensor, while UPSOOK demodulator, GeoHash decoder, altitude measurement module, BP neural network model and positioning module are arranged on the remote server. The functions of each module and the workflow of the positioning system are the same as those in the first embodiment.

In comparison, the first embodiment directly integrates the decoding, demodulation and image processing modules on the mobile terminal without transmitting the light signal image of the LED light collected by the image sensor to the server, which saves network overhead and can quickly realization of positioning.

In the second embodiment, since the main processing process is integrated on the server side, it has larger processing capacity and stronger processing capability than the mobile terminal, can handle a larger amount of data, and can simultaneously handle the positioning requirements of multiple mobile terminals. Other more complex positioning technology applications can be realized on this basis.

As shown in Figure 10, the indoor positioning method based on camera communication and image positioning technology includes:

S1, receiving a modulated optical signal with position coding information to generate image information;

S2. Extract the vertical distance from the receiving end to the light source, the position information of the light source, and the optical signal strength from the image information, and obtain the horizontal distance from the receiving end to the light source;

S3. Perform coordinate positioning and coordinate screening according to the position information of the two light sources in the image information and the horizontal distance of the receiving end from the light source, so as to locate the position coordinates of the receiving end.

The above method corresponds to the receiving end method of the indoor positioning system based on camera communication and image positioning technology. S1 is realized through the receiving end image sensor, and S2 is realized through the receiving end UPSOOK demodulator, GeoHash decoder, height measurement module and BP neural network model. S3 is realized through the positioning module of the receiving end.

The method at the transmitting end corresponding thereto is: modulate the LED light signal and then transmit it, which includes coded information on the position of the LED light.

Said S2 further includes:

S2.1, according to the distance between the two light sources in the image information and the physical focal length of the image sensor, obtain the vertical distance from the receiving end to the light source;

Realized by the altitude measurement module in the system of the present invention, the information needs to be demodulated by the UPSOOK demodulator first.

S2.2, performing corresponding demodulation on the optical signal to obtain the intensity of the optical signal, and further performing corresponding decoding to obtain the position information of the light source;

Realized by the GeoHash decoder in the system of the present invention.

S2.3, based on the BP neural network model, according to the vertical distance from the receiving end to the light source and the optical signal strength, obtain the horizontal distance of the receiving end from different light sources.

Realized by the BP neural network model in the system of the present invention.

Finally, the method of the present application is only a preferred embodiment, and is not intended to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims (10)

1. the indoor locating system of camera communication and framing technology is based on, it is characterised in that including transmitting terminal and receiving terminal,

The transmitting terminal, for by transmitting terminal position information with the optical signal after coded modulation to place spatial emission；

The receiving terminal, for receiving the optical signal, after carrying out decoding demodulation to the optical signal, carries out framing calculating Obtain positional information.

2. the system as claimed in claim 1, it is characterised in that the transmitting terminal includes light source, encoder and modulator,

The encoder, for carrying out coded optical signal to light source position information；

The modulator, for being modulated to the optical signal after coding；

The light source, for launching the optical signal after modulation.

3. the system as claimed in claim 1, it is characterised in that the receiving terminal includes：

Imageing sensor, for receiving optical signal and generating image information；

Demodulator, for carrying out the demodulation opposite with modulated process to described image information, to obtain transmitting terminal light signal strength；

Decoder, for carrying out the decoding opposite with cataloged procedure to the image information after demodulation, to obtain transmitting terminal light source Positional information；

Image processing module, for the positional information according to light signal strength and transmitting terminal light source, obtains receiving terminal range transmission The vertical range and horizontal range of light source are held, with the position coordinates at position receiver end.

4. system as claimed in claim 3, it is characterised in that described image processing module includes：

Elevation carrection module, for distance and the physics focal length of imageing sensor according to two light sources in image information, obtains Vertical range of the receiving terminal apart from light source；

BP neural network model, deviates for the vertical range according to receiving terminal apart from light source with light signal strength output receiving terminal The horizontal range of different light sources；

Locating module, the horizontal range for deviateing light source according to the positional information and receiving terminal of two light sources in image information is entered Row coordinate setting and coordinate are screened, with the position coordinates at position receiver end.

5. system as claimed in claim 4, it is characterised in that the vertical range is：

Wherein, z is vertical range of the receiving terminal apart from light source, and AB is that the distance between upper two light sources, f are pushed up in indoor positioning region It is the physics focal length of mobile terminal camera, CD is two light sources by the distance after mobile terminal camera head shooting on picture.

6. system as claimed in claim 4, it is characterised in that the input of the BP neural network model for one five tie up to Amount, is made up of receiving end signal intensity and receiving terminal height coordinate；It is output as the horizontal range of receiving terminal and corresponding light source；

The light that receiving terminal is received apart from the horizontal range and correspondence receiving terminal that the vertical range of light source, receiving terminal deviate light source Signal strength information trains BP neural network as a data set with these data sets；

BP neural network model after training is according to receiving terminal apart from the vertical range of light source and the optical signal for receiving of receiving terminal Intensity, obtains the horizontal range that receiving terminal deviates light source.

7. system as claimed in claim 4, it is characterised in that the coordinate location algorithm meets following formula：

(x-x_A)²+(y-y_A)²=d_A ²

(x-x_B)²+(y-y_B)²=d_B ²

Wherein, (x_A, y_A) it is A light source coordinates, d_AIt is receiving terminal apart from the horizontal range of A light sources, (x_B, y_B) it is B light source coordinates, d_B Be receiving terminal apart from the horizontal range of B light sources, (x, y) is receiving terminal coordinate, and the receiving terminal coordinate has two values, respectively P1 (x₁, y₁) and P2 (x₂, y₂)。

8. system as claimed in claim 7, it is characterised in that the coordinate screening is included from P1 (x₁, y₁) and P2 (x₂, y₂) in The correct coordinates of receiving terminal are filtered out, is met：

If 0 °<Δθ<180 ° or Δ θ<- 180 °,

X=min (x₁, x₂)

Y=min (y₁, y₂)

If Δ θ>180 ° or -180 °<Δθ<0 °,

X=max (x₁, x₂)

Y=max (y₁, y₂)

Wherein, Δ θ=θ₂-θ₁, θ₁It is the larger light source of coordinate on image and the angle of x-axis, θ₂It is the less light of coordinate on image Source and the angle of x-axis.

9. the indoor orientation method of camera communication and framing technology is based on, it is characterised in that including：

S1, receives with position encoded information and through the optical signal of ovennodulation, generates image information；

S2, extracts receiving terminal strong apart from the vertical range of light source, the positional information of light source and optical signal from described image information Degree, obtains horizontal range of the receiving terminal apart from light source；

S3, the horizontal range that positional information and receiving terminal according to two light sources in image information deviate light source carries out coordinate setting And coordinate screening, with the position coordinates at position receiver end.

10. method as claimed in claim 9, it is characterised in that the S2 is further included：

S2.1, the physics focal length of distance and imageing sensor according to two light sources in image information obtains receiving terminal apart from light The vertical range in source；

S2.2, corresponding demodulation is carried out to optical signal to obtain light signal strength, further carries out corresponding decoding to obtain light The positional information in source；

S2.3, based on BP neural network model, according to receiving terminal apart from the vertical range and light signal strength of light source, obtains and receives Deviate the horizontal range of different light sources in end.