CN115950438B - Ship cabin visible light positioning method based on light intensity compensation - Google Patents
Ship cabin visible light positioning method based on light intensity compensation Download PDFInfo
- Publication number
- CN115950438B CN115950438B CN202310244070.0A CN202310244070A CN115950438B CN 115950438 B CN115950438 B CN 115950438B CN 202310244070 A CN202310244070 A CN 202310244070A CN 115950438 B CN115950438 B CN 115950438B
- Authority
- CN
- China
- Prior art keywords
- light intensity
- data
- compensation
- intensity data
- visible light
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Abstract
A ship cabin visible light positioning method based on light intensity compensation comprises the following steps: detecting light intensity data when receiving ends of all coordinate points of a ship cabin are vertically upwards when the ship is stopped, and collecting the detected light intensity data, attitude angle information and real position coordinate information; constructing a light intensity compensation model; constructing a convolutional neural network visible light positioning model based on light intensity compensation; the method comprises the steps that a target to be positioned is conveyed into a ship cabin, and a receiving end collects and stores detection light intensity data, target attitude angle information and target height information of each LED lamp; and inputting the target attitude angle information and the target height information into a light intensity compensation model to obtain corrected received signal intensity data, inputting the corrected received signal intensity data into a trained convolutional neural network visible light positioning model in a matrix form, and calculating an estimated position coordinate by using a BP algorithm. The invention not only improves the indoor positioning precision of the ship, but also reduces the complexity of the positioning model and improves the training efficiency.
Description
Technical Field
The invention relates to the technical field of ship cabin positioning, in particular to a ship cabin visible light positioning method based on light intensity compensation, which is mainly suitable for improving the indoor positioning precision of a ship.
Background
Indoor visible light positioning is considered to be one of the currently potential indoor positioning technology hot spots, such as an LED positioning system deployed in a large-scale Li supermarket in cooperation with Jiale Fu in Philippine, and a visible light positioning system deployed in a small commodity market in Changzhou city in Jiangsu province in China in Hua Ceguang communication, wherein positioning accuracy of decimeter level can be achieved by the two systems. Compared with other positioning technologies, the indoor visible light positioning technology has higher precision, lower positioning cost and less influence by multipath effects; and the frequency band of visible light does not conflict with other communication frequency bands, and electromagnetic interference is not generated. As a positioning technology which can simultaneously have higher system precision, lower power consumption and lower cost, the method has high research value and application prospect when being applied to the interior of a ship.
The ship is greatly different from the general indoor environment, besides the complex cabin structure and the time variability of indoor environment light, the ship sailing is a non-negligible factor. In the sailing process of the ship, the course, the speed and the movement characteristics of the ship are different, and the factors can lead to the severe fluctuation of the receiving intensity value. Therefore, how to solve the problem of time-varying of the indoor environment light of the ship and overcome the influence of the indoor inclination shaking characteristic of the ship on fingerprint positioning is a problem to be solved urgently. Compared to the terrestrial indoor environment, the marine environment is in dynamic time variation, and the ambient light changes with time. The ship is often in a swaying state in the running process, so that the gesture of the receiving end is changed when the light intensity is detected, and the difference between the measured receiving light intensity and the receiving light intensity of the receiving end in a vertically upward state is too large. The specific problems are as follows:
(1) The indoor visible light positioning of the ship is easy to be interfered by ambient light
Most indoor VLP localization methods directly ignore the effects of interference factors such as ambient light. Compared with the land indoor environment, the ship indoor environment light is in a dynamic time-varying state, and has a great influence on the accuracy of detecting the light intensity.
(2) Inaccurate detection of light intensity of edge area of ship indoor light source
Compared with the land environment, the light intensity detection of the edge area of the ship indoor light source is more inaccurate. The fingerprint positioning method is used in the ship indoor environment, for example, the light intensity detection of the corner position or the central area is inaccurate, the positioning result is not ideal, and the database establishment workload is large.
(3) The detection precision of the indoor light intensity of the ship is greatly influenced by the attitude of the receiver
In the receiving-end-based VLP localization method, the detected light intensity is highly correlated with the pose of the receiving end. Most indoor VLP positioning methods default that a receiving end is vertically upwards when receiving signals, but in practical experiments, the change of the self-posture of the receiving end is found to have great influence on light intensity detection. Compared with the land indoor environment, the ship in the running process is in a shaking state, the gesture change amplitude of the receiving end is larger and more frequent, and the influence on the light intensity detection is larger.
Disclosure of Invention
The invention aims to overcome the defect and problem of low positioning precision in the prior art and provides a ship cabin visible light positioning method with high positioning precision based on light intensity compensation.
In order to achieve the above object, the technical solution of the present invention is: a ship cabin visible light positioning method based on light intensity compensation comprises the following steps:
s1, detecting light intensity data when receiving ends of all coordinate points of a ship cabin are vertically upwards when the ship is berthed, and collecting detected light intensity data, attitude angle information and real position coordinate information;
s2, constructing a light intensity compensation model, wherein the light intensity compensation model comprises an attitude compensation model and a height compensation model;
s3, constructing a convolutional neural network visible light positioning model based on light intensity compensation;
s4, conveying the target to be positioned into a ship cabin, transmitting the target to be positioned into a position coordinate of a receiving end through a visible light communication system, and collecting and storing detection light intensity data, target attitude angle information and target height information of each LED lamp by the receiving end;
s5, inputting the attitude angle information and the target height information of the target into a light intensity compensation model so as to enable the target to be positioned to obtain corrected received signal intensity data;
s6, inputting the corrected received signal strength data into a trained convolutional neural network visible light positioning model in a matrix form, and calculating an estimated position coordinate by using a BP algorithm.
In step S1, the detected light intensity data is preprocessed, which specifically includes the following steps:
performing wavelet decomposition on the detected light intensity data, setting a threshold value, processing wavelet coefficients obtained by decomposition, and performing wavelet reconstruction on the detected light intensity data;
in the method, in the process of the invention,for detecting the light intensity data>For the received signal strength, < >>For noise intensity->Is noise;
average filtering is carried out on the detected light intensity data subjected to noise elimination treatment, and the treated received signal intensity data is obtained;
in the method, in the process of the invention,the first part of the LED lamps is at the same coordinate point>The detected light intensity data subjected to noise elimination processing is adopted next time,is the average value.
In step S2, determining whether there is a light signal deficiency according to measuring the detected light intensity value in a certain plane, then selecting the plane without the signal deficiency area as a standard plane, collecting the detected light intensity value at different heights through each coordinate point of the same plane, fitting out a signal intensity enhancement empirical formula, and constructing a height compensation model:
in the method, in the process of the invention,for the received signal strength after high compensation, < >>For detecting the light intensity +.>Is of standard plane height +.>For the height of the object to be positioned, +.>Is a signal enhancement factor.
In the step S2, the same position coordinates acquire multiple detection light intensity data and attitude angle data, and an attitude compensation empirical formula is fitted according to the change angle values of the pitch angle and the roll angle and the light intensity change value, so as to construct an attitude compensation model:
in the method, in the process of the invention,for the received signal strength after posture compensation, +.>In order for the receiving end to detect the pitch angle when the light intensity is detected,for the roll angle when the receiving end detects the light intensity, < >>And->The compensation factors of pitch angle and roll angle are respectively. />
In step S5, the receiver corrects the received signal strengthThe method comprises the following steps:
the step S3 specifically comprises the following steps:
s31, constructing a data set by utilizing the received signal strength data corrected by the coordinate point receiving end;
assuming the presence of a locating areaPersonal coordinate points, < >>The light sources, the receiving end receives the detection light intensity of each light source as +.>Thus, the->From the coordinate points>Detection light intensity of individual light sources->The method comprises the following steps:
generating virtual data by using a cubic spline interpolation method according to the corrected received signal intensity data acquired by each coordinate point, reducing the step length between the coordinate points, and acquiring each coordinate pointThe detected light intensities or virtual data of the light sources are arranged into a matrix form to be used as the detected light intensity data of the coordinate points; the corrected received signal strength data and the generated virtual data of each coordinate point construct a complete database, and the input form is +.>First, cross and second order terms of the value;
s32, convolutional neural network model selection and parameter setting;
the convolutional neural network model adopts a nine-layer structure and comprises an input layer, two convolutional layers, two maximum pooling layers, three full-connection layers and an output layer;
convolutional layer numberThe filter outputs +.>And maximum pooling layer->Output of individual area->The following is shown:
in the method, in the process of the invention,for the input of the convolution layer, < > for>And->Is->Weights and offsets of the individual filters, +.>Is->The outputs of the filters,/>Is->The outputs of the filters;
s33, collectingThe value is used as fingerprint data to train the visible light positioning model parameters of the convolutional neural network;
the convolutional neural network visible light positioning model learns all input data, outputs a result, calculates a loss error according to the output result by comparing with a real position coordinate, and finally updates parameters by using a BP algorithm;
calculating an error between the output result and the actual value and judging whether iteration times are reached or not; if the specified target is reached, the training is exited; if the specified target is not reached, back propagation is carried out, propagation errors among the neural network connection layers are calculated, and the error gradient is solved for updating the network neuron weight until the set error or iteration number is reached.
In step S32, each convolution layer is followed by using a ReLU function as an activation function, and zero padding is used to ensure that the output and input are the same size; feature extraction is performed twice using a combination of a convolution layer and a max pooling layer; batch normalization is added after the convolutional layer output.
In step S32, for both the maximum pooling levels, the pooling level size and step size are 3×3 and 1, and the output channels are 8 and 32, respectively.
In step S32, after the second maximum pooling layer, the outputs of the plurality of filters are integrated into a row of matrix form and input to the two fully connected layers, the number of neurons being 16 and 8, respectively; the output of each full connection layer passes through a ReLU activation function; the output layer has two neurons representing the estimated coordinates.
In step S4, the detected light intensity data is preprocessed, which specifically includes the following steps:
performing wavelet decomposition on the detected light intensity data, setting a threshold value, processing wavelet coefficients obtained by decomposition, and performing wavelet reconstruction on the detected light intensity data;
in the method, in the process of the invention,for detecting the light intensity data>For the received signal strength, < >>For noise intensity->Is noise;
average filtering is carried out on the detected light intensity data subjected to noise elimination treatment, and the treated received signal intensity data is obtained;
in the method, in the process of the invention,the first part of the LED lamps is at the same coordinate point>The detected light intensity data subjected to noise elimination processing is adopted next time,is the average value.
Compared with the prior art, the invention has the beneficial effects that:
according to the ship cabin visible light positioning method based on light intensity compensation, the wavelet noise reduction method and the mean value filtering method are combined to reduce the influence of interference factors such as ambient light on the detected light intensity, aiming at the problem that the ship cabin visible light positioning is easily interfered by ambient light and the ship cabin ambient light has time variability; meanwhile, aiming at the problem that the light intensity detection precision in the ship cabin is greatly influenced by the attitude of the receiver, a light intensity compensation model is provided, the detected light intensity is compensated according to the height information and the attitude angle information, the ideal detected light intensity is obtained, the average positioning precision in the ship cabin is improved, and the positioning performance is improved; in addition, the convolutional neural network model is provided with a convolutional layer for feature extraction, so that the convolutional neural network model can provide advantages of obtaining representative features for model consideration, the total number of weight parameters used by the convolutional neural network model is smaller than that used by the artificial neural network model, the model complexity is reduced, and the training efficiency is improved.
Drawings
FIG. 1 is a flow chart of a method for positioning visible light of a ship cabin based on light intensity compensation.
Fig. 2 is a view of a ship cabin positioning scene in the present invention.
Detailed Description
The invention is described in further detail below with reference to the accompanying drawings and detailed description.
Referring to fig. 1, a method for positioning visible light of a ship cabin based on light intensity compensation includes an off-line training stage and an on-line positioning stage, and specifically includes the following steps:
s1, detecting light intensity data when receiving ends of all coordinate points of a ship cabin are vertically upwards when the ship is berthed, and collecting detected light intensity data, attitude angle information and real position coordinate information;
preprocessing the detected light intensity data, specifically comprising the following steps:
performing wavelet decomposition on the detected light intensity data, setting a threshold value, processing wavelet coefficients obtained by decomposition, and performing wavelet reconstruction on the detected light intensity data;
step 1: wavelet decomposition of a one-dimensional signal. A wavelet is selected and the level of decomposition is determined, and then a decomposition calculation is performed.
Step 2: threshold quantization of wavelet decomposed high frequency coefficients. And selecting a threshold value for the high-frequency coefficient under each decomposition scale to carry out soft threshold quantization processing.
Step 3: and reconstructing a one-dimensional wavelet. And carrying out one-dimensional wavelet reconstruction according to the lowest-layer low-frequency coefficient and each layer high-frequency coefficient of wavelet decomposition.
In the method, in the process of the invention,for detecting the light intensity data>For the received signal strength, < >>Is the noise intensity; />As noise, ambient light has time variability and thus can be considered gaussian white noise;
average filtering is carried out on the detected light intensity data subjected to noise elimination treatment, and the treated received signal intensity data is obtained;
in the method, in the process of the invention,the first part of the LED lamps is at the same coordinate point>The detected light intensity data subjected to noise elimination processing is adopted next time,is the average value.
Aiming at the time-varying property of the ambient light and the tilting and shaking property of the ship, a wavelet denoising method and an average filtering method are used for reducing the influence of interference factors such as the ambient light on the detection light intensity.
S2, constructing a light intensity compensation model, wherein the light intensity compensation model comprises an attitude compensation model and a height compensation model;
firstly, highly compensating the light intensity, then carrying out attitude compensation on the light intensity, and finally obtaining the corrected received signal intensity;
and (3) judging whether optical signal loss exists or not according to the measured and detected light intensity value in a certain plane, then selecting the plane without the signal loss area as a standard plane, acquiring and detecting the light intensity value at different heights through all coordinate points of the same plane, fitting a signal intensity enhancement empirical formula, and constructing a height compensation model:
in the method, in the process of the invention,for the received signal strength after high compensation, < >>For detecting the light intensity +.>Is of standard plane height +.>For the height of the object to be positioned, +.>Is a signal enhancement factor.
Referring to fig. 2, the same position coordinates collect multiple detection light intensity data and attitude angle data, and fit an attitude compensation empirical formula according to the change angle values of the pitch angle and the roll angle and the light intensity change value to construct an attitude compensation model:
in the method, in the process of the invention,for the received signal strength after posture compensation, +.>In order for the receiving end to detect the pitch angle when the light intensity is detected,for the roll angle when the receiving end detects the light intensity, < >>And->The compensation factors of pitch angle and roll angle are respectively.
and correcting the detected light intensity according to the light intensity compensation model, constructing a fingerprint positioning library of the target position-signal characteristic relation, and estimating the target position in a characteristic recognition and matching mode.
S3, constructing a Convolutional Neural Network (CNN) visible light positioning model based on light intensity compensation; the method specifically comprises the following steps:
s31, constructing a data set by utilizing the received signal strength data corrected by the coordinate point receiving end;
assuming the presence of a locating areaPersonal coordinate points, < >>The light sources, the receiving end receives the detection light intensity of each light source as +.>Thus, the->From the coordinate points>Detection light intensity of individual light sources->The method comprises the following steps: />
Generating virtual data by using a cubic spline interpolation method according to the corrected received signal intensity data acquired by each coordinate point, reducing the step length between the coordinate points, and acquiring each coordinate pointThe detected light intensities or virtual data of the light sources are arranged into a matrix form to be used as the detected light intensity data of the coordinate points; the corrected received signal strength data and the generated virtual data of each coordinate point construct a complete database, and the input form is +.>First, cross and second order terms of the value;
typically, 4 light sources are disposed in the positioning area, so that the target to be positioned receives the detected light intensities from the 4 light sourcesThe method comprises the following steps:
the first order, cross and second order terms of the values are integrated in the form of a matrix as input +.>:
S32, convolutional neural network model selection and parameter setting;
the convolutional neural network model adopts a nine-layer structure and comprises an input layer, two convolutional layers, two maximum pooling layers, three full-connection layers and an output layer;
each convolution layer is followed by a ReLU function as an activation function, zero padding is used to ensure that the output and input are the same size; feature extraction is performed twice using a combination of a convolution layer and a max pooling layer; batch standardization is added after the output of the convolution layer so as to accelerate the convergence speed and prevent the overfitting;
for the two largest pooling layers, the pooling layer sizes and step sizes are 3×3 and 1, and the output channels are 8 and 32, respectively;
after the second maximum pooling layer, integrating the outputs of the plurality of filters into a row of matrix form for input to two fully connected layers, the number of neurons being 16 and 8, respectively; the output of each full connection layer passes through a ReLU activation function; the output layer is provided with two neurons, and the two neurons represent estimated coordinates; setting the rate of the dropout layer to 0.2;
convolutional layer numberThe filter outputs +.>And maximum pooling layer->Output of individual area->The following is shown:
in the method, in the process of the invention,for the input of the convolution layer, < > for>And->Is->Weights and offsets of the individual filters, +.>Is->The outputs of the filters,/>Is->The outputs of the filters;
s33, collectingThe value is used as fingerprint data to train the visible light positioning model parameters of the convolutional neural network;
the convolutional neural network visible light positioning model learns all input data, outputs a result, compares real position coordinates according to the output result to calculate a loss error, and finally performs parameter updating by using a BP algorithm, so that after each round of completion, parameters of the convolutional neural network can be optimized to a certain extent;
calculating an error between the output result and the actual value and judging whether iteration times are reached or not; if the specified target is reached, the training is exited; if the specified target is not reached, back propagation is carried out, propagation errors among the neural network connection layers are calculated, and the error gradient is solved for updating the network neuron weight until the set error or iteration number is reached.
S4, conveying the target to be positioned into a ship cabin, transmitting the target to be positioned into a position coordinate of a receiving end through a visible light communication system, and collecting and storing detection light intensity data, target attitude angle information and target height information of each LED lamp by the receiving end;
preprocessing the detected light intensity data, specifically comprising the following steps:
performing wavelet decomposition on the detected light intensity data, setting a threshold value, processing wavelet coefficients obtained by decomposition, and performing wavelet reconstruction on the detected light intensity data;
in the method, in the process of the invention,for detecting the light intensity data>For the received signal strength, < >>Is the noise intensity; />As noise, ambient light has time variability and thus can be considered gaussian white noise;
average filtering is carried out on the detected light intensity data subjected to noise elimination treatment, and the treated received signal intensity data is obtained;
in the method, in the process of the invention,the first part of the LED lamps is at the same coordinate point>The detected light intensity data subjected to noise elimination processing is adopted next time,is the average value.
Aiming at the time-varying property of the ambient light and the tilting and shaking property of the ship, a wavelet denoising method and an average filtering method are used for reducing the influence of interference factors such as the ambient light on the detection light intensity.
S5, inputting the attitude angle information and the target height information of the target into a light intensity compensation model so as to enable the target to be positioned to obtain corrected received signal intensity data;
s6, inputting the corrected received signal strength data into a trained convolutional neural network visible light positioning model in a matrix form, and calculating an estimated position coordinate by using a BP algorithm.
The visible light positioning system hardware is based on the original indoor lighting facilities and the traditional visible light positioning system, and only the inclination sensor equipment is needed to be added for measuring the attitude angle, so that the visible light positioning system hardware is simple and easy to operate.
Claims (10)
1. The ship cabin visible light positioning method based on light intensity compensation is characterized by comprising the following steps of:
s1, detecting light intensity data when receiving ends of all coordinate points of a ship cabin are vertically upwards when the ship is berthed, and collecting detected light intensity data, attitude angle information and real position coordinate information;
s2, constructing a light intensity compensation model, wherein the light intensity compensation model comprises an attitude compensation model and a height compensation model;
s3, constructing a convolutional neural network visible light positioning model based on light intensity compensation;
s4, conveying the target to be positioned into a ship cabin, transmitting the target to be positioned into a position coordinate of a receiving end through a visible light communication system, and collecting and storing detection light intensity data, target attitude angle information and target height information of each LED lamp by the receiving end;
s5, inputting the attitude angle information and the target height information of the target into a light intensity compensation model so as to enable the target to be positioned to obtain corrected received signal intensity data;
s6, inputting the corrected received signal strength data into a trained convolutional neural network visible light positioning model in a matrix form, and calculating an estimated position coordinate by using a BP algorithm.
2. The ship cabin visible light positioning method based on light intensity compensation according to claim 1, wherein the method comprises the following steps: in step S1, the detected light intensity data is preprocessed, which specifically includes the following steps:
performing wavelet decomposition on the detected light intensity data, setting a threshold value, processing wavelet coefficients obtained by decomposition, and performing wavelet reconstruction on the detected light intensity data;
in the method, in the process of the invention,for detecting the light intensity data>For the received signal strength, < >>For noise intensity->Is noise;
average filtering is carried out on the detected light intensity data subjected to noise elimination treatment, and the treated received signal intensity data is obtained;
3. The ship cabin visible light positioning method based on light intensity compensation according to claim 1, wherein the method comprises the following steps:
in step S2, determining whether there is a light signal deficiency according to measuring the detected light intensity value in a certain plane, then selecting the plane without the signal deficiency area as a standard plane, collecting the detected light intensity value at different heights through each coordinate point of the same plane, fitting out a signal intensity enhancement empirical formula, and constructing a height compensation model:
4. A method for locating the visible light of a ship cabin based on light intensity compensation according to claim 3, wherein:
in the step S2, the same position coordinates acquire multiple detection light intensity data and attitude angle data, and an attitude compensation empirical formula is fitted according to the change angle values of the pitch angle and the roll angle and the light intensity change value, so as to construct an attitude compensation model:
in the method, in the process of the invention,for the received signal strength after posture compensation, +.>For the pitch angle of the receiving end when detecting the light intensity, +.>For the roll angle when the receiving end detects the light intensity, < >>And->The compensation factors of pitch angle and roll angle are respectively. />
6. the ship cabin visible light positioning method based on light intensity compensation according to claim 1, wherein the method comprises the following steps: the step S3 specifically comprises the following steps:
s31, constructing a data set by utilizing the received signal strength data corrected by the coordinate point receiving end;
assuming the presence of a locating areaPersonal coordinate points, < >>The light sources, the receiving end receives the detection light intensity of each light source as +.>Thus, the->From the coordinate points>Detection light intensity of individual light sources->The method comprises the following steps:
generating virtual data by using a cubic spline interpolation method according to the corrected received signal intensity data acquired by each coordinate point, reducing the step length between the coordinate points, and acquiring each coordinate pointThe detected light intensities or virtual data of the light sources are arranged into a matrix form to be used as the detected light intensity data of the coordinate points; the corrected received signal strength data and the generated virtual data of each coordinate point construct a complete database, and the input form is +.>First, cross and second order terms of the value;
s32, convolutional neural network model selection and parameter setting;
the convolutional neural network model adopts a nine-layer structure and comprises an input layer, two convolutional layers, two maximum pooling layers, three full-connection layers and an output layer;
convolutional layer numberThe filter outputs +.>And maximum pooling layer->Output of individual area->The following is shown:
in the method, in the process of the invention,for the input of the convolution layer, < > for>And->Is->Weights and offsets of the individual filters, +.>Is->The outputs of the filters,/>Is->The outputs of the filters;
s33, collectingThe value is used as fingerprint data to train the visible light positioning model parameters of the convolutional neural network;
the convolutional neural network visible light positioning model learns all input data, outputs a result, calculates a loss error according to the output result by comparing with a real position coordinate, and finally updates parameters by using a BP algorithm;
calculating an error between the output result and the actual value and judging whether iteration times are reached or not; if the specified target is reached, the training is exited; if the specified target is not reached, back propagation is carried out, propagation errors among the neural network connection layers are calculated, and the error gradient is solved for updating the network neuron weight until the set error or iteration number is reached.
7. The ship cabin visible light positioning method based on light intensity compensation according to claim 6, wherein the method comprises the following steps: in step S32, each convolution layer is followed by using a ReLU function as an activation function, and zero padding is used to ensure that the output and input are the same size; feature extraction is performed twice using a combination of a convolution layer and a max pooling layer; batch normalization is added after the convolutional layer output.
8. The ship cabin visible light positioning method based on light intensity compensation according to claim 6, wherein the method comprises the following steps: in step S32, for both the maximum pooling levels, the pooling level size and step size are 3×3 and 1, and the output channels are 8 and 32, respectively.
9. The ship cabin visible light positioning method based on light intensity compensation according to claim 6, wherein the method comprises the following steps: in step S32, after the second maximum pooling layer, the outputs of the plurality of filters are integrated into a row of matrix form and input to the two fully connected layers, the number of neurons being 16 and 8, respectively; the output of each full connection layer passes through a ReLU activation function; the output layer has two neurons representing the estimated coordinates.
10. The ship cabin visible light positioning method based on light intensity compensation according to claim 1, wherein the method comprises the following steps: in step S4, the detected light intensity data is preprocessed, which specifically includes the following steps:
performing wavelet decomposition on the detected light intensity data, setting a threshold value, processing wavelet coefficients obtained by decomposition, and performing wavelet reconstruction on the detected light intensity data;
in the method, in the process of the invention,for detecting the light intensity data>For the received signal strength, < >>For noise intensity->Is noise;
average filtering is carried out on the detected light intensity data subjected to noise elimination treatment, and the treated received signal intensity data is obtained;
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310244070.0A CN115950438B (en) | 2023-03-15 | 2023-03-15 | Ship cabin visible light positioning method based on light intensity compensation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310244070.0A CN115950438B (en) | 2023-03-15 | 2023-03-15 | Ship cabin visible light positioning method based on light intensity compensation |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115950438A CN115950438A (en) | 2023-04-11 |
CN115950438B true CN115950438B (en) | 2023-05-30 |
Family
ID=85891416
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310244070.0A Active CN115950438B (en) | 2023-03-15 | 2023-03-15 | Ship cabin visible light positioning method based on light intensity compensation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115950438B (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2010197049A (en) * | 2009-02-20 | 2010-09-09 | Nippon Telegr & Teleph Corp <Ntt> | System for estimating position and device for measuring reference data |
CN104777475A (en) * | 2015-04-16 | 2015-07-15 | 北京理工大学 | Indoor visible light intensity positioning system with error correction function |
CN108732537A (en) * | 2018-05-08 | 2018-11-02 | 北京理工大学 | A kind of indoor visible light localization method based on neural network and received signal strength |
WO2019020200A1 (en) * | 2017-07-28 | 2019-01-31 | Fundació Privada I2Cat, Internet I Innovació Digital A Catalunya | Method and apparatus for accurate real-time visible light positioning |
CN208874572U (en) * | 2018-11-01 | 2019-05-17 | 上海奕行信息科技有限公司 | Communication and positioning system based on visible light wireless communication ship |
CN112423265A (en) * | 2020-11-13 | 2021-02-26 | 武汉理工大学 | CSI-based dual-cycle neural network shipborne environment indoor positioning method |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2012093241A1 (en) * | 2011-01-07 | 2012-07-12 | Toshiba Research Europe Limited | Localisation of electronic equipment |
CN104297726B (en) * | 2014-10-23 | 2017-02-22 | 东南大学 | Multi-receiving-point geometrical center locating method for visible light communication |
US10333620B2 (en) * | 2016-02-11 | 2019-06-25 | Morgan State University | System and method for lighting and building occupant tracking |
CN110261823B (en) * | 2019-05-24 | 2022-08-05 | 南京航空航天大学 | Visible light indoor communication positioning method and system based on single LED lamp |
-
2023
- 2023-03-15 CN CN202310244070.0A patent/CN115950438B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2010197049A (en) * | 2009-02-20 | 2010-09-09 | Nippon Telegr & Teleph Corp <Ntt> | System for estimating position and device for measuring reference data |
CN104777475A (en) * | 2015-04-16 | 2015-07-15 | 北京理工大学 | Indoor visible light intensity positioning system with error correction function |
WO2019020200A1 (en) * | 2017-07-28 | 2019-01-31 | Fundació Privada I2Cat, Internet I Innovació Digital A Catalunya | Method and apparatus for accurate real-time visible light positioning |
CN108732537A (en) * | 2018-05-08 | 2018-11-02 | 北京理工大学 | A kind of indoor visible light localization method based on neural network and received signal strength |
CN208874572U (en) * | 2018-11-01 | 2019-05-17 | 上海奕行信息科技有限公司 | Communication and positioning system based on visible light wireless communication ship |
CN112423265A (en) * | 2020-11-13 | 2021-02-26 | 武汉理工大学 | CSI-based dual-cycle neural network shipborne environment indoor positioning method |
Non-Patent Citations (3)
Title |
---|
Using Data Pre-Processing and Convolutional Neural Network (CNN) to Mitigate Light Deficient Regions in Visible Light Positioning (VLP) Systems;Li-Sheng Hsu;JOURNAL OF LIGHTWAVE TECHNOLOGY;第40卷(第17期);5894-5900 * |
基于卷积神经网络的室内可见光指纹定位方法;许浩等;激光与光电子学进展;第58卷(第17期);1706008-1—1706008-8 * |
融合角度特征的CNN可见光室内定位算法;杨薇等;激光杂志;第43卷(第1期);113-118 * |
Also Published As
Publication number | Publication date |
---|---|
CN115950438A (en) | 2023-04-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN116363133B (en) | Illuminator accessory defect detection method based on machine vision | |
CN111192237B (en) | Deep learning-based glue spreading detection system and method | |
CN107704807B (en) | Dynamic monitoring method based on multi-source remote sensing sequence image | |
CN109635806B (en) | Ammeter value identification method based on residual error network | |
US20140046169A1 (en) | Methods, systems, and devices for spine centrum extraction and intervertebral disk dividing | |
CN113313641A (en) | CT image denoising method with self-adaptive median filtering | |
CN110766058B (en) | Battlefield target detection method based on optimized RPN (resilient packet network) | |
CN109085469A (en) | A kind of method and system of the signal type of the signal of cable local discharge for identification | |
CN108765327A (en) | A kind of image rain removing method based on the depth of field and sparse coding | |
CN116413395B (en) | Intelligent detection method for environmental abnormality | |
CN111189837A (en) | Cigarette appearance online detection method and device | |
CN108596909A (en) | Hollow cylinder Surface Flaw Detection system and method | |
CN111351801A (en) | Wire and cable defect detection system | |
CN117252882B (en) | Cylinder head quality detection method and system | |
CN116541732B (en) | Meteorological monitoring system based on ultrasonic data and optimization algorithm | |
CN108830814A (en) | A kind of relative radiometric correction method of remote sensing image | |
CN112749675A (en) | Potato disease identification method based on convolutional neural network | |
CN111241991B (en) | Water quality classification method based on remote sensing image | |
CN115950438B (en) | Ship cabin visible light positioning method based on light intensity compensation | |
CN115685102A (en) | Target tracking-based radar vision automatic calibration method | |
CN116009065A (en) | Seismic data noise removing method, electronic equipment and medium | |
CN109166127A (en) | A kind of wearable plant phenotype sensory perceptual system | |
CN111089586B (en) | All-day star sensor star point extraction method based on multi-frame accumulation algorithm | |
CN115359258B (en) | Weak and small target detection method and system for component uncertainty measurement | |
CN116843618A (en) | Method for detecting shallow apparent appearance defects of metal parts |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |