A Noise Removal Algorithm for Road State Detection Device
By using multiple working currents in the road state detection device for linear regression and least squares calculation, ambient temperature noise interference is eliminated, and more accurate reflectivity calculation is achieved, which improves detection accuracy and driving safety.
Patent Information
- Application Number
- CN202110621149.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-06-03
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2041-06-03
AI Technical Summary
The existing road surface state detection device has inaccurate detection accuracy due to the influence of ambient temperature and has a large noise interference.
Multiple working current input light emitting diodes with linear proportional relationships are used to perform linear regression through mathematical statistical methods and least squares method to calculate the reflected light intensity coefficient, eliminate noise interference, and accurately calculate the reflectivity.
It improves the accuracy of road state detection, enhances driving safety, ensures that the calculated reflectivity is more accurate, and reduces the impact of temperature on the detection data.
Smart Images

Figure SMS_1 
Figure SMS_3 
Figure SMS_11
Abstract
Description
Technical Field
[0001] The present invention relates to the field of traffic road detection, and in particular to a noise removal algorithm for a road surface state detection device. Background Art
[0002] The existing detection devices for road conditions (such as dry road conditions, flooded road conditions, icy road conditions, snowy road conditions, sewage road conditions and sandy road conditions) are used in different ways. The inventor has developed a detection device that uses a light-emitting diode and a photodiode to achieve detection results. The principle is that the light-emitting diode emits light at a specific wavelength, and the emitted light is reflected to the photodiode by the covering on the road condition. The photodiode reflects the intensity of the reflected light by measuring the number of received photons of the reflected light of a specific wavelength. The intensity of the reflected light is used to calculate the reflectivity of the road condition at a specific wavelength, and finally a threshold is set according to the reflectivity to distinguish different road conditions. However, in actual application, the intensity of the emitted light is greatly affected by the ambient temperature. Therefore, the detection device using a light-emitting diode and a photodiode will cause the problem of inaccurate detection accuracy due to the influence of the ambient temperature.
[0003] To this end, the inventor conducted in-depth research on the above-mentioned issues, which resulted in the present case. Summary of the invention
[0004] An object of the present invention is to provide a noise removal algorithm for a road surface condition detection device that eliminates the influence of noise generated by ambient temperature changes on the detection of an optical sensor.
[0005] In order to achieve the above-mentioned purpose, the solution of the present invention is as follows: A noise removal algorithm for a road surface condition detection device is implemented by the following steps:
[0006] Step 1: Input multiple operating currents I with a linear proportional relationship to the light emitting diode at a specific wavelength. i (i=1, 2, ..., n), the photodiode obtains multiple reflected light intensities P at this specific wavelength i (i=1, 2, ..., n);
[0007] Step 2: Use mathematical statistics to calculate the multiple working currents I in step 1 i The corresponding multiple reflected light intensities P i Perform linear regression and obtain P i =β·I i +ε i (i=1,2,…, n), where the slope of the linear regression line β is recorded as the reflected light intensity coefficient under unit current, ε i is the total noise generated by the ambient temperature at unit current;
[0008] Step 3: Calculate the regression coefficient β using the least squares method. The calculation formula is as follows:
[0009]
[0010] in,
[0011] Step 4: Substituting the reflected light intensity coefficient β obtained in step 3 into the relational formula that the reflectivity is equal to the ratio of the reflected light intensity coefficient to the emitted light intensity coefficient can obtain the reflectivity.
[0012] After adopting the above technical scheme, a noise removal algorithm of a road surface condition detection device of the present invention uses the least squares method to remove noise, so that the calculated reflected light intensity coefficient is a true value without noise interference. When the reflected light intensity coefficient is used to calculate the reflectivity, the calculated reflectivity will be more accurate, thereby making the detection data of the road surface condition detection device more accurate, greatly improving driving safety. DETAILED DESCRIPTION
[0013] In order to further explain the technical solution of the present invention, the present invention is described in detail below through specific embodiments.
[0014] A noise removal algorithm for a road surface condition detection device is implemented by the following steps:
[0015] Step 1: Input multiple operating currents I with a linear proportional relationship to the light emitting diode at a specific wavelength. i (i=1,2,…,n), the photodiode operates at this specific wavelength according to multiple operating currents I i Correspondingly, multiple reflected light intensities P are obtained i (i=1, 2, ..., n); that is, n groups of sample observation values (I1, P1), (I2, P2), ..., (I n , P n );
[0016] Step 2: Use mathematical statistics to calculate the multiple working currents I in step 1 i The corresponding multiple reflected light intensities P i Perform linear regression and obtain the mathematical formula P i =β·I i +ε i (i = 1, 2, ..., n), where the slope of the linear regression line β is recorded as the reflected light intensity coefficient under unit current, ε i is the total noise generated by the ambient temperature at the unit current at that time;
[0017] Step 3: Calculate the regression coefficient β using the least squares method. The calculation formula is as follows:
[0018]
[0019] in,
[0020] Step 4: Substitute the reflected light intensity coefficient β obtained in step 3 into the relational formula that the reflectivity is equal to the ratio of the reflected light intensity coefficient to the emitted light intensity coefficient to obtain the reflectivity. The method for calculating the emitted light intensity coefficient in the present invention adopts the known technology. Specifically, according to the known reflectivity of white paper at a specific wavelength of the light emitted by the light emitting diode, the emitted light intensity of the light emitting diode can be calculated. In this way, when calculating the reflectivity of light under the current road surface conditions, it is only necessary to calculate the ratio of the current reflected light intensity system to the fixed outgoing light intensity through steps one to three.
[0021] A noise removal algorithm for a road surface condition detection device of the present invention uses the least squares method to remove noise, so that the calculated reflected light intensity coefficient is a true value without noise interference. When the reflected light intensity coefficient is used to calculate the reflectivity, the calculated reflectivity will be more accurate, thereby making the detection data of the road surface condition detection device more accurate, greatly improving driving safety.
[0022] In the present invention, the total noise ε generated by the ambient temperature at the unit current in step 1 includes the thermal noise ε1 inside the photoelectric element and the dark noise ε2 generated by the dark current. The thermal noise is generated with the random thermal motion of a large number of electrons, so this large number of random events has a Gaussian distribution probability density. The thermal noise is symmetrical and its value is compensatory. When the number of measurements is large enough, the arithmetic mean is close to 0, so the variance can be Dark noise is shot noise, which is also the result of a large number of independent random events, so the probability distribution of dark noise is also Gaussian distribution, with expectation denoted by μ and variance denoted by Since the expectation of the total noise ε is E(ε)=E(ε1)+E(ε2)=μ, the variance of the total noise is Thus, according to the mathematical structure of step 2, the expectation and variance of the reflected light intensity sample observation values obtained each time under n working currents are calculated, which are E(P i )=β·I i +μ(i=1,2,…,n), Use the least squares method in step 3 to calculate the estimate Seek expectation, that is
[0023]
[0024] From the calculation, we know that the estimated It is an unbiased estimate of β. After that, the data is changed many times and the estimated value of β is repeatedly calculated. There is no systematic tendency of overestimation or underestimation of the estimate. The average value will tend to β and is basically not affected by noise. Then the calculated β is a true value, so that the reflectivity calculated using β basically eliminates the influence of temperature on the experimental data.
[0025] The above embodiments do not limit the method of the present invention, and any appropriate changes or modifications made by ordinary technicians in the relevant technical field should be deemed to be within the patent scope of the present invention.
Claims
1. A noise removal algorithm for a road surface condition detection device, characterized in that : This is accomplished through the following steps: Step 1: Input multiple operating currents I with a linear proportional relationship to the light emitting diode at a specific wavelength. i (i=1, 2, ..., n), the photodiode obtains multiple reflected light intensities P at this specific wavelength i (i=1, 2, ..., n); Step 2: Use mathematical statistics to calculate the multiple working currents I in step 1 i The corresponding multiple reflected light intensities P i Perform linear regression and obtain P i =β·I i +ε i (i = 1, 2, ..., n), where the slope of the linear regression line β is recorded as the reflected light intensity coefficient under unit current, ε i is the total noise generated by the ambient temperature at unit current; Step 3: Calculate the reflected light intensity coefficient β using the least squares method. The calculation formula is as follows: in, Step 4: Substituting the reflected light intensity coefficient β obtained in step 3 into the relational formula that the reflectivity is equal to the ratio of the reflected light intensity coefficient to the emitted light intensity coefficient can obtain the reflectivity.
Citation Information
Patent Citations
Image processing apparatus, image reading apparatus, image forming apparatus and methods therefor
CN101206424A
Coal-rock identification optical fiber sensor
CN103149178A
Motor vehicle retro-reflector testing device
CN105954009A
Data Analysis Method Of Ellipsometer
KR102109540B1
Reflection characteristic measuring apparatus
US6088117A
