Laser ranging method for a phase-based laser ranging system based on negative feedback regulation

Abstract

The present invention discloses a laser ranging method based on a negative feedback regulated phase laser ranging system, specifically relating to the technical field of laser ranging. The present invention uses a phase laser ranging system to obtain original measurement data to be predicted, and performs preprocessing, the preprocessing method being the same as that in step S1. Then, it is loaded into the phase difference prediction neural network model for one forward propagation to obtain a phase difference value, and then an estimated distance is calculated based on the phase difference value. Then, the calibration error obtained by using the negative feedback regulated phase laser ranging system is used to correct the estimated distance to obtain the final measured distance. The final measured distance obtained by the method of the present application can have better accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of laser ranging, and in particular, to a laser ranging method based on a negative feedback regulated phase laser ranging system. Background Art

[0002] The phase laser ranging method is a commonly used technology for distance measurement based on the phase difference of electromagnetic waves. Its core principle is to use the phase difference to determine the propagation time difference between the transmitted wave and the reflected wave, so as to calculate the distance between the object and the measuring device. In the prior art, the digital phase method has limited performance in processing non-linear signals. Especially in a high-noise environment, additional processing steps such as noise filtering and signal enhancement are often required to improve the measurement accuracy. Therefore, researchers have started to apply neural networks to laser ranging. However, the existing methods for laser ranging based on neural networks generally have low accuracy when performing laser ranging in a noisy environment. For this reason, the present application proposes a laser ranging method based on a negative feedback regulated phase laser ranging system with better measurement accuracy. Summary of the Invention

[0003] Based on the above problems, the present invention provides a laser ranging method based on a negative feedback regulated phase laser ranging system.

[0004] The present invention is implemented by the following technical solutions:

[0005] A laser ranging method based on a negative feedback regulated phase laser ranging system includes the following steps:

[0006] S1. Use a negative feedback regulated phase laser ranging system to obtain original measurement data, and perform preprocessing to obtain a phase difference sample, and then obtain a training set and a test set based on the phase difference sample;

[0007] S2. Construct a phase difference prediction neural network for predicting the phase difference; the phase difference prediction neural network includes an input layer, a first convolutional layer, a first normalization layer, a first pooling layer, a second convolutional layer, a second normalization layer, a first LeakyRelu layer, a second pooling layer, a third convolutional layer, a third normalization layer, a second LeakyRelu layer, a first dropout layer, a flattening layer, a first fully connected layer, a fourth normalization layer, a third LeakyRelu layer, a second dropout layer, a second fully connected layer, a fifth normalization layer, a fourth LeakyRelu layer, a third dropout layer, and an output layer connected in sequence;

[0008] S3. Under the guidance of the existing mean square error loss function, use the training set to train the phase difference prediction neural network to obtain a phase difference prediction neural network model;

[0009] S4. Use a phase-type laser ranging system to obtain the original measurement data to be predicted, then preprocess the original measurement data to be predicted in the same way as the preprocessing of the original measurement data in step S1, and then load it into the phase difference prediction neural network model obtained in step S3 for one forward propagation to obtain the phase difference value, and then calculate the estimated distance based on the phase difference value; then, use the calibration error obtained by negative feedback adjustment of the phase-type laser ranging system to correct the estimated distance to obtain the final measurement distance.

[0010] Preferably, in step S1, the negative feedback adjustment phase-type laser ranging system includes a control unit, the control unit is connected to a modulation unit, and the modulation unit is respectively connected to a laser driving unit and an APD driving circuit; the laser driving unit is connected to a laser diode, and the laser diode is respectively connected to a collimating lens and an optical fiber delay line; the APD driving circuit is respectively connected to a first APD detector, a second APD detector, and a third APD detector, the optical fiber delay line is connected to the first APD detector, the present application further includes a converging lens, the converging lens is connected to a beam splitter, and the beam splitter is respectively connected to the second APD detector and the third APD detector. The output ends of the first APD detector and the second APD detector are both connected to a signal conditioning unit, the output end of the third APD detector is connected to a signal processing unit, and the output end of the signal processing unit is connected to the input end of the APD driving circuit; the output end of the signal conditioning unit is connected to an analog-to-digital conversion unit, and the analog-to-digital conversion unit is bidirectionally connected to the control unit.

[0011] Preferably, in step S2, the number of neurons in the first convolutional layer is 24, 64, or 128.

[0012] Preferably, step S3 specifically includes the following steps: input the one-dimensional sample data of the training set into the phase difference prediction neural network. After one forward propagation along the phase difference prediction neural network, the phase difference prediction neural network outputs the mean square error calculated by the mean square error loss function to complete one epoch training process; then, under the guidance of the mean square error, perform backpropagation and update the weights of the phase difference prediction neural network. After iterating 2000 epochs, the training process of the phase difference prediction neural network can be completed; among them, use the phase difference prediction neural network parameters in the epoch training process with the smallest mean square error output during the 2000 epoch training process as the final phase difference prediction neural network model parameters to obtain the phase difference prediction neural network model.

[0013] Preferably, in step S4, the method for calculating the estimated distance based on the phase difference value is: based on the phase difference value, calculate the estimated distance through the formula where D is the estimated distance, is the phase difference, The value ranges from (0, 2π), which is calculated by c×λ / 2, where c is the propagation speed of light and λ is the modulation wavelength of the modulation unit.

[0014] Preferably, in step S1, a training set and a test set are obtained based on the phase difference samples, which specifically include the following steps: 3000 phase difference samples are collected. A part of the 3000 phase difference samples contains full-cycle signals, and the remaining part contains non-full-cycle signals. The phase difference samples containing full-cycle signals and those containing non-full-cycle signals are randomly divided according to a ratio of 8:2 respectively to obtain an original training set and an original test set; then, 10% of the phase difference samples containing full-cycle signals and those containing non-full-cycle signals in the original training set are respectively selected and 10% noise is added. Then, 10% of the remaining phase difference samples containing full-cycle signals and those containing non-full-cycle signals in the original training set are respectively selected and 20% noise is added. The phase difference samples without added noise and those with added noise in the above original training set together constitute the training set for training; then, 10% of the phase difference samples containing full-cycle signals and those containing non-full-cycle signals in the original test set are respectively selected and 10% noise is added. Then, 10% of the remaining phase difference samples containing full-cycle signals and those containing non-full-cycle signals in the original test set are respectively selected and 20% noise is added. The phase difference samples without added noise and those with added noise in the above original test set together constitute the test set for testing.

[0015] Preferably, the preprocessing method in step S4 is the same as that in step S1.

[0016] Compared with the prior art, the beneficial technical effects of this application are as follows:

[0017] The negative feedback regulated phase laser ranging system built in this application uses the third APD detector to perform negative feedback regulation on the output voltage signal of the APD drive circuit, so as to control the voltage signals converted from the current signals output by the first APD detector and the second APD detector to form a complete sine voltage signal, enabling the negative feedback regulated phase laser ranging system in this application to be applicable to more test environments; moreover, since the phase information in the complete sine voltage signal is accurate, the original measurement data obtained by the negative feedback regulated phase laser ranging system in this application has relatively accurate phase information, which is convenient for the subsequent phase difference prediction network to more accurately predict the phase difference;

[0018] In addition, the present application can also use the difference between the distance of the fiber optic delay line measured by the phase laser ranging system with negative feedback regulation and the distance of the fiber optic delay line itself as the calibration error, further improving the accuracy of the measured distance obtained by the present application. In addition, the present application has also built a phase difference prediction neural network. Using the phase difference prediction neural network can not only effectively extract the phase information of the phase difference samples, but also effectively reduce the influence of local noise on the phase information in the feature data, and can also perform denoising processing on the feature data. Through testing the phase difference prediction neural network built by the present application, it is found that the phase difference prediction neural network model obtained by the present application has a small prediction error and high accuracy when testing the phase difference samples without added noise. Moreover, when the phase difference prediction neural network model obtained by the present application tests the phase difference samples with 10% added noise and the phase difference samples with 20% added noise, the prediction error of the phase difference is small and the accuracy is high. In addition, the present application also uses the phase laser ranging system to obtain the original measurement data to be predicted, then preprocesses the original measurement data to be predicted, and then loads it into the phase difference prediction neural network model obtained in step S3 for one forward propagation to obtain the phase difference value. Then, the estimated distance is corrected using the calibration error to obtain the final measured distance. Through testing and comparison, it is found that when the method of the present application, the method of calculating the phase difference according to the discrete Fourier transform, and the full-phase Fourier error representation method of calculating the phase difference according to the full-phase Fourier transform are tested for the same test samples with a phase difference ranging from 4π / 30 to 56π / 30; the ratio of the absolute error value of the phase difference predicted by the method described in Embodiment 1 of the present application to the true phase difference value is at most 2.34%, while the ratio of the absolute error of the phase difference calculated by the method of calculating the phase difference according to the discrete Fourier transform to the true phase difference value is at most 28.51%, and the ratio of the absolute error of the phase difference calculated by the full-phase Fourier error representation method of calculating the phase difference according to the full-phase Fourier transform to the true phase difference value is at most 5.65%. Obviously, the ratio of the absolute error value of the phase difference obtained by the method of the present application to the true phase difference value is smaller, which indicates that the present application predicts the phase difference and calculates the estimated distance based on the phase difference using the formula, and then corrects the estimated distance using the calibration error, and the obtained final measured distance can have better accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a schematic structural diagram of a phase laser ranging system with negative feedback regulation;

[0020] Figure 2 It is a schematic network structure diagram of the phase difference prediction neural network in the present application;

[0021] Figure 3 It is the test result of the present application testing the phase difference samples without added noise in the test set;

[0022] Figure 4 A histogram of the absolute error and average error of the phase difference obtained by the method for visually representing the method described in this application, the method for calculating the phase difference according to the discrete Fourier transform, and the method for calculating the phase difference according to the all-phase Fourier transform. Detailed implementation manners

[0023] Example 1:

[0024] A laser ranging method based on a negative feedback regulated phase laser ranging system, comprising the following steps:

[0025] S1. Use the negative feedback regulated phase laser ranging system to obtain original measurement data, perform preprocessing to obtain phase difference samples, and then obtain a training set and a test set based on the phase difference samples;

[0026] The negative feedback regulated phase laser ranging system includes a control unit, the control unit is connected to a modulation unit, the modulation unit is respectively connected to a laser driving unit and an APD driving circuit; the laser driving unit is connected to a laser diode, the laser diode is respectively connected to a collimating lens and an optical fiber delay line; the APD driving circuit is respectively connected to a first APD detector, a second APD detector, and a third APD detector, the optical fiber delay line is connected to the first APD detector, this application further includes a converging lens, the converging lens is connected to a beam splitter, the beam splitter is respectively connected to the second APD detector and the third APD detector, the output ends of the first APD detector and the second APD detector are both connected to a signal conditioning unit, the output end of the third APD detector is connected to a signal processing unit, the output end of the signal processing unit is connected to the input end of the APD driving circuit; the output end of the signal conditioning unit is connected to an analog-to-digital conversion unit, the analog-to-digital conversion unit is bidirectionally connected to the control unit, and the control unit and the computer are bidirectionally connected.

[0027] The working principle of the negative feedback regulated phase laser ranging system in this application is:

[0028] The computer is connected to the control unit through the serial port, used to send and receive signals to the control unit. The control unit uses an STM controller to send control signals to the modulation unit; the modulation unit emits a sine wave voltage signal according to the received control signal; the laser driving unit receives the sine wave voltage signal sent by the modulation unit and outputs a current signal that controls the brightness change of the laser diode, so that the laser diode generates sine-varying light. Sine-varying light means that the light intensity of the light emitted by the laser diode fluctuates according to the sine wave; after receiving the driving signal, the laser diode generates sine-varying light and outputs the sine-varying light to the fiber optic delay line and the collimating lens; in this application, the fiber optic delay line can be used as the internal optical path. Since the distance of the fiber optic delay line itself is known, when measuring using the phase-based laser ranging system with negative feedback regulation, the difference between the measured distance of the fiber optic delay line and the distance of the fiber optic delay line itself can be used as the calibration error;

[0029] In this application, the phase difference prediction neural network model is used to obtain the phase difference value. Then, based on the phase difference value, through the formula the estimated distance is calculated, where D is the estimated distance, is the phase difference, takes values in (0, 2π), is calculated by c×λ / 2, where c is the propagation speed of light and λ is the modulation wavelength of the modulation unit. Then, the calibration error is used to correct the above estimated distance to obtain the final measured distance;

[0030] After the sine-varying light passes through the fiber optic delay line, it is transmitted to the first APD detector; the collimating lens is used to collimate the sine-varying light so that it is projected onto the target to be measured in the form of parallel light, thereby reducing the beam divergence of the sine-varying light and improving the measurement accuracy; after receiving the sine wave voltage signal sent by the modulation unit, the APD driving circuit unit performs frequency mixing operation and high-voltage bias driving on the sine wave voltage signal, and outputs driving signals and modulation signals to the first APD detector, the second APD detector, and the third APD detector respectively; under the action of the driving signal output by the APD driving circuit, the first APD detector converts the sine-varying light signal output by the fiber optic delay line into a current signal, mixes the current signal with the modulation signal, and then outputs the current signal to the signal conditioning unit;

[0031] In this application, after the sinusoidally varying optical signal projected by the collimating lens onto the target to be measured is reflected by the target to be measured, the converging lens is used to converge the scattered light to obtain converged light, and the converging lens transmits the converged light to the semi-transparent and semi-reflective mirror; the semi-transparent and semi-reflective mirror bisects the converged light to obtain two bisected light beams, and the two bisected light beams are respectively transmitted to the second APD detector and the third APD detector. Among them, the phase difference between the sinusoidal voltage signal emitted by the modulation unit and the bisected light signal received by the second APD detector is (0, 2π); the light intensity of each bisected light beam is 1 / 2 of the light intensity of the converged light; under the action of the drive signal output by the APD drive circuit, the second APD detector converts the bisected light signal output by the semi-transparent and semi-reflective mirror into a current signal, mixes the current signal with the modulation signal, and then outputs the current signal to the signal conditioning unit; under the action of the drive signal output by the APD drive circuit, the third APD detector converts the bisected light signal output by the semi-transparent and semi-reflective mirror into a current signal, mixes the current signal with the modulation signal, and then outputs the current signal to the signal processing unit;

[0032] The signal processing unit is used to convert the current signal output by the third APD detector into a voltage signal and determine whether the voltage signal is within the range of 0 to 3.3V; when the voltage signal is within the range of 0 to 3.3V, the voltage signal converted from the current signal output by the third APD detector is a complete sinusoidal voltage signal, and the signal processing unit does not feedback to the APD drive circuit; when the voltage signal is greater than 3.3V, the voltage signal converted from the current signal output by the third APD detector is an incomplete sinusoidal voltage signal. Specifically, it shows a state of peak clipping at the top of the sine wave on the sine waveform of the sinusoidal voltage signal. At this time, the signal processing unit feedbacks to the APD drive circuit, and the APD drive circuit reduces the drive voltage to facilitate controlling the voltage signals converted from the current signals output by the first APD detector, the second APD detector, and the third APD detector to be a complete sinusoidal voltage signal; when the voltage signal is less than 0V, the voltage signal converted from the current signal output by the third APD detector is also an incomplete sinusoidal voltage signal. Specifically, it shows a state of peak clipping at the bottom of the sine wave on the sine waveform of the sinusoidal voltage signal. At this time, the signal processing unit feedbacks to the APD drive circuit, and the APD drive circuit increases the drive voltage to facilitate controlling the voltage signals converted from the current signals output by the first APD detector, the second APD detector, and the third APD detector to be a complete sinusoidal voltage signal;

[0033] The negative feedback regulated phase laser ranging system built in this application uses a third APD detector to perform negative feedback regulation on the output voltage signal of the APD drive circuit, so as to control the voltage signals converted from the current signals output by the first APD detector and the second APD detector to form a complete sinusoidal voltage signal, enabling the negative feedback regulated phase laser ranging system in this application to be applicable to more test environments; moreover, since the phase information in the complete sinusoidal voltage signal is accurate, it makes the original measurement data obtained by the negative feedback regulated phase laser ranging system in this application have relatively accurate phase information, facilitating the subsequent phase difference prediction network to more accurately predict the phase difference;

[0034] The signal conditioning unit respectively converts the two current signals output by the first APD detector and the second APD detector into analog voltage signals and outputs the two analog voltage signals to the analog-to-digital conversion unit; the analog-to-digital conversion unit is used to sample the two analog voltage signals output by the signal conditioning unit with 6.33 microseconds as a time unit. The samples of the two analog voltage signals collected based on the same time unit are paired sampling samples. For each pair of sampling samples obtained, the pair of sampling samples are both transmitted to the control unit until all the two analog voltage signals are completely sampled according to the time unit; the control unit also stores the paired sampling samples output by the analog-to-digital conversion unit in two arrays respectively, and then transmits the paired sampling sample data to the computer through the serial port, and then transmits the next pair of sampled sample data that has been saved again, repeating in a cycle until all the paired sampling sample data are transmitted to the computer, and these data can be used as the original measurement data; among them, the number of bits of the above array is 1024, and when different sampling samples are stored in the array, they should be stored in the order of time.

[0035] In this application, the original measurement data is obtained by using the negative feedback regulated phase laser ranging system, and the original measurement data is preprocessed, including the following steps: using the np.vstack() function to vertically stack the corresponding sampling sample data in the two arrays (the corresponding sampling sample data is the paired sampling sample data in the previous text) in pairs to obtain a two-dimensional matrix, and the two-dimensional matrix constitutes a phase difference sample.

[0036] In this application, a training set and a test set are obtained based on phase difference samples, which specifically include the following steps: In this application, 3000 phase difference samples are collected. Some of the 3000 phase difference samples contain full-cycle signals, and the remaining part contains non-full-cycle signals. The phase difference samples containing full-cycle signals and those containing non-full-cycle signals are randomly divided according to a ratio of 8:2 respectively to obtain an original training set and an original test set. Then, 10% of the phase difference samples containing full-cycle signals and those containing non-full-cycle signals in the original training set are respectively selected and 10% noise is added. Then, 10% of the remaining phase difference samples containing full-cycle signals and those containing non-full-cycle signals in the original training set are respectively selected and 20% noise is added. The phase difference samples without added noise and those with added noise in the above original training set together constitute the training set for training. Then, 10% of the phase difference samples containing full-cycle signals and those containing non-full-cycle signals in the original test set are respectively selected and 10% noise is added. Then, 10% of the remaining phase difference samples containing full-cycle signals and those containing non-full-cycle signals in the original test set are respectively selected and 20% noise is added. The phase difference samples without added noise and those with added noise in the above original test set together constitute the test set for testing.

[0037] S2. Construct a phase difference prediction neural network:

[0038] In this application, the designed structure of the phase difference prediction neural network is as Figure 2 shown. The phase difference prediction neural network includes an input layer, a first convolutional layer, a first normalization layer, a first pooling layer, a second convolutional layer, a second normalization layer, a first LeakyRelu layer, a second pooling layer, a third convolutional layer, a third normalization layer, a second LeakyRelu layer, a first dropout layer, a flattening layer, a first fully connected layer, a fourth normalization layer, a third LeakyRelu layer, a second dropout layer, a second fully connected layer, a fifth normalization layer, a fourth LeakyRelu layer, a third dropout layer, and an output layer, which are connected in sequence. Among them, the number of neurons in the first convolutional layer is 24, and the output layer is set as a fully connected layer.

[0039] The working principle of the phase difference prediction neural network in this application is as follows:

[0040] In this application, the input layer is used to input one-dimensional sample data in the training set, one-dimensional sample data in the test set, or one-dimensional sample data for prediction; the first convolutional layer is used to perform a convolutional operation on the phase difference samples output by the input layer, extract the phase information of the phase difference samples, and obtain feature data with primary features; the first normalization layer performs a normalization operation on the feature data to prevent overfitting in the training process of the phase difference prediction neural network; the first pooling layer downsamples the feature data output by the first normalization layer to reduce the influence of local noise on the phase information in the feature data; the second convolutional layer performs a convolutional operation on the feature data output by the first pooling layer to further extract the phase information in the feature data and obtain feature data with intermediate features; the second normalization layer performs a normalization operation on the feature data output by the second convolutional layer to further prevent overfitting in the training process of the phase difference prediction neural network; the first LeakyRelu layer denoises the feature data; the second pooling layer downsamples the feature data to effectively reduce the subsequent calculation amount; the third convolutional layer performs a convolutional operation on the feature data to further extract the phase information and obtain feature data with advanced features; the third normalization layer performs a normalization operation on the feature data to further prevent overfitting in the training process of the phase difference prediction neural network; the second LeakyRelu layer further denoises the feature data; during the process of processing the feature data by the first dropout layer, a part of the neurons will be randomly discarded to avoid the influence of some large noises on the phase information; the flattening layer is used to convert the feature data into a one-dimensional vector value to obtain one-dimensional feature data with rich phase information; the first fully connected layer performs a dimensionality reduction operation on the one-dimensional feature data; the fourth normalization layer is used to perform a normalization operation on the feature data output by the first fully connected layer to further prevent overfitting in the training process of the phase difference prediction neural network and improve the stability of the feature data; the third LeakyRelu layer further denoises the feature data; during the process of processing the feature data by the second dropout layer, a part of the neurons are randomly discarded to further avoid the influence of some large noises on the phase information; the second fully connected layer further performs a dimensionality reduction operation on the feature data; the fifth normalization layer performs a normalization operation on the feature data to further prevent overfitting in the training process of the phase difference prediction neural network and improve the stability of the feature data; the fourth LeakyRelu layer further denoises the feature data to make the feature data more convergent; during the process of processing the feature data by the third dropout layer, a part of the neurons are randomly discarded to further avoid the influence of some large noises on the phase information; the output layer performs a feature space reduction operation on the feature data and outputs the phase difference and the mean square error.

[0041] In this application, the number of neurons in the first convolutional layer is 24, and the convolutional kernel size is 5×5; the number of neurons in the second convolutional layer is 128, and the convolutional kernel size is 5×5; the number of neurons in the third convolutional layer is 256, and the convolutional kernel size is 3×3.

[0042] S3. Under the guidance of the existing mean square error loss function, use the training set to train the phase difference prediction neural network to obtain the phase difference prediction neural network model, which specifically includes the following steps:

[0043] Input the one-dimensional sample data of the training set into the phase difference prediction neural network. After one forward propagation along the phase difference prediction neural network, the phase difference prediction neural network outputs the mean square error calculated by the mean square error loss function, completing one epoch training process; then, perform backpropagation under the guidance of the mean square error and update the weights of the phase difference prediction neural network. After iterating 2000 epochs, the training process of the phase difference prediction neural network can be completed; among them, use the phase difference prediction neural network parameters in the epoch training process with the smallest mean square error output during the 2000 epoch training process as the final phase difference prediction neural network model parameters to obtain the phase difference prediction neural network model; among them, the training process of the phase difference prediction neural network is based on the TensorFlow development platform, and in the training process of the phase difference prediction neural network, the batch size is set to 32, and the initial learning rate is set to 0.0001.

[0044] S4. Use the phase laser ranging system to obtain the original measurement data to be predicted, then preprocess the original measurement data to be predicted. The preprocessing method is the same as the method for preprocessing the original measurement data in step S1, and then load it into the phase difference prediction neural network model obtained in step S3 for one forward propagation to obtain the phase difference value, and through the formula calculate the estimated distance, where D is the estimated distance, is the phase difference, takes values in (0, 2π), is calculated by c×λ / 2, c is the propagation speed of light, and λ is the modulation wavelength of the modulation unit; then, use the calibration error to correct the estimated distance to obtain the final measured distance.

[0045] Embodiment 2:

[0046] The difference between this embodiment and Embodiment 1 is that: the number of neurons in the first convolutional layer in Embodiment 2 is 12;

[0047] Embodiment 3:

[0048] The difference between Embodiment 3 and Embodiment 1 is that: the number of neurons in the first convolutional layer in Embodiment 3 is 32;

[0049] Example 4:

[0050] The difference between Example 4 and Example 1 is that: in Example 4, the number of neurons in the first convolutional layer is 64;

[0051] Example 5:

[0052] The difference between Example 5 and Example 1 is that: in Example 5, the number of neurons in the first convolutional layer is 128;

[0053] Example 6:

[0054] The difference between Example 6 and Example 1 is that: in Example 6, the number of neurons in the first convolutional layer is 256.

[0055] Test:

[0056] In order to test the effect of the phase difference prediction neural network model in this application on predicting the phase difference, this application uses a test set for testing. Among them, through testing, it is found that when testing the phase difference samples without added noise in the test set, the test results are as Figure 3 shown. As can be seen from Figure 3 , when Examples 1 to 6 of this application test the phase difference samples without added noise in the test set, the obtained CORR value is greater than or equal to 0.9999, which indicates that the phase difference prediction values obtained by Examples 1 to 6 of this application when testing the phase difference samples without added noise in the test set are very close to the true phase difference values; among them, the phase difference samples without added noise refer to the phase difference samples in the test set except for the phase difference samples with 10% added noise and the phase difference samples with 20% added noise.

[0057] In order to explore the effect of the phase difference prediction neural network model in this application on predicting the phase difference of noisy phase difference samples, this application specifically tests the phase difference samples with 10% added noise and the phase difference samples with 20% added noise in the test set, and the test results are shown in Table 1 and Table 2;

[0058] Table 1 Evaluation Index of 10% Noise

[0059]

[0060] In Table 1, MSE (Mean Squared Error) represents the mean squared error, RMSE (Root Mean Squared Error) represents the root mean squared error, MAE (Mean Absolute Error) represents the mean absolute error, MRE (Mean Relative Error) represents the mean relative error, and CORR (Correlation Coefficient) represents the correlation coefficient. The smaller the error values of the mean squared error, root mean squared error, mean absolute error, and mean relative error, the more accurate the phase difference predicted by the phase difference prediction neural network model in the embodiment. The larger the correlation coefficient, the closer the predicted phase difference value is to the true phase difference value.

[0061] Table 2 20% Noise Evaluation Index

[0062]

[0063] Based on the test results shown in Table 1 and Table 2, when 24 neurons are used in the first convolutional layer of the phase difference prediction neural network, that is, the phase difference prediction neural network model described in Embodiment 1, when testing the phase difference samples with 10% noise and the phase difference samples with 20% noise, considering the test results of the mean squared error, root mean squared error, mean absolute error, mean relative error, and CORR index comprehensively, the phase difference prediction neural network model described in Embodiment 1 can achieve better prediction results, that is, the accuracy of predicting the phase difference is better.

[0064] In addition, this application also uses a phase laser ranging system to obtain the original measurement data to be predicted, then preprocesses the original measurement data to be predicted in the same way as the preprocessing of the original measurement data in step S1, and then loads it into the phase difference prediction neural network model obtained in step S3 for one forward propagation to obtain the phase difference value; then, uses the calibration error to correct the estimated distance to obtain the final measurement distance;

[0065] Since each embodiment of this application is carried out at room temperature (25°C), the calibration error of the negative feedback regulated phase laser ranging system constructed in this application is small. Therefore, this application focuses on considering the effect of predicting the phase difference value; in order to more intuitively show the advantages and disadvantages of the method described in this application compared with traditional spectral analysis methods (such as: the method of calculating the phase difference according to the discrete Fourier transform and the method of calculating the phase difference according to the all-phase Fourier transform) in terms of the absolute error of obtaining the phase difference (the absolute error refers to the absolute value of the difference between the predicted phase difference and the true phase difference) and the average error (the average error refers to the average value of the absolute error of the phase difference), this application uses a histogram to intuitively represent the absolute error and average error of the phase difference, as Figure 4 shown.

[0066] Figure 4 Among them, the average error of Example 1 represents the average error of the phase difference predicted in Example 1; the average error of the discrete Fourier transform represents the average error of the phase difference calculated by the method of calculating the phase difference according to the discrete Fourier transform; the average error of the all-phase Fourier transform represents the average error of the phase difference calculated by the method of calculating the phase difference according to the all-phase Fourier transform; the error of Example 1 represents the absolute error of the phase difference predicted in Example 1; the error of the discrete Fourier transform represents the absolute error of the phase difference calculated by the method of calculating the phase difference according to the discrete Fourier transform; the error of the all-phase Fourier transform represents the absolute error of the phase difference calculated by the method of calculating the phase difference according to the all-phase Fourier transform.

[0067] Figure 4 Among them, the abscissa is the test sample, which means that the test samples are evenly divided into 30 parts in sequence according to the phase difference (0, 2π). Moreover, this application measures by using a phase-based laser ranging system based on negative feedback regulation. During the measurement, the phase difference between the sine wave voltage signal emitted by the modulation unit and the bisected optical signal received by the second APD detector is (0, 2π). Therefore, the original measurement data obtained in this application does not include the case where the phase difference is 0, nor does it include the case where the phase difference is 2π. Therefore, Figure 4 What is shown in are the test results obtained by testing the test samples with phase differences of 2π / 30, 4π / 30, 6π / 30......56π / 30, 58π / 30 respectively;

[0068] For example, Figure 4 The abscissa 2π / 30 in represents the result obtained by testing the test sample with a phase difference of 2π / 30; Figure 4 What is shown in are the test data obtained by testing the test samples with phase differences from 2π / 30 to 58π / 30, as shown in Table 3 and Table 4.

[0069] In Table 3 and Table 4, the DFT method refers to the method of calculating the phase difference according to the discrete Fourier transform, and the ApFFT method refers to the method of calculating the phase difference according to the all-phase Fourier transform; taking the ratio of the absolute error value of the phase difference predicted by the method described in Example 1 to the true phase difference value as 2.91% as an example, the calculation method is introduced as follows: calculate the absolute value of the difference between the predicted phase difference value 0.2033 of the method described in Example 1 and the true phase difference value 2π / 30, and obtain the absolute error value 0.0061 of the phase difference predicted by the method described in Example 1. Then, calculate the ratio of the absolute error value 0.0061 of the phase difference predicted by the method described in Example 1 to the true phase difference value 2π / 30 and then multiply by 100% to obtain 2.91%;

[0070] Table 3 represents Figure 4Test data obtained from testing test samples with a phase difference of 2π / 30 to 34π / 30 shown in

[0071]

[0072] Table 4 shows Figure 4 Test data obtained from testing test samples with a phase difference of 36π / 30 to 58π / 30 shown in

[0073]

[0074] In the prior art, during actual measurement, for example, when measuring the liquid level of a ladle or water, there is usually a relatively accurate measurement range. Specifically in this application, in the first embodiment of this application, the test results obtained from testing test samples with a phase difference of 2π / 30 and 58π / 30 are not accurate enough, while the test results obtained from testing test samples with a phase difference of 4π / 30 to 56π / 30 are better. That is to say, during actual measurement in this application, the phase difference should be obtained based on test samples with a phase difference of 4π / 30 to 56π / 30;

[0075] Therefore, this application focuses on comparing the test results of the method described in this application, the method of calculating the phase difference according to the discrete Fourier transform, and the method of representing the all-phase Fourier error for calculating the phase difference according to the all-phase Fourier transform for the same test samples with a phase difference of 4π / 30 to 56π / 30. From Table 3 and Table 4, when this application predicts the phase difference based on test samples with a phase difference of 4π / 30 to 56π / 30, the ratio of the absolute error value of the phase difference predicted by the method described in the first embodiment of this application to the true phase difference value is at most 2.34%, while the ratio of the absolute error of the phase difference calculated by the method of calculating the phase difference according to the discrete Fourier transform to the true phase difference value is at most 28.51%, and the ratio of the absolute error of the phase difference calculated by the method of representing the all-phase Fourier error for calculating the phase difference according to the all-phase Fourier transform to the true phase difference value is at most 5.65%. Obviously, the ratio of the absolute error value of the phase difference obtained by the method described in this application to the true phase difference value is smaller, which indicates that this application predicts the phase difference and calculates and estimates the distance based on the phase difference using a formula, and then corrects the estimated distance using the calibration error, and the final measured distance can have better accuracy.

Claims

1. A laser ranging method for a phase-based laser ranging system based on negative feedback regulation, characterized in that: Including the following steps: S1. Obtain the original measurement data by using a phase-type laser ranging system and perform preprocessing to obtain phase difference samples, and based on the phase difference samples, obtain a training set and a test set; S2. Construct a phase difference prediction neural network for predicting phase differences; including an input layer, a first convolutional layer, a first normalization layer, a first pooling layer, a second convolutional layer, a second normalization layer, a first LeakyRelu layer, a second pooling layer, a third convolutional layer, a third normalization layer, a second LeakyRelu layer, a first dropout layer, a flattening layer, a first fully connected layer, a fourth normalization layer, a third LeakyRelu layer, a second dropout layer, a second fully connected layer, a fifth normalization layer, a fourth LeakyRelu layer, a third dropout layer, and an output layer connected in sequence; S3. Under the guidance of the existing mean square error loss function, use the training set to train the phase difference prediction neural network to obtain a phase difference prediction neural network model; S4. Obtain the original measurement data to be predicted by using a phase-type laser ranging system and perform preprocessing, load it into the phase difference prediction neural network model for one forward propagation to obtain a phase difference value, calculate an estimated distance based on the phase difference value; use the calibration error obtained by negative feedback adjustment of the phase-type laser ranging system to correct the estimated distance to obtain the final measured distance.

2. The laser ranging method of the phase-based laser ranging system based on negative feedback regulation according to claim 1, wherein: In step S1, the negative feedback adjustment of the phase-type laser ranging system includes a control unit, the control unit is connected to a modulation unit, and the modulation unit is respectively connected to a laser driving unit and an APD driving circuit; the laser driving unit is connected to a laser diode, and the laser diode is respectively connected to a collimating lens and an optical fiber delay line; the APD driving circuit is respectively connected to a first APD detector, a second APD detector, and a third APD detector, the optical fiber delay line is connected to the first APD detector, the present application further includes a converging lens, the converging lens is connected to a semi-transparent and semi-reflective mirror, the semi-transparent and semi-reflective mirror is respectively connected to the second APD detector and the third APD detector, the output ends of the first APD detector and the second APD detector are both connected to a signal conditioning unit, the output end of the third APD detector is connected to a signal processing unit, and the output end of the signal processing unit is connected to the input end of the APD driving circuit; the output end of the signal conditioning unit is connected to an analog-to-digital conversion unit, and the analog-to-digital conversion unit is bidirectionally connected to the control unit.

3. The laser ranging method of the phase-based laser ranging system based on negative feedback regulation according to claim 1, characterized in that: In step S2, the number of neurons in the first convolutional layer is 24, 64, or 128.

4. The laser ranging method of the phase-based laser ranging system based on negative feedback regulation according to claim 1, wherein: Step S3 specifically includes the following steps: Input the one-dimensional sample data of the training set into the phase difference prediction neural network. After one forward propagation along the phase difference prediction neural network, the phase difference prediction neural network outputs the mean square error calculated by the mean square error loss function, completing one epoch training process; then, under the guidance of the mean square error, backpropagation is performed and the weights of the phase difference prediction neural network are updated. After iterating 2000 epochs, the training process of the phase difference prediction neural network can be completed; among them, the phase difference prediction neural network parameters in the epoch training process with the smallest mean square error output during the 2000-epoch training process are used as the final phase difference prediction neural network model parameters to obtain the phase difference prediction neural network model.

5. The laser ranging method of the phase-based laser ranging system based on negative feedback regulation according to claim 1, characterized in that: In step S4, the method for calculating the estimated distance based on the phase difference is as follows: Based on the phase difference, the estimated distance is calculated through the formula where D is the estimated distance, is the phase difference, is calculated by c×λ / 2, where c is the propagation speed of light and λ is the modulation wavelength of the modulation unit.

6. The laser ranging method of the phase-based laser ranging system based on negative feedback regulation according to claim 1, wherein: In step S1, obtaining the training set and the test set based on the phase difference samples specifically includes the following steps: Collect 3000 phase difference samples. Some of the 3000 phase difference samples contain full-cycle signals, and the remaining part contains non-full-cycle signals. The phase difference samples containing full-cycle signals and non-full-cycle signals are randomly divided according to a ratio of 8:2 respectively to obtain the original training set and the original test set; then, 10% noise is added to 10% of the phase difference samples containing full-cycle signals and non-full-cycle signals in the original training set respectively. Then, 10% noise is added to 10% of the remaining phase difference samples containing full-cycle signals and non-full-cycle signals in the original training set respectively. The phase difference samples without added noise and the phase difference samples with added noise in the above original training set together constitute the training set for training; then, 10% noise is added to 10% of the phase difference samples containing full-cycle signals and non-full-cycle signals in the original test set respectively. Then, 10% noise is added to 10% of the remaining phase difference samples containing full-cycle signals and non-full-cycle signals in the original test set respectively. The phase difference samples without added noise and the phase difference samples with added noise in the above original test set together constitute the test set for testing.

7. The laser ranging method of the phase-based laser ranging system based on negative feedback regulation according to claim 1, characterized in that: The preprocessing method in step S4 is the same as that in step S1.

Images