CN109164454B - Pseudo-random code modulation-based medium-long range high-frequency laser radar ranging ambiguity solving method - Google Patents

Pseudo-random code modulation-based medium-long range high-frequency laser radar ranging ambiguity solving method Download PDF

Info

Publication number
CN109164454B
CN109164454B CN201810965428.8A CN201810965428A CN109164454B CN 109164454 B CN109164454 B CN 109164454B CN 201810965428 A CN201810965428 A CN 201810965428A CN 109164454 B CN109164454 B CN 109164454B
Authority
CN
China
Prior art keywords
echo
mta
measurement
interval
waveform
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
Application number
CN201810965428.8A
Other languages
Chinese (zh)
Other versions
CN109164454A (en
Inventor
毛庆洲
刘荣荣
柳晨光
闫保芳
王芳
吴安磊
胡伟
崔昊
董翠军
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan University WHU
Original Assignee
Wuhan University WHU
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Wuhan University WHU filed Critical Wuhan University WHU
Priority to CN201810965428.8A priority Critical patent/CN109164454B/en
Publication of CN109164454A publication Critical patent/CN109164454A/en
Application granted granted Critical
Publication of CN109164454B publication Critical patent/CN109164454B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
    • G01S17/02—Systems using the reflection of electromagnetic waves other than radio waves
    • G01S17/06—Systems determining position data of a target
    • G01S17/08—Systems determining position data of a target for measuring distance only
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Electromagnetism (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Biomedical Technology (AREA)
  • Evolutionary Computation (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Remote Sensing (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Optical Radar Systems And Details Thereof (AREA)

Abstract

The invention provides a pseudo-random code modulation-based medium-and-long-range high-frequency laser radar ranging fuzzy solving method, which comprises the steps of preliminarily calculating potential measuring distance intervals and distribution probabilities of all intervals of a detection target by using a BP neural network and taking the potential measuring distance intervals and the distribution probabilities as initial calculation input; further calculating the measuring interval of the echo by using the complexity of the echo signal in the full-waveform signal, the correlation among the echoes and the continuity characteristics of the ground objects; modulating laser pulse emission time by using a pseudo-random binary sequence, calculating an accurate echo detection interval according to the criterion of minimizing the noise energy of a sequence target reflected echo, determining the corresponding relation between the echo and seed light, and calculating an accurate measurement distance according to the corresponding relation. The invention can effectively and accurately solve the problem of range finding ambiguity in the high-frequency laser radar, and can simultaneously obtain higher measurement precision and accuracy at a longer measurement distance and a denser measurement foot point in the measurement of the high-frequency laser radar.

Description

Pseudo-random code modulation-based medium-long range high-frequency laser radar ranging ambiguity solving method
Technical Field
The invention relates to the field of laser radar measurement, in particular to a pseudo-random code modulation-based medium-long range high-frequency laser radar ranging fuzzy solving method.
Technical Field
The precision, the density and the efficiency are three important indexes for measuring the capability of the airborne laser radar. With the application of full-waveform laser radar technology, people put higher requirements on the measurement distance and the laser foot point density of the airborne laser radar. However, when the high repetition frequency laser transmitter is applied to the airborne laser radar, the problem of ambiguity of association between the laser emission pulse and the echo signal is inevitably brought, that is, in the high frequency laser radar, when the time interval between the echo receiving time and the seed light emission time is more than one or even more than one time interval between the echo receiving time and the seed light emission time, the correct corresponding relation between the echo signal and the seed light signal cannot be directly determined from the time signal sequence, and thus the problem of range finding ambiguity occurs.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a pseudo-random code modulation-based method for solving the ranging ambiguity of the medium-long range high-frequency laser radar, which can reduce the ranging ambiguity problem of the long range high-frequency laser radar and improve the measurement accuracy.
In order to achieve the purpose, the invention provides a pseudo-random code modulation-based medium-long range high-frequency laser radar ranging ambiguity solving method, which is characterized by comprising the following steps of:
establishing an echo waveform characteristic information and echo distance information model by using a BP neural network, and calculating all potential measurement distance intervals and distribution probabilities of all intervals of a single measurement echo, wherein the echo waveform morphological characteristic is an input item of the BP neural network;
decomposing the full waveform data into a plurality of Gaussian echoes by using a waveform decomposition algorithm, and further calculating the optimized measuring distance range of the measuring sequence and calculating the measuring distance interval of the echoes according to the continuity characteristics of the regional target and by using an optimization algorithm;
and modulating the laser pulse emission time by utilizing the pseudo-random binary sequence, and calculating an accurate echo detection interval according to the criterion of minimizing the noise energy of the echo reflected by the sequence target.
Further, the echo waveform morphological characteristics include an echo comprehensive energy parameter, an echo signal-to-noise ratio parameter, an echo average width parameter, a height parameter, an average energy parameter, and a barycentric position parameter.
Further, the specific process of establishing the BP neural network is as follows: the method comprises the steps of collecting a single laser full-waveform echo signal by using a low-frequency laser, taking extracted waveform morphological characteristic information as an input item of a BP (Back propagation) neural network, and calculating a waveform interval MTA [ i ] where the same echo is collected in the high-frequency laser according to a Time scale collected by the low-frequency laser, wherein an MTA (multiple Time around) interval value i is defined as the Time when the echo is received is between the Time of i-1 pulse repetition intervals and i pulse repetition intervals, namely the interval between two transmitted waves. And establishing a 4-layer BP neural network by taking the waveform interval MTA [ i ] of the echo and the probability in different measurement intervals as output items of the neural network, and respectively selecting a linear function and a piecewise function by using the intermediate state transfer function.
Further, the waveform decomposition algorithm is a Levenberg-Marquardt decomposition algorithm.
Further, the optimization process for further optimizing the measurement interval input item of the BP neural network by using the waveform decomposition algorithm specifically includes:
according to the continuity characteristics of regional targets and the adoption of an optimization algorithm, the optimized measurement distance range MTA [ X ] of the measurement sequence can be further calculated](ii) a Assuming a total of N MTA regions, the laser emits the ith pulse at time tT[i]The moment when the detector detects the jth echo is tE[j]The number of laser pulses emitted at this time is Gj]The result output by the front-segment neural network can obtain the distribution probability p of the result in the MTA zone k (k is 1,2 … N)j,kWhich measures the distance Lj,kHas the following relationship:
Ru×(k-1)≤Lj,k≤Ru×k
Figure BDA0001774839700000021
Figure BDA0001774839700000022
then for M consecutive echoes starting from the 1 st, the probability distribution matrix and the measured distance matrix at these N MTA intervals are:
Figure BDA0001774839700000023
Figure BDA0001774839700000031
let the expected vector of the MTA interval of these M consecutive echoes be:
Figure BDA0001774839700000032
the expected vectors for the measured distances are:
Figure BDA0001774839700000033
the geometric shape of the ground object target can be regarded as continuous in a local range, and the situation that the size exceeds one MTA area does not occur; by using the least square method for estimation, the expected measured distances of the M echoes are fitted by a straight line f, and even if the value of formula (14) is the minimum, the optimal estimation value of the measured distance of each echo, namely the optimal estimation value of the MTA interval, can be obtained:
Figure BDA0001774839700000034
furthermore, the specific process of the criterion of minimizing the noise energy of the sequence target reflection echo is as follows:
since the contribution of the low frequency components of the distance signal, which are constant within each MTA interval, to the noise energy is negligible, it is possible to calculate the difference vector ar of successive measurementsMTA,jThe method comprises the following steps:
ΔRMTA,j＝(r2-r1,r3-r2,...,rM-rM-1)T
from each distance vector RMTA,jRemoving high frequency noise components in the middle, thenEach difference vector Δ R associated with all potential MTA intervalsMTA,jNoise energy of
Figure BDA0001774839700000035
Can be given by:
Figure BDA0001774839700000036
minimum noise energy
Figure BDA0001774839700000037
Not affected by PPM, the j value at this time is associated with the correct MTA interval; the MTA interval with the smallest noise energy can be selected as the result by comparing the noise signal energy of each measurement range group (M) assigned to all potential MTA intervals.
Furthermore, the modulation of the pseudo-random binary sequence is determined according to different topographic features, including mountainous areas, cities, rural areas and flat areas, the measurement confidence coefficients of different intervals are determined by using the statistical distance measurement values and the noise energy of different measurement intervals according to the specificity and complexity of a detection target, and the length and the delay length tau of the pseudo-random binary code are adjusted.
Furthermore, the noise energy of the echo reflected by the sequence target is minimized, the obtained waveform measurement interval and the accuracy are calculated, and meanwhile, the morphological characteristic information of the echo is extracted to serve as the prior training sample information of the BP neural network to participate in model training.
The invention has the advantages that:
the method is suitable for solving the ranging ambiguity problem in the high-frequency full-waveform airborne laser radar. Using a BP neural network model, establishing a characteristic information and echo distance information model in a single echo waveform according to waveform form information to calculate all potential measurement intervals and distribution probabilities thereof, and fully considering waveform form information of each waveform component of a full waveform echo; the measuring interval is further determined through the internal associated information of the waveform echo signals and is used as initial measuring interval information, so that the operation accuracy and the reliability are improved; according to the characteristics of different landforms, the statistical ranging values and the noise energy of different measuring intervals are used for determining the measuring confidence degrees of different intervals, and the length and the time delay of the pseudo-random binary code sequence are determined according to the measuring confidence degrees, so that the universality of the method is enhanced, and the method is simple and effective; the pseudo-random binary code is used for adjusting the pulse transmitting time, noise information is represented through the difference vector of continuous measured values, the influence of high-frequency noise components can be well reduced, and the calculation confidence is improved while the calculation amount is reduced.
Drawings
Fig. 1 is a schematic diagram illustrating a problem of range ambiguity in high frequency laser measurement according to the present invention.
Fig. 2 is a schematic diagram of the general design concept of the present invention.
FIG. 3 is a schematic diagram of a BP neural network model established according to echo wave and echo interval.
Fig. 4 is a schematic diagram of pulse transmission after pseudo-random binary code modulation.
Fig. 5 is a schematic diagram of point cloud distribution under different measurement intervals.
Detailed Description
The invention is described in further detail below with reference to the following figures and specific examples:
the invention provides a pseudo-random code modulation-based method for solving the ranging ambiguity of a medium-long range high-frequency laser radar, which is used for solving the ranging ambiguity problem of a high-frequency full-waveform laser radar. In the laser radar measurement, with the continuous improvement of the requirements on the measurement distance and the laser foot point density of the airborne laser radar, the high-frequency laser radar inevitably has a ranging ambiguity problem, as shown in fig. 1, namely before the echo of the first beam of emission pulse is received, because the emission frequency is too high, the middle time may emit one or even multiple beams of laser pulse, and therefore, the incidence relation between the seed light and the echo cannot be correctly corresponded on the time sequence, and the measurement ambiguity problem is generated. When the laser emission frequency (PRR, Pulse Repetition Rate) is 400kHz, and the actual target distance is 1650m, all possible distances calculated directly from the time series are: 150m, 525m, 900m, 1275m and 1650m, which is the problem of range ambiguity.
The overall processing flow of the method for solving the ranging ambiguity problem by using the BP neural network and the pseudo-random binary code modulation is shown in FIG. 2. The method comprises the following steps:
and S1, extracting echo distance information and echo waveform characteristic information by using the single laser full-waveform echo signal. Assuming that a laser pulse signal is transmitted as H (t), the atmospheric attenuation rate is beta, the distance between an observation point and a target is L, the target is provided with N layers, the total area is A, and the corresponding reflectivity, layer area and layer height of the ith layer are respectively alphaiA and liThe echo at the detector surface can then be written as:
Figure BDA0001774839700000051
Figure BDA0001774839700000052
thus, the echo signals received at the probe are:
Figure BDA0001774839700000061
the comprehensive energy of the full-waveform echo is as follows:
Figure BDA0001774839700000062
the full waveform echo average energy is:
Figure BDA0001774839700000063
according to the range equation of the laser radar, the signal-to-noise ratio of the received signal can be obtained as follows:
Figure BDA0001774839700000064
wherein I is the beam intensity; Ω is the solid angle of the target (sr); NEI is receiver equivalent noise irradiance (W/m)2) Determined by the receiver performance; f is the target retroreflectivity (1/sr) of the unit solid angle; etaTIs a pulse broadening equivalent factor; alpha is the atmospheric extinction coefficient (1/m) and includes atmospheric molecular absorption and scattering, aerosol absorption and scattering, and other molecular scattering or absorption.
According to the analysis formulas (4) to (6), the main factors influencing the comprehensive energy, the average energy and the signal-to-noise ratio of the laser echo signal related to the ground object target are the distance between the observation point and the target, the shape of the ground object target and the reflectivity of the ground object target. Therefore, the main factors of comprehensive energy, average energy and signal-to-noise ratio of the laser echo signals and waveform form information including waveform gravity center position, waveform width, waveform height information and the like can be used for establishing the relation with the measurement range between the detection target distance observation point and the measurement range.
S2, establishing an echo waveform characteristic information and echo distance information model by using a BP neural network, and determining the distance range and the distribution probability of the target corresponding to the Nth echo. Using a low-frequency laser to collect a single laser full-waveform echo signal, extracting waveform form characteristic (echo form information such as echo comprehensive energy parameter, echo signal-to-noise ratio parameter, echo average width parameter, height parameter, average energy parameter, gravity center position parameter and the like) information according to S1 and using the information as an input item of a BP neural network, calculating a waveform interval MTA [ i ], an MTA (multiple Time around) interval value i, in which the same echo is collected in the high-frequency laser, according to a Time scale collected by the low-frequency laser, wherein the Time when the echo is received is defined to be the Time when the echo is located between i-1 pulse repetition intervals and i pulse repetition intervals, establishing a 4-layer BP neural network by using the waveform interval MTA [ i ] in which the echo is located and the probabilities in different measurement intervals as output items of the neural network, and selecting a linear function and a piecewise function respectively by an intermediate state transfer function, as shown in fig. 3. And adopting the multiple measured data as a neural network training sample.
And S3, decomposing the waveform signal in the full waveform data (decomposing the waveform signal of the laser radar). Due to the complexity of the ground object target, the incident laser beam may be divided into a plurality of backscattering sources with different intensities by the ground object target, so that the digitized waveform signal received by the laser radar is not a gaussian-like waveform but the superposition of a plurality of gaussian-like waveforms with different amplitudes and widths, and therefore, for airborne full-waveform echo signals, waveform decomposition is required to be carried out to better analyze the ground object target, and the used waveform decomposition algorithm is a Levenberg-Marquardt decomposition algorithm.
Assuming that the full waveform signal consists of N echoes, the waveform can be represented as a superposition of N gaussian components, i.e.:
Figure BDA0001774839700000071
wherein A isk、μk、ωkRespectively representing the maximum amplitude, the center and the half width of the waveform of the kth waveform component. The purpose of waveform decomposition is to find the most suitable series (A)k、μk、ωk) The error of the fit is minimized, i.e., the difference between the fit waveform and the original waveform is made as small as possible or smaller than the error tolerance range e.
Figure BDA0001774839700000072
All gaussian components to be solved are denoted as P (P)1，p2，p3，......pm) And m is 3 × N, the analog waveform can be expressed as: f (x)iP), first a function initial value P is given0F (x)iP) at P0Taylor expansion is performed and the quadratic and higher order terms are omitted. From this, the objective function can be deduced:
Figure BDA0001774839700000081
obtaining an objective function:
Figure BDA0001774839700000082
the ultimate goal of the LM algorithm is to minimize the objective function Q. The first derivative of the objective function is solved to be 0, and a classical iterative formula of the LM algorithm can be obtained through matrix operation:
P＝p0+[H(x，p0)+λE]-1JT(x，p0)[y-f(x，p0)](11)
wherein J is a function f (x)iP), H is a sea plug matrix composed of its second partial derivatives, and for practical calculation and feasibility considerations, a sea plug simulation matrix is used instead, that is:
H＝JTJ (12)
the LM algorithm is a continuous iteration process, when the difference between the result calculated by the formula and the actual quantity is small or the maximum iteration number is reached, the iteration is considered to be finished, wherein the initial parameter information p in the iteration is0May be obtained using a second order peak detection method.
S4, according to the measurement sequence object obtained by decomposition of S3, the optimal measurement distance range MTA [ X ] of the measurement sequence can be further calculated by using the measurement distance and probability distribution information obtained by S2, according to the continuity characteristics of the regional object and adopting an optimization algorithm]. Assuming a total of N MTA regions, the laser emits the ith pulse at time tT[i]The moment when the detector detects the jth echo is tE[j]The number of laser pulses emitted at this time is Gj]The result output by the front-segment neural network can obtain the distribution probability p of the result in the MTA zone k (k is 1,2 … N)j，kWhich measures the distance Lj，kHas the following relationship:
Ru×(k-1)≤Lj，k≤Ru×k (13)
Figure BDA0001774839700000091
Figure BDA0001774839700000092
then for M consecutive echoes starting from the 1 st, the probability distribution matrix and the measured distance matrix at these N MTA intervals are:
Figure BDA0001774839700000093
Figure BDA0001774839700000094
let the expected vector of the MTA interval of these M consecutive echoes be:
Figure BDA0001774839700000095
the expected vectors for the measured distances are:
Figure BDA0001774839700000096
the geometric shape of the ground object can be regarded as continuous in a local range, and the situation that the size exceeds one MTA area does not occur. By adopting the least square method for estimation, the expected measuring distance of the M echoes is fitted by a straight line f, even if the value of the formula (14) is minimum, the optimal estimation value of the measuring distance of each echo, namely the optimal estimation value of the MTA interval can be obtained.
Figure BDA0001774839700000097
And S5, determining the length of the optimized pseudo-random binary code. When the number M of sequence echoes participating in the optimal MTA interval estimation is too large, a plurality of single-transmitted echoes can be contained, and the hypothesis is contrary to the hypothesis of the optimal estimation; when the value of M is too small, a good fitting result can not be obtained, so that according to different topographic features, in mountainous areas, cities, rural areas and plain areas, according to the particularity and complexity of a detection target, the statistical distance measurement value and the noise energy of different measurement intervals are used for determining the measurement confidence coefficients of different intervals, and the length of the pseudo-random binary code and the delay length tau are adjusted. The delay length τ is generally related to the laser pulse width, for example, the delay length τ is 1/3.
S6, adjusting the pulse emission time according to the generated pseudo-random binary code sequence, wherein 0-pulse emission time is not changed, and 1-pulse emission time is delayed by 1 time scale tau, as shown in figure 4;
s7, according to S4, distance vector R of M continuous echo signals in each MTA interval can be calculatedMTA,j＝(r1,r2,...,rM)TWhich represents the M measured distances in the jth MTA interval of a pseudorandom binary sequence in one of the scan lines. Since the contribution of the low frequency components of the distance signal, which are constant within each MTA interval, to the noise energy is negligible, it is possible to calculate the difference vector ar of successive measurementsMTA,jFrom each distance vector R, as in equation 21MTA,jHigh frequency noise components are eliminated. Then, each difference vector Δ R associated with all potential MTA intervalsMTA,jNoise energy of
Figure BDA0001774839700000101
Can be given by equation 22, fig. 5.
ΔRMTA,j＝(r2-r1,r3-r2,...,rM-rM-1)T(21)
Figure BDA0001774839700000102
S8, from the noise figure calculated in S7, it is apparent that the minimum noise energy
Figure BDA0001774839700000103
Not affected by PPM, the value of j at this time has been associated with the correct MTA interval. The MTA interval with the smallest noise energy can be selected as the result by comparing the noise signal energy of each measurement range group (M) assigned to all potential MTA intervals.
And S9, determining the corresponding relation between the seed light and the echoes according to S8, and calculating the measuring distance corresponding to each echo in the full waveform.
The invention relates to a pseudo-random binary code modulation-based method for solving the ranging ambiguity of a medium-long range high-frequency laser radar, which is suitable for solving the ranging ambiguity problem of a high-frequency full-waveform airborne laser radar. Using a BP neural network model, establishing a characteristic information and echo distance information model in a single echo waveform according to waveform form information to calculate all potential measurement intervals and distribution probabilities thereof, and fully considering waveform form information of each waveform component of a full waveform echo; the measuring interval is further determined through the internal associated information of the waveform echo signals and is used as initial measuring interval information, so that the operation accuracy and the reliability are improved; according to the characteristics of different landforms, the statistical ranging values and the noise energy of different measuring intervals are used for determining the measuring confidence degrees of different intervals, and the length and the time delay of the pseudo-random binary code sequence are determined according to the measuring confidence degrees, so that the universality of the method is enhanced, and the method is simple and effective; the pseudo-random binary code is used for adjusting the pulse transmitting time, noise information is represented through the difference vector of continuous measured values, the influence of high-frequency noise components can be well reduced, and the calculation confidence is improved while the calculation amount is reduced.
The above embodiments are only used for illustrating the design idea and features of the present invention, and the purpose of the present invention is to enable those skilled in the art to understand the content of the present invention and implement the present invention accordingly, and the protection scope of the present invention is not limited to the above embodiments. Therefore, all equivalent changes and modifications made in accordance with the principles and concepts disclosed herein are intended to be included within the scope of the present invention.

Claims (7)

1. A pseudo-random code modulation-based medium-long range high-frequency laser radar ranging ambiguity solving method is characterized by comprising the following steps of:
establishing an echo waveform characteristic information and echo distance information model by using a BP neural network, and calculating all potential measurement distance intervals and distribution probabilities of all intervals of a single measurement echo, wherein the echo waveform morphological characteristic is an input item of the BP neural network;
further optimizing the input items of the measurement interval of the BP neural network by utilizing a waveform decomposition algorithm, and calculating the measurement distance interval of the echo;
modulating laser pulse emission time by utilizing a pseudo-random binary sequence, and calculating an accurate echo detection interval according to a criterion of minimizing noise energy of a sequence target reflected echo;
wherein: the specific process of establishing the BP neural network is as follows: the method comprises the steps of collecting a single laser full-waveform echo signal by using a low-frequency laser, using extracted waveform morphological characteristic information as an input item of a BP (Back propagation) neural network, calculating a waveform interval MTA [ i ] where the same echo is collected in the high-frequency laser according to a Time scale collected by the low-frequency laser, defining an interval value i of an MTA (multiple Time around) as the Time when the echo is received to be between the Time of i-1 pulse repetition intervals and the i pulse repetition intervals, establishing a 4-layer BP neural network by using the waveform interval MTA [ i ] where the echo is located and the probability of the echo in different measurement intervals as output items of the neural network, and respectively selecting a linear function and a piecewise function by using an intermediate state transfer function.
2. The pseudo-random code modulation-based medium-long range high frequency lidar ranging ambiguity resolution method of claim 1, wherein:
the echo waveform morphological characteristics comprise an echo comprehensive energy parameter, an echo signal-to-noise ratio parameter, an echo average width parameter, a height parameter, an average energy parameter and a gravity center position parameter.
3. The pseudo-random code modulation-based medium-long range high frequency lidar ranging ambiguity resolution method of claim 1, wherein: the waveform decomposition algorithm is a Levenberg-Marquardt decomposition algorithm.
4. The pseudo-random code modulation-based medium-long range high frequency lidar ranging ambiguity resolution method of claim 1, wherein:
the optimization process for further optimizing the measurement interval input item of the BP neural network by using the waveform decomposition algorithm is as follows:
according to the continuity characteristics of regional targets and the adoption of an optimization algorithm, the optimized measurement distance range MTA [ X ] of the measurement sequence can be further calculated](ii) a Assuming a total of N MTA regions, the laser emits the ith pulse at time tT[i]The moment when the detector detects the jth echo is tE[j]The number of laser pulses emitted at this time is Gj]The result output by the front-segment neural network can obtain the distribution probability p of the result in the MTA zone k (k is 1,2 … N)j,kWhich measures the distance Lj,kHas the following relationship:
Ru×(k-1)≤Lj,k≤Ru×k
Figure FDA0002629663210000021
Figure FDA0002629663210000022
then for M consecutive echoes starting from the 1 st, the probability distribution matrix and the measured distance matrix at these N MTA intervals are:
Figure FDA0002629663210000023
Figure FDA0002629663210000024
let the expected vector of the MTA interval of these M consecutive echoes be:
EMTA＝[E1,E2,...,EM]T,
Figure FDA0002629663210000025
the expected vectors for the measured distances are:
EL＝[E1,E2,...,EM]T,
Figure FDA0002629663210000026
the geometric shape of the ground object target can be regarded as continuous in a local range, and the situation that the size exceeds one MTA area does not occur; the desired measured distances of the M echoes are fitted with a straight line f using least squares estimation, i.e. a formula
Figure FDA0002629663210000027
The value is minimum, and the optimal estimation value of each echo measuring distance, namely the optimal estimation value of the MTA interval can be obtained:
Figure FDA0002629663210000031
5. the pseudo-random code modulation-based medium-long range high frequency lidar ranging ambiguity resolution method of claim 1, wherein:
the pseudo-random binary sequence modulation is determined according to different topographic features, including mountainous areas, cities, rural areas and flat areas, the measurement confidence coefficients of different intervals are determined by using the statistical distance measurement values and the noise energy of different measurement intervals according to the specificity and complexity of a detection target, and the length and the delay length tau of the pseudo-random binary code are adjusted.
6. The pseudo-random code modulation-based medium-long range high frequency lidar ranging ambiguity resolution method of claim 4, wherein:
the specific process of the criterion of minimizing the noise energy of the sequence target reflection echo is as follows:
since the contribution of the low frequency components of the distance signal, which are constant within each MTA interval, to the noise energy is negligible, it is possible to calculate the difference vector Δ R of successive measurementsMTA,jThe method comprises the following steps:
△RMTA,j＝(r2-r1,r3-r2,...,rM-rM-1)T
from each distance vector RMTA,jHigh frequency noise components are eliminated, then each difference vector Δ R associated with all potential MTA intervalsMTA,jNoise energy E of△RMTA,jCan be given by:
Figure FDA0002629663210000032
minimum noise energy
Figure FDA0002629663210000033
Not affected by PPM, the j value at this time is associated with the correct MTA interval; the MTA interval in which the minimum noise energy is located can be selected as a result by comparing the M measurement distance groups in the MTA interval for each measurement distance group assigned to all potential MTA intervals.
7. The pseudo-random code modulation-based medium-long range high frequency lidar ranging ambiguity resolution method of claim 6, wherein:
the waveform measurement interval and the correctness obtained by the noise energy minimization calculation of the sequence target reflection echo can be used for simultaneously extracting the morphological characteristic information of the echo, and further used as the prior training sample information of the BP neural network to participate in model training.
CN201810965428.8A 2018-08-23 2018-08-23 Pseudo-random code modulation-based medium-long range high-frequency laser radar ranging ambiguity solving method Active CN109164454B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810965428.8A CN109164454B (en) 2018-08-23 2018-08-23 Pseudo-random code modulation-based medium-long range high-frequency laser radar ranging ambiguity solving method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810965428.8A CN109164454B (en) 2018-08-23 2018-08-23 Pseudo-random code modulation-based medium-long range high-frequency laser radar ranging ambiguity solving method

Publications (2)

Publication Number Publication Date
CN109164454A CN109164454A (en) 2019-01-08
CN109164454B true CN109164454B (en) 2020-10-30

Family

ID=64896424

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810965428.8A Active CN109164454B (en) 2018-08-23 2018-08-23 Pseudo-random code modulation-based medium-long range high-frequency laser radar ranging ambiguity solving method

Country Status (1)

Country Link
CN (1) CN109164454B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110221241A (en) * 2019-04-29 2019-09-10 西安电子科技大学 A kind of low elevation angle DOA estimation method based on RBF neural
CN110208814B (en) * 2019-05-17 2022-07-08 深圳市速腾聚创科技有限公司 Laser radar and anti-interference method thereof
CN111308489B (en) * 2020-03-16 2022-02-11 杭州爱莱达科技有限公司 Optical domain analysis focusing method and device of pseudo-random code phase modulation laser radar
CN113204027B (en) * 2021-05-06 2024-06-11 武汉海达数云技术有限公司 Pulse type laser radar cross-period ranging method capable of precisely selecting ranging period

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5227784A (en) * 1990-12-10 1993-07-13 Mazda Motor Corporation System for detecting and determining range of target vehicle
CN104931973B (en) * 2015-06-17 2017-05-10 浙江理工大学 Asymmetric phase encoding ranging method applied to laser radar system
CN105738890B (en) * 2016-02-25 2017-12-08 深圳承泰科技有限公司 A kind of method and radar of the measurement range and measurement accuracy for lifting radar
CN106054204B (en) * 2016-07-26 2018-08-17 北京邮电大学 A kind of composite laser distance measuring method and system towards long distance and high precision
CN107402386B (en) * 2017-08-02 2019-11-19 中南大学 A kind of estimation method of underground metalliferous round tube radius and buried depth based on BP neural network

Also Published As

Publication number Publication date
CN109164454A (en) 2019-01-08

Similar Documents

Publication Publication Date Title
CN109164454B (en) Pseudo-random code modulation-based medium-long range high-frequency laser radar ranging ambiguity solving method
JP3798247B2 (en) A radar ice sounder with parallel Doppler processing.
US5252980A (en) Target location system
JP2990095B2 (en) Multiple pulse, multiple reflection, modal distance measurement processing system for clutter elimination
CN107976660B (en) Missile-borne multi-channel radar ultra-low-altitude target analysis and multi-path echo modeling method
CN109541619A (en) The appraisal procedure that speckle coherence influences single-photon laser radar ranging accuracy
CN108445471A (en) A kind of range accuracy appraisal procedure under the conditions of single-photon laser radar multi-detector
CN106872969B (en) Radar target angle estimation method based on MTD pulse accumulation and sliding processing
CN110927735B (en) Multi-target distance measuring method based on multi-channel full-waveform laser radar data
CN101839981A (en) Method and device for acquiring laser imaging echo waveform and level characteristics
CN114089366A (en) Water body optical parameter inversion method of satellite-borne single photon laser radar
CN110146873B (en) Target position and speed estimation method of distributed non-coherent radar
CN114942428B (en) Multi-target tracking hybrid radar data association algorithm
CN110488293A (en) A kind of distributed SAR system of nonuniform space configuration
CN114637021B (en) Sub-centimeter-level full-waveform laser radar ranging method and device
CN107526089A (en) A kind of non-based on time delay second order difference regards radar signal passive location method altogether
CN114235173A (en) Photon counting satellite-borne marine laser radar detection simulation method
CN113589275A (en) Combined observation method for sea surface wind field, sea waves and ocean current
CN111007490A (en) Sky wave over-the-horizon radar coordinate registration method based on buoy geographic information
RU2667484C1 (en) Method for determining the trajectory of movement of low-flying targets
CN116125406A (en) Method for evaluating performance of space-based surveillance radar based on SNR estimation and track-pointing track report
KR102375061B1 (en) Method for Eliminating False Detections due to Spiky Returns in Coherent Radars
CN113359145A (en) Target accurate positioning method in pulse laser ranging and application thereof
RU2392640C1 (en) Method for identification of parametres of trajectory instabilities of small-sized flying object in form of radial acceleration of motion for accompaniment mode with help of signals with per pulse carrier frequency tuning
RU2729886C1 (en) Method for passive jamming suppression with low doppler shift

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