CN104820217A - Calibration method for multi-element linear array detection imaging laser radar with multiple normal planes - Google Patents

Abstract

The invention relates to a calibration method for a multi-element linear array detection imaging laser radar with multiple normal planes, which comprises the following steps: 1) acquiring observation data of a plurality of planar target plates in different directions, fitting a plane equation of the target plates; 2) acquiring point cloud data of a plurality of planar target plates in a calibration area and performing preprocessing, establishing a multi-element laser radar system error calibration model based on the multiple normal planes; 3) according to the system error calibration model, obtaining a preliminary system error calibration parameter through Iterative adjustment calculation; 4) performing rough error elimination on the preliminary system error calibration parameter, for obtaining an optimal calibration parameter, determining whether an observed value which is larger than a given threshold is contained, if yes, eliminating the rough error value and returning to the step 3), and otherwise, performing a step 5); and 5), evaluating the precision of the preliminary system error calibration parameter, and obtaining a final calibration parameter. Compared with the prior art, the calibration method has advantages of realizing a full-parameter calibration model, simplifying establishment of a calibration experiment field, obtaining an accurate test result, etc.

Description

A kind of calibration method of polynary linear array detection imaging laser radar of many Normal planes

Technical field

The present invention relates to a kind of calibration method of radar, especially relate to a kind of calibration method of polynary linear array detection imaging laser radar of many Normal planes.

Background technology

Three-dimensional imaging laser radar has three-D profile and appearance imaging capability, can obtain the multiple image of target, as Range Profile, intensity image and distance-angle picture etc., there is high Distance geometry angular resolution, characteristics of image is stablized, and antijamming capability is strong, and anti-target stealth capabilities is strong.Polynary linear array detection imaging laser radar adopts linear array detector to receive, the Distance geometry echo strength data of row 16 positions can be obtained at every turn, to sweep or two dimensional opti mechanism scan realizes the collection of all pixel datas by pushing away, the multiple laser spots of final acquisition and intensity image, its imaging rate is far above single-element detector two-dimensional scan imaging technique, and due to linear array detector product abundanter, array element number is many, higher resolution image can be obtained, it is a kind of imaging detection technology having much advantage and potential, be suitable for keeping away barrier accurately, navigation, guidance, landing drop point is selected, spacecraft space spacecrafts rendezvous, remote sensing of the earth is observed, topographic mapping, unmanned vehicle and robot keep away barrier, navigation etc., in the large field of the army and the people two, all there is vast potential for future development and application demand.Polynary linear array detection imaging laser radar is subject to the impact of multiple error component in measuring process, comprises multiple laser ranging error, angle error and other errors etc.In order to ensure the quality of data, modeling must be carried out on the error component of the polynary linear array detection imaging laser infrared radar imaging quality of impact, calculating correlated error, improve the quality of polynary linear array detection imaging laser radar.

Summary of the invention

Object of the present invention is exactly providing a kind of population parameter calibration model to overcome defect that above-mentioned prior art exists, simplifying that calibration field experiment is set up, the calibration method of the polynary linear array detection imaging laser radar of the accurate many Normal planes of experimental result.

Object of the present invention can be achieved through the following technical solutions:

A calibration method for the polynary linear array detection imaging laser radar of many Normal planes, comprises the following steps:

1) the High Accuracy Observation data of the plane target plank of multiple different azimuth are obtained, matching target plate plane equation;

2) cloud data of multiple plane target plank in calibration region is obtained by polynary linear array detection imaging laser radar, comprise the polar value of polynary Distance geometry angle, pre-service is carried out to cloud data, sets up the multilasered optical radar systematic error calibration model based on many Normal planes;

3) according to based on the multilasered optical radar systematic error calibration model of many Normal planes, repeating adjustmet is carried out to the systematic error of polynary linear array detection imaging laser radar and elements of exterior orientation and resolves, obtain rudimentary system error calibration parameter;

4) carry out elimination of rough difference to rudimentary system error calibration parameter, the residual distribution according to adjustment result carries out statistical study, judges whether the observed reading containing being greater than given threshold value, if, then rough error value rejected and return step 3), if not, then carry out step 5);

5) evaluate the precision of the rudimentary system error calibration parameter after elimination of rough difference according to the cloud data after correcting and the distance offsets of target plate plane equation, obtain final calibration parameter.

Described step 1) specifically comprise the following steps:

11) by high precision total station, target plate plane is observed, obtain high-precision three-dimensional volume coordinate plane being uniformly distributed discrete point;

12) least square method is utilized to obtain target plate plane equation to three dimensional space coordinate matching.

Step 2 described in 3) in pre-service be by the polar value (x' of cloud data under laser scanning coordinate system, y', z') the rectangular coordinate value (X' based on total powerstation rectangular coordinate system is converted to, Y', Z'), the polar value of described cloud data comprises radial distance, horizontal angle and vertical angle, and belongs to situation according to each plane of some cloud, carry out segmentation to cloud data to extract, set up identiflication number.

The formula that the polar value of described cloud data is converted to rectangular coordinate value (X', Y', Z') is:

$n=\left[\begin{array}{c}-\sin 2{{\mathrm{\θ}}_x}^{\′}\\ -\cos 2{{\mathrm{\θ}}_x}^{\′}\\ \cos 2{{\mathrm{\θ}}_x}^{\′}\·\cos 2{{\mathrm{\θ}}_y}^{\′}-\sin 2{{\mathrm{\θ}}_y}^{\′}\·\tan {\mathrm{\θ}}_k\end{array},\·\sin 2{{\mathrm{\θ}}_y}^{\′}-\cos 2{{\mathrm{\θ}}_y}^{\′}\·\tan {\mathrm{\θ}}_k\right]$

Wherein, (X ₀, Y ₀, Z ₀) be the translation parameters of laser scanning coordinate system and total station instrument coordinate system, for the rotation angle between laser scanning coordinate system and total powerstation rectangular coordinate system, translation parameters and rotation angle are called elements of exterior orientation, ρ ' _ifor the distance measure after correction, θ ' _xfor the horizontal angle surveying value after correction, θ ' _yfor the vertical angle measured value after correction, θ _ifor polynary linear array detection imaging laser radar each probe unit between vertical angle, b and e is known Instrument Design parameter.

Described step 2) in the multilasered optical radar systematic error calibration model based on many Normal planes be:

${{\mathrm{\ρ}}_i}^{\′}={\mathrm{\ρ}}_i+v_{{\mathrm{\ρ}}_i}+\mathrm{\Δ}{\mathrm{\ρ}}_i$

Δρ _i＝A _i

${{\mathrm{\θ}}_x}^{\′}={\mathrm{\θ}}_x+v_{{\mathrm{\θ}}_x}+\mathrm{\Δ}{\mathrm{\θ}}_x$

Δθ _x＝B ₁+B ₂sec(θ _y)+B ₃tan(θ _y)

${{\mathrm{\θ}}_y}^{\′}={\mathrm{\θ}}_y+v_{{\mathrm{\θ}}_y}+\mathrm{\Δ}{\mathrm{\θ}}_y$

Δθ _y＝C ₁

Wherein, ρ _ibe i-th probe unit observed reading, be the stochastic error of i-th probe unit distance measurement value, Δ ρ _ibe the systematic error of i-th probe unit range finding, A _ibe the biased error at zero point of i-th probe unit, θ ' _xfor the horizontal angle value after polynary linear array detection imaging laser radar correction, θ _xfor the observed reading of polynary linear array detection imaging laser radar horizontal angle, for the correction of azimuth observation value, Δ θ _xfor the systematic error of horizontal angle measurement, B ₁for the motor offset error of horizontal pendulum mirror, B ₂for boresight misalignments, B ₃for horizontal axis error, C ₁for vertical pendulum mirror error, θ ' _yfor the vertical angle value after polynary linear array detection imaging laser radar correction, Δ θ _yfor the systematic error of vertical angle, θ _yfor the observed reading of polynary linear array detection imaging laser radar vertical angle, for the correction of Vertical right angle observation value.

Described step 3) in the Iterative of systematic error calibration model of polynary linear array detection imaging laser radar comprise the following steps:

31) supposition unit normal vector of certain plane j under total station instrument coordinate system is n _j=(a _j, b _j, c _j), the initial point of total station instrument coordinate system is d to the vertical range of this plane j _j, any point in this plane can be expressed as:

$[a_j,b_j,c_j]\left[\begin{array}{c}{X}^{\′}\\ Y^{\′}\\ Z^{\′}\end{array}\right]-d_j=0;$

32) the observed reading collection L (ρ of polynary linear array detection imaging laser radar is set ₁, ρ ₂..., ρ _k, θ _x, θ _y) and undetermined parameter collection then step 31) in the Representation Equation be:

$f(\hat{L},\hat{L})=0$

Wherein, for the maximal possibility estimation collection of L, for the maximal possibility estimation collection of T;

33) to step 32) in formula carry out linearization, obtaining the formula after linearization is:

$f(L_0,T_0)+\frac{\∂f}{\∂\hat{L}}{|L_0,T_0\·V+\frac{\∂f}{\∂\hat{T}}|}_{L_0,T_0}\·t=0$

Wherein, V is the correction of observed quantity approximate value, and t is the correction of error parameter approximate value, L ₀for initial value, T ₀for initial value, v _{ρ 1}, v _{ρ 2}v _{ρ k}for the correction of three-dimensional laser scanner k probe unit distance observed reading altogether, Δ A ₁, Δ A ₂... Δ A _kfor the correction of k distance error parameter approximate value, Δ B ₁, Δ B ₂, Δ B ₃, Δ C ₁be respectively the correction of boresight misalignments parameter, the correction of horizontal axis error parameter and the correction of vertical angle error parameter approximate value, Δ ω, Δ κ, Δ X, Δ Y, Δ Z are respectively the correction of 6 elements of exterior orientation error parameter approximate values;

34) set up Iterative equation, Iterative equation is:

$\underset{m\×n}{A}\underset{n\×1}{V}+\underset{m\×u}{B}\underset{u\×1}{t}-\underset{m\×1}{W}=0$

$\left\{\begin{array}{c}A=\frac{\∂f}{\∂\hat{L}}{|}_{L_0,T_0}\\ B=\frac{\∂f}{\∂\hat{T}}{|}_{L_0,T_0}\\ W=-f(L_0,T_0)+A\·(L_0-L)\end{array}\right.$

Wherein, m is the number of the point participating in adjustment, and n is observed reading number and n=k+2, u are the number of undetermined parameter, and u=k+10;

35) carry out Iterative to Iterative equation, iteration is initial, setting initial value equal L, initial value equal 0, in adjustment iterative process each time, L ₀and T ₀upgrade according to the correction after last adjustment, obtain rudimentary system error calibration parameter.

Compared with prior art, the present invention has the following advantages:

One, population parameter calibration model: the calibration method proposing a kind of polynary linear array detection imaging laser radar of many Normal planes, establish the population parameter calibration model of polynary linear array laser radar, the elements of exterior orientation correction overall adjustment of systematic error, stochastic error and two coordinate systems is calculated, consider that it is more accurate to calculate comprehensively.

Two, simplify calibration field experiment to set up: this method uses the plane target plank of multiple different directions as reference plane, sets up calibration place, can fast, the systematic error calibration work completing multilasered optical radar of colleges and universities.

Three, experimental result is accurate: experiment uses systematic error calibration method of the present invention, makes precision improvement be up to 90%, effectively improves image quality and the measuring accuracy of polynary linear array detection imaging laser radar.

Accompanying drawing explanation

Fig. 1 is the method flow diagram of the embodiment of the present invention.

Fig. 2 is the point cloud chart in embodiment before calibration.

Fig. 3 is the point cloud chart in embodiment after calibration.

Embodiment

Below in conjunction with the drawings and specific embodiments, the present invention is described in detail.

Embodiment:

The present invention relates to a kind of calibration method of polynary linear array detection imaging laser radar of many Normal planes, its concrete steps are as follows:

1) acquisition of plane target plank reference data: utilize high precision total station to observe target plate plane, obtain high-precision three-dimensional volume coordinate plane being uniformly distributed discrete point, and utilize least square method to simulate the Datum Equations of plane according to surving coordinate;

2) acquisition of plane target plank scanner cloud data: use polynary linear array detection imaging laser radar, obtain the cloud data in calibration region, and by original polar value (radial distance, horizontal angle and vertical angle), convert rectangular coordinate value (X' to, Y', Z'); According to each plane ownership situation of some cloud, segmentation is carried out to the cloud data of polynary linear array detection imaging laser radar and extracts, and identiflication number is set up to it;

3) the systematic error calibration based on plane is resolved: utilize the multilasered optical radar systematic error calibration model based on plane, adopts the method that repeating adjustmet resolves;

4) resolve the robust estimation of parameter: utilize the residual distribution of adjustment result to carry out statistical study, detect the corresponding raw observation being greater than given threshold value, rough error value is rejected and re-started previous step and resolves, until calculate best calibration parameter;

5) accuracy assessment of parameter is resolved: systematic error parameter will be resolved and carry out after correction and reference value comparative analysis to raw observation, and count the precision of final argument.

This polynary linear array detection of three dimensional laser radar scanner adopts 16 laser acquisition unit, horizontal and vertical angular samples rate is 2mrad, in the scope of 50 ~ 100m, the interval of adjacent analyzing spot can reach 10cm ~ 20cm, is namely about 25 points/m2 in the laser spots density of 100m place scanning.Calibration field puts the surface plate that 11 have different angle and orientation altogether, and each plane is of a size of 4m × 2m, and from left to right successively compile be 1 to No. 11, for the calibration of laser radar scanner systematic error.

Original point cloud as shown in Figure 2, there is range finding and the angle error of multiple probe unit in the polynary linear array detection imaging laser radar due to non-calibration, and the systematic error of each probe unit is not identical, therefore original point cloud cannot reflect real planimetric position and shape, the motor of horizontal scanning simultaneously also exists angular error, and plane also exists obvious inconsistent phenomenon on different scanning band.

The calibration method of the polynary linear array detection imaging laser radar based on plane using the present invention to propose, carries out calibration to polynary linear array detection imaging laser radar.As shown in Figure 3, energy actual response out-of-plane shape, otherness larger between original different probe units existed, also basic correction, the inconsistent phenomenon between each scanning strip is eliminated some cloud after calibration.

Matching is carried out to plane each before and after calibration, and calculates the range deviation (i.e. plane fitting error) of each point cloud to respective planes equation, evaluation calibration quality.Each as can be seen from table 1, the flatness of each plane, medial error before calibration is substantially at about 30cm, after calibration, error significantly reduces, except No. 1 of most edge and No. 11 planes, the flatness of all the other planes is all better than 4cm, precision is the highest improves 90%, due to the Hardware Design impact, the polynary linear array detection imaging laser radar optimum level field angle of experiment is ± 12 °, 1 and No. 11 plane is positioned at the most edge of horizontal field of view, 13 ° are reached with the angle of field of view center line, although not within best field range, but after calibration flatness also can reach 7.5cm, prove that the method effectively can improve measuring accuracy and the image quality of polynary linear array detection imaging laser radar.

Table 1 calibration anterior-posterior plane precision

The calibration method of the polynary linear array detection imaging laser radar of a kind of many Normal planes that this patent proposes, do not needing to use on special hardware device basis, unified Modeling is carried out to the stochastic error of elements of exterior orientation, polynary errors of the distance measurement system, horizontal angle error, vertical angle error and each observed reading, and use the method based on plane, accurate Calculation is carried out to systematic error.Context of methods is used to carry out systematic error calibration to domestic polynary linear array detection imaging laser radar, planar smoothness after calibration can be better than 4cm, precision improvement is up to 90%, effectively improves image quality and the measuring accuracy of polynary linear array detection imaging laser radar.

Claims (6)

1. a calibration method for the polynary linear array detection imaging laser radar of Normal plane more than, is characterized in that, comprise the following steps:

1) the High Accuracy Observation data of the plane target plank of multiple different azimuth are obtained, matching target plate plane equation;

2) cloud data of multiple plane target plank in calibration region is obtained by polynary linear array detection imaging laser radar, comprise the polar value of polynary Distance geometry angle, pre-service is carried out to cloud data, sets up the multilasered optical radar systematic error calibration model based on many Normal planes;

3) according to based on the multilasered optical radar systematic error calibration model of many Normal planes, repeating adjustmet is carried out to the systematic error of polynary linear array detection imaging laser radar and elements of exterior orientation and resolves, obtain rudimentary system error calibration parameter;

4) carry out elimination of rough difference to rudimentary system error calibration parameter, the residual distribution according to adjustment result carries out statistical study, judges whether the observed reading containing being greater than given threshold value, if, then rough error value rejected and return step 3), if not, then carry out step 5);

5) evaluate the precision of the rudimentary system error calibration parameter after elimination of rough difference according to the cloud data after correcting and the distance offsets of target plate plane equation, obtain final calibration parameter.

2. the calibration method of the polynary linear array detection imaging laser radar of a kind of many Normal planes according to claim 1, is characterized in that, described step 1) specifically comprise the following steps:

11) by high precision total station, target plate plane is observed, obtain high-precision three-dimensional volume coordinate plane being uniformly distributed discrete point;

12) least square method is utilized to obtain target plate plane equation to three dimensional space coordinate matching.

3. the calibration method of the polynary linear array detection imaging laser radar of a kind of many Normal planes according to claim 1, it is characterized in that, described step 2) in pre-service be by the polar value (x' of cloud data under laser scanning coordinate system, y', z') the rectangular coordinate value (X' based on total powerstation rectangular coordinate system is converted to, Y', Z'), the polar value of described cloud data comprises radial distance, horizontal angle and vertical angle, and belong to situation according to each plane of some cloud, carry out segmentation to cloud data to extract, set up identiflication number.

4. the calibration method of the polynary linear array detection imaging laser radar of a kind of many Normal planes according to claim 3, is characterized in that, the formula that the polar value of described cloud data is converted to rectangular coordinate value (X', Y', Z') is:

$n=\left[\begin{array}{c}-\sin 2{{\mathrm{\θ}}_x}^{\′}\\ -\cos 2{{\mathrm{\θ}}_x}^{\′}\·\sin 2{{\mathrm{\θ}}_y}^{\′}-\cos 2{{\mathrm{\θ}}_y}^{\′}\·\tan {\mathrm{\θ}}_k\\ \cos 2{{\mathrm{\θ}}_x}^{\′}\·\cos 2{{\mathrm{\θ}}_y}^{\′}-\sin 2{{\mathrm{\θ}}_y}^{\′}\·\tan {\mathrm{\θ}}_k\end{array}\right]$

Wherein, (X ₀, Y ₀, Z ₀) be the translation parameters of laser scanning coordinate system and total station instrument coordinate system, for the rotation angle between laser scanning coordinate system and total powerstation rectangular coordinate system, translation parameters and rotation angle are called elements of exterior orientation, ρ _i' be the distance measure after correction, θ _x' be the horizontal angle surveying value after correction, θ _y' be the vertical angle measured value after correction, θ _ifor polynary linear array detection imaging laser radar each probe unit between vertical angle, b and e is known Instrument Design parameter.

5. the calibration method of the polynary linear array detection imaging laser radar of a kind of many Normal planes according to claim 1, is characterized in that, described step 2) in the multilasered optical radar systematic error calibration model based on many Normal planes be:

${{\mathrm{\ρ}}_i}^{\′}={\mathrm{\ρ}}_i+{v_{\mathrm{\ρ}}}_i+\mathrm{\Δ}{\mathrm{\ρ}}_i$

Δρ _i＝A _i

${{\mathrm{\θ}}_x}^{\′}={\mathrm{\θ}}_x+{v_{\mathrm{\θ}}}_x+\mathrm{\Δ}{\mathrm{\θ}}_x$

Δθ _x＝B ₁+B ₂sec(θ _y)+B ₃tan(θ _y)

${{\mathrm{\θ}}_y}^{\′}={\mathrm{\θ}}_y+{v_{\mathrm{\θ}}}_y+\mathrm{\Δ}{\mathrm{\θ}}_y$

Δθ _y＝C ₁

Wherein, ρ _ibe i-th probe unit observed reading, be the stochastic error of i-th probe unit distance measurement value, Δ ρ _ibe the systematic error of i-th probe unit range finding, A _ibe the biased error at zero point of i-th probe unit, θ _x' be the horizontal angle value after polynary linear array detection imaging laser radar correction, θ _xfor the observed reading of polynary linear array detection imaging laser radar horizontal angle, for the correction of azimuth observation value, Δ θ _xfor the systematic error of horizontal angle measurement, B ₁for the motor offset error of horizontal pendulum mirror, B ₂for boresight misalignments, B ₃for horizontal axis error, C ₁for vertical pendulum mirror error, θ _y' be the vertical angle value after polynary linear array detection imaging laser radar correction, Δ θ _yfor the systematic error of vertical angle, θ _yfor the observed reading of polynary linear array detection imaging laser radar vertical angle, for the correction of Vertical right angle observation value.

6. the calibration method of the polynary linear array detection imaging laser radar of a kind of many Normal planes according to claim 1, it is characterized in that, described step 3) in the Iterative of systematic error calibration model of polynary linear array detection imaging laser radar comprise the following steps:

31) supposition unit normal vector of certain plane j under total station instrument coordinate system is n _j=(a _j, b _j, c _j), the initial point of total station instrument coordinate system is d to the vertical range of this plane j _j, any point in this plane can be expressed as:

$[a_j,b_j,c_j]\left[\begin{array}{c}{X}^{\′}\\ Y^{\′}\\ Z^{\′}\end{array}\right]-d_j=0;$

32) the observed reading collection L (ρ of polynary linear array detection imaging laser radar is set ₁, ρ ₂..., ρ _k, θ _x, θ _y) and undetermined parameter collection then step 31) in the Representation Equation be:

$f(\hat{L},\hat{T})=0$

Wherein, for the maximal possibility estimation collection of L, for the maximal possibility estimation collection of T;

33) to step 32) in formula carry out linearization, obtaining the formula after linearization is:

$f(L_0,T_0)+{\frac{\∂f}{\∂\hat{L}}|}_{L_0,T_0}\·V+{\frac{\∂f}{\∂\hat{T}}|}_{L_0,T_0}\·t=0$

Wherein, V is the correction of observed quantity approximate value, and t is the correction of error parameter approximate value, L ₀for initial value, T ₀for initial value, for the correction of three-dimensional laser scanner k probe unit distance observed reading altogether, Δ A ₁, Δ A ₂... Δ A _kfor the correction of k distance error parameter approximate value, Δ B ₁, Δ B ₂, Δ B ₃, Δ C ₁be respectively the correction of boresight misalignments parameter, the correction of horizontal axis error parameter and the correction of vertical angle error parameter approximate value, Δ ω, Δ κ, Δ X, Δ Y, Δ Z are respectively the correction of 6 elements of exterior orientation error parameter approximate values;

34) set up Iterative equation, Iterative equation is:

$\underset{m\×n}{A}\underset{n\×1}{V}+\underset{m\×u}{B}\underset{u\×1}{t}-\underset{m\×1}{W}=0$

$\left\{\begin{array}{c}A={\frac{\∂f}{\∂\hat{L}}|}_{L_0,T_0}\\ B={\frac{\∂f}{\∂\hat{T}}|}_{L_0,T_0}\\ W=-f(L_0,T_0)+A\·(L_0-L)\end{array}\right.$

Wherein, m is the number of the point participating in adjustment, and n is observed reading number and n=k+2, u are the number of undetermined parameter, and u=k+10;

35) carry out Iterative to Iterative equation, iteration is initial, setting initial value equal L, initial value equal 0, in adjustment iterative process each time, L ₀and T ₀upgrade according to the correction after last adjustment, obtain rudimentary system error calibration parameter.