CN108459314A - A kind of three-dimensional solid-state face battle array laser radar non-uniform correction method - Google Patents
A kind of three-dimensional solid-state face battle array laser radar non-uniform correction method Download PDFInfo
- Publication number
- CN108459314A CN108459314A CN201810189563.8A CN201810189563A CN108459314A CN 108459314 A CN108459314 A CN 108459314A CN 201810189563 A CN201810189563 A CN 201810189563A CN 108459314 A CN108459314 A CN 108459314A
- Authority
- CN
- China
- Prior art keywords
- particle
- weights
- frame
- state
- battle array
- 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.)
- Granted
Links
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
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/48—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S17/00
- G01S7/497—Means for monitoring or calibrating
Abstract
The invention discloses a kind of three-dimensional solid-state face battle array laser radar non-uniform correction method, the problem of nonlinear noise during nonuniformity correction can be well solved.Scheme is：Range image is acquired using three-dimensional solid-state face battle array laser radar, collected range image is divided into K frame block along time shaft.The distribution function of heterogeneity noise satisfaction is determined according to priori and carries out n times particle stochastical sampling, and the initial weight of each particle is 1/N.The new state that each particle is estimated using state transition model makes particle that can carry out state transfer according to corresponding mode.According to time series using maximum Likelihood carry out variance calculating, further more in new frame block each particle weights；One measurement standard Neff is set to particle weights, weights are less than to the particle of Neff, the particle position according to weights more than Neff is redistributed.It calculates heterogeneity noise and actually measured laser signal is combined to carry out real time correction to three-dimensional solid-state face battle array laser radar.
Description
Technical field
The present invention relates to three-dimensional laser radar depth data alignment technique fields, and in particular to a kind of three-dimensional solid-state face battle array is sharp
Optical radar non-uniform correction method.
Background technology
Since manufacturing process limits, it is special to show different responses for detector different pixels in the battle array laser radar of three-dimensional solid-state face
Property, while three-dimensional solid-state face battle array laser radar, since the factors such as temperature change, circuit interference influence, makes after working long hours
Heterogeneity drift occurs for three-dimensional solid-state face battle array laser radar depth data.
Currently, many non-uniformity correction algorithms have been proposed both at home and abroad, it can generally speaking be divided into two classes：One kind is base
In the method for calibration, this method has higher correction accuracy, but correction system bulk is larger, and cost is also higher, and
The necessary break-off of imaging system, seriously constrains the use scope of such algorithm in calibration process.Another kind of is to be based on scene
Method, such as Kalman filtering method, particle filter method, Kalman filtering method state-space model in design is necessary for line
Property and Gauss, and this condition is extremely difficult to during practical nonuniformity correction, and in addition particle filter method is in design when side
Difference using fixed value or only by the variable quantity approximate calculation of scene, prediction error is larger and particle distribution convergence rate compared with
Slowly, it influences to correct real-time.
Therefore it needs to solve non-linear making an uproar when carrying out Nonuniformity Correction to three-dimensional solid-state face battle array laser radar at present
Sound estimates and the slow problem of processing speed.
Invention content
In view of this, the present invention provides a kind of three-dimensional solid-state face battle array laser radar non-uniform correction method, it can be fine
Ground solves the problems, such as that nonlinear noise estimation and processing speed are slow during nonuniformity correction, and accuracy is high, real-time is high, meter
Calculation amount is small.
The technical scheme is that：
A kind of three-dimensional solid-state face battle array laser radar non-uniform correction method, includes the following steps：
Step 1: range image sequence is obtained using the battle array laser radar acquisition of three-dimensional solid-state face, by collected distance map
As sequence is divided into K frame block along time shaft.
Step 2: determining distribution function that heterogeneity noise meets according to priori and carrying out n times particle and adopt at random
The initial weight of sample, each particle is 1/N.
Step 3: estimating the new state of each particle using state transition model, make particle can be according to corresponding side
Formula carries out state transfer.
State transition model is Bk=f (Bk-1)+Wk-1。
Wherein BkFor k-th of frame heterogeneity noise in the block, Wk- 1 is k-th of frame driving noise in the block；F () is
Nonlinear function.
Step 4: the weights of i-th of particle are in k-th of frame block of update：
Wherein δk(n) it is adjacent two field pictures, i.e. difference between the (n+1)th frame image and n-th frame image in k-th of frame block
Value；For δk(n) mean value；Wherein Yk(1) it is the 1st frame image in k-th of frame block
Actually measured laser signal, for -1 frame block of kth, last piece image Yk-1(lk-1) correction result be For the heterogeneity noise component(s) of i-th of particle；δk(n) meet normal distribution, σ is standard deviation；And to institute
There are the weights of particle to be normalized.
Step 5: setting a measurement standard Neff to particle weights, weights are less than to the particle of threshold value, it is big according to weights
It is redistributed in the particle position of Neff.
Step 6: heterogeneity psophometer noise is：
According to heterogeneity noise, pass through observation model Yk(n)=Xk(n)+BkThree-dimensional solid-state face battle array laser radar is carried out
Real time correction.
Further, step 4 comprises the following specific steps that：
Adjacent two field pictures in S401, same frame block, i.e. difference between the (n+1)th frame image and n-th frame image are：
δk(n)=Yk(n+1)-Yk(n)=Xk(n+1)-Xk(n)
Total l in wherein k-th of frame blockkFrame image, n=1~lk；δk(n) meet normal distribution；Yk(n) it is actually measured
Laser signal；Xk(n) it is actual laser signal；
δk(n) mean value is：
S402、Yk(n) and Xk(n) meet：
Yk~N (Xk,σ2), Xk=Yk-Bk
YkAnd XkIt respectively refers to for Yk(n) and Xk(n)。
S403, maximum likelihood estimator are：
Wherein Y1:kRefer to Y1~Yk, X1:kRefer to X1~Xk。
S404, for l (σ | Y1:K,X1:K) seek σ2Partial derivative, and partial derivative is made to be that 0 can obtain：
S405, for linear system, the σ of t moment2It is expressed asThe σ at t+1 moment2It is expressed asWith's
Recurrence relation is as follows,
S406, for -1 frame block of kth, last piece image Yk-1(lk-1) correction result beConstruction becomes
Measure δkSo that：
It is for the variable of i-th of particle then：
δk(n) meet normal distribution
S407, then the weights of each particle are：
Further, measurement standard is：
Advantageous effect：
A kind of method based on recursion maximal possibility estimation disclosed by the invention, on the one hand can well solve non-homogeneous
Nonlinear noise estimation and the slow problem of processing speed, on the other hand make particle distribution rapidly converge to pixel defeated in correction course
The actual value gone out can guarantee that particle estimation is close with actual value in the case where the population of selection is less, improve
The accuracy of Nonuniformity Correction, and reduce calculation amount.This method uses the calculation of recursion, real-time according to time series
Calculating pixel output, have higher real-time, be highly suitable for handling in real time.
Description of the drawings
A kind of three-dimensional solid-state face battle array laser radar asymmetric correction method that Fig. 1 is provided by the embodiment of the present invention.
Specific implementation mode
The present invention will now be described in detail with reference to the accompanying drawings and examples.
The present invention provides a kind of three-dimensional solid-state face battle array laser radar asymmetric correction method, flow as shown in Figure 1,
Include the following steps：
Step 1: range image sequence is obtained using the battle array laser radar acquisition of three-dimensional solid-state face, by collected distance map
As sequence is divided into K frame block along time shaft；
Step 2: determining distribution function that heterogeneity noise meets according to priori and carrying out n times particle and adopt at random
The initial weight of sample, each particle is 1/N；
Step 3: estimating the new state of each particle using state transition model, make particle can be according to corresponding side
Formula carries out state transfer；
State transition model is Bk=f (Bk-1)+Wk-1；
Wherein BkFor k-th of frame heterogeneity noise in the block, Wk- 1 is k-th of frame driving noise in the block；F () is
Nonlinear function；
Step 4: carrying out the calculating of variance using maximum Likelihood according to time series, further update k-th
The weights of i-th of particle are in frame block：
Wherein δk(n) it is adjacent two field pictures, i.e. difference between the (n+1)th frame image and n-th frame image in k-th of frame block
Value；For δk(n) mean value；Wherein Yk(1) it is the 1st frame image in k-th of frame block
Actually measured laser signal, for -1 frame block of kth, last piece image Yk-1(lk-1) correction result be For the heterogeneity noise component(s) of i-th of particle；δk(n) meet normal distribution, σ is standard deviation；And to institute
There are the weights of particle to be normalized.
In the embodiment of the present invention, the weights of particle are sought using following specific steps：
Adjacent two field pictures in S401, same frame block, i.e. difference between the (n+1)th frame image and n-th frame image are：
δk(n)=Yk(n+1)-Yk(n)=Xk(n+1)-Xk(n)
Total l in wherein k-th of frame blockkFrame image, n=1~lk；δk(n) meet normal distribution；Yk(n) it is actually measured
Laser signal；Xk(n) it is actual laser signal；
δk(n) mean value is：
S402、Yk(n) and Xk(n) meet：
Yk~N (Xk,σ2), Xk=Yk-Bk
YkAnd XkIt respectively refers to for Yk(n) and Xk(n)。
S403, maximum likelihood estimator are：
Wherein Y1:kRefer to Y1~Yk, X1:kRefer to X1~Xk。
S404, for l (σ | Y1:K,X1:K) seek σ2Partial derivative, and partial derivative is made to be that 0 can obtain：
S405, for linear system, the σ of t moment2It is expressed asThe σ at t+1 moment2It is expressed asWith's
Recurrence relation is as follows,
S406, for -1 frame block of kth, last piece image Yk-1(lk-1) correction result beConstruction becomes
Measure δkSo that：
It is for the variable of i-th of particle then：
δk(n) meet normal distribution
S407, then the weights of each particle are：
Step 5: setting a measurement standard Neff to particle weights, weights are less than to the particle of threshold value, it is big according to weights
It is redistributed in the particle position of Neff.
In the embodiment of the present invention, measurement standard is set as：
Step 6: heterogeneity psophometer noise is：
According to heterogeneity noise, pass through observation model Yk(n)=Xk(n)+BkThree-dimensional solid-state face battle array laser radar is carried out
Real time correction.
A kind of method based on recursion maximal possibility estimation disclosed by the invention, on the one hand can well solve non-homogeneous
Nonlinear noise estimation and the slow problem of processing speed, on the other hand make particle distribution rapidly converge to pixel defeated in correction course
The actual value gone out can guarantee that particle estimation is close with actual value in the case where the population of selection is less, improve
The accuracy of Nonuniformity Correction, and reduce calculation amount.This method uses the calculation of recursion, real-time according to time series
Calculating pixel output, have higher real-time, be highly suitable for handling in real time.
In conclusion the above is merely preferred embodiments of the present invention, being not intended to limit the scope of the present invention.
All within the spirits and principles of the present invention, any modification, equivalent replacement, improvement and so on should be included in the present invention's
Within protection domain.
Claims (3)
1. a kind of three-dimensional solid-state face battle array laser radar non-uniform correction method, which is characterized in that include the following steps：
Step 1: range image sequence is obtained using the battle array laser radar acquisition of three-dimensional solid-state face, by collected range image sequence
Row are divided into K frame block along time shaft；
Step 2: determining the distribution function of heterogeneity noise satisfaction according to priori and carrying out n times particle stochastical sampling, often
The initial weight of a particle is 1/N；
Step 3: estimate the new state of each particle using state transition model, make particle can according to corresponding mode into
Row state shifts；
The state transition model is Bk=f (Bk-1)+Wk-1；
Wherein BkFor k-th of frame heterogeneity noise in the block, Wk- 1 is k-th of frame driving noise in the block；F () is non-thread
Property function；
Step 4: the weights of i-th of particle are in k-th of frame block of update：
Wherein δk(n) it is adjacent two field pictures, i.e. difference between the (n+1)th frame image and n-th frame image in k-th of frame block；For
δk(n) mean value；Wherein Yk(1) it is the practical survey of the 1st frame image in k-th of frame block
The laser signal obtained, for -1 frame block of kth, last piece image Yk-1(lk-1) correction result be It is
The heterogeneity noise component(s) of i particle；δk(n) meet normal distribution, σ is standard deviation；And the weights of all particles are carried out
Normalized；
Step 5: setting a measurement standard Neff to particle weights, weights are less than to the particle of threshold value, are more than according to weights
The particle position of Neff is redistributed；
Step 6: heterogeneity psophometer noise is：
According to heterogeneity noise, pass through observation model Yk(n)=Xk(n)+BkThree-dimensional solid-state face battle array laser radar is carried out real-time
Correction.
2. the method as described in claim 1, which is characterized in that the step 4 comprises the following specific steps that：
Adjacent two field pictures in S401, same frame block, i.e. difference between the (n+1)th frame image and n-th frame image are：
δk(n)=Yk(n+1)-Yk(n)=Xk(n+1)-Xk(n)
Total l in wherein k-th of frame blockkFrame image, n=1~lk；δk(n) meet normal distribution；Yk(n) it is actually measured laser
Signal；Xk(n) it is actual laser signal；
δk(n) mean value is：
S402、Yk(n) and Xk(n) meet：
Yk~N (Xk,σ2), Xk=Yk-Bk
YkAnd XkIt respectively refers to for Yk(n) and Xk(n)；
S403, maximum likelihood estimator are：
Wherein Y1:kRefer to Y1~Yk, X1:kRefer to X1~Xk；
S404, for l (σ | Y1:K,X1:K) seek σ2Partial derivative, and partial derivative is made to be that 0 can obtain：
S405, for linear system, the σ of t moment2It is expressed asThe σ at t+1 moment2It is expressed asWithRecursion close
System is as follows,
S406, for -1 frame block of kth, last piece image Yk-1(lk-1) correction result beConstructed variable δk,
So that：
It is for the variable of i-th of particle then：
δk(n) meet normal distribution
S407, then the weights of each particle are：
3. the method as described in claim 1, which is characterized in that the measurement standard is：
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810189563.8A CN108459314B (en) | 2018-03-08 | 2018-03-08 | Three-dimensional solid-state area array laser radar non-uniform correction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810189563.8A CN108459314B (en) | 2018-03-08 | 2018-03-08 | Three-dimensional solid-state area array laser radar non-uniform correction method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108459314A true CN108459314A (en) | 2018-08-28 |
| CN108459314B CN108459314B (en) | 2020-05-19 |
Family
ID=63217421
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810189563.8A Active CN108459314B (en) | 2018-03-08 | 2018-03-08 | Three-dimensional solid-state area array laser radar non-uniform correction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108459314B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111208491A (en) * | 2020-01-17 | 2020-05-29 | 岭纬科技（厦门）有限公司 | Method for eliminating miscellaneous points of high-resolution solid-state laser radar point cloud |
| WO2020139332A1 (en) * | 2018-12-24 | 2020-07-02 | Didi Research America, Llc | Real-time estimation of dc biac and noise power of light detection and ranging (lidar) |
| CN112346034A (en) * | 2020-11-12 | 2021-02-09 | 北京理工大学 | Three-dimensional solid-state area array laser radar combination calibration method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101448338A (en) * | 2008-11-27 | 2009-06-03 | 北京理工大学 | Mobile three-dimensional spatial wireless testing network system |
| CN101699313A (en) * | 2009-09-30 | 2010-04-28 | 北京理工大学 | Method and system for calibrating external parameters based on camera and three-dimensional laser radar |
| CN106291512A (en) * | 2016-07-29 | 2017-01-04 | 中国科学院光电研究院 | A kind of method of array push-broom type laser radar range Nonuniformity Correction |
| US20170357003A1 (en) * | 2016-06-08 | 2017-12-14 | U.S.A. As Represented By The Administrator Of The National Aeronautics And Space Administration | Method for the absolute calibration of the location and orientation of large-format detectors using laser radar |
| CN107679279A (en) * | 2017-09-04 | 2018-02-09 | 西安交通大学 | A kind of life-span prediction method of model subjects difference parameter Adaptive matching |
-
2018
- 2018-03-08 CN CN201810189563.8A patent/CN108459314B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101448338A (en) * | 2008-11-27 | 2009-06-03 | 北京理工大学 | Mobile three-dimensional spatial wireless testing network system |
| CN101699313A (en) * | 2009-09-30 | 2010-04-28 | 北京理工大学 | Method and system for calibrating external parameters based on camera and three-dimensional laser radar |
| US20170357003A1 (en) * | 2016-06-08 | 2017-12-14 | U.S.A. As Represented By The Administrator Of The National Aeronautics And Space Administration | Method for the absolute calibration of the location and orientation of large-format detectors using laser radar |
| CN106291512A (en) * | 2016-07-29 | 2017-01-04 | 中国科学院光电研究院 | A kind of method of array push-broom type laser radar range Nonuniformity Correction |
| CN107679279A (en) * | 2017-09-04 | 2018-02-09 | 西安交通大学 | A kind of life-span prediction method of model subjects difference parameter Adaptive matching |
Non-Patent Citations (3)
| Title |
|---|
| SCOTT E, ETAL: "Calibration method for texel images created from fused flash lidar and digital camera images", 《OPTICAL ENGINEERING》 * |
| 刘永进 等: "基于粒子滤波的红外焦平面阵列非均匀校正算法", 《红外与激光》 * |
| 吴奋陟 等: "基于标定场的激光雷达两步标定方法", 《空间控制技术与应用》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020139332A1 (en) * | 2018-12-24 | 2020-07-02 | Didi Research America, Llc | Real-time estimation of dc biac and noise power of light detection and ranging (lidar) |
| US11892573B2 (en) | 2018-12-24 | 2024-02-06 | Guangzhou Woya Laideling Technology Co., Ltd. | Real-time estimation of dc bias and noise power of light detection and ranging (LiDAR) |
| CN111208491A (en) * | 2020-01-17 | 2020-05-29 | 岭纬科技（厦门）有限公司 | Method for eliminating miscellaneous points of high-resolution solid-state laser radar point cloud |
| CN111208491B (en) * | 2020-01-17 | 2022-05-03 | 岭纬科技（厦门）有限公司 | Method for eliminating miscellaneous points of high-resolution solid-state laser radar point cloud |
| CN112346034A (en) * | 2020-11-12 | 2021-02-09 | 北京理工大学 | Three-dimensional solid-state area array laser radar combination calibration method |
| CN112346034B (en) * | 2020-11-12 | 2023-10-27 | 北京理工大学 | Three-dimensional solid-state area array laser radar combined calibration method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108459314B (en) | 2020-05-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7120582B2 (en) | River flow velocity measurement device and method using optical flow image processing | |
| CN107633536A (en) | A kind of camera calibration method and system based on two-dimensional planar template | |
| US10455781B2 (en) | Method for controlling surface irrigation based on surface water depth information | |
| CN108459314A (en) | A kind of three-dimensional solid-state face battle array laser radar non-uniform correction method | |
| CN110033486B (en) | Method and system for monitoring edge and volume of transparent crystal in growth process in real time | |
| CN106197612B (en) | A kind of transparent bottled liquid-level detecting method based on machine vision | |
| CN108627833A (en) | A kind of atmospheric phase compensation method based on GB-InSAR | |
| CN105793892B (en) | A kind of image processing method, device and picture pick-up device | |
| KR102127405B1 (en) | Method and appartus for estimating stream flow discharge using satellite images at streams | |
| CN104917975A (en) | Adaptive automatic exposure method based on object characteristics | |
| CN111160229B (en) | SSD network-based video target detection method and device | |
| CN109003312A (en) | A kind of camera calibration method based on nonlinear optimization | |
| CN107704704A (en) | A kind of relay class unit Estimation of The Storage Reliability method based on bayesian theory | |
| CN107238442B (en) | Without the detector thermometric method for regulating temperature of temperature stabilizer | |
| CN112734858A (en) | Binocular calibration precision online detection method and device | |
| CN111316135B (en) | System for calculating atmospheric precipitation rate in real time according to digital image of environment in which atmospheric precipitation is occurring | |
| CN109190647B (en) | Active and passive data fusion method | |
| CN112819882B (en) | Real-time earth volume calculation method based on binocular vision | |
| CN104407366B (en) | A kind of method being smoothed to pseudorange | |
| CN107945166B (en) | Binocular vision-based method for measuring three-dimensional vibration track of object to be measured | |
| CN110018529A (en) | Rainfall measurement method, device, computer equipment and storage medium | |
| CN113532419A (en) | Sky polarization mode information acquisition method and device, electronic equipment and storage medium | |
| CN113393413B (en) | Water area measuring method and system based on monocular and binocular vision cooperation | |
| CN112802080A (en) | Monocular absolute depth estimation method and device based on deep learning | |
| CN104504700B (en) | A kind of method and system for obtaining image sensor noise level curve |
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 |