CN110533603A - A kind of point cloud noise-reduction method based on entropy and convolutional neural networks - Google Patents
A kind of point cloud noise-reduction method based on entropy and convolutional neural networks Download PDFInfo
- Publication number
- CN110533603A CN110533603A CN201910672215.0A CN201910672215A CN110533603A CN 110533603 A CN110533603 A CN 110533603A CN 201910672215 A CN201910672215 A CN 201910672215A CN 110533603 A CN110533603 A CN 110533603A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- noise
- convolutional neural
- neural networks
- entropy
- 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.)
- Pending
Links
- 238000013527 convolutional neural network Methods 0.000 title claims abstract description 30
- 238000000034 method Methods 0.000 title claims abstract description 24
- 238000001514 detection method Methods 0.000 claims abstract description 3
- 238000005070 sampling Methods 0.000 claims description 21
- 239000011248 coating agent Substances 0.000 claims description 16
- 238000000576 coating method Methods 0.000 claims description 16
- 238000010606 normalization Methods 0.000 claims description 15
- 241001269238 Data Species 0.000 claims description 4
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 230000001537 neural effect Effects 0.000 claims 1
- 230000004069 differentiation Effects 0.000 abstract description 3
- 238000005516 engineering process Methods 0.000 abstract description 3
- 238000013145 classification model Methods 0.000 abstract 1
- 230000008859 change Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000004913 activation Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 150000003839 salts Chemical class 0.000 description 1
Classifications
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G06T5/70—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
- G06T2207/10012—Stereo images
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of point cloud noise-reduction method based on entropy and convolutional neural networks, belongs to field of computer technology.S1, a kind of point cloud noise-reduction method based on entropy and convolutional neural networks collects point cloud data and carries out whether artificial mark point cloud data is noise according to entity；S2, the point cloud data manually marked is normalized, and the point cloud data input entropy after normalized is carried out to the training of discrimination model with convolutional neural networks system；S3, judge whether discrimination model precision meets the requirements, if it is not, system adjust automatically convolutional neural networks model, and S1 is returned, if so, by current class model orientation optimal classification model；S4, point cloud data to be detected is obtained, the differentiation that detection point cloud data is made whether noise is treated using best discrimination model, if noise, then handle noise according to user's subjectivity selection probability noise spot.Noise is judged by artificial and computer fit system, and realizing has the study of supervision, improves study precision.
Description
Technical field
The present invention relates to field of computer technology, and in particular to a kind of point cloud noise reduction side based on entropy and convolutional neural networks
Method.
Background technique
The software and hardware technology of obtaining three-dimensional model was deepening continuously in recent years, and people can pass through a variety of data sampling sides
Method obtains the computer representation of real object, and pre-processes to it, processes, analysis and application.In obtaining data procedures
Because the defect of artificial disturbance or scanner itself makes the three-dimensional data generated often with noise, to make obtained
Measurement data and in kind have certain deviation.
In the prior art, common Denoising Algorithm needs an input parameter, i.e. noise intensity, in order to reach certainly
Dynamic denoising, needs automatically to estimate noise intensity, existing algorithm has often assumed that noise type, such as white Gaussian noise, the spiced salt
Noise, blue noise etc., and noise that actual sensor generates and these statistical laws are not met, therefore this kind of estimation noise is strong
The algorithm role of degree is limited.
Summary of the invention
The present invention provides a kind of point cloud noise-reduction method based on entropy and convolutional neural networks is matched by artificial with computer
Conjunction mode judges noise, and realizing has the study of supervision, improves study precision.
The above-mentioned technical purpose of the present invention is achieved through the following technical solutions, a kind of based on entropy and convolutional neural networks
Point cloud noise-reduction method, which is characterized in that include the following steps,
Step S1, it collects point cloud data and carries out whether artificial mark point cloud data is noise according to entity；
Step S2, the point cloud data manually marked is normalized, and the point cloud data after normalized is inputted
Entropy and convolutional neural networks system carry out the training of discrimination model；
Step S3, judge whether discrimination model precision meets the requirements, if it is not, entropy and convolutional neural networks system Automatic Optimal are rolled up
Product neural network node weight, and return step S1；If so, current discrimination model is positioned best discrimination model；
Step S4, point cloud data to be detected is obtained, detection point cloud data is treated using best discrimination model and is made whether noise
Differentiate, if noise, then noise is handled according to user's subjectivity selection probability noise spot, if not noise, then terminate.
As a preference of the present invention, the normalized includes the meter that all point cloud datas are carried out with comentropy mean value
It calculates, the sample point after the space length for the point cloud data that all people's work marks all is normalized divided by comentropy mean value.
As a preference of the present invention, according to step S3, system Automatic Optimal convolutional neural networks node weights work as node weight
When weight values are greater than or equal to predefined weight value, current discrimination model is best discrimination model.
As a preference of the present invention, according to step S4, according to step S4, by point cloud data to be detected input entropy and convolution mind
Carried out calculating the probability noise spot of noise type in the point cloud data, the maximum noise type of select probability in system.
As a preference of the present invention, noise type includes the probability noise spot greater than 90%, the machine between 90% and 70%
Rate noise spot, between 70% and 50% probability noise spot, between 50% and 30% probability noise spot, between 30% and 10% probability
Noise spot, the probability noise spot lower than 10%.
As a preference of the present invention, system includes network structure, point cloud data passes sequentially through the input of the network structure
Layer, K group's layer, full articulamentum, wherein K is more than or equal to 1.
As a preference of the present invention, input layer is the transverse and longitudinal coordinate and color of image of image.
As a preference of the present invention, group's layer includes convolutional layer, active coating, down-sampling layer, normalization layer；Convolutional layer swashs
Work layer, down-sampling layer, normalization layer.
Compared with prior art, the present invention is based on the point cloud noise-reduction methods of entropy and convolutional neural networks to have the advantage that
1, noise manually is judged with the mutually matched mode of computer, by entropy method for normalizing, improve study precision.
2, train best discrimination model using discrimination model, to whether be noise judgement it is more accurate, and according to use
Person's opinion carries out more suitable denoising.
Detailed description of the invention
Fig. 1 is present system flow chart.
Specific embodiment
Present invention will be described in further detail below with reference to the accompanying drawings.
This specific embodiment is only explanation of the invention, is not limitation of the present invention, those skilled in the art
Member can according to need the modification that not creative contribution is made to the present embodiment after reading this specification, but as long as at this
All by the protection of Patent Law in the scope of the claims of invention.
A kind of point cloud noise-reduction method based on entropy and convolutional neural networks, which is characterized in that include the following steps,
Step S1, it collects point cloud data and carries out whether artificial mark point cloud data is noise according to entity；
Step S2, the point cloud data manually marked is normalized, and the point cloud data after normalized is inputted
Entropy and convolutional neural networks system carry out the training of discrimination model；
Step S3, judge whether discrimination model precision meets the requirements, if it is not, system adjust automatically convolutional neural networks model,
And return step S1, if so, by the best discrimination model of current class model orientation；
Step S4, obtain point cloud data to be detected, using best discrimination model carry out a cloud whether the differentiation of noise, if making an uproar
Point then handles noise according to user's subjectivity selection probability noise spot, if not noise, then terminate.
The sample point cloud data differentiated input convolutional neural networks system is subjected to discrimination model according in step S2
Training mainly enters convolutional neural networks system to the sample point cloud bar with differentiation label and learns；Repeat " training -> tune
The process of whole network structure -> training " is until differentiating correct.
As a preference of the present invention, the normalized includes the meter that all point cloud datas are carried out with comentropy mean value
It calculates, the sample point after the space length for the point cloud data that all people's work marks all is normalized divided by comentropy mean value.
According to step S3, system Automatic Optimal convolutional neural networks node weights, when node weight value is greater than or equal in advance
When determining weighted value, current discrimination model is best discrimination model.
In the technical scheme, it is sampled by the target area in point cloud data to be detected, is put into neural network
Input layer, carry out entirely connect after, finally obtain the probability of each label, the i.e. real value in section [0,1].Select probability
Maximum noise type.
Noise type includes the probability noise spot greater than 90%, the probability noise spot between 90% and 70%, between 70%
With 50% probability noise spot, between 50% and 30% probability noise spot, between 30% and 10% probability noise spot, lower than 10%
Probability noise spot.The noise label of totally 7 seed types, i.e. 7 data, this 7 data and be equal to 1;Each of then, will obtain
The probability of the label of measuring point cloud to be checked is averaged, and the probability of the label of measuring point cloud to be checked, the maximum label of select probability are obtained
The label of noise type as the measuring point cloud to be checked.
Network sequence is input layer _ > K group layer _ > full articulamentum, and wherein K is more than or equal to 1;Group's layer packet
Include convolutional layer, active coating, down-sampling layer, normalization layer；Convolutional layer, active coating, down-sampling layer, the core for normalizing each layer in layer
Size and output size can all carry out any adjusting, and each layer has an input and generates an output,
Each layer of the input exported as next layer.
Wherein, the input size of input layer is Height x Weight x Channel, and wherein Weight, Height are
The width and height of input layer point cloud, Channel are the air coordinates of input layer point cloud.
Convolutional layer；
1) size of core must be odd number, and no more than the width or height of this layer input；
2) wide and height is not changed when intermediate representation passes through convolutional layer, port number is variable can be constant；It theoretically can be any just whole
Number, the present invention using GPU hardware realize due to, here for 16 multiple.
Active coating:
1) active coating does not change width, height or the port number of convolutional layer expression；
2) activation primitive used in active coating includes but is not limited to following type function:
F (x)=l/(1+e-x)
F (x)=a*tanh (b*x), a, b are any non-zero real
f(x)= max(0,x)
f(x)= min(a,max(0,x))
f(x)= log(l+ex)
f(x)= |x|
f(x)= x2
f(x)= √X
f(x)= ax+b
3) active coating is followed after convolutional layer or full connection.
Down-sampling layer:
1) down-sampling layer does not change the port number of intermediate representation；
2) down-sampling layer is the size of core to the drawdown ratio of cloud: the down-sampling layer that i.e. core is mxn will cause intermediate representation contracting
Small is upper one layer (1/m) x (1/n), and theoretically m and n can be random natural number, is realized due to the present invention using GPU hardware
Reason, m=n.For example, becoming 5x5x32 after the down-sampling that 15x15x32 passes through 3x3;After the down-sampling that 15x15x32 passes through 5x5,
Become 3x3x32;But 15x15x32 not can be carried out the down-sampling of 2x2, because 15 cannot be divided exactly by 2；It is not to say that, inputs ruler
Very little must be 2 power, i.e., 16,32,64 etc., as long as input size guarantees to be sampled by all down-sampling layers.
Normalize layer:
1) normalization layer does not change any size of intermediate representation；
2) normalization layer be not required, must shouldn't, addition normalization layer would generally improve precision and increase calculation amount；It is
No addition normalizes layer, to see the speed of the precision and loss that are actually promoted after addition.
- as combination be: convolutional layer -> active coating -> down-sampling layer -> normalization layer.
Following situations is special:
1) layer is smaller to precision improvement when but increasing many operands for addition normalization, cancels normalization layer, i.e., using following
Combination: convolutional layer -> active coating -> down-sampling layer；
2) normalization layer shift to an earlier date, effect is essentially identical, that is, use following combination: convolutional layer -> active coating -> normalization layer -> under adopt
Sample layer.
3) cancel down-sampling layer: convolutional layer -> active coating；Or convolutional layer -> active coating -> normalization layer；Down-sampling essence
It is while to be reduced the effect of the operand of succeeding layer in passing to increase robustness；Usually had in one network it is several layers of under
Sample level, but not all " convolutional layer -> active coating " below will be with down-sampling layer.
Full articulamentum:
1) it will become by the intermediate representation after full articulamentum one-dimensional, be no longer three-dimensional；
2) output of full articulamentum can be with arbitrary value；
3)-denier can not just carry out convolutional layer, down-sampling layer or normalization layer into excessively full articulamentum；
4) active coating can be connect behind full articulamentum, or after the full articulamentum of continued access.
It connects after full articulamentum, effect is the general string real value of full articulamentum generation become between [0,1].
Claims (8)
1. a kind of point cloud noise-reduction method based on entropy and convolutional neural networks, which is characterized in that include the following steps,
Step S1, it collects point cloud data and carries out whether artificial mark point cloud data is noise according to entity；
Step S2, the point cloud data manually marked is normalized, and the point cloud data after normalized is inputted
Entropy and convolutional neural networks system carry out the training of discrimination model；
Step S3, judge whether discrimination model precision meets the requirements, if it is not, entropy and convolutional neural networks system Automatic Optimal are rolled up
Product neural network node weight, and return step S1；If so, current discrimination model is positioned best discrimination model；
Step S4, point cloud data to be detected is obtained, detection point cloud data is treated using best discrimination model and is made whether noise
Differentiate, if noise, then noise is handled according to user's subjectivity selection probability noise spot, if not noise, then terminate.
2. a kind of point cloud noise-reduction method based on entropy and convolutional neural networks according to claim 1, which is characterized in that institute
Stating normalized includes that the calculating of comentropy mean value is carried out to all point cloud datas, by all point cloud datas manually marked
Space length all normalized divided by comentropy mean value after sample point.
3. a kind of point cloud noise-reduction method based on entropy and convolutional neural networks according to claim 1, which is characterized in that root
According to step S3, system Automatic Optimal convolutional neural networks node weights, input model accuracy is more than user's subjectivity setting value
When, current discrimination model is best discrimination model.
4. a kind of point cloud noise-reduction method based on entropy and convolutional neural networks according to claim 1, which is characterized in that root
According to step S4, will calculate noise type in the point cloud data in point cloud data to be detected input entropy and convolutional Neural system
Probability noise spot, the maximum noise type of select probability.
5. a kind of point cloud noise-reduction method based on entropy and convolutional neural networks according to claim 1, which is characterized in that make an uproar
Sound type includes the probability noise spot greater than 90%, the probability noise spot between 90% and 70%, between 70% and 50% probability
Noise spot, the probability noise spot between 50% and 30% probability noise spot, between 30% and 10% probability noise spot, lower than 10%.
6. a kind of point cloud noise-reduction method based on entropy and convolutional neural networks according to claim 1, which is characterized in that be
System includes network structure, and point cloud data passes sequentially through the input layer of the network structure, K group layer, full articulamentum, wherein K
More than or equal to 1.
7. a kind of point cloud noise-reduction method based on entropy and convolutional neural networks according to claim 6, which is characterized in that defeated
Enter the transverse and longitudinal coordinate and color of image that layer is image.
8. a kind of point cloud noise-reduction method based on entropy and convolutional neural networks according to claim 6, which is characterized in that small
Group layer includes convolutional layer, active coating, down-sampling layer, normalization layer；Convolutional layer, active coating, down-sampling layer, normalization layer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910672215.0A CN110533603A (en) | 2019-07-24 | 2019-07-24 | A kind of point cloud noise-reduction method based on entropy and convolutional neural networks |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910672215.0A CN110533603A (en) | 2019-07-24 | 2019-07-24 | A kind of point cloud noise-reduction method based on entropy and convolutional neural networks |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110533603A true CN110533603A (en) | 2019-12-03 |
Family
ID=68661138
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910672215.0A Pending CN110533603A (en) | 2019-07-24 | 2019-07-24 | A kind of point cloud noise-reduction method based on entropy and convolutional neural networks |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110533603A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111539881A (en) * | 2020-04-16 | 2020-08-14 | 南京航空航天大学 | Deep learning-based aerial part point cloud denoising method |
| CN113312946A (en) * | 2020-02-27 | 2021-08-27 | 敦泰电子（深圳）有限公司 | Fingerprint image feature extraction method and device and computer readable storage medium |
| CN113610143A (en) * | 2021-08-02 | 2021-11-05 | 广州文远知行科技有限公司 | Method, device, equipment and storage medium for classifying point cloud noise points |
| CN114821248A (en) * | 2022-07-01 | 2022-07-29 | 深圳大学 | Point cloud understanding-oriented data active screening and labeling method and device |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103971342A (en) * | 2014-05-21 | 2014-08-06 | 厦门美图之家科技有限公司 | Image noisy point detection method based on convolution neural network |
| CN106204468A (en) * | 2016-06-27 | 2016-12-07 | 深圳市未来媒体技术研究院 | A kind of image de-noising method based on ReLU convolutional neural networks |
| CN108230329A (en) * | 2017-12-18 | 2018-06-29 | 孙颖 | Semantic segmentation method based on multiple dimensioned convolutional neural networks |
| CN108921885A (en) * | 2018-08-03 | 2018-11-30 | 南京林业大学 | A kind of method of comprehensive three classes data source joint inversion forest ground biomass |
| CN108961339A (en) * | 2018-07-20 | 2018-12-07 | 深圳辰视智能科技有限公司 | A kind of point cloud gestures of object estimation method, device and its equipment based on deep learning |
| CN109685126A (en) * | 2018-12-17 | 2019-04-26 | 北斗航天卫星应用科技集团有限公司 | Image classification method and image classification system based on depth convolutional neural networks |
-
2019
- 2019-07-24 CN CN201910672215.0A patent/CN110533603A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103971342A (en) * | 2014-05-21 | 2014-08-06 | 厦门美图之家科技有限公司 | Image noisy point detection method based on convolution neural network |
| CN106204468A (en) * | 2016-06-27 | 2016-12-07 | 深圳市未来媒体技术研究院 | A kind of image de-noising method based on ReLU convolutional neural networks |
| CN108230329A (en) * | 2017-12-18 | 2018-06-29 | 孙颖 | Semantic segmentation method based on multiple dimensioned convolutional neural networks |
| CN108961339A (en) * | 2018-07-20 | 2018-12-07 | 深圳辰视智能科技有限公司 | A kind of point cloud gestures of object estimation method, device and its equipment based on deep learning |
| CN108921885A (en) * | 2018-08-03 | 2018-11-30 | 南京林业大学 | A kind of method of comprehensive three classes data source joint inversion forest ground biomass |
| CN109685126A (en) * | 2018-12-17 | 2019-04-26 | 北斗航天卫星应用科技集团有限公司 | Image classification method and image classification system based on depth convolutional neural networks |
Non-Patent Citations (2)
| Title |
|---|
| 刘志青等: "基于信息向量机的机载激光雷达点云数据分类", 《光学精密工程》 * |
| 吴炜等: "基于学习的图像增强技术", 《基于学习的图像增强技术》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113312946A (en) * | 2020-02-27 | 2021-08-27 | 敦泰电子（深圳）有限公司 | Fingerprint image feature extraction method and device and computer readable storage medium |
| CN111539881A (en) * | 2020-04-16 | 2020-08-14 | 南京航空航天大学 | Deep learning-based aerial part point cloud denoising method |
| US11514555B2 (en) | 2020-04-16 | 2022-11-29 | Nanjing University Of Aeronautics And Astronautics | Point cloud denoising method based on deep learning for aircraft part |
| CN113610143A (en) * | 2021-08-02 | 2021-11-05 | 广州文远知行科技有限公司 | Method, device, equipment and storage medium for classifying point cloud noise points |
| CN114821248A (en) * | 2022-07-01 | 2022-07-29 | 深圳大学 | Point cloud understanding-oriented data active screening and labeling method and device |
| CN114821248B (en) * | 2022-07-01 | 2022-10-21 | 深圳大学 | Point cloud understanding-oriented data active screening and labeling method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110533603A (en) | A kind of point cloud noise-reduction method based on entropy and convolutional neural networks | |
| CN103971342B (en) | A kind of image noise detection method based on convolutional neural networks | |
| CN106023298B (en) | Point cloud Rigid Registration method based on local Poisson curve reestablishing | |
| CN112232476B (en) | Method and device for updating test sample set | |
| CN103632158B (en) | Forest fire prevention monitor method and forest fire prevention monitor system | |
| CN105260710B (en) | Water meter calibration method, apparatus based on image procossing and system | |
| CN105938563A (en) | Weld surface defect identification method based on image texture | |
| CN108537808A (en) | A kind of gluing online test method based on robot teaching point information | |
| CN104599286B (en) | A kind of characteristic tracking method and device based on light stream | |
| CN110472611A (en) | Method, apparatus, electronic equipment and the readable storage medium storing program for executing of character attribute identification | |
| CN106970379B (en) | Based on Taylor series expansion to the distance-measuring and positioning method of indoor objects | |
| CN107218894A (en) | A kind of subpixel accuracy thickness of detector detection method of fast and stable | |
| CN106091923A (en) | The central point rapid assay methods of industrial bolt circular hole based on three-dimensional laser scanning technique | |
| CN110008947A (en) | A kind of silo Grain Quantity monitoring method and device based on convolutional neural networks | |
| CN110376290A (en) | Acoustic emission source locating method based on multidimensional Density Estimator | |
| CN116600104B (en) | Phase acquisition quality analysis method and system for IPC network camera | |
| CN106340010A (en) | Corner detection method based on second-order contour difference | |
| CN104050660A (en) | Method for measuring workpiece round edges | |
| CN109618288B (en) | Wireless sensor network distance measuring system and method based on deep convolutional neural network | |
| CN104359847B (en) | Method and device for acquiring centroid set used for representing typical water category | |
| CN106292673B (en) | A kind of method for optimizing route and system | |
| CN105203327B (en) | A kind of gas circuit measurement parameter selection method applied to engine air passage analysis | |
| CN109163674B (en) | Sensor measurement viewpoint planning method in a kind of measurement of area-structure light automatized three-dimensional | |
| CN114419140A (en) | Positioning algorithm for light spot center of track laser measuring device | |
| CN105654053B (en) | Based on the dynamic oscillation signal parameter discrimination method for improving constraint EKF algorithm |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191203 |
|
| RJ01 | Rejection of invention patent application after publication |