CN103971125A - Super-resolution algorithm based on vibration signal of laser echo - Google Patents
Super-resolution algorithm based on vibration signal of laser echo Download PDFInfo
- Publication number
- CN103971125A CN103971125A CN201410185762.3A CN201410185762A CN103971125A CN 103971125 A CN103971125 A CN 103971125A CN 201410185762 A CN201410185762 A CN 201410185762A CN 103971125 A CN103971125 A CN 103971125A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- vibration signal
- vibration
- super
- resolution
- 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
Landscapes
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention discloses a super-resolution imaging method based on a laser echo signal. The super-resolution imaging method mainly solves the problems that a laser echo vibration signal is low in strength and small in amount of information. The super-resolution imaging method mainly comprises the following steps of (1) imaging processing, wherein a laser vibration measurer is used for conducting two-dimensional meshing scanning on a vehicle to be detected, corresponding grey-scale value quantification is conducted on the laser echo signal, and a two-dimensional vibration signal spectrum of the vehicle is built; (2) feature extraction, wherein feature operators are designed, feature extraction is conducted on the vehicle of a known training database, and a vibration feature database of the vehicle is built; (3) super resolution amplification, wherein super resolution amplification is conducted on a vibration spectrum of the vehicle to be detected according to an image super-resolution algorithm and the obtained vehicle vibration feature database, so that a high-resolution vibration spectrum of the vehicle is obtained. The super-resolution imaging method based on the laser echo signal has the advantages of being low in calculation complexity, high in practicability, wide in application range and the like, and can be applied to the fields of object defection, object identification and the like.
Description
Technical field
The present invention relates to the ultra-resolution method to return laser beam vibration signal, belong to Image Super-resolution technical field.
Background technology
Laser detector can be to vehicle, warship, and the object to be identified such as large-scale blindage carries out detection operations.Laser has monochromaticity, coherence, and directivity etc. are different from the advantageous characteristic of other light sources, and these characteristics are for utilizing the laser generator in laser detector to survey theoretical support is provided remote object.Because the mechanical objects such as vehicle, steamer all can have self micro-vibration, the laser that laser generator sends accurately in hit object a bit, then just can to the laser echo signal being reflected back, receive by laser pickoff, utilize PC or other digital processing chips to carry out Treatment Analysis to the signal collecting, to obtain being detected the relevant information of object.
And the key of problem is that the measurement of laser echo signal and collection are relatively complicated, very high to the requirement of instrument and precision, again because of dirt in air, light pollution, the interference of the factors such as water smoke, the accuracy that causes every a branch of echoed signal all can decrease to some degree, and the number of times that we must be by increasing measurement target object is to make up the error that measurement is brought each time.And laser generator generally can only be launched 1-2 bundle laser, be merely able to obtain the vibration information of one to two point of object under test, this just means from the resulting information of collecting terminal very limited, this just makes whole step implement to have produced contradiction.In order to solve this contradiction, in engineering, the general method of hardware or software that adopts is processed, and the cost of raising hardware (optical instrument of laser generator, laser pickoff or intermediate link) is very expensive, and the space of lifting is also little.So in sum, with software approach, echoed signal is carried out to signal processing and just become reasonable manner, and SUPERRESOLUTION PROCESSING FOR ACOUSTIC is exactly the algorithm of a kind of comparative superiority proposed by the invention.
Image Super-resolution is exactly will by infrared thermoviewer or the resulting low-resolution image of visual light imaging instrument, (pixel in image seldom, high frequency edge part is fuzzy) by a series of digitized algorithm process, enable to obtain more pixel, higher picture element density, thus make image more clear.Image Super-resolution has been widely used in a plurality of fields, for example: military infrared acquisition, medical nmr, recognition of face etc.On this basis, the collected methods of vehicle vibration signal of laser pickoff relating in the present invention also can transform into picture signal by relevant treatment, thereby can process to obtain by super-resolution the vibration information of more heterogeneous pass, for ensuing analytical work is provided convenience.Greatly reduced the number of times that on hardware, laser detector is surveyed, saved cost simultaneously.
Summary of the invention
The present invention, in order to reduce the measurement number of times and measurement expense of vibration measurement with laser instrument, has proposed a kind of super-resolution rebuilding algorithm of the methods of vehicle vibration signal based on return laser beam.First by laser generator, treat and survey automobile body and carry out the two-dimensional mesh scanning of formatting, obtain the vibration information of automobile body every bit; Then by laser pickoff, laser echo signal gathered and sample in signal end for process, quantification waits pre-service work, obtains digitized methods of vehicle vibration signal dot matrix; Then digitized methods of vehicle vibration signal dot matrix is carried out to gradation of image value quantification treatment, obtain the vibration signal collection of illustrative plates of vehicle to be measured; Then the vibration signal dot matrix of other vehicles of having surveyed is also passed through to similar gradation of image value quantification treatment, obtain training the two-dimension vibration signal collection of illustrative plates of vehicle, and design filter operator the vibration signal collection of illustrative plates of training vehicle is carried out to feature extraction, create Vehicular vibration property data base, as the priori of super-resolution; Finally, by the priori of obtaining before, the vibration signal collection of illustrative plates that utilizes quadrature coupling track algorithm to treat survey vehicle carries out pixel expansion, has reached the effect of super-resolution.With previously directly utilized vibration signal collection of illustrative plates to analyze to compare, the vibration signal collection of illustrative plates pixel after processing through super-resolution is more, vibrational image is more clear, ensuing analysis meeting is more accurate.
Current two kinds of methods of the general employing of Image Super-resolution Reconstruction algorithm: the Super-resolution Reconstruction based on multiframe and the super-resolution rebuilding based on single frames.And the needed inputted vibration dot matrix image of the super-resolution rebuilding of multiframe is too much, cannot be applied in during laser echo signal processes, and based on single frames super-resolution algorithms, only need the oscillation point system of battle formations of a vehicle to be measured, be more suitable for the application in engineering.What therefore the present invention adopted is the super-resolution rebuilding based on single frames.
The present invention adopts the super-resolution rebuilding of single-frame images, just need to have the support of priori.And the acquisition of priori need to be relevant to testee type several high resolving power oscillation point systems of battle formations as training sample, train.That is to say that the basis that the present invention can realize is exactly to measure a large amount of associated vehicle vibration datas, to obtain vehicle data vibration information, be used as the sample training of training, the meaning of ' training ' can be regarded the process of a feature extraction and integration substantially as here.The present invention combines by designing relevant one dimension and two dimensional filter, training sample is carried out to template filtering, extract signals of vehicles feature, the characteristic signal extracting is formed to characteristic set by operations such as dimensionality reduction, integration, we are referred to as ' dictionary ' this characteristic set.Then by quadrature, mate track algorithm, the vibration signal collection of illustrative plates of the vehicle to be measured obtaining by laser pickoff and training are obtained to vehicle characteristics storehouse (being dictionary) mates, in dictionary, find corresponding high-frequency information, the high-frequency information that these couplings are obtained joins in original vehicle vibration signal dot matrix, just can reach and expand vibration signal point, make Vehicular vibration image object more clearly.
The present invention is achieved by the following technical solutions: the Super-Resolution of Images Based based on dictionary learning carries out Super-resolution Reconstruction to laser echo signal, comprise following 4 large steps: 1) laser echo signal through testee reflection is carried out to pre-service, gradation of image value quantification treatment is carried out in pretreated each oscillation point, form vibration signal collection of illustrative plates, as the pending collection of illustrative plates in experiment; 2) utilize laser vibration measurer to carry out oscillating scanning to the vehicle of different model, different styles, obtain the vibration signal of different vehicle, and vibration signal is done to the gradation of image value quantification treatment in step 1), as training sample database.3) training sample database is carried out to feature extraction and integration, set up vehicular characteristics data storehouse; 4) using vehicular characteristics data storehouse as priori, the quadrature of usining coupling track algorithm carries out super-resolution processing as basic skills to pending vibrorecord spectrum, to reach vibrational lattice, expands, and vibration collection of illustrative plates is object more clearly.
In affiliated step 1), the tested vehicle echo vibration signal first laser pickoff being received is sampled, the pre-service such as quantification, become digitizing vibrational lattice, carry out the gradation of image value quantification treatment of 0 ~ 255 gradient, form vibration signal collection of illustrative plates, as the pending collection of illustrative plates in experiment.
Affiliated step 2) in, utilize laser vibration measurer to carry out oscillating scanning to a large amount of training vehicles in training storehouse, can obtain the vibration signal dot matrix of different vehicle.Vibration data in these micro-vibration signal dot matrix is done to the gradation of image value quantification treatment in step 1, obtain methods of vehicle vibration signal training sample database.
In affiliated step 3), utilize a plurality of one dimensions or the two dimensional filter operator that in the present invention, propose to carry out feature extraction to the methods of vehicle vibration signal training sample database in step 2, and carry out the processing such as dimensionality reduction, integration, and set up vehicular characteristics data storehouse, become the priori of super-resolution.
In affiliated step 4), using the vehicular characteristics data storehouse that obtains in step 3 as priori, use quadrature coupling track algorithm that the pending collection of illustrative plates obtaining in step 1 is mated with the feature in acquired priori, obtain the vehicle minutia information needing, these information are added to original Vehicular vibration collection of illustrative plates to be measured, form final super-resolution image.
Accompanying drawing explanation
Fig. 1 is the methods of vehicle vibration signal super-resolution algorithm flow chart that the present invention is based on return laser beam.
Embodiment
Below in conjunction with vehicle fine motion signal measurement, the present invention is described in further detail:
1, the conversion of vehicle micro-vibration signal dot matrix and vibration signal collection of illustrative plates
Due to the switch regardless of at car engine, vehicle self all can have the vibration of certain frequency, and the different local vibration frequencies of the headstock of vehicle, car door or the tailstock can be had any different.Whenever laser generator sends laser one to any one place of vehicle body, the reflected light by laser just can obtain the vibration information of this vehicle.And when our use laser acquisition as much as possible, vehicle body is carried out to two-dimensional mesh while formatting point by point scanning, just can obtain a Vehicular vibration information dot matrix (for example dot matrix of a 100*100).Certainly, the measurement of laser instrument is very huge to the arrangement of wanting summed data of hardware, and laser scans can not all reach quite high precision at every turn, that is to say that treating survey vehicle can only carry out the detection of general profile.And this vibration information obtaining by scanning is changed by digitizing and just can be obtained digitized vibration information dot matrix, then to carry out quantification treatment to this digitized vibration information dot matrix, quantizing range is 0 ~ 255, forms two-dimension vibration signal collection of illustrative plates
:
Wherein
represent the minimum amplitude point of former vibration information,
represent the maximum amplitude points of former vibration information.The scope that so just vibration signal can be converted into is at 0 ~ 255 image pixel dot matrix
, and
what just can be similar to processes as a secondary two-dimension vibration image.Digitizing vibration information dot matrix through gradation of image value quantification treatment (
) example as follows:
2, the formation of methods of vehicle vibration signal training sample database
In the present invention, main research object is vehicle, and we carry out vibration measurement with laser by a large amount of other vehicles in training storehouse, and we can obtain a large amount of methods of vehicle vibration signal dot matrix.In the present invention using these vibration experiment data as methods of vehicle vibration signal database.With the quantization method in step 1, the vibration signal dot matrix dot matrix of each car is carried out to gradation of image value quantification treatment equally, obtain some width different vehicle vibration signal collection of illustrative plates
, the mode chart of each auxiliary point system of battle formations is as follows:
Wherein
in all data be the image pixel information after quantification,
the row of matrix and the maximal value of row in vibration collection of illustrative plates after representative quantizes.We all regard each oscillation point as a pixel, so just can be by training information dot matrix
as a secondary two dimensional image, process.
3, the foundation in feature extraction and vehicular characteristics data storehouse
The first step, pre-service.With reference to Fig. 1, for each training information dot matrix
, first carry out down-sampled processing, obtain its down-sampled version
.Here to underline, for super-resolution afterwards, process conveniently, make as far as possible the down-sampled version of training information dot matrix
with pending laser echo signal dot matrix
dimension (the information dot matrix to be measured that is consistent
pixel precision be generally far below training information dot matrix
).In like manner,
in each secondary high precision Vehicular vibration info class gray scale dot matrix can process like this, just repeat no more afterwards.
Second step, designing filter, extracts feature.For each secondary training information dot chart
with its down-sampled version
, we all will design suitable filter operator and carry out filtering, extract characteristic of correspondence information, and Frequent Filters is as follows:
Gradient operator:
Laplace operator
:
In the present invention, also according to the feature of the oscillation point system of battle formations, additionally increased by four two dimensional filter operators, the more effective Vehicular vibration characteristics of image that extracts:
These four wave filters add the feature extraction efficiency that can greatly improve the oscillation point system of battle formations, feature extracting method is as follows:
Wherein
representative is from training information dot matrix
in the high frequency vehicle characteristics that extracts,
representative is from training information dot matrix
the low frequency vehicle feature of middle extraction,
represent convolution algorithm.
The 3rd step: application K-SVD algorithm is integrated eigenmatrix.
K-SVD is a kind of a kind of algorithm that is often applied in image processing, feature integration, the iteration renewal thought based on sparse coding, original signal is represented to become to the combination of dictionary and sparse coefficient matrix, then by the thought of iteration, upgrade, until meet the requirement of iteration, till being updated to optimum dictionary and corresponding sparse coefficient matrix, the target equation of K-SVD algorithm is as follows:
Wherein
for permissible error,
according to low frequency vehicle feature
the dictionary of the lower accuracy Vehicular vibration of integrating (being called for short low resolution dictionary), and
to connect vehicle characteristics matrix
with low resolution dictionary
matrix of coefficients, be also the solution amount for the treatment of in target equation.
K-SVD algorithm can upgrade simultaneously
with
, and here
in the thought of Its Sparse Decomposition, play a part tie, in low resolution dictionary, upgrade the sparse coefficient matrix
can be applied in the equationof structure of high resolving power dictionary:
Here
with
be exactly our required low, high resolving power dictionary, our desired priori namely, is wherein comprising the characteristic information of the vehicle overwhelming majority, and the usage of priori will be introduced in step 3.
4, super-resolution is processed.
It is mainly to utilize low, the high resolving power dictionary training in step 2 that super-resolution is processed in this step
with
, come the pretreated vibration signal class gray scale dot chart of process in step 1
carry out dot matrix expansion, namely the process of super-resolution rebuilding.
The present invention is by the low-resolution image obtaining in step 1
as pending image, be divided into little image block
(for example dot matrix of 3*3).By the Vehicular vibration characteristic set obtaining in step 3
with
as priori.The present invention's application quadrature coupling track algorithm is found out it at Vehicular vibration characteristic set to each little image block
in optimum matching represent.According to the conforming principle of sparse coding in the training of high-resolution and low-resolution dictionary, just can be at high resolving power dictionary
in find the detail of the high frequency mating with this image block
, and these detail of the high frequency are added to pending image
in, just can reach vibrational lattice and expand, the object of Image Super-resolution.
be our required result.
Claims (5)
1. the vibration signal super-resolution algorithm based on return laser beam, is characterized in that comprising the following steps:
1) utilize laser vibration measurer to carry out the two-dimensional mesh point by point scanning of formatting to object under test (vehicle), by the return laser beam analog vibration signal obtaining sample, the digitizing pre-service work such as quantification, obtain digitizing vibration signal, and the vibration signal at all grids place is regarded as to the gray-scale value of two dimensional image, build the vibration signal collection of illustrative plates of object under test (vehicle);
2) utilize laser vibration measurer to scan detection to the object (vehicle) in a large amount of training storehouse, the vibration signal dot matrix obtaining is done to same pre-service and gradation of image value quantification treatment, obtain the training sample database that comprises a large amount of training objects (vehicle) vibration signal collection of illustrative plates;
3) design suitable one dimension and two dimensional filter operator, object in training sample database (vehicle) vibration signal collection of illustrative plates is carried out to template filtering, extract object (vehicle) feature needing, utilize K-SVD algorithm that eigenmatrix is integrated into object (vehicle) property data base, the priori of processing as super-resolution;
4) binding object (vehicle) property data base (priori), utilizes quadrature coupling track algorithm that the vibration signal collection of illustrative plates of object under test (vehicle) is carried out to super-resolution, has reached and has expanded vibration signal dot matrix, makes the object of vibration signal collection of illustrative plates sharpening.
2. a kind of vibration signal super resolution ratio reconstruction method based on return laser beam according to claim 1, is characterized in that described step 1) carries out gradation of image value quantification treatment to pretreated digital vibration signal:
Wherein
represent the minimum amplitude point of former vibration information,
represent the maximum amplitude points of former vibration information;
all represent former digital vibration signal point,
digital vibration signal point after representative quantizes, here
the formed image of point set we be called the vibration signal collection of illustrative plates of object under test (vehicle)
.
3. a kind of vibration signal super resolution ratio reconstruction method based on return laser beam according to claim 1, it is characterized in that described step 2), utilize laser vibration measurer to carry out vibration information scanning to the object (vehicle) in training storehouse, and do corresponding pre-service and gray-scale value quantification treatment, obtain the training sample database that comprises voluminous object (vehicle) vibration signal collection of illustrative plates
.
4. a kind of vibration signal super resolution ratio reconstruction method based on return laser beam according to claim 1, is characterized in that described step 3), and designing filter carries out convolutional filtering to the training points system of battle formations, feature extraction, and the present invention's wave filter used is as follows:
Wherein
with
for gradient filter,
with
for Laplace filter,
~
for the wave filter of independent design of the present invention, compare original gradient and Laplace filter, the formula of feature extraction is:
Wherein
with
be respectively high resolving power object (vehicle) vibration signal collection of illustrative plates and low resolution object (vehicle) the vibration signal collection of illustrative plates after down-sampled, then according to K-SVD algorithm, obtain that object (vehicle) is low, high-resolution features database (also claiming dictionary)
with
, the needed priori of super-resolution namely.
5. a kind of vibration signal super resolution ratio reconstruction method based on return laser beam according to claim 1, is characterized in that low, high resolving power dictionary that described step 4) utilization trains in step 4)
with
, find out the vibration signal collection of illustrative plates of the object under test (vehicle) in step 1
the HFS lacking
and add original collection of illustrative plates:
be our required result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410185762.3A CN103971125A (en) | 2014-05-05 | 2014-05-05 | Super-resolution algorithm based on vibration signal of laser echo |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410185762.3A CN103971125A (en) | 2014-05-05 | 2014-05-05 | Super-resolution algorithm based on vibration signal of laser echo |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103971125A true CN103971125A (en) | 2014-08-06 |
Family
ID=51240596
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410185762.3A Pending CN103971125A (en) | 2014-05-05 | 2014-05-05 | Super-resolution algorithm based on vibration signal of laser echo |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103971125A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104778671A (en) * | 2015-04-21 | 2015-07-15 | 重庆大学 | Image super-resolution method based on SAE and sparse representation |
| CN105738894A (en) * | 2016-03-03 | 2016-07-06 | 西安电子科技大学 | Inching group object high resolution imaging method based on augmented Laplace operator |
| CN107024457A (en) * | 2017-03-22 | 2017-08-08 | 华南理工大学 | A kind of far-field optics super-resolution microscopic method |
| CN112373738A (en) * | 2020-11-23 | 2021-02-19 | 北京空间机电研究所 | Thin-wall structure vibration test device and test method considering pressure difference condition |
| CN115993611A (en) * | 2023-03-22 | 2023-04-21 | 清华大学 | Non-visual field imaging method and device based on transient signal super-resolution network |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8054217B2 (en) * | 2006-10-26 | 2011-11-08 | Raytheon Company | Radar imaging system and method using gradient magnitude second moment spatial variance detection |
| CN103714526A (en) * | 2013-12-24 | 2014-04-09 | 西安电子科技大学 | Super-resolution image reconstruction method based on sparse multi-manifold embedment |
-
2014
- 2014-05-05 CN CN201410185762.3A patent/CN103971125A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8054217B2 (en) * | 2006-10-26 | 2011-11-08 | Raytheon Company | Radar imaging system and method using gradient magnitude second moment spatial variance detection |
| CN103714526A (en) * | 2013-12-24 | 2014-04-09 | 西安电子科技大学 | Super-resolution image reconstruction method based on sparse multi-manifold embedment |
Non-Patent Citations (2)
| Title |
|---|
| 任广梅: "稀疏表示在单幅图像超分辨率重建中的应用研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 杨鹏等: "基于EMD分解的车辆振动识别", 《电脑知识与技术》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104778671A (en) * | 2015-04-21 | 2015-07-15 | 重庆大学 | Image super-resolution method based on SAE and sparse representation |
| CN104778671B (en) * | 2015-04-21 | 2017-09-22 | 重庆大学 | A kind of image super-resolution method based on SAE and rarefaction representation |
| CN105738894A (en) * | 2016-03-03 | 2016-07-06 | 西安电子科技大学 | Inching group object high resolution imaging method based on augmented Laplace operator |
| CN105738894B (en) * | 2016-03-03 | 2018-07-06 | 西安电子科技大学 | Fine motion multiple targets high-resolution imaging method based on augmentation Laplace operator |
| CN107024457A (en) * | 2017-03-22 | 2017-08-08 | 华南理工大学 | A kind of far-field optics super-resolution microscopic method |
| CN107024457B (en) * | 2017-03-22 | 2019-05-14 | 华南理工大学 | A kind of far-field optics super-resolution microscopic method |
| US10900895B2 (en) | 2017-03-22 | 2021-01-26 | South China University Of Technology | Far-field optical super-resolution microscopy method |
| CN112373738A (en) * | 2020-11-23 | 2021-02-19 | 北京空间机电研究所 | Thin-wall structure vibration test device and test method considering pressure difference condition |
| CN112373738B (en) * | 2020-11-23 | 2022-07-29 | 北京空间机电研究所 | Thin-wall structure vibration test device and method considering pressure difference condition |
| CN115993611A (en) * | 2023-03-22 | 2023-04-21 | 清华大学 | Non-visual field imaging method and device based on transient signal super-resolution network |
| CN115993611B (en) * | 2023-03-22 | 2023-06-20 | 清华大学 | Non-visual field imaging method and device based on transient signal super-resolution network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103971125A (en) | Super-resolution algorithm based on vibration signal of laser echo | |
| CN104123464B (en) | Method for inversion of ground feature high elevation and number of land subsidence through high resolution InSAR timing sequence analysis | |
| RU2626184C2 (en) | Method, device and system for reconstructing magnetic resonance image | |
| CN102148987B (en) | Compressed sensing image reconstructing method based on prior model and 10 norms | |
| CN102541954B (en) | Method and system for searching trademarks | |
| CN105046651B (en) | A kind of ultra-resolution ratio reconstructing method and device of image | |
| EP2927844A1 (en) | 3d object position and pose estimation | |
| CN111798135B (en) | High-speed rail settlement hazard assessment method based on multi-source data integration | |
| CN111597867B (en) | Face recognition method and system based on millimeter wave image | |
| CN105139339A (en) | Polarization image super-resolution reconstruction method based on multi-level filtering and sample matching | |
| CN109359333A (en) | A kind of body Model construction method comprising multiple dimensioned shape characteristic | |
| KR102309343B1 (en) | Frequency-wavenumber analysis method and apparatus through deep learning-based super resolution ground penetrating radar image generation | |
| CN108663576A (en) | Weak electromagnetic red signal detection method under a kind of complex environment | |
| CN106842198A (en) | A kind of subband fusion of imaging algorithm based on convex optimum theory | |
| CN111222544B (en) | Ground simulation test system for influence of satellite flutter on camera imaging | |
| CN117788379A (en) | Domain transformation-based end-to-end heterogeneous remote sensing image change detection method | |
| Liu et al. | SCBSS signal de-noising method of integrating EEMD and ESMD for dynamic deflection of bridges using GBSAR | |
| CN101666879B (en) | Method for improving resolution of linear-array three-dimensional imaging synthetic aperture radars | |
| CN110231625A (en) | A kind of aperture synthesis imaging method based on Multiscale Fusion | |
| CN102156872A (en) | Multispectral-data-based object identification method and device | |
| CN112819909B (en) | Self-adaptive coding method based on low-resolution priori spectrum image region segmentation | |
| CN103714572A (en) | Three-step hyperspectral image fusion method based on spectrum reconstruction | |
| CN113160063A (en) | Method for denoising tunnel image based on wavelet and median filtering | |
| Marnissi et al. | A data augmentation approach for sampling Gaussian models in high dimension | |
| Peng et al. | Millimeter wave image super resolution using multichannel depth convolution neural network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20140806 |