CN107390259A - A kind of nuclide identification method based on SVD and SVM - Google Patents

A kind of nuclide identification method based on SVD and SVM Download PDF

Info

Publication number
CN107390259A
CN107390259A CN201710575457.9A CN201710575457A CN107390259A CN 107390259 A CN107390259 A CN 107390259A CN 201710575457 A CN201710575457 A CN 201710575457A CN 107390259 A CN107390259 A CN 107390259A
Authority
CN
China
Prior art keywords
dimensional
nucleic
characteristic vector
gamma spectrum
vector
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
Application number
CN201710575457.9A
Other languages
Chinese (zh)
Inventor
张江梅
高宁化
冯兴华
王坤朋
姚娟
张华�
王姮
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Southwest University of Science and Technology
Original Assignee
Southwest University of Science and Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Southwest University of Science and Technology filed Critical Southwest University of Science and Technology
Priority to CN201710575457.9A priority Critical patent/CN107390259A/en
Publication of CN107390259A publication Critical patent/CN107390259A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01T—MEASUREMENT OF NUCLEAR OR X-RADIATION
    • G01T1/00—Measuring X-radiation, gamma radiation, corpuscular radiation, or cosmic radiation
    • G01T1/36—Measuring spectral distribution of X-rays or of nuclear radiation spectrometry
    • G01T1/38—Particle discrimination and measurement of relative mass, e.g. by measurement of loss of energy with distance (dE/dx)
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02—Preprocessing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/12—Classification; Matching

Abstract

The present invention relates to nuclear radiation detection technique field, discloses a kind of nuclide identification method based on SVD and SVM.Including：Step 1：Obtain the one-dimensional gamma spectrum figure of nucleic；Step 2：The two-dimensional matrix according to corresponding to obtaining one-dimensional gamma spectrum figure；Step 3：The characteristic vector of two-dimensional matrix is extracted according to the method for singular value decomposition, obtained characteristic vector represents nucleic；Step 4：Using actually measured multigroup one-dimensional gamma spectrum figure as Sample Storehouse, corresponding characteristic vector in Sample Storehouse is obtained by step 1~3, SVM classifier is obtained as training sample using characteristic vector；Step 5：By unknown nucleic by step 1~3 processing acquisitions characteristic vector to be measured, characteristic vector input SVM classifier to be measured is identified.By measuring the gamma-spectrometric data of a large amount of known nucleic, and grader is built by support vector cassification algorithm, then measure unknown nucleic, this method represents nucleic by obtaining one-dimensional vector and two-dimensional matrix and extracting characteristic value, has prompted nuclide identification efficiency.

Description

A kind of nuclide identification method based on SVD and SVM
Technical field
The present invention relates to nuclear radiation detection technique field, and in particular to a kind of nuclide identification method based on SVD and SVM.
Background technology
Recently as the development of science and technology, nuclear technology is in national defence, industry, agricultural, environmental monitoring, soil constituent point Analysis, research nano material etc., which have had, to be widely applied.Nuclear power is as a kind of novel energy, in International Power resource Shared ratio also increasingly increases, and this is the most typical example that radio isotope is applied to national economy.At the same time, core Safety problem turns into multi-party focus of attention.For caused by the illegal utilization, smuggling and nuclear leakage accident of nuclear material instantly How radiological hazard etc., quickly and accurately detected to radioactive nucleus material and identification is that kernel analysis field is urgent instantly Need to solve the problems, such as.
For same nucleic, count within a certain period of time, the feature of resulting power spectrum is essentially the same, some specific Count value on energy is much greater, causes occur crest in energy spectrum diagram, and these positions are referred to as the feature peak position of power spectrum by we Put, traditional nuclide identification method is that the characteristic peak that will be measured in gained power spectrum is compared with standard nuclide library, is sentenced with this The species of disconnected nucleic.In order to find the characteristic peak of power spectrum, it is necessary to carry out denoising to the power spectrum surveyed, smoothly, filtering, peak-seeking etc. Reason, process is cumbersome, takes relatively long, and is applied only for measuring the partial information of power spectrum, robustness by comparing the method for characteristic peak Poor, when environment changes, power spectrum is changed therewith, and recognition result can be impacted.
The content of the invention
The technical problems to be solved by the invention are：For above-mentioned problem, there is provided one kind is based on SVD and SVM Nuclide identification method.
The technical solution adopted by the present invention is as follows：A kind of nuclide identification method based on SVD and SVM, including procedure below：
Step 1：Obtain the one-dimensional gamma spectrum figure of nucleic；
Step 2：The two-dimensional matrix according to corresponding to obtaining one-dimensional gamma spectrum figure；
Step 3：The characteristic vector of two-dimensional matrix is extracted according to the method for singular value decomposition, obtained characteristic vector represents core Element；
Step 4：Using actually measured multigroup one-dimensional gamma spectrum figure as Sample Storehouse, obtained by step 1~3 in Sample Storehouse Corresponding characteristic vector, SVM classifier is obtained as training sample using characteristic vector；
Step 5：By unknown nucleic to be measured by step 1~3 processing acquisitions characteristic vector to be measured, by spy to be measured Sign vector input SVM classifier is identified.
Further, the step 1 includes procedure below：Detector detects particle and produces signal pulse, and data are adopted Truck collection signal pulse deposit caching, software read calculator memory, the mode of operation of Computercontrolled data acquisition card, Computer analyzes and processes to signal pulse data, obtains one-dimensional gamma spectrum figure.
Further, the step 2 includes procedure below：
Step 21：One-dimensional gamma spectrum figure is chosen, by the count value corresponding to energy calibration on whole one-dimensional gamma spectrum figure Save as an one-dimensional vector；
Step 22：The one-dimensional vector is converted to by two-dimensional matrix by matrixing；
Step 23：Each element of two-dimensional matrix is normalized, 0 to 255 is then mapped to, obtains new Two-Dimensional Moment Battle array.
Compared with prior art, having the beneficial effect that using above-mentioned technical proposal：
(1) by extracting whole one-dimensional gamma spectrum figure, whole spectral informations is applied, tradition is solved and compares characteristics of energy spectrum The shortcomings that method is applied only for part characteristics of energy spectrum information is identified in peak, improves the robustness of nuclide identification.Moreover, pass through The extraction of one-dimensional vector and two-dimensional matrix, it is not necessary to the work such as smooth, denoising, filtering, peak-seeking are carried out to the power spectrum measured, simultaneously Complete pretreatment and feature extraction, it is not necessary to carry out the parameter setting of complexity, simplify the processing to original energy spectrum diagram, significantly Workload is reduced, improves the efficiency of nuclide identification.Secondly, the thought of pattern-recognition is introduced, passes through existing ripe figure As identifying that sorting algorithm carries out Classification and Identification to nucleic, the accuracy rate of nuclide identification is improved, greatly accelerates the identification of nucleic Time.
(2) using the method extraction characteristics of energy spectrum of singular value decomposition (SVD), subtract as far as possible while information required for reservation Lack redundancy, serve denoising, the effect of filtering, and effect is more preferable, reduces environmental exact details shadow to caused by power spectrum Ring.Simultaneously because it have compressed information content so that identification process is quicker, is easy to judge, and then improves the identification speed of nucleic Degree and efficiency.
Brief description of the drawings
Fig. 1 is the schematic flow sheet of the nuclide identification method of the invention based on SVD and SVM.
Fig. 2 is the data acquisition flow figure of the one-dimensional gamma spectrum figure of the present invention.
Fig. 3 is the original one-dimensional gamma spectrum figures of nucleic Co60 of the present invention.
Fig. 4 is the nucleic Co60 energy spectrum diagram pictures that the present invention is obtained later by singular value decomposition.
Embodiment
The present invention is described further below in conjunction with the accompanying drawings.
As shown in figure 1, a kind of nuclide identification method based on SVD and SVM, including procedure below：
Step 1：Obtain the one-dimensional gamma spectrum figure of nucleic；
Obtaining the idiographic flow of one-dimensional gamma spectrum figure is：As shown in Fig. 2 one-dimensional gamma spectrum figure measurement apparatus is visited including NaI Device, data collecting card, computer are surveyed, NaI detectors first measure in the place for having radioactive source, probe detection to particle, Pulse signal is produced, real-time sampling is then carried out to the pulse signal of NaI detectors using data collecting card, deposit computer delays Deposit, read calculator memory by software and data are analyzed and processed by computer, finally by output computer equipment Output obtains one-dimensional gamma spectrum figure.The mode of operation of data collecting card by computer software control and is realized in whole process.With It is the one-dimensional gamma spectrum figure obtained according to the above method as shown in Figure 3 exemplified by nucleic Elements C o60.
Step 2：The two-dimensional matrix according to corresponding to obtaining one-dimensional gamma spectrum figure；
Detailed process is：
Step 21：One-dimensional gamma spectrum figure is chosen, by the count value corresponding to energy calibration on whole one-dimensional gamma spectrum figure An one-dimensional vector is saved as, realizes using whole energy spectrum diagram as integrally, improves the robustness of nuclide identification；Step 22：It is logical Cross matrixing and the one-dimensional vector is converted into two-dimensional matrix；Step 23：Each element of two-dimensional matrix is normalized, returned One change after each element value between [0,1], minimum in element to be changed into 0, maximum is changed into 1, then by the value after normalization 0 to 255 are mapped to, obtains new two-dimensional matrix.New two-dimensional matrix is shown as a secondary gray level image in MATLAB, this ash Degree figure just reflects the information that whole gamma spectrum is included.0-255 represents different colouring informations, and 0 represents black, and 255 represent White, color change from deep to shallow.
Step 3：The characteristic vector of two-dimensional matrix, obtained characteristic vector are extracted according to the method for singular value decomposition (SVD) Represent nucleic；Because the data volume of two-dimensional matrix is larger, directly it is identified that efficiency is low, and speed is slow, so in order to improve nucleic Recognition efficiency, feature extraction is carried out to two-dimensional matrix using the method for singular value decomposition (SVD), reaches the purpose of dimensionality reduction, is protecting The complexity for compressed data of being tried one's best while staying data characteristics.
Singular value decomposition (SVD) is a kind of method of a decomposition that can be suitably used for Arbitrary Matrix, i.e., by a matrix A ∈ RN×MRepresented by formula (1)：
A=U ∑s VT (1)
Orthogonal matrix U=[u in formula1…uN]∈RN×N, V=[v1…vM]∈RM×MMeet UTU=IN, VTV=IM, ∑ is pair Angular moment battle array satisfaction, ∑=diag (σ1…σP)∈RN×M, wherein p=min (M, N), singular value σ1≥σ2≥…≥σp>=0, one As in the case of, preceding 10% or even 5% singular value sum just occupies more than the 99% of all singular value sums, that is, Say, we can also eliminate noise and phase with singular value big preceding r come approximate description matrix while data dimension is reduced The influence of closing property.
Usual gamma spectrum can be regarded as one wide steady random vector, and the vector can regard the row vector that length is L as, such as Shown in formula (2)：
Y=[y1..., yL] (2)
Wherein yiFor the counting on i-th.It is analysis object with full spectrum when extracting feature with singular value decomposition, it is not necessary to Initial parameter is set, automated intelligent analysis is realized in the case where ensureing accuracy.Formula (2) is expressed as such as formula by we (3) matrix
L=m × n in formula, i.e., one-dimensional gamma-spectrometric data is for conversion into the view data of two dimension.Next Co60 power spectrums are used Exemplified by figure, singular value decomposition is carried out to matrix Y, makes Y=U σ VT, solve Y singular value vector σ.And preceding 9 singular value bags in σ In the vector 90% energy is contained, has made preceding 9 row that U^ is U respectively, V^ is V preceding 9 row, 9 before singular value σ, thus σ ^ are Obtain the estimation Y^ of initial data
Y^=U^ σ ^V^T (4)
And recover the Co60 of gained gamma spectrum figure as shown in figure 4, comparison diagram 2 and Fig. 4 is, it is apparent that by unusual The signal statistics fluctuation that recovers reduces after value is decomposed, and the FWHM of full energy peak broadens and more obvious.
Pretreatment and feature extraction are so completed simultaneously using whole power spectrum as overall, it is not necessary to carry out complex parameters Setting, so as to substantially increase the efficiency of nuclide identification, shortens the time of nuclide identification.
Step 4：Using actually measured multigroup one-dimensional gamma spectrum figure as Sample Storehouse, obtained by step 1~3 in Sample Storehouse Corresponding characteristic vector, SVM classifier is obtained as training sample using characteristic vector；
The known nucleic species that measures is Am241, Ba133, Co60 in the present embodiment, Cs137, I131, Ra226 totally six Kind, the gamma-spectrometric data of 200 groups of above-mentioned known nucleic is obtained with step 1 measurement under various circumstances, and pass through step 2, step 3 The characteristic vector of known nucleic is obtained, primordial nuclide is characterized with this feature vector, then using these characteristic vectors as training Sample, obtain a suitable SVM classifier.
SVM methods are primarily directed to classification problem, defining classification function：
F (x)=ωTx+b (5)
X represents the attribute of sample, and ω is vectorial with x dimension identical, and b is a constant, and f (x) represents the classification of sample Label, be Optimal Separating Hyperplane by f (x) planes determined, as f (x)=0, x is point on hyperplane, and it is two that x is divided to by sgn [f (x)] Class (1 and -1).Make sample completely separable in the presence of one or more f (x), and a most optimal sorting in all Optimal Separating Hyperplanes be present Class face makes to obtain the distance between the nearest sample of hyperplane and this hyperplane maximum, i.e.,It is minimum.It is translated into two Suboptimization problem is：
In formula, ξiFor the slack variable of each sample point, yiIt is the class label of training sample, C is punishment parameter.For solution The double optimization problem, by constructing Suzanne Lenglen day multiplier：
In formula：αi>=0, μi>=0 is Lagrange multipliers,For class interval.By being solved to formula (7), obtain To optimal solution (ω*, b*), so as to obtain the maximum optimal classification surface in interval.So far, SVM classifier structure is completed.
In order to verify validity of the singular value vector σ ^ as input, by Am241, Ba133, Co60, Cs137, I131, The σ ^ of totally six kinds of nucleic are projected its distribution are observed into space Ra226, and the description of multidimensional data is relatively difficult, therefore we are only First three singular value is chosen in σ ^ to observe its distribution characteristics, every group of nucleic randomly selects 100 sample points, six kinds of nucleic of acquisition Singular value space distribution map, experiment prove six kinds of nucleic singular value can divide in three dimensions, with reference to more higher-dimension Characteristic information, identification can be more effectively completed.
Step 5：By unknown nucleic to be measured by step 1~3 processing acquisitions characteristic vector to be measured, by spy to be measured Sign vector input SVM classifier is identified, and draws the species of nucleic to be identified.
The invention is not limited in foregoing embodiment.The present invention, which expands to, any in this manual to be disclosed New feature or any new combination, and disclose any new method or process the step of or any new combination.If this Art personnel, it is altered or modified in the unsubstantiality that the spirit for not departing from the present invention is done, should all belongs to power of the present invention The claimed scope of profit.

Claims (3)

  1. A kind of 1. nuclide identification method based on SVD and SVM, it is characterised in that including procedure below：
    Step 1：Obtain the one-dimensional gamma spectrum figure of nucleic；
    Step 2：The two-dimensional matrix according to corresponding to obtaining one-dimensional gamma spectrum figure；
    Step 3：The characteristic vector of two-dimensional matrix is extracted according to the method for singular value decomposition, obtained characteristic vector represents nucleic；
    Step 4：Using actually measured multigroup one-dimensional gamma spectrum figure as Sample Storehouse, obtained by step 1~3 corresponding in Sample Storehouse Characteristic vector, using characteristic vector as training sample obtain SVM classifier；
    Step 5：Unknown nucleic to be measured is obtained into characteristic vector to be measured by the processing of step 1~3, by feature to be measured to Amount input SVM classifier is identified.
  2. 2. the nuclide identification method based on SVD and SVM as claimed in claim 1, it is characterised in that
    The step 1 includes procedure below：Detector detects particle and produces signal pulse, data collecting card collection signal arteries and veins Punching deposit caching, software read calculator memory, and the mode of operation of Computercontrolled data acquisition card, computer is to signal arteries and veins Rush data to be analyzed and processed, obtain one-dimensional gamma spectrum figure.
  3. 3. the nuclide identification method based on SVD and SVM as claimed in claim 2, it is characterised in that the step 2 include with Lower process：
    Step 21：One-dimensional gamma spectrum figure is chosen, the count value corresponding to energy calibration on whole one-dimensional gamma spectrum figure is preserved For an one-dimensional vector；
    Step 22：The one-dimensional vector is converted to by two-dimensional matrix by matrixing；
    Step 23：Each element of two-dimensional matrix is normalized, 0 to 255 is then mapped to, obtains new two-dimensional matrix.
CN201710575457.9A 2017-07-14 2017-07-14 A kind of nuclide identification method based on SVD and SVM Pending CN107390259A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710575457.9A CN107390259A (en) 2017-07-14 2017-07-14 A kind of nuclide identification method based on SVD and SVM

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710575457.9A CN107390259A (en) 2017-07-14 2017-07-14 A kind of nuclide identification method based on SVD and SVM

Publications (1)

Publication Number Publication Date
CN107390259A true CN107390259A (en) 2017-11-24

Family

ID=60339763

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710575457.9A Pending CN107390259A (en) 2017-07-14 2017-07-14 A kind of nuclide identification method based on SVD and SVM

Country Status (1)

Country Link
CN (1) CN107390259A (en)

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109063741A (en) * 2018-07-05 2018-12-21 南京航空航天大学 A kind of energy spectrum analysis method based on hibert curve transformation and deep learning
CN109635650A (en) * 2018-11-06 2019-04-16 中国电子科技集团公司电子科学研究院 The recognition methods of the nucleic type of gamma-spectrometric data
CN111008356A (en) * 2019-11-13 2020-04-14 成都理工大学 WTSVD algorithm-based background-subtracted gamma energy spectrum set analysis method
CN111539324A (en) * 2020-04-23 2020-08-14 重庆建安仪器有限责任公司 Novel nuclide identification method
CN112101058A (en) * 2020-08-17 2020-12-18 武汉诺必答科技有限公司 Method and device for automatically identifying test paper bar code
CN112613484A (en) * 2021-01-06 2021-04-06 西南石油大学 Gas transmission pipeline leakage identification method based on singular spectrum analysis and support vector machine
CN114371494A (en) * 2022-03-22 2022-04-19 西南科技大学 Radioactive source scene simulation method for autonomous sourcing robot

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3988586A (en) * 1975-06-24 1976-10-26 The United States Of America As Represented By The United States Energy Research And Development Administration Combination neutron-gamma ray detector
CN103424766A (en) * 2013-03-19 2013-12-04 中国人民解放军第二炮兵工程大学 Nuclide rapid identification method based on pattern recognition
CN105607111A (en) * 2014-11-05 2016-05-25 中国科学院高能物理研究所 Gamma nuclide identification method
JP2017058355A (en) * 2015-09-14 2017-03-23 国立研究開発法人産業技術総合研究所 Radiation meter with uncertainty evaluation function

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3988586A (en) * 1975-06-24 1976-10-26 The United States Of America As Represented By The United States Energy Research And Development Administration Combination neutron-gamma ray detector
CN103424766A (en) * 2013-03-19 2013-12-04 中国人民解放军第二炮兵工程大学 Nuclide rapid identification method based on pattern recognition
CN105607111A (en) * 2014-11-05 2016-05-25 中国科学院高能物理研究所 Gamma nuclide identification method
JP2017058355A (en) * 2015-09-14 2017-03-23 国立研究開発法人産業技術総合研究所 Radiation meter with uncertainty evaluation function

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
任俊松 等: "《基于SVD和SVM的核素识别算法》", 《兵工自动化》 *

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109063741A (en) * 2018-07-05 2018-12-21 南京航空航天大学 A kind of energy spectrum analysis method based on hibert curve transformation and deep learning
CN109063741B (en) * 2018-07-05 2021-12-10 南京航空航天大学 Energy spectrum analysis method based on Hilbert curve transformation and deep learning
CN109635650A (en) * 2018-11-06 2019-04-16 中国电子科技集团公司电子科学研究院 The recognition methods of the nucleic type of gamma-spectrometric data
CN111008356A (en) * 2019-11-13 2020-04-14 成都理工大学 WTSVD algorithm-based background-subtracted gamma energy spectrum set analysis method
CN111539324A (en) * 2020-04-23 2020-08-14 重庆建安仪器有限责任公司 Novel nuclide identification method
CN112101058A (en) * 2020-08-17 2020-12-18 武汉诺必答科技有限公司 Method and device for automatically identifying test paper bar code
CN112613484A (en) * 2021-01-06 2021-04-06 西南石油大学 Gas transmission pipeline leakage identification method based on singular spectrum analysis and support vector machine
CN114371494A (en) * 2022-03-22 2022-04-19 西南科技大学 Radioactive source scene simulation method for autonomous sourcing robot

Similar Documents

Publication Publication Date Title
CN107390259A (en) A kind of nuclide identification method based on SVD and SVM
Choma et al. Graph neural networks for icecube signal classification
CN107451614B (en) Hyperspectral classification method based on fusion of space coordinates and space spectrum features
CN104966085B (en) A kind of remote sensing images region of interest area detecting method based on the fusion of more notable features
CN109063741B (en) Energy spectrum analysis method based on Hilbert curve transformation and deep learning
Liang et al. Rapid nuclide identification algorithm based on convolutional neural network
CN102867183B (en) Method and device for detecting littered objects of vehicle and intelligent traffic monitoring system
CN104834942A (en) Remote sensing image change detection method and system based on mask classification
CN104408705A (en) Anomaly detection method of hyperspectral image
CN103955926A (en) Method for remote sensing image change detection based on Semi-NMF
CN109948593A (en) Based on the MCNN people counting method for combining global density feature
CN102945374B (en) Method for automatically detecting civil aircraft in high-resolution remote sensing image
Yuan et al. Systematic benchmarking of aerial image segmentation
CN104268556A (en) Hyperspectral image classification method based on nuclear low-rank representing graph and spatial constraint
CN102158486A (en) Method for rapidly detecting network invasion
CN107219510B (en) Radar target identification method based on unlimited largest interval linear discriminant projection model
CN110488368A (en) A kind of contraband recognition methods and device based on dual intensity X-ray screening machine
CN104408472A (en) Wishart and SVM (support vector machine)-based polarimetric SAR (synthetic aperture radar) image classification method
CN103093243A (en) High resolution panchromatic remote sensing image cloud discriminating method
Zhang et al. On using XMC R-CNN model for contraband detection within X-ray baggage security images
CN105809092A (en) Population target detection method and device thereof
CN104318224A (en) Face recognition method and monitoring equipment
Fu et al. A noise-resistant superpixel segmentation algorithm for hyperspectral images
CN103268476B (en) A kind of Remote Sensing Target monitoring method
CN116385957A (en) X-ray image contraband detection method, device, equipment and medium

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
RJ01 Rejection of invention patent application after publication

Application publication date: 20171124