CN111412987A - Dynamic scanning detection device and method based on spectral characteristics - Google Patents
Dynamic scanning detection device and method based on spectral characteristics Download PDFInfo
- Publication number
- CN111412987A CN111412987A CN201910433146.8A CN201910433146A CN111412987A CN 111412987 A CN111412987 A CN 111412987A CN 201910433146 A CN201910433146 A CN 201910433146A CN 111412987 A CN111412987 A CN 111412987A
- Authority
- CN
- China
- Prior art keywords
- feature
- characteristic
- data
- deflection angle
- data set
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 150
- 230000003595 spectral effect Effects 0.000 title claims abstract description 99
- 238000000034 method Methods 0.000 title claims abstract description 34
- 239000000126 substance Substances 0.000 claims abstract description 102
- 238000001228 spectrum Methods 0.000 claims abstract description 52
- 239000013598 vector Substances 0.000 claims description 54
- 230000000875 corresponding effect Effects 0.000 claims description 51
- 238000004140 cleaning Methods 0.000 claims description 24
- 238000012545 processing Methods 0.000 claims description 18
- 238000009826 distribution Methods 0.000 claims description 12
- 239000000463 material Substances 0.000 claims description 12
- 230000002596 correlated effect Effects 0.000 claims description 10
- 239000011159 matrix material Substances 0.000 claims description 7
- 238000001914 filtration Methods 0.000 claims description 4
- 230000001678 irradiating effect Effects 0.000 claims description 4
- 239000010421 standard material Substances 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 2
- 238000012360 testing method Methods 0.000 claims description 2
- 238000007689 inspection Methods 0.000 claims 2
- 230000008569 process Effects 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 16
- 238000004458 analytical method Methods 0.000 description 10
- 238000000605 extraction Methods 0.000 description 7
- 239000000203 mixture Substances 0.000 description 6
- 238000012986 modification Methods 0.000 description 5
- 230000004048 modification Effects 0.000 description 5
- 238000013461 design Methods 0.000 description 4
- 239000000523 sample Substances 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000005215 recombination Methods 0.000 description 2
- 230000006798 recombination Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000010183 spectrum analysis Methods 0.000 description 2
- 238000010521 absorption reaction Methods 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 229910052792 caesium Inorganic materials 0.000 description 1
- TVFDJXOCXUVLDH-UHFFFAOYSA-N caesium atom Chemical compound [Cs] TVFDJXOCXUVLDH-UHFFFAOYSA-N 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 210000000887 face Anatomy 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 229910052734 helium Inorganic materials 0.000 description 1
- 239000001307 helium Substances 0.000 description 1
- SWQJXJOGLNCZEY-UHFFFAOYSA-N helium atom Chemical compound [He] SWQJXJOGLNCZEY-UHFFFAOYSA-N 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 210000000554 iris Anatomy 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 210000001747 pupil Anatomy 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 229910052701 rubidium Inorganic materials 0.000 description 1
- IGLNJRXAVVLDKE-UHFFFAOYSA-N rubidium atom Chemical compound [Rb] IGLNJRXAVVLDKE-UHFFFAOYSA-N 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000011524 similarity measure Methods 0.000 description 1
- 238000004611 spectroscopical analysis Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J3/00—Spectrometry; Spectrophotometry; Monochromators; Measuring colours
- G01J3/28—Investigating the spectrum
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J3/00—Spectrometry; Spectrophotometry; Monochromators; Measuring colours
- G01J3/02—Details
- G01J3/06—Scanning arrangements arrangements for order-selection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/12—Fingerprints or palmprints
- G06V40/13—Sensors therefor
- G06V40/1318—Sensors therefor using electro-optical elements or layers, e.g. electroluminescent sensing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/12—Fingerprints or palmprints
- G06V40/1347—Preprocessing; Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/12—Fingerprints or palmprints
- G06V40/1365—Matching; Classification
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J3/00—Spectrometry; Spectrophotometry; Monochromators; Measuring colours
- G01J3/02—Details
- G01J3/06—Scanning arrangements arrangements for order-selection
- G01J2003/064—Use of other elements for scan, e.g. mirror, fixed grating
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J3/00—Spectrometry; Spectrophotometry; Monochromators; Measuring colours
- G01J3/28—Investigating the spectrum
- G01J2003/283—Investigating the spectrum computer-interfaced
- G01J2003/284—Spectral construction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/58—Extraction of image or video features relating to hyperspectral data
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A40/00—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production
- Y02A40/10—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production in agriculture
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Human Computer Interaction (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention provides a dynamic scanning detection device and a method based on spectral characteristics, wherein the device comprises: the light emitter is connected with the spectral characteristic data processor; the micro-electro-mechanical system scanning galvanometer is connected with the spectral characteristic data processor; the spectrum collector is connected with the spectrum characteristic data processor; the spectral characteristic data processor is connected with the light emitter, the micro-electro-mechanical system scanning galvanometer, the spectrum collector and the substance spectrum detector; extracting each feature data from the relevant feature data set; and the substance spectrum detector is connected with the spectral characteristic data processor. The invention comprehensively processes and analyzes a plurality of groups of hyperspectral data in an integrated manner, obtains the component characteristics of the detected substance from the integrated level, and realizes efficient and accurate substance detection and identification.
Description
Technical Field
The invention relates to the technical field of material detection, in particular to a dynamic scanning detection device and method based on spectral characteristics.
Background
The spectral analysis has the advantages of high sensitivity and rapid detection when used for detecting substances, and refers to a method for identifying the substances and determining the chemical compositions and relative contents of the substances according to the spectra of the substances. Historically, many new elements have been discovered by spectroscopic analysis, such as rubidium, cesium, helium, and the like. The chemical composition and relative content of a substance are identified by measuring the wavelength and intensity of emission, absorption or scattering resulting from energy level transitions occurring within the substance when the substance is exposed to radiation. Therefore, the collection of data in the spectral analysis is very important, and the accuracy of the analysis result of the substance is directly influenced by the quality of the data.
The traditional spectral feature identification method is to collect hyperspectral data of a certain fixed area of an object to be detected, establish new features reflecting comprehensive information of substances by recombination and optimization, and then complete identification and classification of different objects by utilizing clustering of different detection objects or analysis modeling of large-scale data. However, the method has certain limitations, and since the structure and material distribution of the detection object are generally uneven, there is a certain difference in the hyperspectral data acquired in different areas, that is, when performing spectrum detection on an object with a complex structure or composition, there is an obvious difference between the actually acquired data and the data acquired when establishing a model, and finally the originally established model cannot be accurately identified with respect to the object.
The accuracy in detecting the substance by using the spectrum is one of important indexes for measuring the quality of a detection system, for an object with a complex structure or component composition, only spectrum data of a certain fixed region is collected to perform overall judgment on the detected object, which is very one-sided and not strict, and the collection of the fixed region information easily leads to wrong judgment, possibly affecting the subsequent judgment on the substance characteristic.
Therefore, how to provide a spectrum detection system for increasing effective information acquisition on a detection object is a technical problem to be solved in the field, which is to solve the problem of low recognition accuracy.
Disclosure of Invention
The invention aims to provide a dynamic scanning detection device and method based on spectral characteristics, and solves the technical problem that in the prior art, the whole judgment of a detected object is carried out by only acquiring spectral data of a certain fixed area of an object, so that the detection result of a substance is one-sided and not strict.
To achieve the above object, the present invention provides
A dynamic scan test device based on spectral characteristics, comprising: a light emitter, a micro-electro-mechanical system scanning galvanometer, a spectrum collector, a spectrum characteristic data processor and a substance spectrum detector, wherein,
the illuminator is connected with the spectral characteristic data processor and emits incident light to irradiate the micro-electro-mechanical system scanning galvanometer;
the micro electro mechanical system scanning galvanometer is connected with the spectral characteristic data processor and adjusts the deflection angle of the micro electro mechanical system scanning galvanometer in a preset deflection angle adjusting mode;
the spectrum collector is connected with the spectrum characteristic data processor and is used for collecting hyperspectral data of the scanning detection area under each deflection angle;
the spectral characteristic data processor is connected with the illuminator, the micro-electro-mechanical system scanning galvanometer, the spectrum collector and the substance spectrum detector, establishes a spatial relationship hyperspectral data set corresponding to the scanning detection area and the hyperspectral data, and extracts characteristic data to form a multidimensional characteristic vector set;
the substance spectrum detector is connected with the spectrum characteristic data processor and matches each characteristic point in the multi-dimensional characteristic vector set with the characteristic point of the gradient layer corresponding to the standard substance to form a characteristic point matching pair; and calculating the similarity of the feature point matching pairs, and determining the material characteristics in the scanning detection area according to the similarity.
Optionally, the mems scanning galvanometer is configured to deflect the incident light to a scanning detection area at a preset initial deflection angle; adjusting the deflection angle of the scanning galvanometer of the micro electro mechanical system in a preset deflection angle adjusting mode;
the spectral characteristic data processor is used for establishing a spatial relationship hyperspectral data set corresponding to the scanning detection area and the hyperspectral data based on the scanning detection area, the deflection angle and the hyperspectral data; according to a preset uncorrelated noise characteristic, cancelling uncorrelated hyperspectral data in the spatial relationship hyperspectral data set to obtain a correlated characteristic data set;
extracting each feature data from the relevant feature data set, generating a scale space according to the feature data, constructing a differential pyramid, and searching extreme points which are not influenced by the scale in the scale space as feature data set key points; distributing a direction for each key point by utilizing the gradient distribution of the neighborhood characteristics of the key points to form a multi-dimensional characteristic vector set for describing the local gradient characteristics of the related characteristic data set;
the substance spectrum detector is used for matching each characteristic point in the multi-dimensional characteristic vector set with the characteristic point of the corresponding gradient layer of the standard substance in a scale space to form a characteristic point matching pair; and calculating the similarity of the feature point matching pairs, and determining that the substance in the scanning detection area is the standard substance when the similarity reaches or exceeds a preset similarity threshold value.
Optionally, the mems scanning galvanometer, comprising: a deflection angle adjustment mode acquisition unit and a deflection angle adjustment unit; wherein,
the deflection angle adjustment mode acquisition unit is connected with the deflection angle adjustment unit, receives shape information of an object to be detected in the scanning detection area, and compares preset object shape information with a corresponding relation of a deflection angle adjustment mode according to the shape information of the object to be detected to obtain a deflection angle adjustment mode for detecting the object to be detected;
the deflection angle adjusting unit is connected with the deflection angle adjusting mode acquiring unit and the spectral characteristic data processor to detect the deflection angle adjusting mode of the object to be detected and adjust the deflection angle of the scanning galvanometer of the micro-electro-mechanical system.
Optionally, the spectral feature data processor comprises: the system comprises a spatial relation hyperspectral data set processing unit, a related feature data set cleaning unit and a multi-dimensional feature vector set generating unit; wherein,
the spatial relationship hyperspectral data set processing unit is connected with the illuminator, the micro-electro-mechanical system scanning galvanometer, the spectrum collector and the related characteristic data set cleaning unit, and establishes a spatial relationship hyperspectral data set corresponding to the scanning detection area and the hyperspectral data based on the scanning detection area, the deflection angle and the hyperspectral data;
the relevant feature data set cleaning unit is connected with the spatial relationship hyperspectral data set processing unit and the multi-dimensional feature vector set generating unit and eliminates irrelevant hyperspectral data in the spatial relationship hyperspectral data set according to a preset irrelevant noise feature;
the multi-dimensional feature vector set generating unit is connected with the relevant feature data set cleaning unit and the substance spectrum detector, the same type of features in different areas are used as a matrix, all layers of different features form a three-dimensional feature data set, and the feature layers in the three-dimensional feature data set are subjected to fuzzy and downsampling processing by utilizing local characteristics to obtain a relevant feature data set;
extracting each feature data from the relevant feature data set, generating a scale space according to the feature data, constructing a differential pyramid, and searching extreme points which are not influenced by the scale in the scale space as feature data set key points; and distributing a direction for each key point by utilizing the gradient distribution of the neighborhood characteristics of the key points to form a multi-dimensional characteristic vector set for describing the local gradient characteristics of the related characteristic data set.
Optionally, the substance spectral detector comprises: the system comprises a spectral data characteristic point matching pair unit, a spectral data characteristic point matching pair cleaning unit and a substance judgment unit; wherein,
the spectrum data feature point matching pair unit is connected with the spectrum data feature point matching pair cleaning unit and the spectrum feature data processor, and matches each feature point in the multi-dimensional feature vector set with the feature point of the corresponding gradient layer of the standard substance in the scale space to form a feature point matching pair;
the spectral data feature point matching pair cleaning unit is connected with the spectral data feature point matching pair unit and the substance judgment unit, compares the positions, angles and scales of the feature points with the corresponding feature points of the standard substance, and filters out mismatching feature points with errors exceeding a preset error threshold value to obtain accurate feature point matching pairs;
and the substance judgment unit is connected with the spectral data characteristic point matching pair cleaning unit, calculates the similarity of the accurate characteristic point matching pair, and determines that the substance in the scanning detection area is the standard substance when the similarity reaches or exceeds a preset similarity threshold value.
In another aspect, the present invention further provides a dynamic scanning detection method based on spectral features, including:
irradiating the incident light emitted by the light emitter on the scanning galvanometer of the micro-electro-mechanical system;
adjusting the deflection angle of the scanning galvanometer of the micro electro mechanical system in a preset deflection angle adjusting mode;
collecting hyperspectral data of the scanning detection area under each deflection angle;
establishing a spatial relationship hyperspectral data set corresponding to the scanning detection area and the hyperspectral data, and extracting characteristic data to form a multi-dimensional characteristic vector set;
matching each characteristic point in the multi-dimensional characteristic vector set with the characteristic point of the corresponding gradient layer of the standard substance in a scale space to form a characteristic point matching pair; and calculating the similarity of the feature point matching pairs, and determining the material characteristics in the scanning detection area according to the similarity.
Optionally, the incident light emitted by the light emitter is irradiated on the mems scanning galvanometer, and the mems scanning galvanometer deflects the incident light to the scanning detection area at a preset initial deflection angle;
the method comprises the steps of establishing a spatial relationship hyperspectral data set corresponding to a scanning detection area and hyperspectral data, extracting characteristic data to form a multi-dimensional characteristic vector set, and establishing a spatial relationship hyperspectral data set corresponding to the scanning detection area and the hyperspectral data based on the scanning detection area, a deflection angle and the hyperspectral data; according to a preset uncorrelated noise characteristic, cancelling uncorrelated hyperspectral data in the spatial relationship hyperspectral data set to obtain a correlated characteristic data set;
extracting each feature data from the relevant feature data set, generating a scale space according to the feature data, constructing a differential pyramid, and searching extreme points which are not influenced by the scale in the scale space as feature data set key points; distributing a direction for each key point by utilizing the gradient distribution of the neighborhood characteristics of the key points to form a multi-dimensional characteristic vector set for describing the local gradient characteristics of the related characteristic data set;
matching each characteristic point in the multi-dimensional characteristic vector set with the characteristic point of the corresponding gradient layer of the standard substance in a scale space to form a characteristic point matching pair; calculating the similarity of the feature point matching pairs, determining the material characteristics in the scanning detection area according to the similarity, and matching each feature point in the multi-dimensional feature vector set with the feature point of the corresponding gradient layer of the standard material in a scale space to form the feature point matching pairs; and calculating the similarity of the feature point matching pairs, and determining that the substance in the scanning detection area is the standard substance when the similarity reaches or exceeds a preset similarity threshold value.
Optionally, adjusting the deflection angle of the mems scanning galvanometer in a preset deflection angle adjustment manner, where the deflection angle is:
receiving shape information of an object to be detected in the scanning detection area, and comparing preset object shape information with a corresponding relation of a deflection angle adjustment mode according to the shape information of the object to be detected to obtain a deflection angle adjustment mode for detecting the object to be detected;
and adjusting the deflection angle of the scanning galvanometer of the micro-electro-mechanical system in a manner of detecting the deflection angle of the object to be detected.
Optionally, according to a preset uncorrelated noise characteristic, cancelling uncorrelated hyperspectral data in the spatial relationship hyperspectral data set to obtain a correlated characteristic data set, where:
according to a preset uncorrelated noise characteristic, cancelling uncorrelated hyperspectral data in the spatial relationship hyperspectral data set;
the same type of features in different areas are used as a matrix, all layers with different features form a three-dimensional feature data set, and the feature layers in the three-dimensional feature data set are subjected to fuzzy and downsampling processing by utilizing local characteristics to obtain a related feature data set.
Optionally, matching each feature point in the multi-dimensional feature vector set with a feature point of a gradient layer corresponding to the standard substance in a scale space to form a feature point matching pair, where:
matching each characteristic point in the multi-dimensional characteristic vector set with the characteristic point of the corresponding gradient layer of the standard substance in a scale space to form a characteristic point matching pair;
and comparing the positions, angles and scales of the feature points with the corresponding feature points of the standard substance, filtering out mismatching feature points with errors exceeding a preset error threshold value, and obtaining accurate feature point matching pairs.
The dynamic scanning detection device and method based on the spectral characteristics have the following beneficial effects:
(1) the dynamic scanning detection device and method based on the spectral characteristics are used for realizing scanning type hyperspectral data acquisition of the detected substances by combining with the MEMS galvanometer, and the component characteristics of the detected substances are obtained from the whole level by integrally and comprehensively processing and analyzing a plurality of groups of acquired hyperspectral data, so that efficient and accurate substance detection and identification are realized.
(2) The dynamic scanning detection device and method based on the spectral characteristics change the light path by changing the rotation angle of the MEMS galvanometer, realize the detection of scanning type hyperspectral data acquisition of a detection object, and can realize the data acquisition of areas with different densities and different shapes by modifying the deflection parameters of the MEMS galvanometer, thereby achieving the effect of acquiring the detection data of substances from multiple angles in the most convenient mode.
(3) According to the dynamic scanning detection device and method based on the spectral characteristics, according to the characteristics of an analysis object, feature extraction is firstly carried out on feature data of scanning detection, then a hyperspectral feature data set is formed, invalid data such as noise and the like are filtered at the beginning, and the efficiency and accuracy of subsequent calculation and analysis are improved.
(4) According to the dynamic scanning detection device and method based on the spectral characteristics, the identification conclusion is not obtained through single hyperspectral data analysis, but the local characteristics of a plurality of groups of hyperspectral data obtained through scanning are analyzed, the collected data are subjected to multiple feature extraction to construct a multidimensional characteristic vector with robustness and uniqueness, and finally the substance detection result is obtained through analysis, so that the subsequent matching efficiency and performance are improved; in addition, various sensors can be combined to effectively fuse information, the structure or the components of the detection object can be effectively classified and identified, the relation among the data of the object can be utilized for further research and application, and the application scene of spectral feature scanning detection is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments described in the present invention, and other drawings can be obtained by those skilled in the art according to the drawings.
FIG. 1 is a schematic diagram of the difference of hyperspectral data acquired by detecting different areas of an object in the prior art;
FIG. 2 is a schematic structural diagram of a dynamic scanning detection apparatus based on spectral characteristics according to an embodiment of the present invention;
FIG. 3 is a schematic diagram illustrating a principle of extracting feature data from the relevant feature data set in FIG. 2 according to an embodiment of the present invention;
FIG. 4 is a schematic diagram illustrating a principle of multi-dimensional feature vector set extraction in FIG. 2 according to an embodiment of the present invention;
FIG. 5 is a schematic illustration of substance identification by spectral data feature matching in FIG. 2 in accordance with an embodiment of the present invention;
FIG. 6 is a schematic structural diagram of another dynamic scanning detection apparatus based on spectral characteristics according to an embodiment of the present invention;
FIG. 7 is a schematic structural diagram of another dynamic scanning detection apparatus based on spectral characteristics according to an embodiment of the present invention;
FIG. 8 is a schematic structural diagram of another dynamic scanning detection apparatus based on spectral characteristics according to an embodiment of the present invention;
FIG. 9 is a schematic flow chart of a dynamic scanning detection method based on spectral characteristics according to an embodiment of the present invention;
FIG. 10 is a schematic flow chart of another dynamic scanning detection method based on spectral characteristics according to an embodiment of the present invention;
FIG. 11 is a schematic flow chart illustrating another method for dynamic scanning detection based on spectral characteristics according to an embodiment of the present invention;
FIG. 12 is a schematic flow chart illustrating a dynamic scanning detection method based on spectral characteristics according to another embodiment of the present invention;
fig. 13 is a schematic diagram of the dynamic scanning detection method based on spectral features applied to an artificial neural network model in the embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention are clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Examples
The traditional spectral feature identification method is to collect hyperspectral data of a certain fixed area of an object to be detected, establish new features reflecting comprehensive information of substances by recombination and optimization, and then complete identification and classification of different objects by utilizing clustering of different detection objects or analysis modeling of large-scale data. However, the method has a certain limitation, that is, because the structure and material distribution of the detection object are generally uneven, as shown in fig. 1, which is a schematic diagram of differences in the hyperspectral data acquired by different regions of the detection object in the prior art, the hyperspectral data acquired by three different regions of the detection object A, B, C have a certain difference, that is, when an object with a complex structure or composition is detected, the actually acquired data and the data acquired when the model is built have a significant difference, so that the originally built model cannot be accurately identified with respect to the object.
As shown in fig. 2 to 5, fig. 2 is a schematic diagram of a dynamic scanning detection apparatus 200 based on spectral features in the present embodiment, and fig. 3 is a schematic diagram of a principle of extracting feature data from a related feature data set in fig. 2 in the present embodiment; FIG. 4 is a schematic diagram illustrating a principle of multi-dimensional feature vector set extraction in FIG. 2 according to this embodiment; FIG. 5 is a schematic diagram of substance identification by matching spectral data features in FIG. 2 according to the present embodiment; to solve the substance detection problem shown in fig. 1, the apparatus includes: the system comprises a light emitter 201, a micro-electro-mechanical system scanning galvanometer 202, a spectrum collector 203, a spectral feature data processor 204 and a substance spectrum detector 205.
The light emitter 201 is connected with the spectral characteristic data processor 204 and irradiates the emitted incident light on the scanning galvanometer of the micro-electro-mechanical system; the micro-electro-mechanical system scanning galvanometer 202 is connected with the spectral characteristic data processor 204 and deflects incident light to a scanning detection area at a preset initial deflection angle; and adjusting the deflection angle of the scanning galvanometer of the micro electro mechanical system in a preset deflection angle adjusting mode.
The spectrum collector 203 is connected with the spectrum characteristic data processor 204 and is used for collecting hyperspectral data of the scanning detection area under each deflection angle; the spectral characteristic data processor 204 is connected with the illuminator 201, the micro-electro-mechanical system scanning galvanometer 202, the spectrum collector 203 and the substance spectrum detector 205, and establishes a spatial relationship hyperspectral data set corresponding to the scanning detection area and the hyperspectral data based on the scanning detection area, the deflection angle and the hyperspectral data; and according to a preset uncorrelated noise characteristic, removing uncorrelated hyperspectral data in the spatial relationship hyperspectral data set to obtain a correlated characteristic data set.
Extracting each feature data from the related feature data set, generating a scale space according to the feature data, constructing a differential pyramid, and searching extreme points which are not influenced by the scale in the scale space as key points of the feature data set; distributing a direction for each key point by utilizing the gradient distribution of the neighborhood characteristics of the key points to form a multi-dimensional characteristic vector set for describing the local gradient characteristics of the related characteristic data set;
the substance spectrum detector 205 is connected with the spectral feature data processor 204 and matches each feature point in the multi-dimensional feature vector set with the feature point of the corresponding gradient layer of the standard substance in the scale space to form a feature point matching pair; and calculating the similarity of the feature point matching pairs, and determining that the substances in the scanning detection area are standard substances when the similarity reaches or exceeds a preset similarity threshold.
In this embodiment, a design of the MEMS scanning galvanometer is adopted to achieve multi-region data acquisition, and the micro electro Mechanical Systems (MEMS for short) realizes deflection of a light path by changing a deflection angle of the MEMS galvanometer, so that light irradiates different regions of a detection object. The probe type hyperspectral sensor obtains high-resolution spectral data of each area and forms a hyperspectral data set according to the spatial relationship. The existing design for realizing multi-region spectral data acquisition is multiple, and the design based on the MEMS scanning galvanometer has the advantages that: the MEMS galvanometer is combined with data acquisition of the probe type hyperspectral sensor, and compared with a planar array type hyperspectral sensor, the MEMS galvanometer has the advantages of higher spectral resolution, smaller structure, larger scanning field range and higher utilization rate. In addition, the deflection parameters of the MEMS galvanometer are modified, the distance between the acquisition points is flexibly changed, and dense or sparse data acquisition and irregular RO I data acquisition in the same scanning area can be realized, so that different requirements of different detection objects are met, data acquisition is performed in a targeted manner, and the identification precision and efficiency can be effectively improved.
In addition, the information quantity is enriched by multi-region data acquisition, but the information is not all effective data, and the accuracy and the efficiency of identification can be ensured only by accurate filtering and extraction. As shown in fig. 3 below, the spectral curves collected in the respective regions are subjected to feature extraction to eliminate noise features irrelevant to subsequent analysis, feature forming feature sets capable of better describing the detection object are retained, the same type of features in different regions of the detection object are regarded as a matrix, and the composition of different features can be regarded as a three-dimensional feature data set.
Although the detection object is well described to a certain extent, the characteristics still have partial redundancy, so that the accuracy of characteristic matching is reduced, the storage space is wasted, and the speed of identification and matching is slow. The feature data of each region of the detection object does not exist independently, the relationship between the regions provides richer information about the detection object, and the local characteristics represented by the multi-region feature data are not changed along with the changes of the scale, the rotation and the visual angle of the detection object. Therefore, each layer in the feature set is further subjected to feature extraction by using local characteristics, and a feature descriptor with robustness and uniqueness is constructed to improve matching performance.
As shown in fig. 4, each layer of the feature set is subjected to fuzzy and downsampling processing to generate a scale space and construct a differential pyramid, in order to ensure that a subsequently generated multi-dimensional feature vector has scale, rotation and view invariance, extreme points in the space which are not affected by the scale are searched and determined as key points of the feature set, a direction is allocated to each key point by using the gradient distribution characteristics of the neighborhood features of the key points, and finally, multi-dimensional vectors describing local gradient features of the feature set are generated, and the vectors of each layer form a multi-dimensional feature vector set to more specifically and accurately represent a detection object.
The identification of the detection object is substantially a process of realizing feature matching, the identification process is as shown in fig. 5, the matching degree between features is measured by calculating the similarity degree between the detection object and the feature points in the corresponding layer of the candidate object, the features with high matching degree form a matching point pair, mismatching may exist in the matching point pair, the mismatching is filtered by combining the position, angle, scale and other information of the features through a space verification method, and the similarity calculation is performed on the remaining matching points.
Optionally, the currently commonly used similarity measure method is euclidean distance and residueChord similarity, histogram intersection distance, etc. Among them, Euclidean distance is the most common similarity measurement method due to its simple operation, assuming XiIs the characteristic point set of the ith object in the candidate objects, Y is the characteristic point set of the target object, xij，yjAre respectively an object XiAnd the jth feature point on the Y match. m is the scale of the candidate object, and n is the number of the final matching point pairs of the candidate object and the detection object. The concrete formula is as follows:
and comparing the similarity and the threshold of the target object and each candidate object, and finally realizing the matching and the identification of the multi-dimensional spectral features.
In some alternative embodiments, as shown in fig. 6, which is a schematic structural diagram of another dynamic scanning detection apparatus 600 based on spectral characteristics in an embodiment, different from fig. 2, the mems scanning galvanometer 202 includes: a yaw angle adjustment mode acquisition means 221 and a yaw angle adjustment means 222. The deflection angle adjustment mode obtaining unit 221 is connected to the deflection angle adjustment unit 222, and is configured to receive shape information of an object to be detected in the scan detection area, and compare preset object shape information with a corresponding relationship between a deflection angle adjustment mode and the shape information of the object to be detected to obtain a deflection angle adjustment mode for detecting the object to be detected.
The deflection angle adjusting unit 222 is connected to the deflection angle adjusting mode obtaining unit 221 and the spectral feature data processor 204, and adjusts the deflection angle of the mems scanning galvanometer in a deflection angle adjusting mode for detecting an object to be detected.
In some alternative embodiments, as shown in fig. 7, which is a schematic structural diagram of another dynamic scanning detection apparatus 700 based on spectral characteristics in the embodiments, different from fig. 2, the spectral characteristic data processor 204 includes: a spatial relationship hyperspectral dataset processing unit 241, a related feature dataset cleaning unit 242, and a multidimensional feature vector set generation unit 243. The spatial relationship hyperspectral data set processing unit 241 is connected with the light emitter 201, the micro-electromechanical system scanning galvanometer 202, the spectrum collector 203 and the related characteristic data set cleaning unit 242, and establishes a spatial relationship hyperspectral data set corresponding to the scanning detection area and the hyperspectral data based on the scanning detection area, the deflection angle and the hyperspectral data.
The related feature data set cleaning unit 242 is connected to the spatial relationship hyperspectral data set processing unit 241 and the multidimensional feature vector set generating unit 243, and removes the uncorrelated hyperspectral data in the spatial relationship hyperspectral data set according to a preset uncorrelated noise feature.
And a multi-dimensional feature vector set generating unit 243 connected to the relevant feature data set cleaning unit 242 and the material spectrum detector 205, wherein the same type of features in different regions are used as a matrix, and different layers of different features form a three-dimensional feature data set, and the feature layers in the three-dimensional feature data set are subjected to fuzzy and downsampling processing by using local characteristics to obtain the relevant feature data set. Extracting each feature data from the related feature data set, generating a scale space according to the feature data, constructing a differential pyramid, and searching extreme points which are not influenced by the scale in the scale space as key points of the feature data set; and distributing a direction for each key point by utilizing the gradient distribution of the neighborhood characteristics of the key points to form a multi-dimensional characteristic vector set for describing the local gradient characteristics of the related characteristic data set.
In some alternative embodiments, as shown in fig. 8, which is a schematic structural diagram of another dynamic scanning detection apparatus 800 based on spectral characteristics in an embodiment, different from fig. 2, the substance spectrum detector 205 includes: a spectral data characteristic point matching pair unit 251, a spectral data characteristic point matching pair cleaning unit 252, and a substance determination unit 253. The spectral data feature point matching pair unit 251 is connected with the spectral data feature point matching pair cleaning unit 252 and the spectral feature data processor 204, and matches each feature point in the multi-dimensional feature vector set with the feature point of the corresponding gradient layer of the standard substance in the scale space to form a feature point matching pair.
The spectral data feature point matching pair cleaning unit 252 is connected to the spectral data feature point matching pair unit 251 and the substance determining unit 253, compares the positions, angles and scales of the feature points with the corresponding feature points of the standard substance, and filters out feature mismatching feature points with errors exceeding a preset error threshold value to obtain accurate feature point matching pairs.
And the material judging unit 253 is connected with the spectral data characteristic point matching pair cleaning unit 252, calculates the similarity of the accurate characteristic point matching pairs, and determines that the material in the scanning detection area is the standard material when the similarity reaches or exceeds a preset similarity threshold.
Optionally, the substance determination unit 253 is configured to calculate a similarity of the feature point matching pairs by using a euclidean distance, a cosine similarity, or a histogram intersection distance; and when the similarity reaches or exceeds a preset similarity threshold, determining that the substance in the scanning detection area is the standard substance.
In some optional embodiments, as shown in fig. 9, a flow chart of a dynamic scan detection method based on spectral characteristics in an embodiment is shown, where the method may be implemented by the above dynamic scan detection apparatus based on spectral characteristics, and the method includes the following steps:
And 903, collecting hyperspectral data of the scanning detection area at each deflection angle through a spectrum collector.
904, establishing a spatial relationship hyperspectral data set corresponding to the scanning detection area and the hyperspectral data based on the scanning detection area, the deflection angle and the hyperspectral data; and according to a preset uncorrelated noise characteristic, removing uncorrelated hyperspectral data in the spatial relationship hyperspectral data set to obtain a correlated characteristic data set.
In some optional embodiments, as shown in fig. 10, which is a schematic flow chart of another dynamic scan detection method based on spectral features in the embodiments, different from fig. 9, in step 902, adjusting a deflection angle of a scanning galvanometer of a mems in a preset deflection angle adjustment manner is that:
The deflection angle adjusting mode of some common object shapes, such as a cube, is preset, and the deflection angle adjusting mode of scanning from five faces of the cube is preset, so that the specific shape characteristics can be favorably adjusted in a targeted scanning deflection angle, and the efficiency of detecting substances is improved.
In some optional embodiments, as shown in fig. 11, which is a schematic flow chart of another dynamic scanning detection method based on spectral features in an embodiment, different from fig. 9, in step 904, a spatial relationship hyperspectral data set corresponding to a scanning detection area and hyperspectral data is established based on the scanning detection area, a deflection angle, and the hyperspectral data; according to a preset uncorrelated noise characteristic, cancelling uncorrelated hyperspectral data in a spatial relationship hyperspectral data set to obtain a correlated characteristic data set, wherein the correlated characteristic data set comprises the following steps:
In some optional embodiments, as shown in fig. 12, which is a schematic flowchart of another dynamic scanning detection method based on spectral features in the embodiments, unlike in fig. 9, in step 906, each feature point in the multi-dimensional feature vector set is matched with a feature point of a gradient layer corresponding to the standard substance in the scale space to form a feature point matching pair; calculating the similarity of the feature point matching pairs, and when the similarity reaches or exceeds a preset similarity threshold, determining that the substances in the scanning detection area are standard substances, wherein the similarity is as follows:
And 1202, comparing the positions, angles and scales of the feature points with corresponding feature points of the standard substance, and filtering feature mismatching feature points with errors exceeding a preset error threshold value to obtain accurate feature point matching pairs. And calculating the similarity of the feature point matching pairs, and determining that the substances in the scanning detection area are standard substances when the similarity reaches or exceeds a preset similarity threshold.
Optionally, calculating the similarity of the feature point matching pairs by using Euclidean distance, cosine similarity or histogram intersection distance; and when the similarity reaches or exceeds a preset similarity threshold, determining that the substance in the scanning detection area is the standard substance.
In some alternative embodiments, as shown in fig. 13, the dynamic scan detection method based on spectral features in this embodiment is applied to an artificial neural network model. Except for the original algorithm model, algorithms such as a neural network and a cluster, which can effectively analyze multi-dimensional spectral information, are all suitable for realizing more accurate detection and identification of the object by combining the design of the MEMS galvanometer in the embodiment. If an end-to-end artificial intelligence algorithm is adopted, based on the fact that data of all areas collected by an MEMS galvanometer are used as input, a recognition result is used as output, a large amount of data are collected for a plurality of known objects for training, and an effective multi-dimensional spectral information detection system can be constructed.
Alternatively, the method can be applied to the improvement of the conventional spectrum analyzer. By adopting the application of the MEMS galvanometer in the embodiment, a plurality of groups of linked hyperspectral data are obtained by data acquisition each time, and the model is trained by data containing more information, so that the capability of distinguishing confusable objects is enhanced, and the model is more complete; the algorithm aspect adopts a framework combining global search and local search, and adopts an intelligent analysis algorithm in the global search strategy, so that the local search strategy is added to carry out more precise search on possible solutions in the neighborhood of the global search result, and the identification accuracy and the estimation accuracy are improved. Similarly, the method can also be applied to industrial detection equipment, daily detection equipment and laboratory detection equipment.
Optionally, the method is also suitable for building various sensors to realize effective fusion of various information. For example, the biological identification accuracy is improved by the integration with a fingerprint identification system. In traditional fingerprint identification, the people who have both hands to operate by hands for a long time often because the fingerprint is damaged or in the wet environment futilely, be stained with the foreign matter and discern inefficacy, and the fingerprint imitation is easy in addition, and it is still not enough to pass through fingerprint identification identity security performance simply. If the hyperspectral data of the finger of the person is acquired while the fingerprint information is acquired, the fingerprint information and the hyperspectral information are combined to be used as the identity information code, the problems of identity identification failure and identity counterfeiting under special conditions are solved, the identification accuracy is improved, and the application occasions are enriched. Similarly, the method can also be applied to detection systems of pupils, irises, face recognition and the like.
The method can also be combined with a laser scanner to realize data acquisition of multi-dimensional point cloud, and the high spectral data of each point is acquired to form multi-dimensional point cloud data while the three-dimensional point cloud data of the surface of the object is acquired through the scanner. By the combination, the scanning instrument can obtain the three-dimensional size data of the sample and can also obtain the component content and the structural information of the sample. The data are directly interfaced with software, processed and recorded on the software, so that the manufacturing period of the product can be shortened, and the method can be used for intelligent detection of the product and the like.
The scanning type hyperspectral data acquisition of the detection object is realized by combining with the MEMS galvanometer, and efficient and accurate identification is realized by processing and analyzing a plurality of groups of hyperspectral data. The application of collecting more effective hyperspectral information of an object to carry out accurate estimation and identification is needed in multiple fields. Such as: article identification in daily life, identification in biometric systems, product detection in industry, etc. The method can solve the problems of low object identification accuracy such as complex structure and partially shielded surface of spectral feature identification.
In the dynamic scanning detection method or device based on the spectral characteristics in this embodiment, the rotation angle of the MEMS galvanometer is changed to change the optical path, so as to implement a detection system for scanning hyperspectral data acquisition on a detection object, and data acquisition in areas with different densities and different shapes can be implemented by modifying the deflection parameters of the MEMS galvanometer. The hyperspectral data of each area is not directly formed into a hyperspectral data set of the detection object, and the hyperspectral data set is formed by firstly extracting the features and then forming the hyperspectral feature data set according to the characteristics of the analysis object. Invalid data such as noise and the like are filtered at the beginning, and the subsequent calculation and analysis efficiency is improved. The identification conclusion is not obtained by analyzing single hyperspectral data, but a multi-dimensional feature vector with robustness and uniqueness is constructed by analyzing the local characteristics of a plurality of groups of hyperspectral data obtained by scanning and extracting features for a plurality of times so as to improve the subsequent matching efficiency and performance. And the information of the algorithm model and various sensors is effectively fused, so that the structure or the component of the detection object can be effectively classified and identified, and the relationship among various data of the object can be utilized for further research and application.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all such alterations and modifications as fall within the scope of the invention. It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Claims (10)
1. A dynamic scan test device based on spectral characteristics, comprising: a light emitter, a micro-electro-mechanical system scanning galvanometer, a spectrum collector, a spectrum characteristic data processor and a substance spectrum detector, wherein,
the illuminator is connected with the spectral characteristic data processor and emits incident light to irradiate the micro-electro-mechanical system scanning galvanometer;
the micro electro mechanical system scanning galvanometer is connected with the spectral characteristic data processor and adjusts the deflection angle of the micro electro mechanical system scanning galvanometer in a preset deflection angle adjusting mode;
the spectrum collector is connected with the spectrum characteristic data processor and is used for collecting hyperspectral data of the scanning detection area under each deflection angle;
the spectral characteristic data processor is connected with the illuminator, the micro-electro-mechanical system scanning galvanometer, the spectrum collector and the substance spectrum detector, establishes a spatial relationship hyperspectral data set corresponding to the scanning detection area and the hyperspectral data, and extracts characteristic data to form a multidimensional characteristic vector set;
the substance spectrum detector is connected with the spectrum characteristic data processor and matches each characteristic point in the multi-dimensional characteristic vector set with the characteristic point of the gradient layer corresponding to the standard substance to form a characteristic point matching pair; and calculating the similarity of the feature point matching pairs, and determining the material characteristics in the scanning detection area according to the similarity.
2. The dynamic scanning detection device based on spectral characteristics according to claim 1,
the micro electro mechanical system scanning galvanometer is used for deflecting the incident light to a scanning detection area at a preset initial deflection angle; adjusting the deflection angle of the scanning galvanometer of the micro electro mechanical system in a preset deflection angle adjusting mode;
the spectral characteristic data processor is used for establishing a spatial relationship hyperspectral data set corresponding to the scanning detection area and the hyperspectral data based on the scanning detection area, the deflection angle and the hyperspectral data; according to a preset uncorrelated noise characteristic, cancelling uncorrelated hyperspectral data in the spatial relationship hyperspectral data set to obtain a correlated characteristic data set;
extracting each feature data from the relevant feature data set, generating a scale space according to the feature data, constructing a differential pyramid, and searching extreme points which are not influenced by the scale in the scale space as feature data set key points; distributing a direction for each key point by utilizing the gradient distribution of the neighborhood characteristics of the key points to form a multi-dimensional characteristic vector set for describing the local gradient characteristics of the related characteristic data set;
the substance spectrum detector is used for matching each characteristic point in the multi-dimensional characteristic vector set with the characteristic point of the corresponding gradient layer of the standard substance in a scale space to form a characteristic point matching pair; and calculating the similarity of the feature point matching pairs, and determining that the substance in the scanning detection area is the standard substance when the similarity reaches or exceeds a preset similarity threshold value.
3. The dynamic scanning inspection device based on spectral characteristics of claim 2, wherein said mems scanning galvanometer comprises: a deflection angle adjustment mode acquisition unit and a deflection angle adjustment unit; wherein,
the deflection angle adjustment mode acquisition unit is connected with the deflection angle adjustment unit, receives shape information of an object to be detected in the scanning detection area, and compares preset object shape information with a corresponding relation of a deflection angle adjustment mode according to the shape information of the object to be detected to obtain a deflection angle adjustment mode for detecting the object to be detected;
the deflection angle adjusting unit is connected with the deflection angle adjusting mode acquiring unit and the spectral characteristic data processor to detect the deflection angle adjusting mode of the object to be detected and adjust the deflection angle of the scanning galvanometer of the micro-electro-mechanical system.
4. The dynamic scan detection device based on spectral characteristics according to claim 2, wherein the spectral characteristics data processor comprises: the system comprises a spatial relation hyperspectral data set processing unit, a related feature data set cleaning unit and a multi-dimensional feature vector set generating unit; wherein,
the spatial relationship hyperspectral data set processing unit is connected with the illuminator, the micro-electro-mechanical system scanning galvanometer, the spectrum collector and the related characteristic data set cleaning unit, and establishes a spatial relationship hyperspectral data set corresponding to the scanning detection area and the hyperspectral data based on the scanning detection area, the deflection angle and the hyperspectral data;
the relevant feature data set cleaning unit is connected with the spatial relationship hyperspectral data set processing unit and the multi-dimensional feature vector set generating unit and eliminates irrelevant hyperspectral data in the spatial relationship hyperspectral data set according to a preset irrelevant noise feature;
the multi-dimensional feature vector set generating unit is connected with the relevant feature data set cleaning unit and the substance spectrum detector, the same type of features in different areas are used as a matrix, all layers of different features form a three-dimensional feature data set, and the feature layers in the three-dimensional feature data set are subjected to fuzzy and downsampling processing by utilizing local characteristics to obtain a relevant feature data set;
extracting each feature data from the relevant feature data set, generating a scale space according to the feature data, constructing a differential pyramid, and searching extreme points which are not influenced by the scale in the scale space as feature data set key points; and distributing a direction for each key point by utilizing the gradient distribution of the neighborhood characteristics of the key points to form a multi-dimensional characteristic vector set for describing the local gradient characteristics of the related characteristic data set.
5. The dynamic scanning inspection device based on spectral characteristics of any one of claims 1 to 4, wherein the substance spectral detector comprises: the system comprises a spectral data characteristic point matching pair unit, a spectral data characteristic point matching pair cleaning unit and a substance judgment unit; wherein,
the spectrum data feature point matching pair unit is connected with the spectrum data feature point matching pair cleaning unit and the spectrum feature data processor, and matches each feature point in the multi-dimensional feature vector set with the feature point of the corresponding gradient layer of the standard substance in the scale space to form a feature point matching pair;
the spectral data feature point matching pair cleaning unit is connected with the spectral data feature point matching pair unit and the substance judgment unit, compares the positions, angles and scales of the feature points with the corresponding feature points of the standard substance, and filters out mismatching feature points with errors exceeding a preset error threshold value to obtain accurate feature point matching pairs;
and the substance judgment unit is connected with the spectral data characteristic point matching pair cleaning unit, calculates the similarity of the accurate characteristic point matching pair, and determines that the substance in the scanning detection area is the standard substance when the similarity reaches or exceeds a preset similarity threshold value.
6. A dynamic scanning detection method based on spectral characteristics is characterized by comprising the following steps:
irradiating the incident light emitted by the light emitter on the scanning galvanometer of the micro-electro-mechanical system;
adjusting the deflection angle of the scanning galvanometer of the micro electro mechanical system in a preset deflection angle adjusting mode;
collecting hyperspectral data of the scanning detection area under each deflection angle;
establishing a spatial relationship hyperspectral data set corresponding to the scanning detection area and the hyperspectral data, and extracting characteristic data to form a multi-dimensional characteristic vector set;
matching each characteristic point in the multi-dimensional characteristic vector set with the characteristic point of the corresponding gradient layer of the standard substance in a scale space to form a characteristic point matching pair; and calculating the similarity of the feature point matching pairs, and determining the material characteristics in the scanning detection area according to the similarity.
7. The dynamic scanning detection method based on spectral features of claim 6, wherein the incident light emitted from the light emitter is irradiated on the mems scanning galvanometer for irradiating the incident light emitted from the light emitter on the mems scanning galvanometer, and the mems scanning galvanometer deflects the incident light to the scanning detection area by a preset initial deflection angle;
the method comprises the steps of establishing a spatial relationship hyperspectral data set corresponding to a scanning detection area and hyperspectral data, extracting characteristic data to form a multi-dimensional characteristic vector set, and establishing a spatial relationship hyperspectral data set corresponding to the scanning detection area and the hyperspectral data based on the scanning detection area, a deflection angle and the hyperspectral data; according to a preset uncorrelated noise characteristic, cancelling uncorrelated hyperspectral data in the spatial relationship hyperspectral data set to obtain a correlated characteristic data set;
extracting each feature data from the relevant feature data set, generating a scale space according to the feature data, constructing a differential pyramid, and searching extreme points which are not influenced by the scale in the scale space as feature data set key points; distributing a direction for each key point by utilizing the gradient distribution of the neighborhood characteristics of the key points to form a multi-dimensional characteristic vector set for describing the local gradient characteristics of the related characteristic data set;
matching each characteristic point in the multi-dimensional characteristic vector set with the characteristic point of the corresponding gradient layer of the standard substance in a scale space to form a characteristic point matching pair; calculating the similarity of the feature point matching pairs, determining the material characteristics in the scanning detection area according to the similarity, and matching each feature point in the multi-dimensional feature vector set with the feature point of the corresponding gradient layer of the standard material in a scale space to form the feature point matching pairs; and calculating the similarity of the feature point matching pairs, and determining that the substance in the scanning detection area is the standard substance when the similarity reaches or exceeds a preset similarity threshold value.
8. The dynamic scan detection method based on spectral features of claim 7, wherein the adjusting the deflection angle of the mems scan galvanometer in a preset deflection angle adjustment manner comprises:
receiving shape information of an object to be detected in the scanning detection area, and comparing preset object shape information with a corresponding relation of a deflection angle adjustment mode according to the shape information of the object to be detected to obtain a deflection angle adjustment mode for detecting the object to be detected;
and adjusting the deflection angle of the scanning galvanometer of the micro-electro-mechanical system in a manner of detecting the deflection angle of the object to be detected.
9. The dynamic scanning detection method based on spectral features of claim 7, wherein the uncorrelated hyperspectral data in the spatial relationship hyperspectral data sets are eliminated according to preset uncorrelated noise features to obtain correlated feature data sets, and the method comprises the following steps:
according to a preset uncorrelated noise characteristic, cancelling uncorrelated hyperspectral data in the spatial relationship hyperspectral data set;
the same type of features in different areas are used as a matrix, all layers with different features form a three-dimensional feature data set, and the feature layers in the three-dimensional feature data set are subjected to fuzzy and downsampling processing by utilizing local characteristics to obtain a related feature data set.
10. The method according to any one of claims 6 to 9, wherein each feature point in the multi-dimensional feature vector set is matched with a feature point of a gradient layer corresponding to a standard substance in a scale space to form a feature point matching pair, which is:
matching each characteristic point in the multi-dimensional characteristic vector set with the characteristic point of the corresponding gradient layer of the standard substance in a scale space to form a characteristic point matching pair;
and comparing the positions, angles and scales of the feature points with the corresponding feature points of the standard substance, filtering out mismatching feature points with errors exceeding a preset error threshold value, and obtaining accurate feature point matching pairs.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910433146.8A CN111412987B (en) | 2019-05-23 | 2019-05-23 | Dynamic scanning detection device and method based on spectral characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910433146.8A CN111412987B (en) | 2019-05-23 | 2019-05-23 | Dynamic scanning detection device and method based on spectral characteristics |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111412987A true CN111412987A (en) | 2020-07-14 |
| CN111412987B CN111412987B (en) | 2023-06-06 |
Family
ID=71492681
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910433146.8A Active CN111412987B (en) | 2019-05-23 | 2019-05-23 | Dynamic scanning detection device and method based on spectral characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111412987B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105424190A (en) * | 2015-09-30 | 2016-03-23 | 广州超音速自动化科技股份有限公司 | Grayscale detection method of product appearance |
| CN205879774U (en) * | 2016-04-12 | 2017-01-11 | 北京世纪桑尼科技有限公司 | High -speed scanning that can be used to spectral analysis is total to focusing imaging system |
| CN107221010A (en) * | 2017-07-12 | 2017-09-29 | 中国科学院上海技术物理研究所 | Airborne hyperspectral geometric image correction method and device based on three area array cameras |
| CN107941334A (en) * | 2017-10-13 | 2018-04-20 | 中国科学院上海技术物理研究所 | A kind of standard reflection formula super continuous spectrums Image-forming instrument |
| CN108268902A (en) * | 2018-01-26 | 2018-07-10 | 厦门大学 | High spectrum image transformation and substance detection identifying system and method based on recurrence plot |
-
2019
- 2019-05-23 CN CN201910433146.8A patent/CN111412987B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105424190A (en) * | 2015-09-30 | 2016-03-23 | 广州超音速自动化科技股份有限公司 | Grayscale detection method of product appearance |
| CN205879774U (en) * | 2016-04-12 | 2017-01-11 | 北京世纪桑尼科技有限公司 | High -speed scanning that can be used to spectral analysis is total to focusing imaging system |
| CN107221010A (en) * | 2017-07-12 | 2017-09-29 | 中国科学院上海技术物理研究所 | Airborne hyperspectral geometric image correction method and device based on three area array cameras |
| CN107941334A (en) * | 2017-10-13 | 2018-04-20 | 中国科学院上海技术物理研究所 | A kind of standard reflection formula super continuous spectrums Image-forming instrument |
| CN108268902A (en) * | 2018-01-26 | 2018-07-10 | 厦门大学 | High spectrum image transformation and substance detection identifying system and method based on recurrence plot |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111412987B (en) | 2023-06-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11861889B2 (en) | Analysis device | |
| JP7170037B2 (en) | Multi-Step Image Alignment Method for Large Offset Die Inspection | |
| Hu et al. | Accurate automatic quantification of taxa-specific plankton abundance using dual classification with correction | |
| CN101975818B (en) | Detection system and method of characteristic substance | |
| CN109858477A (en) | The Raman spectrum analysis method of object is identified in complex environment with depth forest | |
| CN105842173A (en) | Method for identifying hyperspectral material | |
| CN102279146A (en) | Blood cell five classification method based on laser sheath flow technology | |
| WO2016080442A1 (en) | Quality evaluation method and quality evaluation device | |
| CN106340007A (en) | Image processing-based automobile body paint film defect detection and identification method | |
| Kumar et al. | Deep remote sensing methods for methane detection in overhead hyperspectral imagery | |
| CN109115719A (en) | A kind of Citrus Huanglongbing pathogen Band fusion rapid detection method based on high light spectrum image-forming technology | |
| Devassy et al. | Ink classification using convolutional neural network | |
| CN105223140A (en) | The method for quickly identifying of homology material | |
| Van den Broek et al. | Application of a spectroscopic infrared focal plane array sensor for on-line identification of plastic waste | |
| CN111412987B (en) | Dynamic scanning detection device and method based on spectral characteristics | |
| CN114662594B (en) | Target feature recognition analysis system | |
| Ghorpade et al. | 3D shape representation with spatial probabilistic distribution of intrinsic shape keypoints | |
| Boldeanu et al. | Automatic pollen classification using convolutional neural networks | |
| Gampa | A data-driven approach for detecting stress in plants using hyperspectral imagery | |
| CN113567417A (en) | Method for identifying peanut oil production place based on Raman spectrum fingerprint analysis technology | |
| CN112414973A (en) | Speckle fingerprint-based target material identification method and system | |
| CN111222543A (en) | Substance identification method and apparatus, and computer-readable storage medium | |
| Shahbeigi et al. | A Novel Score-based LiDAR Point Cloud Degradation Analysis Method | |
| Concea et al. | Texture Analysis for Images with Forested Areas | |
| US11841314B2 (en) | Method and system for characterizing particles using an angular detection in a flow cytometer |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |