WO2022168594A1 - スペクトルデータキューブを改変する方法および装置 - Google Patents
スペクトルデータキューブを改変する方法および装置 Download PDFInfo
- Publication number
- WO2022168594A1 WO2022168594A1 PCT/JP2022/001493 JP2022001493W WO2022168594A1 WO 2022168594 A1 WO2022168594 A1 WO 2022168594A1 JP 2022001493 W JP2022001493 W JP 2022001493W WO 2022168594 A1 WO2022168594 A1 WO 2022168594A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- matrix
- data
- decoding
- spectral
- compressed image
- Prior art date
Links
- 230000003595 spectral effect Effects 0.000 title claims abstract description 136
- 238000000034 method Methods 0.000 title claims abstract description 93
- 239000011159 matrix material Substances 0.000 claims abstract description 230
- 238000012545 processing Methods 0.000 claims description 170
- 238000003860 storage Methods 0.000 claims description 61
- 238000003384 imaging method Methods 0.000 claims description 59
- 230000008569 process Effects 0.000 claims description 43
- 230000003287 optical effect Effects 0.000 claims description 25
- 238000001228 spectrum Methods 0.000 claims description 19
- 238000009826 distribution Methods 0.000 claims description 16
- 238000000411 transmission spectrum Methods 0.000 claims description 14
- 238000007689 inspection Methods 0.000 description 61
- 238000002834 transmittance Methods 0.000 description 53
- 238000010586 diagram Methods 0.000 description 48
- 238000005516 engineering process Methods 0.000 description 11
- 238000000701 chemical imaging Methods 0.000 description 10
- 230000006870 function Effects 0.000 description 9
- 238000003780 insertion Methods 0.000 description 9
- 230000037431 insertion Effects 0.000 description 9
- 230000006835 compression Effects 0.000 description 7
- 238000007906 compression Methods 0.000 description 7
- 238000012986 modification Methods 0.000 description 7
- 230000004048 modification Effects 0.000 description 7
- 238000012935 Averaging Methods 0.000 description 5
- 235000013305 food Nutrition 0.000 description 5
- 239000000463 material Substances 0.000 description 5
- 238000004891 communication Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 230000010354 integration Effects 0.000 description 4
- 229910052751 metal Inorganic materials 0.000 description 4
- 239000002184 metal Substances 0.000 description 4
- 239000004065 semiconductor Substances 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 238000004590 computer program Methods 0.000 description 3
- 239000000470 constituent Substances 0.000 description 3
- 230000002194 synthesizing effect Effects 0.000 description 3
- 238000005259 measurement Methods 0.000 description 2
- 239000011368 organic material Substances 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 230000000704 physical effect Effects 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 241000238631 Hexapoda Species 0.000 description 1
- 240000007594 Oryza sativa Species 0.000 description 1
- 235000007164 Oryza sativa Nutrition 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000005856 abnormality Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000001574 biopsy Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000009509 drug development Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 229910052500 inorganic mineral Inorganic materials 0.000 description 1
- 238000012966 insertion method Methods 0.000 description 1
- 238000010030 laminating Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 229910044991 metal oxide Inorganic materials 0.000 description 1
- 150000004706 metal oxides Chemical class 0.000 description 1
- 150000002739 metals Chemical class 0.000 description 1
- 239000011707 mineral Substances 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 239000000049 pigment Substances 0.000 description 1
- 239000011295 pitch Substances 0.000 description 1
- 239000004033 plastic Substances 0.000 description 1
- 229920003023 plastic Polymers 0.000 description 1
- 238000000513 principal component analysis Methods 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 235000009566 rice Nutrition 0.000 description 1
- 238000004611 spectroscopical analysis Methods 0.000 description 1
- 230000006641 stabilisation Effects 0.000 description 1
- 238000011105 stabilization Methods 0.000 description 1
- 230000000087 stabilizing effect Effects 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/194—Terrestrial scenes using hyperspectral data, i.e. more or other wavelengths than RGB
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T1/00—General purpose image data processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T1/00—General purpose image data processing
- G06T1/0021—Image watermarking
- G06T1/0028—Adaptive watermarking, e.g. Human Visual System [HVS]-based watermarking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/10—Image acquisition
- G06V10/12—Details of acquisition arrangements; Constructional details thereof
- G06V10/14—Optical characteristics of the device performing the acquisition or on the illumination arrangements
- G06V10/143—Sensing or illuminating at different wavelengths
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
- H04N1/32—Circuits or arrangements for control or supervision between transmitter and receiver or between image input and image output device, e.g. between a still-image camera and its memory or between a still-image camera and a printer device
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/70—Circuitry for compensating brightness variation in the scene
- H04N23/75—Circuitry for compensating brightness variation in the scene by influencing optical camera components
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/80—Camera processing pipelines; Components thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2201/00—General purpose image data processing
- G06T2201/005—Image watermarking
- G06T2201/0051—Embedding of the watermark in the spatial domain
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2201/00—General purpose image data processing
- G06T2201/005—Image watermarking
- G06T2201/0053—Embedding of the watermark in the coding stream, possibly without decoding; Embedding of the watermark in the compressed domain
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
- G06T2207/10036—Multispectral image; Hyperspectral image
Definitions
- the present disclosure relates to methods and apparatus for modifying spectral data cubes.
- Spectral information includes information that characterizes the material, state, or type of an object. Therefore, from the viewpoint of security protection or data concealment, it may be required to encode or encrypt the spectrum information.
- Patent Document 1 discloses an example of a hyperspectral imaging device using compressed sensing.
- the imaging device includes an encoding element that is an array of a plurality of optical filters arranged two-dimensionally, an image sensor that detects light transmitted through the encoding element, and a signal processing circuit.
- the encoding element is arranged on the optical path connecting the subject and the image sensor.
- the transmission spectra of the multiple filters in the encoding element differ from filter to filter.
- An image sensor acquires one two-dimensional image by simultaneously detecting light in which components of a plurality of wavelength bands that have passed through a filter are superimposed for each pixel.
- Patent Literature 1 discloses a technique of encoding or encrypting spectral information of an object using an encoding element when photographing.
- Patent Document 2 discloses an example of a technique for encoding and compressing data such as multispectral images or hyperspectral images after shooting.
- Patent Document 3 discloses an example of technology for embedding digital watermark data in unit color data in multicolor image data.
- the present disclosure provides novel methods for making modifications such as inserting identifiers into spectral data cubes such as hyperspectral or multispectral images.
- a method is a computer-implemented method.
- the method includes an encoding matrix used to encode a spectral datacube containing image information in multiple wavelength bands to generate a compressed image, or for decoding the spectral datacube from the encoded compressed image. and editing the matrix data such that when the spectral data cube is decoded, image information for at least one wavelength band in the spectral data cube is altered. and outputting the edited matrix data.
- Computer-readable recording media include, for example, non-volatile recording media such as CD-ROMs (Compact Disc-Read Only Memory).
- a device may consist of one or more devices. When the device is composed of two or more devices, the two or more devices may be arranged in one device, or may be divided and arranged in two or more separate devices. As used herein and in the claims, a "device" can mean not only one device, but also a system of multiple devices.
- image information of at least one wavelength band in the spectral datacube is modified during encoding or decoding of the spectral datacube. This makes it possible, for example, to keep the contents of the spectrum data cube confidential, and to improve traceability when the spectrum data cube is illegally leaked. Alternatively, it is also possible to artificially change the characteristics of the imaging device that acquires the compressed image.
- FIG. 1 is a diagram for explaining a method of inserting an identifier into a datacube.
- FIG. 2A is a diagram showing an example of editing matrix data.
- FIG. 2B is a diagram showing another example of editing matrix data.
- FIG. 2C is a diagram showing still another example of editing matrix data.
- FIG. 3A is a diagram schematically showing a configuration example of a hyperspectral imaging system.
- FIG. 3B is a schematic diagram of a first variant of a hyperspectral imaging system.
- FIG. 3C schematically illustrates a second variant of the hyperspectral imaging system.
- FIG. 3D schematically illustrates a third variant of the hyperspectral imaging system.
- FIG. 4A is a diagram schematically showing an example of a filter array;
- FIG. 4A is a diagram schematically showing an example of a filter array
- FIG. 4B is a diagram showing an example of a spatial distribution of light transmittance in each of a plurality of wavelength bands included in the target wavelength band.
- FIG. 4C is a diagram showing an example of spectral transmittance of area A1 included in the filter array shown in FIG. 4A.
- FIG. 4D is a diagram showing an example of spectral transmittance of area A2 included in the filter array shown in FIG. 4A.
- FIG. 5A is a diagram for explaining the relationship between a target wavelength band and a plurality of wavelength bands included therein.
- FIG. 5B is a diagram for explaining the relationship between a target wavelength band and a plurality of wavelength bands included therein.
- FIG. 6A is a diagram for explaining spectral transmittance characteristics in a certain area of the filter array.
- FIG. 6B is a diagram showing the results of averaging the spectral transmittances shown in FIG. 6A for each wavelength band.
- FIG. 7 is a diagram showing a configuration example of an inspection system according to the first embodiment.
- FIG. 8 is a block diagram showing a configuration example related to data processing in the inspection system.
- FIG. 9 is a diagram for explaining identifier insertion processing according to the first embodiment.
- FIG. 10A is a flow chart showing an example of processing of a processing circuit.
- FIG. 10B is a flowchart illustrating an example of processing for editing a decoding table according to identifiers.
- FIG. 11 is a diagram for explaining an identifier insertion process according to the second embodiment.
- FIG. 12 is a block diagram for explaining the identifier insertion process in the second embodiment.
- FIG. 13 is a block diagram showing the configuration of the system according to the third embodiment.
- FIG. 14 is a block diagram showing the configuration of the system according to the fourth embodiment.
- FIG. 15 is a block diagram showing the configuration of an inspection system according to the fifth embodiment.
- FIG. 16 is a block diagram showing a modification of the inspection system according to the fifth embodiment.
- FIG. 17 is a block diagram showing the configuration of the system according to the sixth embodiment.
- FIG. 18 is a block diagram showing the configuration of the system according to the seventh embodiment.
- all or part of a circuit, unit, device, member or section, or all or part of a functional block in a block diagram is, for example, a semiconductor device, a semiconductor integrated circuit (IC), or an LSI (large scale integration). ) may be performed by one or more electronic circuits.
- An LSI or IC may be integrated on one chip, or may be configured by combining a plurality of chips.
- functional blocks other than memory elements may be integrated into one chip.
- LSIs or ICs may be called system LSIs, VLSIs (very large scale integration), or ULSIs (ultra large scale integration) depending on the degree of integration.
- FPGAs Field Programmable Gate Arrays
- RLDs reconfigurable logic devices
- circuits, units, devices, members or parts can be executed by software processing.
- the software is recorded on one or more non-transitory storage media, such as ROMs, optical discs, hard disk drives, etc., and when the software is executed by a processor, the functions specified in the software are performed. is performed by the processor and peripherals.
- a system or apparatus may comprise one or more non-transitory storage media on which software is recorded, a processing unit, and required hardware devices such as interfaces.
- Imaging devices such as hyperspectral cameras or multispectral cameras can capture image information in more wavelength bands than cameras that typically capture RGB images.
- the image data acquired by those cameras is referred to herein as a “spectral datacube” or simply a “datacube.”
- the data representing the hyperspectral image acquired by the hyperspectral camera is referred to as the "hyperspectral datacube” or "HS datacube”.
- a spectral datacube contains image information for each of a plurality of wavelength bands.
- a spectral datacube contains important information that characterizes the material, condition, or type of object. Therefore, strict management of the spectrum data cube is required.
- problems may occur if data cubes generated in processes such as product inspection are illegally taken out or leaked to the outside. Therefore, it may be required to modify the data cube to make it confidential. It may be possible to trace even if the datacube is leaked illegally.
- a countermeasure of embedding an identifier such as a digital watermark in the data is conceivable. For example, it is possible to embed in the data cube an identifier that identifies the camera or inspection device that generated the data cube, or an identifier that identifies when the data cube was generated.
- the step of generating the data cube and the step of modifying the data cube, such as inserting an identifier into the data cube are separated, so there is a possibility that the unaltered data cube will be leaked.
- rice field For example, in a product inspection process using a hyperspectral camera, there is a possibility that an operator or a malicious third party illegally removes or leaks the HS data cube before it has been modified.
- a method is a computer-implemented signal processing method.
- the method comprises (a) an encoding matrix used to encode a spectral datacube containing image information in multiple wavelength bands to generate a compressed image, and/or the spectral data from the encoded compressed image; obtaining matrix data indicative of a decoding matrix used to decode a cube; and (b) image information of at least one wavelength band in said spectral data cube is altered when said spectral data cube is decoded. and (c) outputting the edited matrix data.
- the "compressed image” means image data in which image information of multiple wavelength bands is compressed as one two-dimensional image.
- a compressed image may be generated by encoding a spectral datacube generated by an imaging device, such as a hyperspectral camera, based on an encoding matrix.
- a compressed image can also be generated by an imaging device that utilizes compressed sensing, such as that disclosed in US Pat.
- Spectral data cube is data with two-dimensional coordinates (x, y) and three-dimensional values of wavelength ⁇ .
- a spectral datacube contains image information for each of a plurality of wavelength bands.
- the number of wavelength bands in the spectral data cube is any number greater than or equal to four.
- the number of bands can be 10 or more, in some cases 100 or more, in some examples. As the number of bands increases, more spectral information can be obtained, and the physical properties, characteristics, types, etc. of the object can be inspected in more detail.
- the above method can be executed, for example, by a computer in an inspection system that inspects products, or by a server computer that delivers data necessary for inspection to such a computer.
- the matrix data indicating the encoding matrix used to generate the compressed image and/or the decoding matrix used to decode the spectral data cube from the compressed image is the modified content of the spectral data cube. edited accordingly. Specifically, the value of one or more matrix elements in the matrix data representing the encoding matrix or the decoding matrix is rewritten to a value different from the original value according to the contents of modification.
- the values of one or more specific matrix elements in matrix data representing an encoding matrix or a decoding matrix are written to different values from the original values depending on the content of the identifier. be replaced.
- the values of some or all of the matrix elements in the matrix data are changed to values different from the original values, such as superimposing random noise on the spectrum data cube. can be rewritten.
- the compressed image generated based on the encoding matrix is modified, and as a result, the decoded spectral data cube is also modified.
- the decoding matrix is compiled, the compressed image is not altered, but the decoded spectral datacube is altered. In either case, the final decoded spectral data cube is modified according to the compilation of the matrix data.
- the spectrum data cube before modification is not generated, so it is possible to eliminate data security vulnerabilities. For example, it is possible to make it difficult for a third party who does not know the compilation of the matrix data to read the original spectral data cube. By embedding an identifier in the spectral data cube, it becomes possible to determine whether the spectral data cube was legitimately generated or illegally leaked.
- both the encoding matrix and the decoding matrix may be edited appropriately, and as a result, modifications such as giving an identifier to the finally decoded spectral data cube may be performed.
- the matrix data includes first matrix data indicating the encoding matrix and second matrix data indicating the decoding matrix
- Editing the matrix data may include rewriting the matrix data such that the image information of the at least one wavelength band includes an identifier when the spectral data cube is decoded.
- the identifier is embedded in the image data of at least one wavelength band in the finally generated spectral data cube.
- the identifier may be embedded in such a manner that it can be visually recognized by a person, or may be embedded in a manner such as steganography that is not recognizable by a human but is recognizable by a computer.
- Editing the matrix data may include rewriting the matrix data such that noise is imparted to the spectral data cube that interferes with reading image information of the at least one wavelength band.
- noise that hinders reading of image information refers to noise that makes it difficult for a person or computer to recognize the original image information. Such noise may be imparted to the image information of multiple wavelength bands in the spectral datacube. According to such a configuration, since the content of the original spectrum data cube is kept confidential, vulnerability in data security can be eliminated.
- Editing the matrix data may include rewriting the matrix data such that image information for the plurality of wavelength bands in the spectral data cube is altered when the spectral data cube is decoded. .
- image information for the plurality of wavelength bands in the spectral data cube is altered when the spectral data cube is decoded.
- the above signal processing method is not limited to the purpose of eliminating vulnerabilities in data security, and may be used for the purpose of artificially changing the characteristics of an imaging device that acquires image data for generating a spectrum data cube. good.
- editing the matrix data may rewrite the matrix data such that when the spectral datacube is decoded, the image information of the at least one wavelength band changes in gray level.
- by superimposing random noise on the decoding matrix it is possible to make it appear that the noise of the imaging device has increased in a pseudo manner.
- editing the matrix data may rewrite the matrix data such that when the spectral datacube is decoded, the resolution of image information in the at least one wavelength band changes.
- the device that edits the matrix data, the device that generates the compressed image, and the device that decodes the spectral data cube may be separate devices or the same device.
- an imager or inspection device that produces compressed images may have the ability to edit matrix data and the ability to decode spectral datacubes.
- a signal processing device separate from the imaging device or inspection device that generates the compressed image may have the function of editing the matrix data and the function of decoding the spectral data cube.
- the matrix data indicates the decoding matrix.
- the matrix data representing the decoding matrix is edited and output according to the modified content of the spectrum data cube. This allows the modified spectral datacube to be decoded using the edited matrix data.
- Outputting the edited matrix data may include transmitting the matrix data to a device that decodes the spectral data cube from the compressed image based on the matrix data representing the decoded matrix.
- the device may be, for example, an inspection device installed in a factory or the like. In that case, the edited matrix data can be delivered to the inspection device for use.
- Outputting the matrix data may include storing the matrix data indicating the decoding matrix in a storage medium.
- the method may further comprise obtaining the compressed image and using the edited matrix data to decode the spectral data cube from the compressed image.
- the compressed image may be generated by an imaging device equipped with a filter array.
- the filter array may include multiple types of optical filters having different transmission spectra.
- the plurality of types of optical filters may be arranged in a two-dimensional plane.
- the encoding matrix corresponds to a two-dimensional distribution of transmission spectra of the filter array.
- a matrix representing the two-dimensional distribution of the transmission spectrum of the filter array can be used as the encoding matrix.
- the generation of image data by the imager based on the light passed through the filter array corresponds to encoding the spectral data cube with the encoding matrix to generate a compressed image.
- Decoding the spectral datacube may include decoding the spectral datacube from the compressed image by a compressed sensing process based on the decoding matrix. According to such a configuration, by using compressed sensing, a spectrum data cube can be decoded with higher accuracy from a compressed image encoded using a filter array corresponding to the encoding matrix.
- a compressed image may be generated by an imaging device having a filter array as described above, as well as a device having another structure.
- a compressed image may be generated by performing an encoding process based on an encoding matrix on a spectral data cube generated by an imaging device such as a hyperspectral camera.
- an imaging device such as a hyperspectral camera.
- Such a form is employed, for example, when the spectral data cube needs to be compressed to reduce its data size in order to record it on a storage medium or transmit it over a communication circuit. obtain.
- the matrix data indicates the encoding matrix.
- the matrix data representing the encoding matrix is edited and output according to the modified contents of the spectral data cube. This allows the edited matrix data to be used to generate a modified compressed image, resulting in a modified spectral data cube.
- Outputting the matrix data may comprise transmitting the matrix data to a device that encodes the spectral data cube based on the matrix data representing the encoding matrix to generate the compressed image. .
- This allows a device that generates a compressed image to generate a modified compressed image using the edited encoding matrix. Using such a modified compressed image and the decoding matrix, a modified spectral datacube can be generated.
- Outputting the edited matrix data may include storing the matrix data representing the encoded matrix in a storage medium.
- the method may further comprise obtaining the spectral datacube and encoding the spectral datacube based on the edited matrix data to produce the compressed image. This allows to generate a modified compressed image. Using such a modified compressed image and the decoding matrix, a modified spectral datacube can be generated.
- the identifier directs apparatus for generating the compressed image based on the encoding matrix or apparatus for decoding the spectral datacube based on the decoding matrix. It may contain identifying information.
- the identifier may include information specifying when the identifier was assigned. Including such information in the identifier can improve the traceability of the spectral data cube.
- Obtaining the matrix data includes obtaining, from the compressed image, a decoding table for decoding an image for each of a plurality of wavelength bands included in the target wavelength band, and based on the decoding table, the Reduction for decoding as the spectral datacube an image for each of the fewer wavelengthbands than the plurality of wavelengthbands, wherein two or more of the plurality of wavelengthbands are combined as one wavelengthband.
- Editing the matrix data may include editing the reduced decoding table. By generating the reduced decoding table as the matrix data, it is possible to reduce the computational load of the decoding process. Such processing is particularly useful when it is sufficient to acquire image information in a small number of bands, for example with a relatively wide bandwidth.
- Outputting the edited matrix data may include transmitting the edited reduced decoding table to another device. This allows other devices to decode the modified spectral data cube from the compressed image based on the edited reduced decoding table with low computational load.
- An apparatus comprises a storage device and processing circuitry.
- the storage device stores an encoding matrix used to encode a spectral data cube containing image information in multiple wavelength bands to generate a compressed image, and/or stores the spectral data cube from the encoded compressed image.
- the processing circuit edits the matrix data such that image information of at least one wavelength band in the spectral data cube is modified when the spectral data cube is decoded, and outputs the edited matrix data. .
- Such a device may, for example, eliminate security vulnerabilities associated with the generation or transmission of unaltered spectral datacubes.
- Such a device can be, for example, an inspection device that inspects products, or a server computer that delivers queue data to an inspection device.
- the above-described device can be used not only for the purpose of eliminating security vulnerabilities, but also for the purpose of artificially changing the characteristics of an imaging device that acquires compressed images, for example.
- a device includes a storage device that stores the edited matrix data output by the device that edits the matrix data, and a processing circuit.
- the processing circuitry encodes the spectral data cube based on the matrix data to generate the compressed image and/or decodes the spectral data cube from the compressed image based on the matrix data. Run.
- Such a device can generate a modified compressed image or spectral datacube based on the edited matrix data.
- the matrix data may indicate the decoding matrix.
- the apparatus may further comprise an imaging device that produces the compressed image.
- the processing circuitry may decode the spectral data cube from the compressed image based on the matrix data indicative of the decoding matrix. With such a configuration, a modified spectral datacube can be generated.
- the imaging device may include a filter array including a plurality of types of optical filters with different transmission spectra, and an image sensor that detects light that has passed through the filter array and generates the compressed image.
- the encoding matrix corresponds to a two-dimensional distribution of transmission spectra of the filter array.
- Fig. 1 is a diagram for explaining a method of inserting an identifier into a data cube.
- a datacube 20 containing image information for each of four or more (nine in the example of FIG. 1) wavelength bands is generated from one compressed image 10 .
- the compressed image 10 is data obtained by compressing image information of a plurality of wavelength bands as one monochrome image.
- Compressed image 10 may be generated by a hyperspectral camera using compressed sensing (hereinafter also referred to as "compressed HS camera”), such as that disclosed in US Pat.
- a compression HS camera includes a filter array in which a plurality of types of optical filters with different transmission spectra are arranged in a two-dimensional plane, and an image sensor that acquires an image of light transmitted through the filter array. At least some of the filters in the filter array may have a characteristic that the transmittance is locally high in at least two bands among a plurality of narrow bands included in the preset target wavelength band. A two-dimensional distribution of transmission spectra in the filter array is determined based on a predetermined encoding matrix. Each filter in the filter array modulates the intensity of incident light with a different transmittance for each wavelength. Such a filter array is called a "coding element". Compressed image 10 may be produced by a device different from the compressed HS camera.
- a data processor may generate compressed image 10 by encoding a data cube generated by any hyperspectral or multispectral camera with an encoding matrix.
- Datacube 20 is decoded from compressed image 10 by a decoding process using a decoding matrix corresponding to the encoding matrix.
- Compressed sensing technology as disclosed in Patent Document 1, for example, can be used for the decoding process.
- At least one of the encoding matrix and the decoding matrix is edited according to the contents of the identifier 22 so that the identifier 22 is included in the decoded datacube 20.
- the identifier 22 can be automatically added to the compressed image 10 or the datacube 20 by performing the encoding process or the decoding process.
- a separate process of inserting the identifier is required after obtaining the data cube before the identifier is inserted.
- a data cube without an inserted identifier is generated or transmitted, which may result in an unsafe state in terms of data security.
- the identifier 22 is inserted at the same time as the compressed image 10 or datacube 20 is generated. Data security is superior to the prior art because data cubes without identifiers inserted are never generated or transmitted.
- the identifier 22 shown in FIG. 1 is character information that can be visually distinguished by humans, it may be other types of information such as numbers or symbols.
- the identifier may be information such as steganography that is decoded by a computer and is difficult to distinguish visually.
- the letters Panasonic shown in FIG. 1 are an example of the identifier 22 .
- FIG. 2A is a diagram showing an example of editing matrix data.
- an example of editing matrix data representing a decoding matrix will be described.
- the matrix data representing the encoding matrix is called "encoding table”
- the matrix data representing the decoding matrix is called "decoding table”.
- FIG. 2A shows an example of some numerical values of the encoding table 30 and an example of some numerical values of the decoding table 40 .
- FIG. 2A shows a decoded image 50 for one band obtained when decoding using the same table as the encoding table 30, and a decoded image 50 for the band obtained when decoding using the edited decoding table 40.
- An example of a decoded image 60 of is also shown.
- Decoding of the data cube is performed, for example, as disclosed in Japanese Unexamined Patent Application Publication No. 2002-200012, by calculating so that the product of the decoding matrix acting on the vector representing the data cube substantially matches the vector representing the compressed image.
- n doubled in the example of FIG. 2
- the luminance value of the decoded image 60 in that area is the original luminance of the decoded image 50.
- the decoding matrix can be edited so that certain regions have luminance values that are smaller or larger than those of the original decoded image 50 .
- the identifier 22 can be inserted at the same time as the spectrum data cube is generated.
- a similar process can be performed for the encoding matrix.
- a desired identifier 22 can be inserted into the spectral data cube by rewriting the values of a partial region of the encoding matrix and/or the decoding matrix according to the content of the identifier 22 .
- Editing of matrix data in this way can be applied not only when assigning the identifier 22 to the spectrum data cube, but also when modifying the spectrum data cube in other manners.
- random noise may be imparted to the decoded spectral data cube by randomly rewriting the matrix elements of at least one of the encoding and decoding matrices, as schematically illustrated in FIG. 2B.
- Such alterations can make the spectral datacube difficult to read and conceal the information.
- Editing the matrix data is equivalent to pseudo-editing the characteristics of the image sensor. For example, as shown in FIG. 2C, multiplying each matrix element in the decoding matrix by a constant is equivalent to multiplying each pixel value of the captured compressed image by its reciprocal. Therefore, by multiplying each matrix element in the decoding matrix by a constant, the dynamic range of the image sensor can be expanded or reduced in a pseudo manner. In other words, by rewriting the matrix data, it is possible to change the gradation of the image information of at least one wavelength band. By superimposing random noise on the decoding matrix, it is possible to make it appear that noise in the compressed image has increased.
- FIG. 3A is a diagram schematically showing a configuration example of a hyperspectral imaging system.
- This system includes an imaging device 100 and a processing device 200 .
- the imaging device 100 has a configuration similar to that of the imaging device disclosed in Patent Document 1.
- the imaging device 100 includes an optical system 140 , a filter array 110 and an image sensor 160 .
- the optical system 140 and the filter array 110 are arranged on the optical path of light incident from the object 70, which is a subject. Filter array 110 is placed between optical system 140 and image sensor 160 .
- FIG. 3A illustrates an apple as an example of the target object 70 .
- Object 70 is not limited to an apple, and can be any object that can be inspected.
- the image sensor 160 generates data of the compressed image 10 in which information of multiple wavelength bands is compressed as a two-dimensional monochrome image.
- the processing device 200 Based on the data of the compressed image 10 generated by the image sensor 160, the processing device 200 generates image data for each of the multiple wavelength bands included in the target wavelength range.
- the generated image data of multiple wavelength bands is called a "hyperspectral (HS) datacube" or "hyperspectral image data".
- HS hyperspectral
- N is an integer equal to or greater than 4.
- decoded images 20W 1 , 20W 2 , . may be collectively referred to as
- data or signals representing an image that is, a set of data or signals representing pixel values of pixels may be simply referred to as an "image.”
- the filter array 110 is an array of a plurality of translucent filters arranged in rows and columns.
- the multiple filters include multiple types of filters with different transmission spectra (also called “spectral transmittance”), that is, wavelength dependence of light transmittance.
- the filter array 110 modulates the intensity of incident light for each wavelength and outputs the modulated light. This process by filter array 110 is referred to herein as “encoding,” and filter array 110 is sometimes referred to as the "encoding element.”
- the filter array 110 is arranged near or directly above the image sensor 160 .
- “near” means that the image of the light from the optical system 140 is close enough to be formed on the surface of the filter array 110 in a somewhat clear state.
- “Directly above” means that they are so close to each other that there is almost no gap. Filter array 110 and image sensor 160 may be integrated.
- the optical system 140 includes at least one lens. Although optical system 140 is shown as a single lens in FIG. 3A, optical system 140 may be a combination of multiple lenses. Optical system 140 forms an image on the imaging surface of image sensor 160 via filter array 110 .
- the filter array 110 may be arranged away from the image sensor 160 .
- 3B to 3D are diagrams showing configuration examples of the imaging device 100 in which the filter array 110 is arranged away from the image sensor 160.
- FIG. In the example of FIG. 3B, filter array 110 is positioned between optical system 140 and image sensor 160 and at a distance from image sensor 160 .
- FIG. 3C filter array 110 is positioned between object 70 and optics 140 .
- the imaging device 100 comprises two optical systems 140A and 140B with the filter array 110 positioned therebetween.
- an optical system including one or more lenses may be arranged between filter array 110 and image sensor 160 .
- the image sensor 160 is a monochrome photodetector having a plurality of two-dimensionally arranged photodetection elements (also referred to as "pixels" in this specification).
- the image sensor 160 can be, for example, a CCD (Charge-Coupled Device), a CMOS (Complementary Metal Oxide Semiconductor) sensor, or an infrared array sensor.
- the photodetector includes, for example, a photodiode.
- Image sensor 160 does not necessarily have to be a monochrome type sensor. For example, color type sensors with R/G/B, R/G/B/IR, or R/G/B/W filters may be used.
- the wavelength range to be acquired may be arbitrarily determined, and is not limited to the visible wavelength range, and may be the ultraviolet, near-infrared, mid-infrared, or far-infrared wavelength ranges.
- the processing device 200 is a computer that includes a processor and a storage medium such as memory. Based on the compressed image 10 acquired by the image sensor 160, the processing device 200 generates data of a plurality of decoded images 20W1 , 20W2 , . . . 20WN each including information of a plurality of wavelength bands.
- FIG. 4A is a diagram schematically showing an example of the filter array 110.
- FIG. Filter array 110 has a plurality of regions arranged two-dimensionally. In this specification, the area may be referred to as a "cell".
- An optical filter having an individually set spectral transmittance is arranged in each region.
- the spectral transmittance is represented by a function T( ⁇ ), where ⁇ is the wavelength of incident light.
- the spectral transmittance T( ⁇ ) can take a value of 0 or more and 1 or less.
- the filter array 110 has 48 rectangular regions arranged in 6 rows and 8 columns. This is only an example and in actual applications more areas may be provided. The number may be about the same as the number of pixels of the image sensor 160, for example. The number of filters included in the filter array 110 is determined depending on the application, for example, within the range of tens to tens of millions.
- FIG. 4B is a diagram showing an example of spatial distribution of transmittance of light in each of a plurality of wavelength bands W 1 , W 2 , . . . , W N included in the target wavelength range.
- the difference in shading in each region represents the difference in transmittance.
- a lighter area has a higher transmittance, and a darker area has a lower transmittance.
- the spatial distribution of light transmittance differs depending on the wavelength band.
- FIGS. 4C and 4D are diagrams respectively showing examples of spectral transmittance of area A1 and area A2 included in filter array 110 shown in FIG. 4A.
- the spectral transmittance of the area A1 and the spectral transmittance of the area A2 are different from each other.
- the spectral transmittance of filter array 110 differs depending on the region. However, it is not necessary that all regions have different spectral transmittances.
- Filter array 110 includes two or more filters having different spectral transmittances.
- the number of spectral transmittance patterns in the plurality of regions included in the filter array 110 can be equal to or greater than the number N of wavelength bands included in the wavelength range of interest.
- the filter array 110 may be designed such that more than half of the regions have different spectral transmittances.
- the target wavelength band W can be set in various ranges depending on the application.
- the target wavelength range W can be, for example, a visible light wavelength range from about 400 nm to about 700 nm, a near-infrared wavelength range from about 700 nm to about 2500 nm, or a near-ultraviolet wavelength range from about 10 nm to about 400 nm.
- the target wavelength range W may be a wavelength range such as mid-infrared or far-infrared.
- the wavelength range used is not limited to the visible light range.
- the term “light” refers to radiation in general, including not only visible light but also infrared and ultraviolet rays.
- N is any integer equal to or greater than 4, and wavelength bands W 1 , W 2 , .
- a plurality of wavelength bands included in the target wavelength band W may be set arbitrarily.
- different wavelength bands may have different bandwidths.
- the wavelength bands have different bandwidths and there is a gap between two adjacent wavelength bands. In this way, the plurality of wavelength bands may be different from each other, and the method of determining them is arbitrary.
- FIG. 6A is a diagram for explaining spectral transmittance characteristics in a certain area of the filter array 110.
- the spectral transmittance has multiple maxima P1 to P5 and multiple minima for wavelengths within the wavelength range W of interest.
- normalization is performed so that the maximum value of the light transmittance within the target wavelength range W is 1 and the minimum value is 0.
- the spectral transmittance has maximum values in wavelength bands such as the wavelength band W 2 and the wavelength band W N ⁇ 1 .
- the spectral transmittance of each region has maximum values in at least two of the plurality of wavelength bands W1 to WN.
- local maxima P1, P3, P4 and P5 are greater than or equal to 0.5.
- the filter array 110 transmits a large amount of components in a certain wavelength band and transmits less components in other wavelength bands among the incident light. For example, for light in k wavelength bands out of N wavelength bands, the transmittance is greater than 0.5, and for light in the remaining Nk wavelength bands, the transmittance is 0.5. can be less than k is an integer that satisfies 2 ⁇ k ⁇ N. If the incident light is white light that evenly includes all wavelength components of visible light, the filter array 110 converts the incident light into light having a plurality of discrete intensity peaks with respect to wavelength. , and superimposes and outputs these multi-wavelength lights.
- FIG. 6B is a diagram showing, as an example, the result of averaging the spectral transmittance shown in FIG. 6A for each wavelength band W 1 , W 2 , . . . , W N .
- the averaged transmittance is obtained by integrating the spectral transmittance T( ⁇ ) for each wavelength band and dividing by the bandwidth of that wavelength band.
- the transmittance value averaged for each wavelength band is defined as the transmittance in that wavelength band.
- the transmittance is remarkably high in the three wavelength regions having the maximum values P1, P3 and P5. In particular, the transmittance exceeds 0.8 in the two wavelength regions having the maximum values P3 and P5.
- a grayscale transmittance distribution is assumed in which the transmittance of each region can take any value between 0 and 1 inclusive.
- a binary-scale transmittance distribution may be employed in which the transmittance of each region can take either a value of approximately 0 or approximately 1.
- each region transmits a majority of light in at least two wavelength bands of the plurality of wavelength bands included in the wavelength band of interest and transmits a majority of light in the remaining wavelength bands. don't let Here, "most" refers to approximately 80% or more.
- Part of the total cells may be replaced with transparent areas.
- Such a transparent region transmits light in all wavelength bands W1 to WN contained in the wavelength range W of interest with a similarly high transmittance, eg, a transmittance of 80% or more.
- the multiple transparent areas may be arranged in a checkerboard, for example. That is, in the two directions in which the plurality of regions in the filter array 110 are arranged, the regions having different light transmittances depending on the wavelength and the transparent regions can be alternately arranged.
- Such data indicating the spatial distribution of the spectral transmittance of the filter array 110 is obtained in advance based on design data or actual measurement calibration, and stored in a storage medium included in the processing device 200. This data is used for arithmetic processing to be described later.
- the filter array 110 can be constructed using, for example, a multilayer film, an organic material, a diffraction grating structure, or a microstructure containing metal.
- a multilayer film for example, a dielectric multilayer film or a multilayer film containing a metal layer can be used.
- each cell is formed such that at least one of the thickness, material, and stacking order of each multilayer film is different. Thereby, different spectral characteristics can be realized depending on the cell.
- a multilayer film a sharp rise and fall in spectral transmittance can be realized.
- a configuration using an organic material can be realized by differentiating the pigment or dye contained in each cell, or by laminating different materials.
- a configuration using a diffraction grating structure can be realized by providing diffraction structures with different diffraction pitches or depths for each cell.
- a microstructure containing metal it can be produced using spectroscopy due to the plasmon effect.
- the processing device 200 reconstructs a multi-wavelength hyperspectral image 20 based on the compressed image 10 output from the image sensor 160 and the spatial distribution characteristics of transmittance for each wavelength of the filter array 110 .
- multiple wavelengths means a wavelength range greater than the three color wavelength ranges of RGB acquired by a normal color camera, for example.
- the number of wavelength bands may be on the order of 4 to 100, for example. This number of wavelength regions is referred to as the "number of bands". Depending on the application, the number of bands may exceed 100.
- the data to be obtained is the data of the hyperspectral image 20, and the data is assumed to be f.
- f is data obtained by integrating image data f 1 , f 2 , . . . , fN of each band.
- the horizontal direction of the image is the x direction
- the vertical direction of the image is the y direction.
- the data f is three-dimensional data having n ⁇ m ⁇ N elements.
- This three-dimensional data is called “hyperspectral image data” or “hyperspectral datacube”.
- the number of elements of the data g of the compressed image 10 obtained by being encoded and multiplexed by the filter array 110 is n ⁇ m.
- Data g can be represented by the following equation (1).
- each of f 1 , f 2 , . . . , f N is data having n ⁇ m elements. Therefore, the vector on the right side is strictly a one-dimensional vector of n ⁇ m ⁇ N rows and 1 column.
- the vector g is converted into a one-dimensional vector of n ⁇ m rows and 1 column, and calculated.
- the matrix H encodes and intensity - modulates each component f 1 , f 2 , . represents a transform that adds Therefore, H is a matrix with n ⁇ m rows and n ⁇ m ⁇ N columns.
- the processing device 200 utilizes the redundancy of the images included in the data f and obtains the solution using the method of compressed sensing. Specifically, the desired data f is estimated by solving the following equation (2).
- f' represents the estimated data of f.
- the first term in parentheses in the above formula represents the amount of deviation between the estimation result Hf and the acquired data g, ie, the so-called residual term.
- the sum of squares is used as the residual term here, the absolute value or the square root of the sum of squares may be used as the residual term.
- the second term in parentheses is the regularization or stabilization term. Equation (2) means finding f that minimizes the sum of the first and second terms.
- the processing unit 200 can converge the solution by recursive iterative computation and calculate the final solution f'.
- the first term in the parentheses of Equation (2) means an operation for obtaining the sum of squares of the difference between the obtained data g and Hf obtained by transforming f in the estimation process using the matrix H.
- the second term ⁇ (f) is a constraint condition for regularization of f, and is a function reflecting sparse information of estimated data. This function has the effect of smoothing or stabilizing the estimated data.
- the regularization term may be represented by, for example, the Discrete Cosine Transform (DCT), Wavelet Transform, Fourier Transform, or Total Variation (TV) of f. For example, when the total variation is used, it is possible to acquire stable estimated data that suppresses the influence of noise in the observed data g.
- the sparsity of the object 70 in the space of each regularization term depends on the texture of the object 70 .
- a regularization term may be chosen that makes the texture of the object 70 more spars in the space of regularization terms.
- multiple regularization terms may be included in the operation.
- ⁇ is a weighting factor. The larger the weighting factor ⁇ , the larger the reduction amount of redundant data and the higher the compression rate. The smaller the weighting factor ⁇ , the weaker the convergence to the solution.
- the weighting factor ⁇ is set to an appropriate value with which f converges to some extent and does not become over-compressed.
- the image encoded by the filter array 110 is obtained in a blurred state on the imaging surface of the image sensor 160.
- FIG. Therefore, the hyperspectral image 20 can be reconstructed by storing the blur information in advance and reflecting the blur information on the matrix H described above.
- blur information is represented by a point spread function (PSF).
- PSF is a function that defines the degree of spread of a point image to peripheral pixels. For example, when a point image corresponding to one pixel in an image spreads over a region of k ⁇ k pixels around that pixel due to blurring, the PSF is a coefficient group that indicates the effect on the brightness of each pixel in that region. can be defined as a matrix.
- the hyperspectral image 20 can be reconstructed by reflecting the influence of blurring of the encoding pattern by the PSF on the matrix H.
- FIG. The position where the filter array 110 is placed is arbitrary, but a position can be selected where the coding pattern of the filter array 110 is not too diffuse and disappears.
- the hyperspectral image 20, that is, the hyperspectral data cube can be decoded from the compressed image 10 acquired by the image sensor 160.
- the calculation of the above equation (2) is similar to the calculation disclosed in Japanese Unexamined Patent Application Publication No. 2002-200012.
- the same matrix H is used for encoding and decoding for Equation (2).
- at least partially different matrices are used during encoding and during decoding. As a result, as will be described later, modification such as adding a desired identifier to the decoded hyperspectral image can be performed.
- FIG. 7 is a diagram schematically showing a configuration example of an inspection system 1000 according to this embodiment.
- This inspection system 1000 includes an imaging device 100 , a processing device 200 , a display device 300 and a conveyor 400 .
- the display device 300 in the example shown in FIG. 7 is a display. Devices such as speakers or lamps may be provided instead of or in addition to the display.
- Conveyor 400 is a belt conveyor. In addition to the belt conveyor, a picking device may be provided to remove the abnormal objects 70 .
- the object 70 to be inspected is placed on the belt of the conveyor 400 and conveyed.
- Object 70 is any article, such as an industrial product or a food product.
- the inspection system 1000 acquires a hyperspectral image of the object 70 and detects foreign matter mixed in the object 70 based on the image information.
- Foreign objects to be detected can be anything, such as certain metals, plastics, insects, debris, or hair.
- the foreign matter is not limited to these objects, and may be a portion of the object 70 whose quality has deteriorated. For example, if the object 70 is food, a rotten part of the food may be detected as a foreign object.
- the inspection system 1000 can output information indicating that the foreign object has been detected to the display device 300 or an output device such as a speaker, or remove the object 70 including the foreign object by a picking device. can.
- the imaging device 100 is a camera capable of the aforementioned hyperspectral imaging.
- the imaging device 100 generates the above-mentioned compressed image by photographing the object 70 continuously flowing on the conveyor.
- Processing device 200 is any computer such as, for example, a personal computer, a server computer, or a laptop computer.
- the processing device 200 generates a decoded image for each of a plurality of bands by performing decoding processing based on the above equation (2) based on the compressed image generated by the imaging device 100 .
- the processing device 200 detects foreign matter or abnormalities contained in the object 70 based on the images of those bands, and outputs the detection results to the display device 300 .
- FIG. 8 is a block diagram showing a configuration example related to data processing in the inspection system 1000.
- This inspection system 1000 includes an imaging device 100, a processing device 200, a display device 300, and a storage device 500 as components related to data processing.
- the imaging device 100 is a compression HS camera including an image sensor, a filter array, and an optical system such as a lens, as described with reference to FIGS. 1A to 1D.
- the imaging device 100 captures an image of the object 70 , generates compressed image data in which image information of a plurality of bands is superimposed, and sends the compressed image data to the processing device 200 .
- the processing device 200 generates an image for each band based on the compressed image generated by the imaging device 100.
- the processing device 200 comprises processing circuitry 210 .
- the processing circuit 210 includes a processor such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit).
- the processing circuit 210 determines whether or not the target object 70 contains a specific foreign substance based on the compressed image generated by the imaging device 100, and outputs information indicating the determination result.
- the storage device 500 includes arbitrary storage media such as semiconductor memory, magnetic storage device, and optical storage device.
- Storage device 500 stores computer programs executed by processing circuitry 210, data used by processing circuitry 210 in the course of processing, and data generated by processing circuitry 210 in the course of processing.
- the storage device 500 stores, for example, compressed image data generated by the imaging device 100, a decoding table that is matrix data representing a decoding matrix, a decoded HS data cube generated by the processing circuit 210, and foreign matter determination results.
- Stores data indicating FIG. 8 exemplifies the decoding table and the HS data cube among those data.
- these data include identifier information.
- the identifier can be, for example, unique identification information for each imaging device 100 (that is, camera).
- a decoding table is compiled according to the content of the identifier. Using such a compiled decoding table will result in the identifier being included in the decoded HS datacube. Based on the identifier, the camera that took the image can be identified.
- the decoding table edited to generate the data cube containing the identifier may be expressed as "the decoding table containing the identifier" for convenience.
- the encoding table edited to generate the compressed image and the data cube including the identifier is expressed as "the encoding table including the identifier" for convenience. I have something to do.
- FIG. 9 is a diagram for explaining the identifier insertion processing in this embodiment.
- a Macbeth Color Checker is used as the subject, but the actual subject is the object of inspection.
- the spectral information of the HS data cube 20A without inserted identifiers is encoded by the encoding element (that is, the filter array 110 described above) corresponding to the encoding table 30, and the compressed image 10 without inserted identifiers is recorded.
- the encoding element that is, the filter array 110 described above
- the compressed image 10 without inserted identifiers is recorded.
- compression is performed by hardware by the encoding element in the imaging device 100, and the compressed image 10 is generated by the image sensor. It is also possible to generate a compressed image 10 by compressing a pre-generated HS datacube using software, as in the example disclosed in US Pat.
- the processing circuitry 210 performs the decoding process using a partially modified decoding table 40 compared to the encoding table 30 to decode the HS datacube 20B from the compressed image 10 .
- the HS data cube 20B in which the identifier 22 is inserted can be generated by transferring the modified portion to a part of the decoded image in the decoded HS data cube 20B.
- FIG. 10A is a flowchart showing the processing of the processing circuit 210.
- the processing circuitry 210 first acquires the compressed image 10 generated by the imaging device 100 (step S110).
- the processing circuit 210 acquires the edited decoding table 40 recorded in the storage device 500 (step S120). Note that the order of steps S110 and S120 may be reversed.
- the processing circuit 210 uses the edited decoding table 40 to decode the HS data cube 20B from the compressed image 10 (step S130). This decoding process is performed, for example, according to the above equation (2).
- the HS datacube 20B including the identifier 22 is generated.
- the processing circuit 210 outputs the HS data cube 20B including the generated identifier to the storage device 500 for storage (step S140).
- the edited decoding table 40 is generated in advance and recorded in the storage device 500 .
- the processing circuit 210 may edit the decoding table 40 according to the identifier, without being limited to such a form.
- FIG. 10B is a flowchart showing an example of processing when the processing circuit 210 edits the decoding table 40 according to the identifier.
- the processing circuitry 210 first acquires the compressed image 10 generated by the imaging device 100 (step S210).
- the processing circuit 210 acquires the decoding table before editing from the storage device 500 (step S220).
- the decoding table before editing is the same as the encoding table, and is recorded in the storage device 500 in advance.
- the processing circuit 210 edits the decoding table according to the content of the identifier to be assigned (step S230). For example, as described with reference to FIG. 2, the decoding table is edited by rewriting the values in a part of the decoding table by multiplying them by a constant.
- the processing circuit 210 outputs the edited decoding table to the storage device 500 for storage. Note that the processes of steps S220 and S230 may be performed before step S210.
- processing circuit 210 decodes HS data cube 20B from compressed image 10 using the edited decoding table (step S240). This decoding process is performed, for example, according to the above equation (2). As a result, the HS datacube 20B including the identifier 22 is generated.
- the processing circuit 210 outputs the HS data cube 20B including the generated identifier to the storage device 500 for storage (step S250).
- the processing circuit 210 in this embodiment decodes the HS data cube 20B while inserting the identifier.
- data exists only in a state in which the spectrum information is compressed as the compressed image 10 .
- the data exists as the HS data cube 20B in which the identifier 22 is inserted.
- an HS datacube in which the identifier 22 is not inserted does not exist in any process. Therefore, high data security can be maintained.
- the identifier is inserted after decoding the compressed image into the HS datacube.
- the process of generating the HS data cube and the process of inserting the identifier are separated, so that the HS data cube in which the identifier is not inserted exists. If the processing circuit that generates the HS data cube and the processing circuit that inserts the identifier are separate, the HS data cube without the identifier inserted will be transmitted, resulting in a state in which data security is not guaranteed. Occur.
- an HS data cube in which no identifier is inserted is never generated, so data security can be improved. Furthermore, by using the technology of the present embodiment, it is possible to unify the HS datacube acquisition process and the identifier insertion process. As a result, the data processing load can be reduced as compared with the conventional identifier insertion technique, so that the HS data cube including identifiers can be generated with a simpler circuit configuration.
- the HS data cubes may be modified by other methods. For example, random noise may be added to the decoding table to make it difficult to read the original HS datacube, as shown in FIG. 2B.
- the dynamic range of the HS datacube may be expanded or contracted by uniformly multiplying each matrix element of the decoding table by a constant. Such modifications can be similarly applied to each of the following embodiments.
- FIG. 11 is a diagram for explaining the identifier insertion process in this embodiment.
- the spectral information of the HS datacube 20A with no inserted identifier is encoded by the encoding table 30 with the identifier information or its corresponding encoding element.
- the compressed image 10 in which the identifier information is inserted is recorded.
- Compression by encoding may be performed by hardware or by software.
- compression is performed in hardware using encoding elements.
- a filter that reflects or absorbs only the light of a specific wavelength band among the plurality of wavelength bands included in the target wavelength band is arranged only in a partial area corresponding to the position of the identifier in the filter array 110.
- a compressed image 10 including identifier information can be obtained.
- the compressed image 10 may be generated by software, as in the device disclosed in Patent Document 2.
- FIG. In that case, arbitrary identifier information can be inserted into the encoding table 30 by software processing. In this way, decoding processing is performed based on the compressed image 10 in which identifier information and spectrum information are compressed.
- the decoding table 40 in which the identifier information is removed from the encoding table 30 (that is, the identifier information is not inserted) is used. This decoding table 40 is the same as the original encoding table.
- the modified portion of the encoding table 30 is transferred to the decoded image of some bands in the HS datacube 20B. Thereby, the HS data cube 20B in which the identifier is inserted can be generated.
- FIG. 12 is a block diagram for explaining the identifier insertion process in this embodiment.
- the processing unit 200 in this example includes a first processing circuit 210A for generating the compressed image and a second processing circuit 210B for decoding the HS datacube.
- the first processing circuit 210A generates the compressed image 10 including the identifier information by compressing the HS data cube generated by the imaging device 100A using the encoding table 30 in which the identifier information is inserted.
- the first processing circuit 210A outputs the generated compressed image 10 to the storage device 500 for storage.
- the second processing circuit 210B acquires the identifier-inserted compressed image 10 and the identifier-uninserted decoding table 40 from the storage device 500, and performs the decoding process based on the above equation (2). Thereby, the second processing circuit 210B generates the HS data cube 20B into which the identifier is inserted. The second processing circuit 210B causes the storage device 500 to store the generated identifier-containing HS data cube 20B.
- the HS data cube exists without an identifier inserted only in the part where the HS data cube is sent from the imaging device 100A, which is the HS camera.
- the data always exists with the identifier inserted. This makes it possible to eliminate the vulnerability of data.
- an HS data cube acquired by an arbitrary HS camera is compressed using an encoding table in which identifier information is not inserted, and the identifier is inserted after decoding. Compressed images are recorded without an identifier inserted.
- the compressed image is processed using the same decoding table as the encoding table used for compression.
- the HS datacube with no inserted identifier is decoded.
- an identifier is inserted into the HS datacube.
- an HS data cube with no identifier inserted is generated and transmitted from the time the HS data cube is decoded until the identifier is inserted.
- a problem may arise in terms of data security.
- the identifier insertion method using the conventional technology if the identifier is inserted in a stage prior to the compressed image generation process, only the process in which the HS data cube is sent from the HS camera is the HS data in which the identifier is not inserted. Accompanied by moving cubes. In this case, the data security risk is the same as in this embodiment, but more processing processes are required than in this embodiment, and the cost required for data processing increases.
- FIG. 13 is a block diagram showing the system configuration according to the third embodiment of the present disclosure.
- the system of this embodiment comprises a data processing device 700 and one or more inspection systems 1000 . Although three inspection systems 1000 are illustrated in FIG. 13, the number of inspection systems 1000 is arbitrary.
- Each inspection system 1000 comprises the same components as the inspection system 1000 in the first embodiment. However, each inspection system 1000 also includes a communicator 600 for communicating with the data processor 700 .
- the data processing device 700 is a computer that distributes necessary information to each inspection system 1000 .
- the data processing device 700 can be, for example, a server computer owned by the manufacturer of the imaging device 100 .
- the data processing device 700 can communicate with each inspection system 1000 via a network such as the Internet.
- the data processing device 700 comprises a communicator 710 , a processing circuit 720 and a storage device 730 .
- the communicator 710 is a circuit that communicates with the communicator 600 in each inspection system 1000 .
- Processing circuitry 720 is circuitry that includes a processor that generates a decoding table that includes identifier information.
- Storage device 730 stores various data such as a program executed by processing circuit 720 , a decoding table referred to by processing circuit 720 during processing, and a decoding table including identifier information generated by processing circuit 720 .
- the processing circuit 720 edits the decoding table recorded in the storage device 730 according to the content of the identifier to be inserted into the HS data cube, thereby generating a decoding table including the identifier information and recording it in the storage device 730. do.
- the processing circuit 720 delivers the decoding table containing the generated identifier information to the inspection system 1000 .
- the communication device 600 in the inspection system 1000 records the received decoding table containing the identifier in the storage device 500 .
- the processing device 200 generates an HS data cube including an identifier based on the distributed decoding table and the compressed image generated by the imaging device 100 and records it in the storage device 500 .
- the processing of the processing device 200 is the same as the processing in the first embodiment described with reference to FIG. 10A.
- the data processing device 700 may distribute an individual identifier for each inspection system 1000 . By changing the identifier periodically or irregularly, it is possible to identify or track which device was used for decoding or when the decoding occurred.
- FIG. 14 is a block diagram showing the system configuration according to the fourth embodiment of the present disclosure.
- the system in this embodiment comprises a data processing device 700 and a plurality of inspection systems 1000 .
- the processing circuit 720 in the data processing device 700 edits the encoding table instead of the decoding table according to the identifier, and the processing device 200 in each inspection system 1000 uses the edited encoding table.
- the processing device 200 in each inspection system 1000 performs the same operation as the processing device 200 in the second embodiment.
- the processing circuit 720 of the data processing device 700 edits the encoding table recorded in the storage device 730 according to the content of the identifier to be inserted into the HS datacube, thereby creating a code containing identifier information.
- a conversion table is generated and recorded in the storage device 730 .
- the processing circuit 720 delivers the encoding table containing the generated identifier information to the inspection system 1000 .
- the communication device 600 in the inspection system 1000 records the received identifier-containing encoding table in the storage device 500 .
- the processing device 200 includes a first processing circuit 210A and a second processing circuit 210B, as in the second embodiment.
- the first processing circuit 210A generates a compressed image with an identifier by encoding the hyperspectral data cube generated by the imaging device 100 using the coding table with the identifier distributed from the data processing device 700. do.
- the first processing circuit 210A records the generated compressed image with the identifier in the storage device 500.
- FIG. The second processing circuit 210B acquires the identifier-containing compressed image and the identifier-free decoding table from the storage device 500, generates an HS data cube containing the identifier based on them, and records them in the storage device 500.
- the data processing device 700 may distribute an individual identifier for each inspection system 1000 . By changing the identifier periodically or irregularly, it is possible to identify or track which device was used for decoding or when the decoding occurred.
- FIG. 15 is a block diagram showing the configuration of an inspection system 1000 according to the fifth embodiment of the present disclosure.
- An inspection system 1000 in this embodiment has the same hardware configuration as in the first embodiment.
- the processing device 200 in this embodiment synthesizes a part of a plurality of wavelength bands included in the target wavelength band, organizes them into a smaller number of bands, and then performs decoding.
- a decoding table for decoding images for each of a plurality of wavelength bands included in the target wavelength band will be referred to as a "complete decoding table".
- a decoding table for decoding an image for each of a plurality of new wavelength bands in which some of the plurality of wavelength bands included in the target wavelength band are integrated as one synthetic band is called a "reduced decoding table".
- a reduced decoding table has a smaller data size than a full decoding table. By generating and using the reduced decoding table, it is possible to reduce the calculation cost of the decoding process. For example, by integrating a plurality of continuous bands that are not important in inspection as one band, it is possible to reduce the computational processing load without reducing the accuracy of inspection.
- the processing device 200 in this embodiment generates a reduced decoding table based on the full decoding table.
- the processing device 200 comprises three processing circuits 201 , 202 , 203 .
- the processing circuit 201 generates a reduced decoding table by synthesizing some bands based on the complete decoding table. For example, by averaging the values of two or more elements respectively corresponding to two or more consecutive bands in the complete decoding table, those elements are integrated as one element. Such processing produces a reduced decoding table that is compressed in size relative to the full decoding table.
- the processing circuit 202 adds identifier information to the reduced code table.
- the identifier information is embedded in the reduced decoding table by multiplying the value of a part of the reduced decoding table by a constant.
- the processing circuitry 203 generates an HS datacube including identifiers based on the identifier-free compressed image generated by the imaging device 100 and the reduced decoding table including identifier information generated by the processing circuitry 202 .
- the generated HS data cube is recorded in the storage device 500.
- the processing circuit 203 may cause the display device 300 to display the decoded image of each band.
- the order of the band synthesizing process by the processing circuit 201 and the identifier inserting process by the processing circuit 202 may be changed.
- FIG. 16 is a block diagram showing a modified example of the inspection system 1000 of this embodiment.
- the inspection system 1000 in this example has the same hardware configuration as the inspection system 1000 of the second embodiment.
- the processing device 200 in this example comprises processing circuitry 211 , 212 , 213 .
- the processing circuit 211 generates a compressed image including the identifier information by encoding the HS data cube generated by the imaging device 100A using the encoding table including the identifier information recorded in the storage device 500. and record it in the storage device 500 .
- the processing circuit 212 acquires the complete decoding table from the storage device 500 , generates a reduced decoding table by integrating information of some bands, and records it in the storage device 500 .
- the reduced decoding table in this modified example does not include identifier information.
- the processing circuit 213 generates an HS data cube containing the identifier based on the compressed image containing the identifier information generated by the processing circuit 211 and the reduced decoding table not containing the identifier information generated by the processing circuit 212. do.
- the processing circuit 213 records the generated HS data cube in the storage device 500 and causes the display device 300 to display the image of each band. Such an operation can also generate an HS datacube containing identifiers.
- the bands to be synthesized may be determined by referring to a statistical learning model or the like generated in advance using machine learning or the like.
- FIG. 17 is a block diagram showing the system configuration according to the sixth embodiment of the present disclosure.
- the system in this embodiment includes a data processing device 700 and a plurality of inspection systems 1000 as in the third embodiment.
- the configuration of each inspection system 1000 is the same as the configuration in the third embodiment.
- the data processing device 700 generates a reduced decoding table from the complete decoding table, adds identifier data to the reduced decoding table, and distributes the reduced decoding table to each inspection system 1000.
- different from A data processing device 700 in this embodiment comprises two processing circuits 721 and 722 .
- the processing circuit 721 acquires the complete decoding table recorded in the storage device 730 and generates a reduced decoding table by synthesizing information of two or more consecutive wavelength bands with relatively low importance in the complete decoding table. In this process, the processing circuit 721 generates a reduced decoding table according to the model data.
- the model data includes data such as a statistical model based on principal component analysis or a nonlinear model such as a neural network, which includes a weight for each wavelength band as a parameter, and is generated from a large amount of learning data.
- the processing circuit 722 generates a reduced decoding table by adding identifier information to the reduced decoding table generated by the processing circuit 721 and stores it in the storage device 730 . Note that the process of generating the reduced decoding table by the processing circuit 721 and the process of inserting the identifier by the processing circuit 722 may be performed in the reverse order.
- FIG. 18 is a block diagram showing the system configuration according to the seventh embodiment of the present disclosure.
- the system in this embodiment also includes a data processing device 700 and a plurality of inspection systems 1000 .
- the configuration of each inspection system 1000 is the same as the configuration in the fourth embodiment.
- the data processing device 700 generates a reduced decoding table from the full decoding table and distributes it to each inspection system 1000 .
- An encoding table containing identifier information is prerecorded in the storage device 500 .
- the data processing device 700 may generate an encoding table including identifier information and distribute it to the inspection system 1000 .
- the processing circuit 720 in this embodiment generates a reduced decoding table based on the complete decoding table recorded in the storage device 730 and model learning data prepared in advance. This reduced decoding table does not contain identifier information. Processing circuitry 720 distributes the generated reduced decoding table to inspection system 1000 .
- the processing device 200 in the inspection system 1000 includes a processing circuit 210A and a processing circuit 210B, as in the example of FIG.
- the processing circuit 210A encodes the HS datacube generated by the imaging device 100A using the encoding table containing the identifier information recorded in the storage device 500.
- FIG. As a result, a compressed image including identifier information is generated.
- the processing circuit 210B decodes the HS data cube containing the identifier from the compressed image using the reduced decoding table which does not contain the information of the identifier distributed from the data processing device 700 .
- the decoded HS datacube is recorded in the storage device 500, and the image of each band is displayed on the display device 300.
- the technology of the present disclosure is useful, for example, for cameras and measurement equipment that acquire multi-wavelength images.
- the technology of the present disclosure can be used, for example, for detecting foreign matter mixed in articles such as industrial products or foods.
- Compressed Image 20 Data Cube 22 Identifier 30 Encoding Table 40 Decoding Table 50, 60 Decoded Image 70 Object 100 Imaging Device 110 Filter Array 140 Optical System 160 Image Sensor 200 Processing Device 300 Display Device 400 Conveyor 500 Storage Device 600 Communicator 700 Data processor 710 Communication device 720 Processing circuit 730 Storage device 1000 Inspection system
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Remote Sensing (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Image Processing (AREA)
Abstract
Description
ハイパースペクトルカメラまたはマルチスペクトルカメラなどの撮像装置は、一般的なRGB画像を取得するカメラよりも多くの波長バンドの画像情報を取得することができる。それらのカメラによって取得される画像データを、本明細書において「スペクトルデータキューブ」または単に「データキューブ」と称する。特に、ハイパースペクトルカメラによって取得されるハイパースペクトル画像を示すデータを「ハイパースペクトルデータキューブ」または「HSデータキューブ」と称する。スペクトルデータキューブは、複数の波長バンドのそれぞれの画像情報を含む。スペクトルデータキューブは、対象物の素材、状態、または種別を特徴付ける重要な情報を含む。このため、スペクトルデータキューブを厳重に管理することが求められる。例えば製品の検査などの工程において生成されたデータキューブが不正に持ち出されたり外部に流出したりすると問題が生じる可能性がある。このため、データキューブを改変して秘匿化することが求められる場合がある。データキューブが不正に流出した場合であっても追跡が可能であってもよい。そのような目的のために、データに電子透かしなどの識別子を埋め込む対策が考えられる。例えば、そのデータキューブを生成したカメラまたは検査装置を特定する識別子、またはそのデータキューブが生成された時期を特定する識別子をデータキューブに埋め込む対策が考えられる。
まず、本開示の第1の実施形態において用いられるハイパースペクトル撮像システムの構成例を説明する。その後、ハイパースペクトル撮像システムを利用した検査システムの例を説明する。
図3Aは、ハイパースペクトル撮像システムの構成例を模式的に示す図である。このシステムは、撮像装置100と、処理装置200とを備える。撮像装置100は、特許文献1に開示された撮像装置と同様の構成を備える。撮像装置100は、光学系140と、フィルタアレイ110と、イメージセンサ160とを備える。光学系140およびフィルタアレイ110は、被写体である対象物70から入射する光の光路上に配置されている。フィルタアレイ110は、光学系140とイメージセンサ160との間に配置される。
次に、上記の撮像システムを利用した検査システムの例を説明する。
次に、本開示の第2の実施形態を説明する。本実施形態では、復号テーブルではなく符号化テーブルが識別子に応じて編集される点で実施形態1とは異なっている。以下、実施形態1と異なる点を中心に説明し、共通する事項についての説明は省略する。
図13は、本開示の第3の実施形態によるシステムの構成を示すブロック図である。本実施形態のシステムは、データ処理装置700と、1つ以上の検査システム1000とを備える。図13には、3つの検査システム1000が例示されているが、検査システム1000の数は任意である。各検査システム1000は、実施形態1における検査システム1000と同様の構成要素を備える。ただし、各検査システム1000は、データ処理装置700と通信するための通信器600も備えている。データ処理装置700は、各検査システム1000に必要な情報を配信するコンピュータである。データ処理装置700は、例えば撮像装置100のメーカが保有するサーバコンピュータであり得る。データ処理装置700は、例えばインターネットなどのネットワークを介して各検査システム1000と通信することができる。データ処理装置700は、通信器710と、処理回路720と、記憶装置730とを備える。通信器710は、各検査システム1000における通信器600と通信を行う回路である。処理回路720は、識別子の情報を含む復号テーブルを生成するプロセッサを含む回路である。記憶装置730は、処理回路720が実行するプログラム、処理回路720が処理の過程で参照する復号テーブル、および処理回路720によって生成される識別子の情報を含む復号テーブルなどの各種のデータを記憶する。
図14は、本開示の第4の実施形態によるシステムの構成を示すブロック図である。本実施形態におけるシステムは、データ処理装置700と、複数の検査システム1000とを備える。本実施形態においては、データ処理装置700における処理回路720が、復号テーブルではなく符号化テーブルを識別子に応じて編集し、各検査システム1000における処理装置200が、編集された符号化テーブルを用いて識別子を含む圧縮画像を生成する点で、実施形態3とは異なる。各検査システム1000における処理装置200は、実施形態2における処理装置200と同様の動作を実行する。
図15は、本開示の第5の実施形態による検査システム1000の構成を示すブロック図である。本実施形態における検査システム1000は、実施形態1と同様のハードウェア構成を備える。本実施形態における処理装置200は、対象波長域に含まれる複数の波長バンドの一部を合成し、より少数のバンドに編成した上で復号を行う。以下の説明において、対象波長域に含まれる複数の波長バンドのそれぞれについての画像を復号するための復号テーブルを「完全復号テーブル」と称する。対象波長域に含まれる複数の波長バンドの一部が1つの合成バンドとして統合された新たな複数の波長バンドのそれぞれについての画像を復号するための復号テーブルを「縮小復号テーブル」と称する。縮小復号テーブルは、完全復号テーブルよりも小さいデータサイズを有する。縮小復号テーブルを生成して利用することにより、復号処理の演算コストを低減することができる。例えば検査において重要でない複数の連続するバンドを1つのバンドとして統合することにより、検査の精度を落とさずに演算処理の負荷を低減することができる。本実施形態における処理装置200は、完全復号テーブルに基づいて、縮小復号テーブルを生成する。このとき、HSデータキューブに挿入すべき識別子に対応する情報を含む縮小復号テーブルを生成する。処理装置200は、3つの処理回路201、202、203を備える。処理回路201は、完全復号テーブルに基づき、一部のバンドを合成する処理を行うことにより、縮小復号テーブルを生成する。例えば、完全復号テーブルにおける連続する2つ以上のバンドにそれぞれ対応する2つ以上の要素の値を平均化するなどの処理により、それらの要素を1つの要素として統合する。そのような処理により、完全復号テーブルよりもサイズが圧縮された縮小復号テーブルが生成される。処理回路202は、縮小符号テーブルに識別子の情報を付加する。例えば、縮小復号テーブルにおける一部の領域の値を定数倍するなどの方法により、縮小復号テーブルに識別子の情報を埋め込む。処理回路203は、撮像装置100によって生成された識別子を含まない圧縮画像と、処理回路202によって生成された識別子の情報を含む縮小復号テーブルとに基づき、識別子を含むHSデータキューブを生成する。生成されたHSデータキューブは、記憶装置500に記録される。処理回路203は、表示装置300に、復号した各バンドの画像を表示させてもよい。なお、処理回路201によるバンドの合成処理と、処理回路202による識別子の挿入処理の順序は入れ替わってもよい。
図17は、本開示の第6の実施形態によるシステムの構成を示すブロック図である。本実施形態におけるシステムは、実施形態3と同様、データ処理装置700と、複数の検査システム1000とを備える。各検査システム1000の構成は、実施形態3における構成と同様である。本実施形態においては、データ処理装置700が、完全復号テーブルから縮小復号テーブルを生成し、その縮小復号テーブルに識別子のデータを付加して各検査システム1000に配信する点で、実施形態3のシステムとは異なる。本実施形態におけるデータ処理装置700は、2つの処理回路721、722を備える。処理回路721は、記憶装置730に記録された完全復号テーブルを取得し、完全復号テーブルにおいて相対的に重要度の低い連続する2つ以上の波長バンドの情報を合成した縮小復号テーブルを生成する。この処理において、処理回路721は、モデルデータに従って縮小復号テーブルを生成する。モデルデータは、例えば、波長バンドごとの重みをパラメータに含む、主成分分析などによる統計モデルまたはニューラルネットワーク等の非線形モデルなどのデータを含み、多数の学習用データから生成される。処理回路722は、処理回路721によって生成された縮小復号テーブルに、識別子の情報を付加した縮小復号テーブルを生成して記憶装置730に記憶させる。なお、処理回路721による縮小復号テーブルの生成処理と、処理回路722による識別子の挿入処理とが逆の順序で行われてもよい。
図18は、本開示の第7の実施形態によるシステムの構成を示すブロック図である。本実施形態におけるシステムも、データ処理装置700と、複数の検査システム1000とを備える。各検査システム1000の構成は、実施形態4における構成と同様である。本実施形態においては、データ処理装置700が、完全復号テーブルから縮小復号テーブルを生成し、各検査システム1000に配信する。識別子の情報を含む符号化テーブルが記憶装置500に予め記録されている。なお、実施形態4と同様に、データ処理装置700が識別子の情報を含む符号化テーブルを生成して検査システム1000に配信してもよい。
20 データキューブ
22 識別子
30 符号化テーブル
40 復号テーブル
50、60 復号画像
70 対象物
100 撮像装置
110 フィルタアレイ
140 光学系
160 イメージセンサ
200 処理装置
300 表示装置
400 コンベア
500 記憶装置
600 通信器
700 データ処理装置
710 通信器
720 処理回路
730 記憶装置
1000 検査システム
Claims (21)
- コンピュータによって実行される方法であって、
複数の波長バンドの画像情報を含むスペクトルデータキューブを符号化して圧縮画像を生成するために用いられる符号化行列、および/または符号化された前記圧縮画像から前記スペクトルデータキューブを復号するために用いられる復号行列を示す行列データを取得することと、
前記スペクトルデータキューブが復号された場合に前記スペクトルデータキューブにおける少なくとも1つの波長バンドの画像情報が改変されるように前記行列データを編集することと、
編集後の前記行列データを出力することと、
を含む方法。 - 前記行列データを編集することは、前記スペクトルデータキューブが復号された場合に前記少なくとも1つの波長バンドの画像情報が識別子を含むように前記行列データを書き替えることを含む、請求項1に記載の方法。
- 前記行列データを編集することは、前記少なくとも1つの波長バンドの画像情報の読み取りを妨げるノイズが前記スペクトルデータキューブに付与されるように、前記行列データを書き換えることを含む、請求項1または2に記載の方法。
- 前記行列データを編集することは、前記少なくとも1つの波長バンドの画像情報の階調、および解像度の少なくとも一方が変化するように、前記行列データを書き換えることを含む、請求項1から3のいずれかに記載の方法。
- 前記行列データを編集することは、前記スペクトルデータキューブが復号された場合に前記スペクトルデータキューブにおける前記複数の波長バンドの画像情報が改変されるように前記行列データを書き換えることを含む、請求項1から4のいずれかに記載の方法。
- 前記行列データは、前記復号行列を示す、請求項1から5のいずれかに記載の方法。
- 編集後の前記行列データを出力することは、前記復号行列を示す前記行列データに基づいて前記圧縮画像から前記スペクトルデータキューブを復号する装置に前記行列データを送信することを含む、請求項6に記載の方法。
- 前記行列データを出力することは、前記復号行列を示す前記行列データを記憶媒体に記憶させることを含み、
前記方法は、さらに、
前記圧縮画像を取得することと、
編集後の前記行列データを用いて前記圧縮画像から前記スペクトルデータキューブを復号することと、
を含む、請求項6に記載の方法。 - 前記圧縮画像は、フィルタアレイを備えた撮像装置によって生成され、
前記フィルタアレイは、透過スペクトルが互いに異なる複数種類の光学フィルタを含み、
前記符号化行列は、前記フィルタアレイの透過スペクトルの二次元分布に対応し、
前記スペクトルデータキューブを復号することは、前記復号行列に基づく圧縮センシング処理によって前記圧縮画像から前記スペクトルデータキューブを復号することを含む、請求項1から8のいずれかに記載の方法。 - 前記行列データは、前記符号化行列を示す、請求項1から5のいずれかに記載の方法。
- 前記行列データを出力することは、前記符号化行列を示す前記行列データに基づいて前記スペクトルデータキューブを符号化して前記圧縮画像を生成する装置に前記行列データを送信することを含む、請求項10に記載の方法。
- 編集後の前記行列データを出力することは、前記符号化行列を示す前記行列データを記憶媒体に記憶させることを含み、
前記方法は、さらに、
前記スペクトルデータキューブを取得することと、
編集後の前記行列データに基づいて前記スペクトルデータキューブを符号化して前記圧縮画像を生成することと、
を含む、請求項10に記載の方法。 - 前記識別子は、前記符号化行列に基づいて前記圧縮画像を生成する装置、または前記復号行列に基づいて前記スペクトルデータキューブを復号する装置を特定する情報を含む、請求項2に記載の方法。
- 前記識別子は、前記識別子が付与された時期を特定する情報を含む、請求項2または13に記載の方法。
- 前記行列データを取得することは、
前記圧縮画像から、対象波長域に含まれる複数の波長バンドのそれぞれについての画像を復号するための復号テーブルを取得することと、
前記復号テーブルに基づいて、前記複数の波長バンドのうちの2つ以上の波長バンドが1つの波長バンドとして統合された、前記複数の波長バンドよりも少数の波長バンドのそれぞれについての画像を前記スペクトルデータキューブとして復号するための縮小復号テーブルを、前記復号行列を示す前記行列データとして生成することと、
を含み、
前記行列データを編集することは、前記縮小復号テーブルを編集することを含む、
請求項1から14のいずれかに記載の方法。 - 編集後の前記行列データを出力することは、編集後の前記縮小復号テーブルを他の装置に送信することを含む、請求項15に記載の方法。
- 前記行列データは、前記符号化行列を示す第1行列データと、前記復号行列を示す第2行列データとを含み、
前記行列データを編集することは、前記第1行列データおよび前記第2行列データを編集することを含む、請求項1から5のいずれかに記載の方法。 - 複数の波長バンドの画像情報を含むスペクトルデータキューブを符号化して圧縮画像を生成するために用いられる符号化行列、および/または符号化された前記圧縮画像から前記スペクトルデータキューブを復号するために用いられる復号行列を示す行列データを記憶する記憶装置と、
前記スペクトルデータキューブが復号された場合に前記スペクトルデータキューブにおける少なくとも1つの波長バンドの画像情報が改変されるように前記行列データを編集し、編集後の前記行列データを出力する処理回路と、
を備える装置。 - 請求項18に記載の装置によって出力された編集後の前記行列データを記憶する記憶装置と、
前記行列データに基づいて前記スペクトルデータキューブを符号化して前記圧縮画像を生成する処理、および/または、前記行列データに基づいて前記圧縮画像から前記スペクトルデータキューブを復号する処理を実行する処理回路と、
を備える装置。 - 前記行列データは前記復号行列を示し、
前記装置は、前記圧縮画像を生成する撮像装置をさらに備え、
前記処理回路は、前記復号行列を示す前記行列データに基づいて前記圧縮画像から前記スペクトルデータキューブを復号する、
請求項19に記載の装置。 - 前記撮像装置は、
透過スペクトルが互いに異なる複数種類の光学フィルタを含むフィルタアレイと、
前記フィルタアレイを通過した光を検出して前記圧縮画像を生成するイメージセンサと、
を備え、
前記符号化行列は、前記フィルタアレイの透過スペクトルの二次元分布に対応する、
請求項20に記載の装置。
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202280009937.8A CN116762330A (zh) | 2021-02-03 | 2022-01-18 | 改变光谱数据立方体的方法和装置 |
JP2022579422A JPWO2022168594A1 (ja) | 2021-02-03 | 2022-01-18 | |
US18/353,883 US20230360393A1 (en) | 2021-02-03 | 2023-07-18 | Method and device of altering spectral data cube |
Applications Claiming Priority (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2021016153 | 2021-02-03 | ||
JP2021-016153 | 2021-02-03 | ||
JP2021-203149 | 2021-12-15 | ||
JP2021203149 | 2021-12-15 |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US18/353,883 Continuation US20230360393A1 (en) | 2021-02-03 | 2023-07-18 | Method and device of altering spectral data cube |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2022168594A1 true WO2022168594A1 (ja) | 2022-08-11 |
Family
ID=82740635
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2022/001493 WO2022168594A1 (ja) | 2021-02-03 | 2022-01-18 | スペクトルデータキューブを改変する方法および装置 |
Country Status (3)
Country | Link |
---|---|
US (1) | US20230360393A1 (ja) |
JP (1) | JPWO2022168594A1 (ja) |
WO (1) | WO2022168594A1 (ja) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9599511B2 (en) * | 2014-11-19 | 2017-03-21 | Panasonic Intellectual Property Management Co., Ltd. | Imaging apparatus comprising coding element and spectroscopic system comprising the imaging apparatus |
US20170205337A1 (en) * | 2014-07-24 | 2017-07-20 | Ecole Polytechnique Federale De Lausanne (Epfl) | Compact Multifunctional System for Imaging Spectroscopy |
JP2019012869A (ja) * | 2017-06-29 | 2019-01-24 | ソニーセミコンダクタソリューションズ株式会社 | 画像送受信システム、サーバおよび撮像装置 |
-
2022
- 2022-01-18 JP JP2022579422A patent/JPWO2022168594A1/ja active Pending
- 2022-01-18 WO PCT/JP2022/001493 patent/WO2022168594A1/ja active Application Filing
-
2023
- 2023-07-18 US US18/353,883 patent/US20230360393A1/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20170205337A1 (en) * | 2014-07-24 | 2017-07-20 | Ecole Polytechnique Federale De Lausanne (Epfl) | Compact Multifunctional System for Imaging Spectroscopy |
US9599511B2 (en) * | 2014-11-19 | 2017-03-21 | Panasonic Intellectual Property Management Co., Ltd. | Imaging apparatus comprising coding element and spectroscopic system comprising the imaging apparatus |
JP2019012869A (ja) * | 2017-06-29 | 2019-01-24 | ソニーセミコンダクタソリューションズ株式会社 | 画像送受信システム、サーバおよび撮像装置 |
Also Published As
Publication number | Publication date |
---|---|
JPWO2022168594A1 (ja) | 2022-08-11 |
US20230360393A1 (en) | 2023-11-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP6952277B2 (ja) | 撮像装置および分光システム | |
Mihoubi et al. | Multispectral demosaicing using pseudo-panchromatic image | |
CN105791640B (zh) | 摄像装置 | |
WO2021246192A1 (ja) | 信号処理方法、信号処理装置、および撮像システム | |
WO2022163421A1 (ja) | 検査対象に含まれる異物を検出する方法および装置 | |
CN115211106A (zh) | 信号处理方法、信号处理装置及摄像系统 | |
JP6945195B2 (ja) | 光学フィルタ、光検出装置、および光検出システム | |
EP4092397A1 (en) | Optical filter array, optical detection device, and optical detection system | |
Farid | Image forensics | |
JP6975933B2 (ja) | 画像処理装置、撮像システム、画像処理方法およびプログラム | |
EP4332877A1 (en) | Image processing device, image capturing system, and method for estimating error in restored image | |
CN115038945A (zh) | 摄像装置 | |
WO2022168594A1 (ja) | スペクトルデータキューブを改変する方法および装置 | |
CN115668008A (zh) | 滤波器阵列及光检测系统 | |
CN116762330A (zh) | 改变光谱数据立方体的方法和装置 | |
CN117460947A (zh) | 信号处理装置及信号处理方法 | |
Sippel et al. | Cross spectral image reconstruction using a deep guided neural network | |
Mathur et al. | An efficient spatial domain based image watermarking using shell based pixel selection | |
WO2023106143A1 (ja) | 分光画像を生成するシステムに用いられる装置およびフィルタアレイ、分光画像を生成するシステム、ならびにフィルタアレイの製造方法 | |
WO2023106142A1 (ja) | 信号処理方法、プログラム、およびシステム | |
Kaarna et al. | Multiwavelets in watermarking spectral images | |
JP2024020922A (ja) | 復元画像の評価方法および撮像システム、 | |
CN118251898A (zh) | 在生成分光图像的系统中使用的装置及滤光器阵列、生成分光图像的系统以及滤光器阵列的制造方法 | |
JP7122636B2 (ja) | フィルタアレイおよび光検出システム | |
WO2023100660A1 (ja) | 撮像システムおよび撮像方法 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 22749470 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 202280009937.8 Country of ref document: CN |
|
ENP | Entry into the national phase |
Ref document number: 2022579422 Country of ref document: JP Kind code of ref document: A |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 22749470 Country of ref document: EP Kind code of ref document: A1 |