KR101498096B1 - Apparatus and method for discriminating of geographical origin of agricutural products using hyperspectral imaging - Google Patents

Abstract

A hyperspectral image analyzing apparatus according to the present invention includes: a sample plate which arranges target crops for discriminating a production place; a sample transfer unit which transfers the sample plate to a first place and a second place; and an optical fiber cable to radiate a light source in a line shape. The present invention also includes: at least one light oscillation unit which radiates light to the sample plate; a standard board measuring unit which corrects the intensity of irradiation on the surface of the sample; a light blocking unit which blocks external light; a lens unit which receives reflected light of the light radiated to target crops; an image sensor unit which generates an image spectrograph dividing reflection light, received from the lens unit, according to frequencies and a hyperspectral reflection light image in which light is divided according to frequencies; and a data acquisition calculating unit which acquires the hyperspectral reflection light image of each spectroscopic frequency from the image sensor unit, corrects noise caused by uneven irradiation in the hyperspectral reflection light image of each spectroscopic frequency, acquires an interest area by excluding remaining areas except the interest area from the acquired hyperspectral image of each frequency, and discriminates the production place of the target crops by applying a prediction model using a partial least square discriminating model. The system of the present invention is able to discriminate a production place of target crops nondestructively.

Description

[0001] APPARATUS AND METHOD FOR DISCRIMINATING OF GEOGRAPHICAL ORIGINAL AGRICULTURAL PRODUCTS [0002] USING HYPERSPECTRAL IMAGING [0003]

The present invention relates to an apparatus and method for discriminating the origin of grains by analyzing spectra, and more particularly, to an apparatus and method for discriminating origin of grains using ultrasound and fluorescence image processing.

As the interest in health has increased recently, the demand of consumers has changed from the quantitative aspect to the stability and quality. Especially, the reliability of the production site is important.

In addition, agricultural products from many countries of the world are imported into the country and widely circulated due to the opening of agricultural imports. In particular, it is a fact that it is difficult for the administrative investigating authorities to discriminate realistically about agricultural products that are sold as domestic low-priced agricultural products or genetically modified agricultural products, or are mixed with domestic products.

In this situation, various methods for determining origin of origin have been developed, for example, methods such as DNA analysis and isotope analysis, which can distinguish domestic pork from imported pork.

In addition to the above-mentioned methods, infrared spectral analysis, nuclear magnetic resonance analysis (NMR), or near-infrared spectrum analysis have been developed.

However, the destructive discrimination methods and apparatuses for discriminating origin are disadvantageous in that it is impossible to reuse samples to be discriminated. Spectral analysis, which is a non-destructive analysis method, has some problems to be improved in accuracy.

SUMMARY OF THE INVENTION Accordingly, the present invention has been made in order to solve the above problems, and it is an object of the present invention to provide a method and apparatus for irradiating UV-A or crystalline tungsten halogen light to a target grain for nondestructive determination of origin, It is an object of the present invention to provide an apparatus and a method for finally determining a country of origin by applying one prediction model.

In order to achieve the above object, the ultrasonic image analyzing apparatus according to the present invention includes a sample receiver capable of disposing a target grain for discrimination of origin, a sample transfer unit for transferring the sample receiver from a first position to a second position, And a lens unit for receiving reflected light and fluorescence of the light source irradiated to the object grains and a lens unit for receiving the reflected light and fluorescence from the light source, (IMAGE SPECTROGRAPH) for spectroscopically reflecting the reflected light and fluorescence received from the light source and the ultrasound spectroscopic image, and an image sensor unit And the noise due to the uneven irradiation light is observed in the ultra-spectral image by the above-mentioned spectroscopic wavelength. A data acquiring operation unit for acquiring a region of interest by excluding regions other than the region of interest in the ultraspectral image acquired for each wavelength and applying a prediction model using the partial least squares discriminant model to determine the origin of the object grains; .

According to the present invention, the origin of a target sample, i.e., a target grain, can be discriminated nondestructively through an ultraspectral image analysis system and method.

In addition, the apparatus and method of the present invention can be utilized for analyzes requiring promptness, accuracy, and real-time, and it is possible to reuse grain samples after analysis because of the advantage of the present system that no pre-processing is required.

In addition, the application range of the ultra-spectroscopic image analysis technique applied to the present invention can be expanded to be applied to nondestructive inspection in various fields, and further, it is expected to have an effect on related industries.

FIG. 1 shows a configuration of an ultrasound image analyzing apparatus according to the present invention.
FIG. 2 is a flowchart illustrating analysis of an ultrasound reflected light image according to an embodiment of the present invention.
3 is a schematic view of a reference plate measuring section of the present invention.
FIG. 4 shows a result of a principal component model for determining the country of origin of red pepper powder according to an embodiment of the present invention.
FIG. 5A shows a result of a principal component model for discriminating the origin of domestic rice and US rice according to an embodiment of the present invention.
FIG. 5B shows a result of a principal component model for discriminating the origin of domestic rice and Chinese rice according to an embodiment of the present invention.
FIGS. 6A and 6B illustrate PLSR models for predicting red pepper powder incorporation using ultrasound fluorescence images and mixed domestic and imported red pepper powder ultrasonic fluorescence images according to an embodiment of the present invention.
FIGS. 7A and 7B illustrate PLSR models for predicting mixing of red pepper powder using ultraspectral reflected light images, and mixed ultraviolet spectra of domestic and imported red pepper powder according to an embodiment of the present invention.
FIGS. 8A and 8B illustrate a PLSR model for predicting rice incorporation using ultrasound fluorescence images and mixed domestic and imported rice ultrasonic spectroscopic fluorescence images according to an embodiment of the present invention.

The present invention is capable of various modifications and various forms, and specific embodiments are illustrated in the drawings and described in detail in the text.

It is to be understood, however, that the invention is not intended to be limited to the particular forms disclosed, but on the contrary, is intended to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the invention. The terms first, second, etc. may be used to describe various elements, but the elements should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The singular expressions include plural expressions unless the context clearly dictates otherwise. In the present application, the terms "comprising" or "having ", and the like, are intended to specify the presence of stated features, integers, steps, operations, elements, parts, or combinations thereof, But do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, parts, or combinations thereof.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Reference will now be made in detail to the preferred embodiments of the present invention, examples of which are illustrated in the accompanying drawings.

FIG. 1 shows a configuration of an ultrasound image analyzing apparatus according to the present invention.

Referring to FIG. 1, the ultrasonic image analyzing apparatus of the present invention includes a sample support 600 capable of placing object grains for discrimination of origin; A sample transfer unit 700 for transferring the sample support 600 from a first position to a second position; A reference plate 903 for correcting the difference between the output of the line light irradiated to the measurement surface and a reference plate 903 for irradiating the irradiated light source in a line shape Fiber optic cable for; A lens unit 400 for receiving reflected light of the light source irradiated on the object grain; An image spectrograph (SPECTROGRAPH) which spectrally reflects the reflected light and the fluorescence received from the lens unit 400 for each wavelength; An image sensor unit 200 for generating spectroscopic ultra-spectral reflection light and fluorescence image for each wavelength; And acquiring the spectroscopic reflected light and fluorescence image by the spectroscopic wavelength from the image sensor unit (200), correcting the noise due to the uneven irradiation light in the spectroscopic reflected light image by the spectroscopic wavelength, And a data acquisition and calculation unit (100) for acquiring a region of interest by excluding regions other than the region of interest in the ultrasound image and determining a country of origin of the target grains by applying a prediction model using a partial least squares discriminant model .

The light source oscillating units 510 and 520 may be selected from any one of UV-A light and modified tungsten halogen light having a wavelength of 365 nm.

The light irradiation direction of the optical fiber cable may be 15 degrees about the normal direction of a vertical line connecting the first position and the second position.

When the light source of the light source oscillating units 510 and 520 is the UV-A light having the wavelength of 365 nm, it is possible to discriminate the origin of the object grains by comparing images of a specific wavelength. It is preferable that the wavelength range of the UV-A light has a wavelength range of 400 nm to 700 nm, and when the light source of the light source oscillating unit 510 or 520 is a quartz tungsten halogen light, the measurement wavelength range is preferably 400 nm to 1000 nm .

The image measuring unit includes an image measuring unit height regulating unit 1000 and an image measuring unit 200 including the image sensor unit 200, the image spectrograph 300, the lens unit 400, and the light intercepting unit 800, Not shown) can be adjusted.

In addition, the light blocking unit 800 is provided to prevent light from entering the lens unit. In the dark state, the ultrasound image is obtained by closing the light blocking unit 800 before the ultrasound image of the sample is measured, And the reference plate 903 and the ultrasound image of the sample are measured while the light shielding portion is open. It is used to remove the noise of the equipment by removing the hyperspectral image in the dark state in the ultraspectral image, and to measure the ultrasound fluorescence image and the reflected light image.

The reference plate 903 for correcting the difference in the output of the line light irradiated to the measurement surface is provided on one surface of the sample support 600 so that the height of the measurement surface of the reference plate 903 is equal to the height of the measurement surface of the sample A reference plate height adjusting device 901 is provided. The reference plate 903 uses a material having a wavelength of 400 nm to 1000 nm and a reflectance of 99% or more for 1000 nm to 1700 nm. A reference plate protective cover opening / closing device 902 for preventing contamination of the reference plate 903 is provided at the upper end of the reference plate 903. The superspectral image of the reference plate 903 is measured before the measurement of the ultrasound reflected light image to correct the surface output of the reflected light.

In the case of domestic red pepper powder and red pepper powder from Vietnam, the correlation coefficient between the two origins was analyzed as non-image of fluorescence wavelengths of 583 nm and 540 nm (F583 nm / F540 nm) And discrimination was possible.

In the case of domestic red pepper powder and Chinese red pepper powder, the correlation coefficient between the two origins was analyzed as 655 nm and 573 nm difference image (F655nm - F573nm), and the correlation coefficient was as high as 0.9085. Respectively.

In the case of domestic red pepper powder and red pepper powder imported from Vietnam and China, correlation analysis between domestic and imported samples was performed with non-image of 979nm and 974nm in visible and near infrared wavelengths (F979nm / F974nm) It was possible to discriminate whether it was domestic or not.

The correction in the spectroscopic reflected light image by the spectroscopic wavelength is corrected using the following equation (1).

는

파장에서의 반사광 영상이고, 상기

는 상기

파장에서 획득된 초분광 반사광 영상이며, 상기

는 장치의 노이즈를 제거하기 위해서 상기 광을 차단한 상태에서 획득된 초분광 영상이고, 상기

는 반사율 99 퍼센트의 백색 테프론에 상기 광을 조사하여 획득된 초분광 영상이다.Here,

The

Is a reflection light image at a wavelength,

Quot;

An ultrasound reflection light image obtained at a wavelength,

Is an ultra-spectroscopic image obtained in a state in which the light is blocked to remove noise of the apparatus,

Is an ultra-spectroscopic image obtained by irradiating the above-mentioned light to white Teflon with a reflectance of 99 percent.

The principal component analysis model was developed by using the obtained ultrasound image data to develop a principal component model to discriminate the origin of red pepper powder from domestic red pepper powder and imported red pepper powder from China and Vietnam. As a result, as shown in FIG. 4, Origin of Chinese red pepper powder was distinguished. In addition, a principal component analysis model was developed as shown in FIGS. 5A and 5B for discriminating the origin of domestic rice and imported rice (Chinese origin, US origin). As a result, domestic and imported rice could be discriminated.

Next, a partial least squares regression model was developed to determine whether red pepper powder was mixed with domestic red pepper powder. The partial least squares discrimination model is characterized by assigning a predetermined weight to the ultrasound spectra acquired for each wavelength. Six domestic red pepper powder, 6 Chinese red pepper powder (Kwangyang, Kwangtap, Beijing Hong, Braid, 1 second) and one red pepper powder (Vietnam first) produced in Nonghyup, Were used. A total of 36 sets of domestic red pepper powder and imported red pepper powder were prepared in six steps (100: 0, 80:20, 60:40, 40:60, 20:80, 0: 100) And developed a prediction model. As shown in FIG. 6, the determination coefficient (R _C ² ) of the calibration model (R _C ² ) was 0.910 and the calibration model prediction error (SEC) was ± 8.953% As a result of cross validation of this calibration model, the determination coefficient (R _V ² ) was found to be 0.897 and the prediction error (SEP) was found to be ± 9.637%. In addition, as a result of determining the degree of incorporation of imported acid into the domestic product as an ultraspectral reflected light image, the determination coefficient (R _C ² ) of the calibration model is 0.958 and the calibration model prediction error (SEC) is ± 6.092% As a result of cross validation of the calibration model, the determination coefficient (R _V ² ) was 0.954 and the prediction error (SEP) was ± 6.452%, and the incorporation amount was detectable to less than 7%.

Next, a partial least squares regression model was developed to determine whether imported rice was incorporated into domestic rice. The mixing ratio of domestic rice and imported rice was divided into six stages (100: 0, 80:20, 60:40, 40:60, 20: 80, 0: 100), and three models were prepared to obtain hyperspectral image data. As shown in Fig. 8, the determination coefficient (R _C ² ) of the calibration model (R _C ² ) was 0.996 and the calibration model prediction error (SEC) was ± 1.950% As a result of cross-validation of this calibration model, the determination coefficient (R _V ² ) was 0.958 and the prediction error (SEP) was ± 6.369%, and the incorporation amount was detectable to less than 7%.

FIG. 2 is a flowchart illustrating analysis of an ultrasound reflected light image according to an embodiment of the present invention.

Referring to FIG. 2, the method for discriminating origin using ultra-spectral image analysis according to an embodiment of the present invention includes the steps of measuring a dark state ultrasound image, and irradiating the corrected tungsten halogen light from the light source oscillators 510, ; Receiving reflected light of the light irradiated to the reference plate through a lens unit (400); Acquiring a superspectral image for each wavelength of the reflected light received from the lens unit (400) using an image spectrograph (IMAGE SPECTROGRAPH); Irradiating the object grain with modified tungsten halogen light from the light source oscillating parts (510, 520); Receiving reflected light of the light irradiated to the object grain through a lens unit (400); Acquiring a superspectral image for each wavelength of the reflected light received from the lens unit (400) using an image spectrograph (IMAGE SPECTROGRAPH); Removing noise due to irregular irradiation light on the obtained ultra-spectral image for each wavelength; Acquiring an interest image region excluding an image region other than the object grain in the ultra-spectroscopic image from which the noise is removed; And applying a partial least squares discriminant model to the region of interest image to predict whether the domestic and imported grains of the subject grains are mixed.

It is preferable that the wavelength range of the modified tungsten halogen light is 400 nm to 1000 nm.

The step of removing the noise is preferably corrected using the equation (1).

The Partial Least Square Regression (PLSR) is a step of assigning a predetermined weight to an ultrasound image acquired for each wavelength.

In one embodiment, the obtaining (S1) of the ultrasound spectral image comprises irradiating the object grain with UV-A or quartz tungsten halogen light at the light source oscillator 510, 520 and irradiating the object with the irradiated UV-A or modified tungsten And a step in which the halogen light is reflected and received through the lens unit 400.

(S3) of removing an image region other than the image region of the object grain for final discrimination through the step (S2) of removing the noise of the ultrasonic reflected light image and removing the noise of the apparatus, (PLSR) that gives a predetermined weight to the ultrasound reflected light image acquired at step S3 may be applied to the ultrasound reflected light image obtained at step S3 to predict the country of origin. (S5)

The present invention has been described in detail with respect to an apparatus and method for discriminating origin of grains using ultrasound and fluorescence image processing according to a preferred embodiment of the present invention. However, it will be understood by those skilled in the art that various modifications and changes may be made thereto without departing from the spirit and scope of the invention as set forth in the following claims. It will be understood that the present invention can be changed.

Accordingly, the scope of the present invention is limited only by the claims that follow.

100: data acquisition and calculation unit 200: image sensor
300: image spectrograph 400: lens part
510: Crystal Tungsten Halogen Light Source 520: UV-A Light Source
600: sample holder 700: sample transfer part
800: light blocking part 900: reference plate measuring part
901: Reference plate height adjusting device 902: Reference plate protecting cover opening / closing device

Claims (17)

A sample holder capable of placing the subject grain for discrimination of origin;
A sample transfer unit for transferring the sample holder from the first position to the second position;
At least one light source oscillating part including an optical fiber cable for irradiating the light source with the line light and irradiating light to the sample receiver;
A lens unit for receiving reflected light of the light source irradiated to the object grain;
An image spectrograph (SPECTROGRAPH) for spectroscopically reflecting the light received from the lens unit for each wavelength;
An image sensor unit for generating a spectroscopic ultra-spectral reflection light image for each wavelength;
A reference plate measuring unit including a reference plate having no wavelength dependence to correct a difference in output of the line light irradiated to the surface of the object grain;
And
Wherein the spectroscopic reflection image of the spectroscopic wavelength is obtained from the image sensor unit, the noise due to the non-uniform irradiation light is corrected in the spectroscopic reflection image of the spectroscopic wavelength, A data acquiring operation unit for acquiring a region of interest by excluding the remaining regions except for the region of interest, and determining a country of origin of the object grains by applying a prediction model using a partial least squares discriminant model;
And a second spectroscopic image analyzing system. The method according to claim 1,
Further comprising a light blocking unit disposed on one surface of the lens unit to remove noise from at least one component of the ultrasound spectral image analysis system. The light source apparatus according to claim 1,
Wherein the ultraviolet light is a light source of any one of UV-A light and modified tungsten halogen light having a wavelength of 365 nm. The optical fiber cable according to claim 1,
Wherein the angle between the first position and the second position is 15 degrees about a normal direction of a vertical line connecting the first position and the second position. The method of claim 3,
Wherein when the light source of the light source oscillating unit is UV-A light having a wavelength of 365 nm, images of a specific wavelength are compared with each other to discriminate the origin of the object grains. 6. The method of claim 5,
Wherein the measurement wavelength range of the UV-A light is 400 nm to 700 nm. 6. The method according to claim 5,
And comparing the images obtained at wavelengths of 578 nm and 616 nm when the target grain is red pepper powder, to discriminate the origin of the target grains. 6. The method according to claim 5,
Wherein when the target grain is rice, images obtained at wavelengths of 477 nm and 679 nm are compared to discriminate the origin of the target grains. The method of claim 3,
Wherein the measurement wavelength range is 400 nm to 1000 nm when the light source oscillator is a quartz tungsten halogen light source. The method according to claim 1,
The correction in the spectroscopic reflected light image by the spectroscopic wavelength,
The reflected light image at the i-th wavelength is obtained by irradiating the white Teflon with the light by subtracting the ultra-spectral image obtained at the i-wavelength and the ultra-spectral image obtained by blocking the light to remove the noise of the apparatus Wherein the correction is performed using an equation obtained by multiplying a value obtained by dividing the super-spectroscopic image obtained in a state in which the light is cut off and the super-spectroscopic image obtained in a state in which the light is cut off by 100 to remove noise of the apparatus. The method of claim 1, wherein the partial least squares discriminant model comprises:
Wherein a predetermined weight is assigned to the ultrasound image acquired for each wavelength. 3. The light-emitting device according to claim 2,
Wherein the lens is provided on one surface of the lens unit to prevent external light from entering the lens. The apparatus according to claim 1, wherein the reference plate of the reference plate measuring unit includes white Teflon having a reflectance of 99%
Wherein the reference plate measuring unit comprises:
A reference plate height adjusting device which is provided on one surface of the sample holder and adjusts the height so that the height of the measurement surface of the target grain is equal to the height of the measurement surface of the reference plate; And
A reference plate protective cover opening / closing device provided on the reference plate to prevent foreign matter or contamination;
Further comprising an ultrasound image analyzing system. In a method for discriminating the origin of a target grain using ultrasound image analysis,
Irradiating the object grain with a crystal tungsten halogen light from a light source oscillation portion;
Receiving reflected light of the light irradiated to the object grain through a lens unit;
Obtaining a superspectral image for each wavelength by using the image spectrograph (IMAGE SPECTROGRAPH), the reflected light received from the lens unit;
Removing noise due to irregular irradiation light on the obtained ultra-spectral image for each wavelength;
Acquiring an interest image region excluding an image region other than the object grain in the ultra-spectroscopic image from which the noise is removed; And
Applying a partial least squares discriminant model to the region of interest image to predict the mixing ratio of domestic and imported grains of the subject grains;
Lt; / RTI >
The step of removing the noise includes:
The reflected light image at the i-th wavelength is obtained by irradiating the white Teflon with the light by subtracting the ultra-spectral image obtained at the i-wavelength and the ultra-spectral image obtained by blocking the light to remove the noise of the apparatus Wherein the correction is performed by using an equation obtained by multiplying the value obtained by dividing the ultrasonic image obtained by subtracting the light from the ultrasonic image obtained by cutting off the light to 100 by the difference between the ultrasonic image and the device, Way. 15. The method of claim 14,
Wherein the wavelength range of the modified tungsten halogen light is 400 nm to 1000 nm. delete 15. The method of claim 14, wherein the partial least squares discriminant model comprises:
And a predetermined weight is assigned to the ultrasound image acquired for each wavelength.

Images