CN110487737B - Image information extraction and calculation method and system for spectrum detection of smart phone - Google Patents

Abstract

The invention discloses an image information extraction and calculation method for spectrum detection of a smart phone, which comprises the following steps: obtaining RGB images of the spectrum images, and rotating the images to a uniform angle; selecting an effective image area of the image according to different samples; extracting RGB values in the selected region, and converting the RGB values in the region into gray values; calculating the absorbance of the sample through an image gray value inversion model, obtaining the absorbance of standard samples with different concentrations according to the test, drawing a concentration-absorbance scattergram to establish a sample standard curve by taking the concentration of the sample as an abscissa and the absorbance as an ordinate, and calculating the concentration of the actual sample. The invention is widely applicable to samples which can be tested by a spectrophotometry method and is applicable to various types of smart phones, and the algorithm is simple to operate and high in accuracy.

Description

Image information extraction and calculation method and system for smartphone spectral detection

technical field

The invention relates to the technical field of spectral detection and image processing, in particular to an image information extraction and calculation method and system for smart phone spectral detection.

Background technique

Spectroscopic analysis can be used to determine the chemical composition and relative content of substances, and plays an important role in the fields of food safety, biosecurity, environmental monitoring, and healthcare. Large-scale spectrometers used in the laboratory are heavy and expensive, and cannot meet the needs of real-time, on-site detection of target samples. Therefore, portable spectrometers have continued to develop. Modern smartphones contain different sensor technologies and can be used in a wide range of fields as stand-alone measuring instruments. In spectral detection, the optical spectroscopic part can be integrated into an external device of the mobile phone, and it can be matched with the mobile phone, and the CMOS sensor of the smart phone can be used to convert the light signal passing through the sample into an electrical signal, which is displayed on the screen of the mobile phone by interpreting the electrical signal. Quantitative analysis of the sample to be tested can be realized by using the mobile phone software with a color quantification model to output the image.

At present, many scholars at home and abroad have conducted research on the color quantification model. Abbaspour et al. converted the RGB color values of the image into absorbance, and calculated the Fe ⁺² and Fe ³⁺ concentrations; Oncescu proposed to use the H value in the HSV model to represent color to detect biomarkers in sweat and saliva; Suzuki et al. The chromaticity coordinates of CIE XYZ were determined for Li+, NH4+ and protein. The results show that the concentration of the analyte tested by using the image color information is similar to the test result of the large-scale instrument, which has a strong feasibility. However, different color models have different ways of quantifying images. For substances with different color reactions, a specific color model needs to be used, so the above color model is difficult to be widely used.

Chinese patent document CN 107084790A discloses a spectrum detection method of a portable spectrometer based on a smart phone, including: 1) collecting the light signal to be measured; (2) performing collimation shaping and dispersion light separation on the light signal to be measured to form a sequence of wavelengths (3) photographing the dispersion fringes obtained in step (2) by a smartphone to form a color fringe picture arranged in order of wavelength; (4) acquiring each pixel position point of the color fringe picture obtained in step (3) and calculate the light intensity value I corresponding to each pixel position point to obtain an array I(x), where x is the coordinate of the pixel position point of the picture; (5) According to the wavelength-pixel position calibration data λ(x), the The x in the array I(x) is replaced with the corresponding λ to obtain the corresponding relationship between the wavelength and the light intensity I(λ), and the spectral curve corresponding to the data I(λ) is drawn to complete the spectral detection. The above color models are difficult to use widely. In addition, there are differences in the image display of different models of mobile phones. The pixel fixing method can be applied to the same model of mobile phones. However, when used on different models of mobile phones, there may be problems such as failure to load image information or excessive data errors. Therefore, it is very important to develop an image information extraction and calculation method that can be applied to different types of mobile phones.

SUMMARY OF THE INVENTION

In order to solve the above-mentioned technical problems, the present invention provides an image information extraction and calculation method and system for smartphone spectral detection, which improves the image processing accuracy of the mobile phone spectrometer and the compatibility of the APP on various types of mobile phones.

The technical scheme of the present invention is:

An image information extraction and calculation method for smartphone spectral detection, comprising the following steps:

S01: Obtain the RGB image of the spectral image, and rotate the image to a uniform angle;

S02: According to different samples, select the effective image area of the image;

S03: Extract the RGB values in the selected area, and convert the RGB values in the area into grayscale values;

S04: Calculate the absorbance of the sample through the image gray value inversion model, obtain the absorbance of standard samples with different concentrations according to the test, take the sample concentration as the abscissa and the absorbance as the ordinate, draw a concentration-absorbance scatter diagram to establish a sample standard curve, Calculate the concentration of the actual sample.

In a preferred technical solution, the effective image area of the image selected in the step S02 includes the following steps:

S21: Set a plurality of color block information greater than or equal to 2×2, obtain color blocks composed of consecutive color points that meet the conditions on the RGB image, and record the position information (x, y) of the color blocks;

S22: Take the union of the positions of all color blocks to obtain x _max , x _min , y _max , y _min , determine a rectangular area through these four points, and scale the coordinate size of the vertical direction y of the spectral diffraction on the basis of this rectangular area, The scaling formula is as follows:

(y ₁ ,y ₂ )=(y _max -(y _max -y _min )/a,y _min +(y _max -y _min )/a)

Among them, y ₁ and y ₂ are the coordinates after scaling; a is the set scaling factor, and a cannot be less than 2;

S23: The effective image area is a rectangular area determined by four points (x _max , y ₁ ), (x _min , y ₁ ), (x _max , y ₂ ), and (x _min , y ₂ ).

In a preferred technical solution, the formula for converting RGB values into gray values in step S03 is as follows:

Gray=(R _value +G _value +B _value )/3

Among them, Gray is the grayscale value, and R _value , G _value , and B _value are the values of each component of R, G, and B;

In a preferred technical solution, the obtained two-dimensional matrix of gray values is reduced to one dimension, and the average value of gray values in the vertical direction of spectral diffraction is calculated. The calculation formula of the average value of gray values is as follows:

为灰度平均值，n为y₁-y₂，i为y₂到y₁之间的坐标值；in,

is the average value of grayscale, n is y ₁ -y ₂ , i is the coordinate value between y ₂ and y ₁ ;

The gray value-pixel plot is plotted with the calculated one-dimensional matrix along the spectral diffraction direction.

In a preferred technical solution, in the step S04, the maximum gray value in the spectral image region generated without the sample is taken as the incident light intensity, and the gray maximum value in the spectral image region generated by the sample is taken as the outgoing light intensity, and the image gray scale The calculation formula of the value inversion model is as follows:

A=lg(1/T)=lg(gray ₁ /gray ₂ )

Among them, A is the absorbance, T is the transmittance, gray ₁ is the maximum gray value of the spectral image area generated without the sample, and gray ₂ is the maximum gray value of the spectral image area generated by the sample.

In a preferred technical solution, according to the least squares method in the step S04, a linear function curve is fitted as a sample standard curve, and the standard curve formula is as follows:

Y=aX+b

Among them, Y is the sample concentration, X is the absorbance of the sample, a is the fitted slope, and b is the fitted intercept.

The invention also discloses an image information extraction and calculation system for smart phone spectrum detection, comprising:

The spectral image processing module obtains the RGB image of the spectral image, and rotates the image to a uniform angle;

The effective image area extraction module selects the effective image area of the image according to different samples;

The conversion module extracts the RGB values in the selected area, and converts the RGB values in the area into grayscale values;

The sample standard curve building module, through the image gray value inversion model, calculate the absorbance of the sample, obtain the absorbance of standard samples with different concentrations according to the test, take the sample concentration as the abscissa and the absorbance as the ordinate, draw the concentration-absorbance scatter diagram to establish Sample standard curve to calculate the concentration of the actual sample.

In a preferred technical solution, the effective image area of the selected image includes the following steps:

S21: Set a plurality of color block information greater than or equal to 2×2, obtain color blocks composed of consecutive color points that meet the conditions on the RGB image, and record the position information (x, y) of the color blocks;

S22: Take the union of the positions of all color blocks to obtain x _max , x _min , y _max , y _min , determine a rectangular area through these four points, and scale the coordinate size of the vertical direction y of the spectral diffraction on the basis of this rectangular area, The scaling formula is as follows:

(y ₁ ,y ₂ )=(y _max -(y _max -y _min )/a,y _min +(y _max -y _min )/a)

Among them, y ₁ and y ₂ are the coordinates after scaling; a is the set scaling factor, and a cannot be less than 2;

S23: The effective image area is a rectangular area determined by four points (x _max , y ₁ ), (x _min , y ₁ ), (x _max , y ₂ ), and (x _min , y ₂ ).

In a preferred technical solution, the formula for converting the RGB values into grayscale values is as follows:

Gray=(R _value +G _value +B _value )/3

Among them, Gray is the grayscale value, and R _value , G _value , and B _value are the values of each component of R, G, and B;

In a preferred technical solution, the obtained two-dimensional matrix of gray values is reduced to one dimension, and the average value of gray values in the vertical direction of spectral diffraction is calculated. The calculation formula of the average value of gray values is as follows:

为灰度平均值，n为y₁-y₂，i为y₂到y₁之间的坐标值；in,

is the average value of grayscale, n is y ₁ -y ₂ , i is the coordinate value between y ₂ and y ₁ ;

The gray value-pixel plot is plotted with the calculated one-dimensional matrix along the spectral diffraction direction.

Compared with the prior art, the advantages of the present invention are:

According to different substances to be tested, the invention can frame and select the spectral image area to be used, which can reduce the calculation amount of the mobile phone, optimize the running speed of the APP, and improve the accuracy of the data in the subsequent calculation process. The invention is widely applicable to samples that can be tested by spectrophotometry, as well as to various types of smart phones, and the algorithm is simple to run and has high accuracy.

Description of drawings

Below in conjunction with accompanying drawing and embodiment, the present invention is further described:

Fig. 1 is the flow chart of the image information extraction and calculation method used for smart phone spectrum detection of the present invention;

Fig. 2 is the effective image area of frame selection;

FIG. 3 is a gray value-pixel curve diagram;

Fig. 4 is the gray value curve of different concentrations of ammonia nitrogen standard samples;

Figure 5 is the ammonia nitrogen standard curve.

Detailed ways

In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the specific embodiments and the accompanying drawings. It should be understood that these descriptions are exemplary only and are not intended to limit the scope of the invention. Also, in the following description, descriptions of well-known structures and techniques are omitted to avoid unnecessarily obscuring the concepts of the present invention.

Example:

The preferred embodiments of the present invention will be further described below with reference to the accompanying drawings.

This embodiment takes the ammonia nitrogen test as an example.

As shown in Figure 1, an image information extraction and calculation method for smartphone spectral detection, the algorithm of the present invention is simple to operate and has high accuracy, and can be run by a mobile phone with a general configuration, and of course other terminals such as PAD can also be used. The device includes identifying and obtaining the spectral image generated by the external device, adjusting the picture angle based on the placement angle of the mobile phone, obtaining the effective image area by setting the color block parameters, and finally obtaining the grayscale of the spectral diffraction direction image through the RGB color model. The value is brought into the established model to calculate the concentration of the analyte (sample). Specifically includes the following steps:

1) Project the image captured by the mobile phone onto a two-dimensional canvas to obtain an RGB image;

2) Retrieve the image rotation angle information from the mobile phone camera, and rotate the RGB image on the canvas to the uniformly set angle;

3) For different substances to be tested, the spectral image area to be used can be selected by box, which can reduce the calculation amount of the mobile phone, optimize the running speed of the APP, and improve the accuracy of the data in the subsequent calculation process. The core of this process is to use the tracking.js database to determine the position of the color block on the entire image based on the set color block parameters. The specific color block parameters can be set through a configuration file;

3.1) The configuration file obtains color blocks composed of consecutive color points that meet the conditions on the RGB image by setting multiple color block information (RGB values) greater than or equal to 2×2, and records the location information of these colors (x, y), x is the coordinate of the spectral diffraction direction, and y is the vertical coordinate of the spectral diffraction. Take the union of the positions of these color blocks to obtain x _max , x _min , y _max , y _min , and a rectangular area can be determined through these four points. Computational error due to the difference between the two sides and the middle part of the spectral image. The specific scaling formula is as follows:

(y ₁ ,y ₂ )=(y _max -(y _max -y _min )/a,y _min +(y _max -y _min )/a)

Among them, y ₁ and y ₂ are the coordinates after scaling; a is the scaling factor set according to the specific situation, the smaller a is, the larger the scaling factor is; the larger a is, the smaller the scaling factor is, and a cannot be less than 2;

3.2) (x _max , y ₁ ), (x _min , y ₁ ), (x _max , y ₂ ), (x _min , y ₂ ) are the four vertices of the rectangular image area that are finally involved in the calculation, as shown in Figure 2 Show.

4) converting the image information obtained in step (3) into digital information;

4.1) Extract the RGB values of the effective image area, and convert the RGB values in the area into grayscale values. There are many algorithms for RGB to grayscale conversion. After trying this method, the average algorithm with the best effect is selected. The specific formula of RGB to grayscale value is as follows:

Gray=(R _value +G _value +B _value )/3

Among them, Gray is the gray value, and R _value , G _value , and B _value are the values of each component of R, G, and B.

4.2) Reduce the obtained two-dimensional matrix of gray values to one dimension, calculate the average value of gray values in the vertical direction of spectral diffraction, y is the coordinate perpendicular to the direction of spectral diffraction, and the specific calculation formula of the average gray value is as follows:

为灰度平均值，n为y₁-y₂，i为y₂到y₁之间的坐标值；in,

is the average value of grayscale, n is y ₁ -y ₂ , i is the coordinate value between y ₂ and y ₁ ;

Use the calculated one-dimensional matrix along the spectral diffraction direction to make a gray value-pixel curve, as shown in Figure 3, this curve includes the gray average value within the above selection range;

5) Use the image gray value inversion model to establish the sample standard curve;

5.1) The image gray value inversion model simulates the Lambert-Beer law. In order to maximize the signal response value, the maximum gray value in the spectral image area generated without the sample is taken as the incident light intensity, and in the spectral image area generated by the sample. The maximum gray value is used as the outgoing light intensity, and the model calculation formula is as follows:

A=lg(1/T)=lg(gray ₁ /gray ₂ )

Among them, A is the absorbance, T is the transmittance, gray ₁ is the maximum gray value of the spectral image region generated without the sample, and gray ₂ is the maximum gray value of the spectral image region generated by the sample;

5.2) Test the absorbance of standard samples with different concentrations, as shown in Figure 4, where the black box is the effective area selected by the box, I0 is the gray value curve of the spectral image generated without the sample, 0-2 is the after Gray value curves of spectral images generated from standard samples with different concentrations.

Taking the sample concentration as the abscissa and the absorbance as the ordinate, draw a concentration-absorbance scattergram, and fit a linear function curve according to the least squares method, which is the standard curve of the sample, as shown in Figure 5. The standard curve formula is as follows:

Y=aX+b

Among them, Y is the sample concentration, X is the absorbance of the sample, a is the fitted slope, and b is the fitted intercept;

Substitute the absorbance value of the actual sample obtained by the test into the standard curve to calculate the concentration of the actual sample.

It should be understood that the above-mentioned specific embodiments of the present invention are only used to illustrate or explain the principle of the present invention, but not to limit the present invention. Therefore, any modifications, equivalent replacements, improvements, etc. made without departing from the spirit and scope of the present invention should be included within the protection scope of the present invention. Furthermore, the appended claims of this invention are intended to cover all changes and modifications that fall within the scope and boundaries of the appended claims, or the equivalents of such scope and boundaries.

Claims (8)

1. a kind of image information extraction and calculation method for smart phone spectrum detection, is characterized in that, comprises the following steps: S01: Obtain the RGB image of the spectral image, and rotate the image to a uniform angle; S02: According to different samples, select the effective image area of the image, including the following steps: S21: Set a plurality of color block information greater than or equal to 2×2, obtain color blocks composed of consecutive color points that meet the conditions on the RGB image, and record the position information (x, y) of the color blocks; S22: Take the union of the positions of all color blocks to obtain x _max , x _min , y _max , y _min , determine a rectangular area through these four points, and scale the coordinate size of the vertical direction y of the spectral diffraction on the basis of this rectangular area, The scaling formula is as follows: (y ₁ ,y ₂ )=(y _max -(y _max -y _min )/a,y _min +(y _max -y _min )/a) Among them, y ₁ and y ₂ are the coordinates after scaling; a is the set scaling factor, and a cannot be less than 2; S23: The effective image area is a rectangular area determined by four points (x _max , y ₁ ), (x _min , y ₁ ), (x _max , y ₂ ), and (x _min , y ₂ ); S03: Extract the RGB values in the selected area, and convert the RGB values in the area into grayscale values; S04: Calculate the absorbance of the sample through the image gray value inversion model, obtain the absorbance of standard samples with different concentrations according to the test, take the sample concentration as the abscissa and the absorbance as the ordinate, draw a concentration-absorbance scatter diagram to establish a sample standard curve, Calculate the concentration of the actual sample. 2. The image information extraction and calculation method used for smart phone spectrum detection according to claim 1, is characterized in that, in described step S03, the formula that RGB value is converted into gray value is as follows: Gray=(R _value +G _value +B _value )/3 Among them, Gray is the gray value, and R _value , G _value , and B _value are the values of each component of R, G, and B. 3. The image information extraction and calculation method for smartphone spectral detection according to claim 2, wherein the obtained two-dimensional matrix of grayscale values is reduced to one dimension, and the grayscale values in the vertical direction of spectral diffraction are calculated The calculation formula of the average value and the gray average value is as follows:

in,

is the average value of grayscale, n is y ₁ -y ₂ , i is the coordinate value between y ₂ and y ₁ ; The gray value-pixel plot is plotted with the calculated one-dimensional matrix along the spectral diffraction direction. 4. The image information extraction and calculation method for smartphone spectral detection according to claim 1, wherein in the step S04, the maximum gray value in the spectral image region not generated by the sample is used as the incident light intensity , the maximum gray value in the spectral image area generated by the sample is taken as the outgoing light intensity, and the calculation formula of the image gray value inversion model is as follows: A=lg(1/T)=lg(gray ₁ /gray ₂ ) Among them, A is the absorbance, T is the transmittance, gray ₁ is the maximum gray value of the spectral image area generated without the sample, and gray ₂ is the maximum gray value of the spectral image area generated by the sample. 5. The image information extraction and calculation method for smartphone spectral detection according to claim 1, wherein in the step S04, a linear function curve is fitted according to the least squares method, as a sample standard curve, The standard curve formula is as follows: Y=aX+b Among them, Y is the sample concentration, X is the absorbance of the sample, a is the fitted slope, and b is the fitted intercept. 6. An image information extraction and computing system for smart phone spectrum detection, characterized in that, comprising: The spectral image processing module obtains the RGB image of the spectral image, and rotates the image to a uniform angle; The effective image area extraction module, according to different samples, selects the effective image area of the image, including the following steps: S21: Set a plurality of color block information greater than or equal to 2×2, obtain color blocks composed of consecutive color points that meet the conditions on the RGB image, and record the position information (x, y) of the color blocks; S22: Take the union of the positions of all color blocks to obtain x _max , x _min , y _max , y _min , determine a rectangular area through these four points, and scale the coordinate size of the vertical direction y of the spectral diffraction on the basis of this rectangular area, The scaling formula is as follows: (y ₁ ,y ₂ )=(y _max -(y _max -y _min )/a,y _min +(y _max -y _min )/a) Among them, y ₁ and y ₂ are the coordinates after scaling; a is the set scaling factor, and a cannot be less than 2; S23: The effective image area is a rectangular area determined by four points (x _max , y ₁ ), (x _min , y ₁ ), (x _max , y ₂ ), and (x _min , y ₂ ); The conversion module extracts the RGB values in the selected area, and converts the RGB values in the area into grayscale values; The sample standard curve building module, through the image gray value inversion model, calculate the absorbance of the sample, obtain the absorbance of standard samples with different concentrations according to the test, take the sample concentration as the abscissa and the absorbance as the ordinate, draw the concentration-absorbance scatter diagram to establish Sample standard curve to calculate the concentration of the actual sample. 7. The image information extraction and computing system for smartphone spectral detection according to claim 6, wherein the formula that the RGB value is converted into a gray value is as follows: Gray=(R _value +G _value +B _value )/3 Among them, Gray is the gray value, and R _value , G _value , and B _value are the values of each component of R, G, and B. 8 . The image information extraction and calculation system for smartphone spectral detection according to claim 7 , wherein the obtained two-dimensional matrix of grayscale values is reduced to one dimension, and the grayscale values in the vertical direction of spectral diffraction are calculated. 9 . The calculation formula of the average value and the gray average value is as follows:

in,

is the average value of grayscale, n is y ₁ -y ₂ , i is the coordinate value between y ₂ and y ₁ ; The gray value-pixel plot is plotted with the calculated one-dimensional matrix along the spectral diffraction direction.

Images