CN101271522A - A method for automatic identification of yellow-grained rice in rice - Google Patents

Abstract

The invention discloses a method for automatic identification of yellow-grained rice, which belongs to the technical field of computer image processing. Take the original chromaticity information R, G, and B values of each grain of rice, and convert the R, G, and B chromaticity values into the uniform color model chromaticity value L ^* a ^* b ^* through the ideal color model XYZ; select the judgment feature color If the characteristic chromaticity value b ^* in all pixels of each grain of rice exceeds the chromaticity threshold and ^the proportion is greater than the area threshold, it is determined that the rice is yellow-grained rice. Among them, the chromaticity threshold and the area threshold are set by the operator according to the variety and origin of the rice. The method of the invention can be widely used in the detection process of rice quality in the on-site purchase of rice and market transactions, so that the detection is fast, objective and accurate.

Description

A method for automatic identification of yellow-grained rice in rice

technical field

The invention belongs to the technical field of computer image processing, and in particular relates to an automatic identification method for yellow-grained rice in rice.

Background technique

Yellow-grained rice refers to grains whose endosperm is yellow when brown rice is ground into first-class rice according to the national standard, which is obviously different in color from normal rice grains. The percentage of yellow-grained rice weight in the sample weight is called yellow-grained rice rate. The national standard GB/T 17891-1999 "High Quality Paddy" stipulates that the detection of yellow-grained rice is carried out according to the method specified in the national standard GB 1350-1999 "Paddy", but there is no yellow-grained rice detection in the national standard GB 1350-1999 "Paddy" In the description of the method, it is pointed out that the detection of yellow-grained rice is carried out according to the national standard GB 5496-1985 "Inspection of Yellow-grained Rice and Cracked Grains in Grain and Oil Seeds".

The national standard GB 5496-1985 "Inspection Method for Yellow-grained Rice and Cracked Grains of Grain and Oil Plants" stipulates that the detection method of yellow-grained rice is: after the roughness of the rice is tested, the brown rice sample is ground with a small rice mill to approximately For the accuracy of standard second-class rice, remove bran powder, weigh W, and use it as the sample weight, then pick out yellow-grained rice according to regulations, and weigh W ₁ . The content of yellow-grained rice in paddy is calculated according to the formula: yellow-grained rice (%)=(W ₁ /W)×100%. Among them, the identification method of yellow-grained rice is: take about 50g of rice samples or sort out the yellow-grained rice according to the regulations while inspecting the broken rice. The basis for picking out is completely based on human eyes. Due to the defects of long detection time, strong subjectivity, low accuracy, poor operability and repeatability, and being easily affected by the environment, this method is difficult to meet the requirements of rapid, objective and high-accuracy quality detection in on-site rice purchases and market transactions. requirements.

In view of the above problems, the present invention proposes a yellow-grained rice detection method based on the L ^* a ^* b ^* color model, and introduces computer image processing technology into rice quality evaluation, so that the detection technology is accurate, objective, operable, and highly automated.

Contents of the invention

The invention provides a method for automatic identification of yellow-grained rice based on the L ^* a ^* b ^* color model, which is characterized in that it comprises the following steps:

Obtain the image information of rice grains;

Analyze the color information of rice grains;

Identify yellow grain rice.

The described acquisition of rice grain image information specifically includes the following steps:

Place a certain number of rice samples in the image collector to collect the original image information;

Read the original image information, segment the background and rice grains, and set the background color to a color that is different from the color of the rice grains.

The quantity of the rice sample is 10-1000 grains.

The method for segmenting the background and rice grains is an iterative method.

The iterative method specifically includes the following steps:

Calculate the maximum and minimum grayscale values Z ₁ and Z _k in the image, and set the threshold initial value T _k =(Z ₁ +Z _k )/2;

According to the threshold T _k, the image is divided into two parts, the target and the background, and the average gray value Z ₀ and Z _B of the two parts are calculated;

Calculate the new threshold T _k+1 = (Z ₀ +Z _B )/2;

If T _k =T _k+1 , then the threshold value is obtained, otherwise, the threshold value is continuously calculated according to the calculated T _k value, and the calculation is iterative.

When the background and rice grains are divided, the background color is pure black, ie RGB (0, 0, 0).

The concrete steps of described analysis rice grain chromaticity information comprise:

Convert the original RGB chromaticity information into L ^* a ^* b ^* chromaticity information, where L ^* is the lightness index, the value range is 0 to 100, and a ^* is the chromaticity index, indicating that the color is from dark green to bright red, taking The value range is -120~120, b ^* is the chromaticity index, which means the color is from bright blue to burnt yellow, and the value range is -120~120;

Analyze the L ^* a ^* b ^* chromaticity information of each pixel in each grain of rice, and select b ^* in the L ^* a ^* b ^* chromaticity information as the characteristic chromaticity value for analyzing yellow-grained rice.

The characteristic chromaticity value b ^* in the L ^* a ^* b ^* color model is used to determine, and the specific conversion formula from the RGB color model to the L ^* a ^* b ^* color model through the ideal color model XYZ is:

X＝0.4124f(R/255)+0.3576f(G/255)+0.1805f(B/255)

Y＝0.2126f(R/255)+0.7152f(G/255)+0.0722f(B/255)

Z＝0.0193f(R/255)+0.1192f(G/255)+0.9505f(B/255)

Among them: R, G, B are the original chromaticity information values of the rice grains read,

function $f\left(r\right)=\left\{\begin{array}{c}{[(r+0.055)/1.055]}^{2.4},r>0.04045\\ r/12.92,r\≤0.04045\end{array}\right.$

The conversion formula from ideal XYZ color model to L ^* a ^* b ^* color model is:

L ^* ＝116f(Y/Y ₀ )-16

a ^* ＝500[f(X/X ₀ )-f(Y/Y ₀ )]

b ^* =200[f(Y/Y ₀ )-f(Z/Z ₀ )]

$\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="z"\; filenames="A20081011208400111.png"\; state="\{\{state\}\}">z</figure-callout></span>}}^{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; ></span>\mathrm{<span\; class="notranslate">\; 0.008856</span>}\\ \mathrm{<span\; class="notranslate">\; 7.787</span>}\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 16</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 116</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; 0.008856</span>}\end{array}\right.<span\; class="notranslate">\; .</span>$

The concrete steps of described identifying yellow-grained rice comprise:

Use the yellow-grained rice detection software to set the chromaticity threshold and area threshold for judging yellow-grained rice;

Based on b ^* in the L ^* a ^* b ^* color model, combined with the chromaticity threshold and area threshold, determine whether each grain of rice is yellow-grained rice, draw its boundary, and count the rate of yellow-grained rice.

When judging whether the rice is yellow-grained rice, the chromaticity threshold and the area threshold are set by the operator according to different rice varieties and production places, and the selection range is chromaticity threshold 15-30; area threshold 5%-40%.

The beneficial effect of the present invention is: the use of computer image processing technology to replace the human eye discrimination method stipulated in GB/T 17891-1999 "High Quality Rice", can quickly, objectively and accurately identify the yellow rice in the polished rice, overcome the The existing technical solutions have defects such as long detection time, strong subjectivity, low accuracy, poor operability and repeatability. It meets the requirements of rapid, objective and accurate quality inspection in the on-site purchase of rice and market transactions. Moreover, the computer image recognition system compiled according to the method of the present invention also has the function of simultaneously completing the detection of multiple indicators such as chalkiness, chalkiness grain rate, polished rice rate, grain type and liquid substance, so that the implementation of national The detection of multiple indicators that are independent of each other in the standard can be completed by a set of systems, and a maximum of 1,000 grains of rice can be detected each time. It has the characteristics of high degree of automation, fast and easy operation.

Description of drawings

Fig. 1 is a schematic diagram of the connection of the image acquisition and processing device of the present invention.

Labels in the figure:

1-scanner; 2-computer; 3-printer.

Detailed ways

The present invention will be further described below in conjunction with accompanying drawing:

Fig. 1 is the connection diagram of image acquisition and processing device of the present invention,

1. Use the counting board and sampling board to take about 10 to 1000 grains of rice samples from the batch of rice samples to be tested, place the sampler on the scanner 1, collect images, and store them as 24-bit bmp format files. The brightness and contrast of the scanner 1 are set between -15 and 25.

2. Utilize the yellow-grained rice detection system to identify the yellow-grained rice in the image on the computer 2, specifically:

a) read the original image information, store the chromaticity information of each pixel in each grain of rice, and its original chromaticity information is RGB color information;

b) Use the iterative method to segment the background and rice grains, and set the background to pure black, namely RGB (0, 0, 0),

The specific steps of the iterative method are:

(1) Calculate the maximum and minimum grayscale values Z ₁ and Z _k in the image, and set the threshold initial value T _k = (Z ₁ +Z _k )/2;

(2) Segment the image into two parts, target and background, according to the threshold T _k , and calculate the average gray value Z ₀ and Z _B of the two parts;

(3) Calculate the new threshold T _k+1 = (Z ₀ +Z _B )/2;

(4) If T _k =T _k+1 , the result is the threshold value, otherwise go to (2) for iterative calculation;

c) Convert the RGB color information of rice grains into L ^* a ^* b ^* color information, where L ^* is the lightness index, the value range is 0 to 100, and a ^* is the chromaticity index, indicating that the color is from dark green to bright red, The value range is -120~120, b ^* is the chromaticity index, which means the color is from bright blue to burnt yellow, and the value range is -120~120;

The specific conversion formula from the RGB color model to the L ^* a ^* b ^* color model through the ideal color model XYZ is:

X＝0.4124f(R/255)+0.3576f(G/255)+0.1805f(B/255)

Y＝0.2126f(R/255)+0.7152f(G/255)+0.0722f(B/255)

Z＝0.0193f(R/255)+0.1192f(G/255)+0.9505f(B/255)

Among them: R, G, B are the original chromaticity information values of the rice grains read,

function $f\left(r\right)=\left\{\begin{array}{c}{[(r+0.055)/1.055]}^{2.4},r>0.04045\\ r/12.92,r\&le;0.04045\end{array}\right.$

The conversion formula from ideal XYZ color model to L ^* a ^* b ^* color model is:

L ^* ＝116f(Y/Y ₀ )-16

a ^* ＝500[f(X/X ₀ )-f(Y/Y ₀ )]

b ^* =200[f(Y/Y ₀ )-f(Z/Z ₀ )]

$\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="z"\; filenames="A20081011208400111.png"\; state="\{\{state\}\}">z</figure-callout></span>}}^{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; ></span>\mathrm{<span\; class="notranslate">\; 0.008856</span>}\\ \mathrm{<span\; class="notranslate">\; 7.787</span>}\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 16</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 116</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; 0.008856</span>}\end{array}\right.<span\; class="notranslate">\; ;</span>$

d) analyze the L ^* a ^* b ^* chromaticity information of every grain of rice, and select the chromaticity value b ^* as the characteristic chromaticity value of analyzing yellow-grained rice;

e) Use the yellow-grained rice detection software to set the chromaticity threshold and area threshold. The threshold setting is set by the operator according to the origin and variety of the rice sample to be tested. The selection range is chromaticity threshold 15-30; area threshold 5% ~40%;

f) Based on the chromaticity value b ^* , combined with the chromaticity threshold and the area threshold, the specific judgment principle for judging whether each grain of rice is yellow-grained rice is: when the chromaticity value b ^* in each grain of rice is greater than the chromaticity threshold pixel If it exceeds the specified value of the area threshold, the grain of rice is yellow grain rice. And draw the outline of the rice grain with a yellow line, and count the yellow grain rice grain rate at the same time;

g) The detection result is output by the human-computer interaction interface. If you need to know the average chromaticity information of each grain of rice, or the chromaticity information of each pixel in each grain of rice, you can input it into EXCEL through the menu item of the detection interface for viewing;

h) Output the obtained result to the printer 3 on the computer 2 for printing.

The embodiments described above are only preferred specific implementations of the present invention, and those skilled in the art can make various modifications within the scope of the appended claims.

Claims (10)

1. an automatic identification method of yellow-grained rice in a kind of rice, it is characterized in that comprising the following steps: Obtain the image information of rice grains; Analyze the color information of rice grains; Identify yellow grain rice. 2. according to the automatic identification method of yellow grain rice in the described rice of claim 1, it is characterized in that, described obtaining grain of rice image information specifically comprises the following steps: Place the rice sample in the image collector to collect the original image information; Read the original image information, set the background color to a color different from the color of the rice grains, and separate the background and the rice grains. 3. according to the automatic identification method of yellow-grained rice in the described rice of claim 2, it is characterized in that, described rice sample quantity is 10-1000 grains. 4. according to the automatic recognition method of yellow grain rice in the described rice of claim 2, it is characterized in that, the method for described segmentation background and rice grain is an iterative method. 5. according to the automatic identification method of yellow grain rice in the described rice of claim 4, it is characterized in that, described iterative method specifically comprises the following steps: Calculate the maximum and minimum grayscale values Z ₁ and Z _k in the image, and set the threshold initial value T _k =(Z ₁ +Z _k )/2; According to the threshold T _k, the image is divided into two parts, the target and the background, and the average gray value Z ₀ and Z _B of the two parts are calculated; Calculate the new threshold T _k+1 = (Z ₀ +Z _B )/2; If T _k =T _k+1 , then the threshold value is obtained, otherwise, the threshold value is continuously calculated according to the calculated T _k value, and the calculation is iterative. 6. The method for automatic identification of yellow-grained rice according to claim 2, characterized in that, when said segmenting the background and rice grains, the background color is selected pure black, i.e. RGB (0,0,0). 7. according to the automatic identification method of yellow-grained rice in the described rice of claim 1, it is characterized in that, the concrete steps of described analysis rice grain chromaticity information comprise: Convert the original RGB chromaticity information into L ^* a ^* b ^* chromaticity information, where L ^* is the lightness index, the value range is 0 to 100, and a ^* is the chromaticity index, indicating that the color is from dark green to bright red, taking The value range is -120~120, b ^* is the chromaticity index, which means the color is from bright blue to burnt yellow, and the value range is -120~120; Analyze the L ^* a ^* b ^* chromaticity information of each pixel in each grain of rice, and select b ^* in the L ^* a ^* b ^* chromaticity information as the characteristic chromaticity value for analyzing yellow-grained rice. 8. according to the automatic identification method of yellow-grained rice in the described rice of claim 7, it is characterized in that, described utilizes the feature chromaticity value b ^* in the L ^* a ^* b ^* color model to judge, by RGB color model through ideal The specific conversion formula from color model XYZ to L ^* a ^* b ^* color model is: X＝0.4124f(R/255)+0.3576f(G/255)+0.1805f(B/255) Y＝0.2126f(R/255)+0.7152f(G/255)+0.0722f(B/255) Z＝0.0193f(R/255)+0.1192f(G/255)+0.9505f(B/255) Among them: R, G, B are the original chromaticity information values of the rice grains read, function $f\left(r\right)=\left\{\begin{array}{c}{[(r+0.055)/1.055]}^{2.4},r>0.04045\\ r/12.92,r\&le;0.04045\end{array}\right.$ The conversion formula from ideal XYZ color model to L ^* a ^* b ^* color model is: L ^* ＝116f(Y/Y ₀ )-16 a ^* ＝500[f(X/X ₀ )-f(Y/Y ₀ )] b ^* =200[f(Y/Y ₀ )-f(Z/Z ₀ )] $\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; z</span>}}^{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; ></span>\mathrm{<span\; class="notranslate">\; 0.008856</span>}\\ \mathrm{<span\; class="notranslate">\; 7.787</span>}\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 16</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 116</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; 0.008856</span>}\end{array}\right.<span\; class="notranslate">\; .</span>$ 9. according to the automatic identification method of yellow-grained rice in the described rice of claim 1, it is characterized in that, the concrete steps of described identification yellow-grained rice comprise: Use the yellow-grained rice detection software to set the chromaticity threshold and area threshold for judging yellow-grained rice; Based on b ^* in the L ^* a ^* b ^* color model, combined with the chromaticity threshold and area threshold, determine whether each grain of rice is yellow-grained rice, draw its boundary, and count the rate of yellow-grained rice. 10. according to the automatic identification method of yellow-grained rice in the described rice of claim 9, it is characterized in that when said judging whether rice is yellow-grained rice, chromaticity threshold value and area threshold value are voluntarily determined by the operator according to different rice varieties and production places. Setting, the selection range is chroma threshold 15-30; area threshold 5%-40%.

Images