Flue-cured tobacco leaf raw material appearance quality evaluation system based on machine vision

Abstract

The invention relates to the technical field of flue-cured tobacco leaf raw material appearance screening, and discloses a flue-cured tobacco leaf raw material appearance quality evaluation system based on machine vision. The flue-cured tobacco leaf raw material appearance quality evaluation system comprises a flue-cured tobacco leaf appearance data acquisition module, a flue-cured tobacco leaf appearance data analysis module and a flue-cured tobacco leaf raw material appearance quality evaluation module. According to the system, the flue-cured tobacco leaf appearance data which is currently conveyed is collected, preprocessing is carried out based on the flue-cured tobacco leaf appearance data, the preprocessed tobacco leaf color, the tobacco leaf area and the tobacco leaf damage degree are obtained, and meanwhile the tobacco leaf color, the tobacco leaf area and the tobacco leaf damage degree are obtained; the color evaluation score of the flue-cured tobacco leaves, the area evaluation score of the flue-cured tobacco leaves and the damage degree evaluation score of the flue-cured tobacco leaves are obtained and are integrated to obtain the appearance quality score of the flue-cured tobacco leaf raw materials, so that the problem that the quality of different samples cannot be intuitively judged when the appearance quality is close or different quality factors are poor is avoided; therefore, the appearance quality of the flue-cured tobacco leaf raw material is objectively evaluated.

Description

Technical Field

[0001] The invention relates to the technical field of flue-cured tobacco leaf raw material appearance screening, and in particular to a flue-cured tobacco leaf raw material appearance quality evaluation system based on machine vision. Background Art

[0002] The appearance quality of tobacco leaves refers to the external characteristics of tobacco leaves that can be felt and identified by human senses, that is, by touching and observing, the grade and quality of tobacco leaves are ultimately determined. Currently, the existing technology still relies on senses and experience to determine. The appearance characteristics of tobacco leaves are closely related to their internal quality. It is generally believed that the appearance characteristics of high-quality tobacco leaves are good maturity, orange-yellow color, loose leaf structure, moderate thickness, good oil content, and strong color.

[0003] The qualitative description method can intuitively describe the characteristics of a certain factor of the appearance quality of tobacco leaves, but it has great flexibility and cannot intuitively judge the quality between different samples when the appearance quality is similar or different quality factors are good or bad. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In view of the deficiencies in the prior art, the present invention provides a flue-cured tobacco leaf raw material appearance quality evaluation system based on machine vision, which has the advantages of objectively evaluating the appearance quality of flue-cured tobacco leaf raw materials and solves the above-mentioned technical problems.

[0006] (II) Technical solution

[0007] To achieve the above object, the present invention provides the following technical solutions: a flue-cured tobacco leaf raw material appearance quality evaluation system based on machine vision, comprising a flue-cured tobacco leaf appearance data acquisition module, a flue-cured tobacco leaf appearance data analysis module and a flue-cured tobacco leaf raw material appearance quality evaluation module;

[0008] The flue-cured tobacco leaf appearance data acquisition module includes a flue-cured tobacco leaf appearance data acquisition unit and a flue-cured tobacco leaf appearance data preprocessing unit, the flue-cured tobacco leaf appearance data acquisition unit is used to collect the flue-cured tobacco leaf appearance data currently conveyed, and send the flue-cured tobacco leaf appearance data to the flue-cured tobacco leaf appearance data preprocessing unit, the flue-cured tobacco leaf appearance data preprocessing unit performs preliminary processing on the received flue-cured tobacco leaf appearance data to obtain preprocessed flue-cured tobacco leaf appearance data, the flue-cured tobacco leaf appearance data preprocessing unit sends the preprocessed flue-cured tobacco leaf appearance data to the flue-cured tobacco leaf appearance data processing module, and the preprocessed flue-cured tobacco leaf appearance data includes tobacco leaf color, tobacco leaf area and tobacco leaf damage degree;

[0009] The flue-cured tobacco leaf appearance data processing module includes a tobacco leaf color analysis unit, a tobacco leaf area analysis unit and a tobacco leaf damage degree analysis unit. The tobacco leaf color analysis unit performs analysis and calculation based on the tobacco leaf color in the flue-cured tobacco leaf appearance data, and obtains a color evaluation score of the flue-cured tobacco leaf. The tobacco leaf area analysis unit performs analysis based on the tobacco leaf area in the flue-cured tobacco leaf appearance data, and obtains an area evaluation score of the flue-cured tobacco leaf. The tobacco leaf damage degree analysis unit performs calculation and analysis based on the tobacco leaf damage degree in the flue-cured tobacco leaf appearance data, and obtains a tobacco leaf damage degree evaluation score. The flue-cured tobacco leaf appearance data processing module sends the calculated tobacco leaf color evaluation score, tobacco leaf area evaluation score and tobacco leaf damage degree evaluation score to the flue-cured tobacco leaf raw material appearance quality evaluation module;

[0010] The flue-cured tobacco leaf raw material appearance quality evaluation module finally calculates the flue-cured tobacco leaf raw material appearance quality score based on the color evaluation score of the flue-cured tobacco leaves, the area evaluation score of the flue-cured tobacco leaves and the tobacco leaf damage degree evaluation score.

[0011] As a preferred technical solution of the present invention, the specific steps of the flue-cured tobacco leaf appearance data preprocessing unit performing preliminary processing on the received flue-cured tobacco leaf appearance data are as follows:

[0012] Step A1: The tobacco leaf appearance data acquisition unit acquires multiple tobacco leaf appearance data of the same group of tobacco leaves, and uses the Tenengrad gradient method to select the clarity of all images [QXD1, ..., QXD i ,…,QXD I ] and execute step A2;

[0013] Among them, QXD1, ..., QXD i ,·…,QXD I Respectively represent the clarity of the first image, ..., the clarity of the ith image, ..., the clarity of the Ith image calculated by the Tenengrad gradient method from multiple flue-cured tobacco leaf appearance data of the same group of flue-cured tobacco leaves collected;

[0014] Step A2: cropping the first J images with the highest definition among all the selected images, cropping the image background to obtain a cropped image, and performing area calculation on the cropped image to obtain the tobacco leaf area, tobacco leaf color and leaf damage degree of the flue-cured tobacco leaves.

[0015] As a preferred technical solution of the present invention, the specific steps of calculating the tobacco leaf color in step A2 are as follows:

[0016] Step A2.1a: Output the color histogram H(j) of the jth cropped image = [h r (j), h g (j), hb (j)];

[0017] Among them, h r (j), h g (j), h b (j) represents the histogram of the red, green and blue channels of the jth cropped image, respectively;

[0018] Step A2.2a: Calculate the color histogram of the cropped image, and finally obtain the color histogram H of the cropped image = [h r ,h g ,h b ];

[0019] Among them, h r ,h g ,h b Represents the histograms of the red, green, and blue channels of the cropped image, respectively.

[0020] As a preferred technical solution of the present invention, the color histogram of the cropped image is calculated in step A2.2a, and finally the color histogram H of the cropped image is obtained. r ,h g ,h b The specific expression of ] is as follows:

[0021]

[0022] Where x=r, g, b represents the index of r, g, b, h x represents the histogram of channel x, j∈J represents the jth image among the top J images with the highest clarity.

[0023] As a preferred technical solution of the present invention, the specific expression for obtaining the tobacco leaf area in step A2 is as follows:

[0024]

[0025] Among them, S represents the tobacco leaf area of ​​the cropped image, represents the summation, S j represents the j-th tobacco leaf area.

[0026] As a preferred technical solution of the present invention, the specific expression of the blade damage degree is as follows:

[0027]

[0028] Among them, YOPS represents the degree of leaf breakage, S represents the tobacco leaf area of ​​the cropped image, and PS j，k represents the damaged area of ​​the kth region on the jth flue-cured tobacco leaf, K represents the total damaged area, It means to sum all damaged areas in J images.

[0029] As a preferred technical solution of the present invention, the tobacco leaf color analysis unit performs analysis and calculation based on the tobacco leaf color in the flue-cured tobacco leaf appearance data, and obtains a specific expression of the color evaluation score of the flue-cured tobacco leaf as follows:

[0030] YSPF=1_DB[H,H0]

[0031] Among them, DB[H, H0] represents the Bhattacharyya distance between H and H0, H represents the color histogram of the cropped image, H0 represents the color histogram of the standard flue-cured tobacco leaves, and YSPF represents the color evaluation score of the flue-cured tobacco leaves.

[0032] As a preferred technical solution of the present invention, the tobacco leaf area analysis unit analyzes the tobacco leaf area in the flue-cured tobacco leaf appearance data, and obtains the specific expression of the area evaluation score of the flue-cured tobacco leaf as follows:

[0033]

[0034] Among them, MJPF represents the area evaluation score of flue-cured tobacco leaves, S0 represents the area of ​​standard flue-cured tobacco leaves, and S represents the area of ​​tobacco leaves in the cropped image.

[0035] As a preferred technical solution of the present invention, the tobacco leaf damage degree analysis unit performs calculation and analysis based on the tobacco leaf damage degree in the flue-cured tobacco leaf appearance data, and obtains a specific expression for the tobacco leaf damage degree evaluation score as follows:

[0036] PSPF=1-YOPS

[0037] Among them, YOPS represents the degree of leaf breakage, and PSPF represents the assessment score of the degree of tobacco leaf damage.

[0038] As a preferred technical solution of the present invention, the specific expression of the appearance quality score of the flue-cured tobacco leaf raw material is finally calculated by the flue-cured tobacco leaf raw material appearance quality evaluation module based on the color evaluation score of the flue-cured tobacco leaf, the area evaluation score of the flue-cured tobacco leaf and the tobacco leaf damage degree evaluation score as follows:

[0039] ZLPF=θ1*PSPF+θ2*MJPF+θ3*YSPF

[0040] Among them, PSPF represents the evaluation score of tobacco leaf damage degree, MJPF represents the area evaluation score of flue-cured tobacco leaves, YSPF represents the color evaluation score of flue-cured tobacco leaves, ZLPF represents the appearance quality score of flue-cured tobacco leaf raw materials, and θ1, θ2, and θ3 represent three different weight coefficients respectively.

[0041] Compared with the prior art, the present invention provides a system for evaluating the appearance quality of flue-cured tobacco leaf raw materials based on machine vision, which has the following beneficial effects:

[0042] The present invention collects the appearance data of the flue-cured tobacco leaves currently transported, and performs preprocessing based on the appearance data of the flue-cured tobacco leaves to obtain the color, area and damage degree of the preprocessed tobacco leaves. At the same time, the color evaluation score, area evaluation score and damage degree evaluation score of the flue-cured tobacco leaves are obtained based on the color, area and damage degree of the preprocessed tobacco leaves. The appearance quality score of the flue-cured tobacco leaf raw materials is obtained by combining the three, thereby avoiding the inability to intuitively judge the quality of different samples when the appearance qualities are similar or different quality factors are good or bad, thereby objectively evaluating the appearance quality of the flue-cured tobacco leaf raw materials. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 Schematic diagram of the system framework of the present invention. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] See also Figure 1 , a flue-cured tobacco leaf raw material appearance quality evaluation system based on machine vision, including a flue-cured tobacco leaf appearance data acquisition module, a flue-cured tobacco leaf appearance data analysis module and a flue-cured tobacco leaf raw material appearance quality evaluation module;

[0046] The flue-cured tobacco leaf appearance data acquisition module includes a flue-cured tobacco leaf appearance data acquisition unit and a flue-cured tobacco leaf appearance data preprocessing unit. The flue-cured tobacco leaf appearance data acquisition unit is used to collect the appearance data of the flue-cured tobacco leaves currently conveyed, and send the appearance data of the flue-cured tobacco leaves to the flue-cured tobacco leaf appearance data preprocessing unit. The flue-cured tobacco leaf appearance data preprocessing unit performs preliminary processing on the received flue-cured tobacco leaf appearance data to obtain the preprocessed flue-cured tobacco leaf appearance data. The flue-cured tobacco leaf appearance data preprocessing unit sends the preprocessed flue-cured tobacco leaf appearance data to the flue-cured tobacco leaf appearance data processing module. The preprocessed flue-cured tobacco leaf appearance data includes tobacco leaf color, tobacco leaf area and tobacco leaf damage degree. The specific steps of the flue-cured tobacco leaf appearance data preprocessing unit performing preliminary processing on the received flue-cured tobacco leaf appearance data are as follows:

[0047] Step A1: The tobacco leaf appearance data acquisition unit acquires multiple tobacco leaf appearance data of the same group of tobacco leaves, and uses the Tenengrad gradient method to select the clarity of all images [QXD1, ..., QXD i ,…,QXD I ] and execute step A2;

[0048] Among them, QXD1, ..., QXD i ,…,QXD I Respectively represent the clarity of the first image, ..., the clarity of the ith image, ..., the clarity of the Ith image calculated by the Tenengrad gradient method from multiple flue-cured tobacco leaf appearance data of the same group of flue-cured tobacco leaves collected;

[0049] The Tenengrad gradient method is a method for calculating the brightness gradient of an image. It determines the clarity of an image by analyzing the brightness changes of pixels in the image. The following is a detailed introduction to the Tenengrad gradient method:

[0050] Local variance calculation: The core of the Tenengrad gradient method is to calculate the local variance of each pixel in the image. The local variance reflects the intensity of the brightness change in the area around a pixel.

[0051] Gradient size determination: Based on the calculated local variance, the gradient size of each pixel can be determined. Usually, the square root or absolute value of the gradient is used as the gradient size value.

[0052] Then grayscale conversion is performed: the original color image is converted into a grayscale image to simplify calculation and better reflect brightness changes. Local variance calculation: for each pixel, a fixed-size window is selected and the variance of the pixel values ​​in the window is calculated. Gradient magnitude calculation: the gradient magnitude of each pixel is calculated based on the local variance and converted to a grayscale value. Result output: the calculated gradient image is used as output, in which the grayscale value of each pixel represents the gradient magnitude of the point.

[0053] Step A2: crop the first J images with the highest definition among all the selected images, crop the image background to obtain a cropped image, calculate the area of ​​the cropped image to obtain the tobacco leaf area, tobacco leaf color and leaf damage degree of the flue-cured tobacco leaf. The specific steps of calculating the tobacco leaf color in step A2 are as follows:

[0054] Step A2.1a: Output the color histogram H(j) of the jth cropped image = [h r (j), h g (j), h b (j)];

[0055] Among them, h r (j), h g (j), h b (j) represents the histogram of the red, green and blue channels of the jth cropped image, respectively;

[0056] Step A2.2a: Calculate the color histogram of the cropped image, and finally obtain the color histogram H of the cropped image = [h r ,h g ,h b ], the color histogram of the cropped image is calculated in step A2.2a, and finally the color histogram of the cropped image is obtained. r ,h g ,h b The specific expression of ] is as follows:

[0057]

[0058] Where x=r, g, b represents the index of r, g, b, h x represents the histogram of channel x, j∈J represents the jth image among the top J images with the highest clarity, h r ,h g ,h b Represents the histograms of the red, green, and blue channels of the cropped image, respectively.

[0059] The specific expression for the tobacco leaf area obtained in step A2 is as follows:

[0060]

[0061] Among them, S represents the tobacco leaf area of ​​the cropped image, represents the summation, S j represents the jth tobacco leaf area. The specific expression of the leaf damage degree obtained in step A2 is as follows:

[0062]

[0063] Among them, YOPS represents the degree of leaf breakage, S represents the tobacco leaf area of ​​the cropped image, and PS j，k represents the damaged area of ​​the kth region on the jth flue-cured tobacco leaf, K represents the total damaged area, It means to sum all damaged areas in the images;

[0064] The tobacco leaf appearance data processing module includes a tobacco leaf color analysis unit, a tobacco leaf area analysis unit and a tobacco leaf damage degree analysis unit. The tobacco leaf color analysis unit analyzes and calculates based on the tobacco leaf color in the tobacco leaf appearance data, and obtains a color evaluation score of the tobacco leaf. The tobacco leaf color analysis unit analyzes and calculates based on the tobacco leaf color in the tobacco leaf appearance data, and obtains a specific expression of the color evaluation score of the tobacco leaf as follows:

[0065] YSPF=1-DB[H,H0]

[0066] Among them, DB[H, H0] represents the Bhattacharyya distance between H and H0, H represents the color histogram of the cropped image, H0 represents the color histogram of the standard flue-cured tobacco leaves, YSPF represents the color evaluation score of the flue-cured tobacco leaves, the tobacco leaf area analysis unit analyzes the tobacco leaf area in the flue-cured tobacco leaf appearance data to obtain and obtain the area evaluation score of the flue-cured tobacco leaves, and the tobacco leaf area analysis unit analyzes the tobacco leaf area in the flue-cured tobacco leaf appearance data to obtain and obtain the specific expression of the area evaluation score of the flue-cured tobacco leaves as follows:

[0067]

[0068] Among them, MJPF represents the area evaluation score of flue-cured tobacco leaves, S0 represents the area of ​​standard flue-cured tobacco leaves, S represents the area of ​​tobacco leaves in the cropped image, and the tobacco leaf damage degree analysis unit calculates and analyzes the tobacco leaf damage degree in the flue-cured tobacco leaf appearance data to obtain the tobacco leaf damage degree evaluation score. The tobacco leaf damage degree analysis unit calculates and analyzes the tobacco leaf damage degree in the flue-cured tobacco leaf appearance data to obtain the specific expression of the tobacco leaf damage degree evaluation score as follows:

[0069] PSPF=1-YOPS

[0070] Among them, YOPS represents the degree of leaf breakage, PSPF represents the evaluation score of tobacco leaf damage, and the flue-cured tobacco leaf appearance data processing module sends the calculated color evaluation score of the flue-cured tobacco leaf, the area evaluation score of the flue-cured tobacco leaf, and the tobacco leaf damage evaluation score to the flue-cured tobacco leaf raw material appearance quality evaluation module;

[0071] The appearance quality evaluation module of flue-cured tobacco leaf raw materials finally calculates the appearance quality score of flue-cured tobacco leaf raw materials based on the color evaluation score of flue-cured tobacco leaf, the area evaluation score of flue-cured tobacco leaf and the degree of damage evaluation score of tobacco leaf. The specific expression of the appearance quality score of flue-cured tobacco leaf raw materials finally calculated by the appearance quality evaluation module of flue-cured tobacco leaf raw materials based on the color evaluation score of flue-cured tobacco leaf, the area evaluation score of flue-cured tobacco leaf and the degree of damage evaluation score of tobacco leaf is as follows:

[0072] ZLPF=θ1*PSPF+θ2*MJPF+θ3*YSPF

[0073] Among them, PSPF represents the evaluation score of tobacco leaf damage, MJPF represents the area evaluation score of flue-cured tobacco leaves, YSPF represents the color evaluation score of flue-cured tobacco leaves, ZLPF represents the appearance quality score of flue-cured tobacco leaf raw materials, θ1, θ2, and θ3 represent three different weight coefficients, which can be adjusted according to production needs. If the degree of tobacco leaf damage is required to be as low as possible, θ1 needs to be increased. If the color of the tobacco leaves needs to be as consistent as possible with the standard library, θ3 needs to be increased. If the area of ​​the tobacco leaves needs to be as large as possible, θ2 needs to be increased. Therefore, the appearance quality score of the flue-cured tobacco leaf raw materials can be adjusted according to production needs, so as to ensure The comprehensiveness of the appearance quality scoring of flue-cured tobacco leaf raw materials is verified. By collecting the appearance data of the flue-cured tobacco leaves currently conveyed and preprocessing them based on the appearance data of the flue-cured tobacco leaves, the color, area and degree of damage of the pretreated tobacco leaves are obtained. At the same time, the color evaluation score, area evaluation score and degree of damage of the pretreated tobacco leaves are obtained, and the appearance quality score of the flue-cured tobacco leaf raw materials are obtained by combining the three. This avoids the inability to intuitively judge the quality of different samples when the appearance quality is close or different quality factors are good or bad, thereby objectively evaluating the appearance quality of the flue-cured tobacco leaf raw materials.

[0074] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A system for evaluating the appearance quality of flue-cured tobacco leaves based on machine vision, characterized in that: It includes a flue-cured tobacco leaf appearance data collection module, a flue-cured tobacco leaf appearance data analysis module and a flue-cured tobacco leaf raw material appearance quality evaluation module; The flue-cured tobacco leaf appearance data acquisition module includes a flue-cured tobacco leaf appearance data acquisition unit and a flue-cured tobacco leaf appearance data preprocessing unit, the flue-cured tobacco leaf appearance data acquisition unit is used to collect the flue-cured tobacco leaf appearance data currently conveyed, and send the flue-cured tobacco leaf appearance data to the flue-cured tobacco leaf appearance data preprocessing unit, the flue-cured tobacco leaf appearance data preprocessing unit performs preliminary processing on the received flue-cured tobacco leaf appearance data to obtain preprocessed flue-cured tobacco leaf appearance data, the flue-cured tobacco leaf appearance data preprocessing unit sends the preprocessed flue-cured tobacco leaf appearance data to the flue-cured tobacco leaf appearance data processing module, and the preprocessed flue-cured tobacco leaf appearance data includes tobacco leaf color, tobacco leaf area and tobacco leaf damage degree; The flue-cured tobacco leaf appearance data processing module includes a tobacco leaf color analysis unit, a tobacco leaf area analysis unit and a tobacco leaf damage degree analysis unit. The tobacco leaf color analysis unit performs analysis and calculation based on the tobacco leaf color in the flue-cured tobacco leaf appearance data, and obtains a color evaluation score of the flue-cured tobacco leaf. The tobacco leaf area analysis unit performs analysis based on the tobacco leaf area in the flue-cured tobacco leaf appearance data, and obtains an area evaluation score of the flue-cured tobacco leaf. The tobacco leaf damage degree analysis unit performs calculation and analysis based on the tobacco leaf damage degree in the flue-cured tobacco leaf appearance data, and obtains a tobacco leaf damage degree evaluation score. The flue-cured tobacco leaf appearance data processing module sends the calculated tobacco leaf color evaluation score, tobacco leaf area evaluation score and tobacco leaf damage degree evaluation score to the flue-cured tobacco leaf raw material appearance quality evaluation module; The flue-cured tobacco leaf raw material appearance quality evaluation module finally calculates the flue-cured tobacco leaf raw material appearance quality score based on the color evaluation score of the flue-cured tobacco leaves, the area evaluation score of the flue-cured tobacco leaves and the tobacco leaf damage degree evaluation score.

2. The machine vision-based flue-cured tobacco leaf raw material appearance quality evaluation system according to claim 1, characterized in that: The specific steps of the flue-cured tobacco leaf appearance data preprocessing unit performing preliminary processing on the received flue-cured tobacco leaf appearance data are as follows: Step A1: The tobacco leaf appearance data acquisition unit acquires multiple tobacco leaf appearance data of the same group of tobacco leaves, and uses the Tenengrad gradient method to select the clarity of all images [QXD1,…,QXD i ,…,QXD I ] and execute step A2; Among them, QXD1,…,QXD i ,…,QXD I Respectively represent the clarity of the first image, ..., the clarity of the i-th image, ..., the clarity of the I-th image calculated by the Tenengrad gradient method from multiple flue-cured tobacco leaf appearance data of the same group of flue-cured tobacco leaves collected; Step A2: cropping the first J images with the highest definition among all the selected images, cropping the image background to obtain a cropped image, and performing area calculation on the cropped image to obtain the tobacco leaf area, tobacco leaf color and leaf damage degree of the flue-cured tobacco leaves.

3. The machine vision-based flue-cured tobacco leaf raw material appearance quality evaluation system according to claim 2, characterized in that: The specific steps of calculating the tobacco leaf color in step A2 are as follows: Step A2.1a: Output the color histogram H(j) of the jth cropped image = [h r (j),h g (j),h b (j)]; Among them, h r (j),h g (j),h b (j) represents the histogram of the red, green and blue channels of the jth cropped image, respectively; Step A2.2a: Calculate the color histogram of the cropped image, and finally obtain the color histogram H of the cropped image = [h r ,h g ,h b ]; Among them, h r ,h g ,h b Represents the histograms of the red, green, and blue channels of the cropped image, respectively.

4. The machine vision-based flue-cured tobacco leaf raw material appearance quality evaluation system according to claim 3, characterized in that: In step A2.2a, the color histogram of the cropped image is calculated, and finally the color histogram H of the cropped image is obtained. r ,h g ,h b The specific expression of ] is as follows: Where x=r,g,b represents the index of r,g,b, and h x represents the histogram of channel x, j∈J represents the jth image among the top J images with the highest clarity.

5. The machine vision-based flue-cured tobacco leaf raw material appearance quality evaluation system according to claim 3, characterized in that: The specific expression for the tobacco leaf area obtained in step A2 is as follows: Among them, S represents the tobacco leaf area of ​​the cropped image, represents the summation, S j represents the j-th tobacco leaf area.

6. The machine vision-based flue-cured tobacco leaf raw material appearance quality evaluation system according to claim 5, characterized in that: The specific expression of the blade damage degree is as follows: Among them, YOPS represents the degree of leaf breakage, S represents the tobacco leaf area of ​​the cropped image, and PS j,k represents the damaged area of ​​the kth region on the jth flue-cured tobacco leaf, K represents the total damaged area, It means to sum all damaged areas in J images.

7. The machine vision-based flue-cured tobacco leaf raw material appearance quality evaluation system according to claim 6, characterized in that: The tobacco leaf color analysis unit performs analysis and calculation based on the tobacco leaf color in the flue-cured tobacco leaf appearance data, and obtains a specific expression for the color evaluation score of the flue-cured tobacco leaf as follows: YSPF=1-DB[H,H0] Among them, DB[H,H0] represents the Bhattacharyya distance between H and H0, H represents the color histogram of the cropped image, H0 represents the color histogram of the standard flue-cured tobacco leaves, and YSPF represents the color evaluation score of the flue-cured tobacco leaves.

8. The machine vision-based flue-cured tobacco leaf raw material appearance quality evaluation system according to claim 7, characterized in that: The tobacco leaf area analysis unit analyzes the tobacco leaf area in the flue-cured tobacco leaf appearance data to obtain a specific expression for the area evaluation score of the flue-cured tobacco leaf as follows: Among them, MJPF represents the area evaluation score of flue-cured tobacco leaves, S0 represents the area of ​​standard flue-cured tobacco leaves, and S represents the area of ​​tobacco leaves in the cropped image.

9. The machine vision-based flue-cured tobacco leaf raw material appearance quality evaluation system according to claim 8, characterized in that: The tobacco leaf damage degree analysis unit performs calculation and analysis based on the tobacco leaf damage degree in the flue-cured tobacco leaf appearance data, and obtains a specific expression for the tobacco leaf damage degree evaluation score as follows: PSPF=1-YOPS Among them, YOPS represents the degree of leaf breakage, and PSPF represents the assessment score of the degree of tobacco leaf damage.

10. The machine vision-based flue-cured tobacco leaf raw material appearance quality evaluation system according to claim 9, characterized in that: The specific expression of the appearance quality score of the flue-cured tobacco leaf raw material calculated by the flue-cured tobacco leaf raw material appearance quality evaluation module based on the color evaluation score of the flue-cured tobacco leaf, the area evaluation score of the flue-cured tobacco leaf, and the tobacco leaf damage degree evaluation score is as follows: ZLPF=θ1*PSPF+θ2*MJPF+θ3*YSPF Among them, PSPF represents the evaluation score of tobacco leaf damage degree, MJPF represents the area evaluation score of flue-cured tobacco leaves, YSPF represents the color evaluation score of flue-cured tobacco leaves, ZLPF represents the appearance quality score of flue-cured tobacco leaf raw materials, and θ1, θ2, and θ3 represent three different weight coefficients respectively.

Images