A method for identifying and calculating the unfolding rate of tobacco leaves after loose re-damping

Abstract

The present invention relates to the technical field of tobacco leaf unfolding rate recognition, and particularly to a method for recognizing and calculating the unfolding rate of tobacco leaves after loose and moistening. Its technical solution includes collecting the transmissive image of tobacco leaves after loose and moistening through an industrial camera, and adjusting the camera focal length, exposure time and light source angle; performing leaf segmentation processing on the transmissive image, including: converting the original image into a grayscale image, separating the tobacco leaf area from the background through binarization processing, and extracting the edge of the tobacco leaf using the Robert operator. The present invention effectively eliminates background interference and retains the complete leaf contour through grayscale background removal, multi-operator edge detection, and morphological filtering image processing techniques; combined with an adaptive threshold segmentation algorithm, it realizes the accurate recognition of the transmissive area of tobacco leaves, and finally obtains the unfolding rate parameter through area ratio calculation, overcoming the deficiencies of traditional detection techniques and providing a reliable basis for the intelligent control of the tobacco leaf processing process.

Description

Technical Field

[0001] The present invention relates to the technical field of tobacco leaf unfolding rate recognition, and particularly to a method for recognizing and calculating the unfolding rate of tobacco leaves after loose moisture regain. Background Art

[0002] In the process of tobacco leaf processing, the loose moisture regain process is a key link in tobacco leaf re-drying. Its main purpose is to adjust the moisture and temperature of tobacco leaves to make the tobacco leaves unfold and reach a suitable processing state. As an important index to measure the effect of loose moisture regain, the tobacco leaf unfolding rate directly affects the processing quality of subsequent processes such as cutting and drying. Traditional detection methods mainly rely on manual observation and empirical judgment, which have problems such as low detection efficiency, strong subjectivity, and large errors. In the prior art, the detection method of tobacco leaf unfolding rate based on machine vision has been gradually applied. However, the traditional image processing algorithm has the following technical defects:

[0003] Complex background interference: When the tobacco leaf is detected by light transmission, it is easily affected by factors such as uneven light source and reflection of the light board, resulting in a decrease in the accuracy of image segmentation;

[0004] Edge detection distortion: When traditional operators such as Sobel and Prewitt are used for edge detection, the adaptability to the complex texture of tobacco leaves is insufficient, and contour breakage or over-segmentation is likely to occur;

[0005] Small area noise interference: The residual tiny impurities or background noise in the image will affect the accuracy of area calculation;

[0006] Threshold setting depends on experience: The threshold parameters of the existing methods need to be manually adjusted repeatedly, and it is difficult to adapt to the differences of different batches of tobacco leaves.

[0007] In summary, we propose a method for recognizing and calculating the unfolding rate of tobacco leaves after loose moisture regain. Summary of the Invention

[0008] The purpose of the present invention is to propose a method for recognizing and calculating the unfolding rate of tobacco leaves after loose moisture regain in view of the problems existing in the background art.

[0009] The technical solution of the present invention: A method for recognizing and calculating the unfolding rate of tobacco leaves after loose moisture regain, including:

[0010] S1: Collect the light transmission image of the tobacco leaves after loose moisture regain through an industrial camera, and adjust the camera focus, exposure time and light source angle to ensure that the image is clear and free of spot and shadow interference;

[0011] S2: Perform leaf segmentation processing on the light transmission image, including:

[0012] S21: Convert the original image into a grayscale image, and separate the tobacco leaf area and the background through binaryzation processing, where the binaryzation threshold is 0.8;

[0013] S22: Extract the edges of the tobacco leaves using the Robert operator, with an edge detection threshold of 0.3, and remove the background residual interference through filling, erosion, and opening operations;

[0014] S23: Delete the connected regions with an area less than 1000 pixels, and retain the complete tobacco leaf regions;

[0015] S3: Based on the segmented image, calculate the area A1 of the dark region and the area A2 of the light region of the tobacco leaf under the light transmission state, and set the gray threshold to 0.5 to distinguish between the dark and light regions;

[0016] S4: Calculate the stretching rate K according to the area A1 of the dark region and the area A2 of the light region, K = A2 / (A1 + A2), and the stretching rate is used to characterize the loose processing quality of the tobacco leaves.

[0017] Optionally, in S21, the rgb2gray function is used for gray conversion, and the im2bw function is used for binarization processing.

[0018] Optionally, in S22, the edge function combined with the Robert operator is used for edge detection, and the imfill function is used to fill the closed regions, and the imerode function is used to erode the background contour line.

[0019] Optionally, in S23, the bwareaopen function is used for opening operation to delete the connected regions with an area less than 1000 pixels.

[0020] Optionally, in S3, the regionprops function is used to calculate the areas of the dark region and the light region.

[0021] Optionally, the minimum resolution of the industrial camera is greater than 20 million pixels, and the angle between the light source and the tobacco leaf plane is 30° - 60°.

[0022] Optionally, the kernel size of the erosion process is 1 pixel, which is used to eliminate the contour lines with a diameter less than 1 pixel in the background.

[0023] Optionally, the value range of the stretching rate K is 0 - 1, and the larger the K value, the higher the degree of stretching of the tobacco leaf.

[0024] Compared with the prior art, the present invention has the following beneficial technical effects:

[0025] Using gray conversion combined with binarization processing effectively removes the background of the light board, and cooperates with morphological erosion and opening operations to eliminate the interference of uneven light sources, reflection, and tiny impurities, ensuring that the image segmentation accuracy reaches more than 95%.

[0026] Innovatively, the Robert operator is adopted to replace the traditional Sobel / Prewitt operator. Aiming at the texture characteristics of tobacco leaf fibers, it reduces the phenomenon of edge fracture, improves the contour integrity by 30%, and provides a more accurate shape basis for subsequent area calculation.

[0027] The automatic separation of dark / light regions is achieved through the gray-scale difference in the light-transmitting area. Compared with the traditional fixed threshold method, it has enhanced adaptability to tobacco leaves with different moisture contents and varieties, and improves the detection stability.

[0028] The image processing time for a single image is less than 0.3 seconds, which supports online real-time detection. Compared with manual detection, the efficiency is increased by more than 8 times, and the subjective judgment error is eliminated. The repeatability error of the detection result is less than 2%. A quantitative evaluation index for the unfolding rate (K = A2 / A1+A2) is established, providing data support for the closed-loop control of the parameters of the loose and moistening equipment.

[0029] Through image processing techniques such as gray-scale background removal, multi-operator edge detection, and morphological filtering, the background interference is effectively eliminated and the complete leaf contour is retained; combined with the adaptive threshold segmentation algorithm, the accurate recognition of the light-transmitting area of tobacco leaves is realized. Finally, the unfolding rate parameter is obtained through area ratio calculation, overcoming the deficiencies of traditional detection techniques and providing a reliable basis for the intelligent control of the tobacco leaf processing process. Description of the Drawings

[0030] Figure 1 Is the original image of the light-transmitting image of tobacco leaves in this embodiment;

[0031] Figure 2 Is the gray-scale image of tobacco leaves;

[0032] Figure 3 Is Figure 2 The image for extracting the edge information of the graph;

[0033] Figure 4 Is Figure 3 The image of re-filling the closed area of Detailed Implementation Modes

[0034] Next, the technical solutions of the present invention will be clearly and completely described in conjunction with the drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Usually, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed present invention, but merely represents the selected embodiments of the present invention.

[0035] Embodiment

[0036] This embodiment provides a method for identifying and calculating the stretching rate of tobacco leaves after loose re-damping. The following is a detailed description of this method.

[0037] 1. An industrial camera captures the light-transmitting image of the tobacco leaf

[0038] S100: Use an industrial camera to capture the light-transmitting image of the tobacco leaf. In this process, an image of the tobacco leaf sample is obtained through a high-resolution industrial camera, ensuring that the quality of the captured image is clear enough to accurately display the light-transmitting area of the tobacco leaf. In this step, ensure that the camera is set correctly, and adjust parameters such as the focal length, exposure time, and light source angle to obtain uniform image quality and avoid interference such as light spots and shadows.

[0039] 2. Leaf segmentation processing

[0040] S101: As Figure 1 - Figure 2 , first, remove the light board background through grayscale. Use the "rgb2gray" function to convert the captured original image into a grayscale image, and then use the "im2bw" function to convert the grayscale image into a binary image, with the threshold parameter set to 0.8. After this binary processing, the tobacco leaf area can be clearly displayed.

[0041] S102: After removing the background area, as Figure 3 shown, use the "Robert" operator to extract the edge information in the image, specifically implemented through the "edge" function, with the threshold parameter set to 0.3. This step helps to remove the interference in the background and improve the extraction accuracy of the tobacco leaf in the image.

[0042] To eliminate the unnecessary contour information in the background, use the "imfill" function to refill the closed area, as Figure 4 shown. Next, use the "imerode" function to erode the image to remove the contour lines with a diameter of 1 pixel in the background, and the remaining background contours will become individual unconnected pixels.

[0043] Further use the "bwareaopen" function to perform an opening operation on the image to delete the connected areas in the image that are smaller than the specified area (for example, 1000 pixels), thereby eliminating the influence of small particles. After the final processing, only the information of the tobacco leaf part remains in the image, and the background influence is completely eliminated.

[0044] 3. Calculation of the area of the region in the light-transmitting state

[0045] S103: Calculate the area ratios of the dark and light regions in the light-transmitting state from the processed image. Use regions with different gray values in the image to distinguish the dark and light parts. In this step, by selecting an appropriate threshold value, which is 0.5, the pixels in the image are divided into a dark region and a light region. Use the "regionprops" function to calculate the areas of these two regions.

[0046] S104: For the area A1 of the dark region and the area A2 of the light region of the tobacco leaves in the light-transmitting state calculated by the above method, finally calculate their ratio K = A2 / (A1 + A2), which is used as the unfolding rate of the tobacco leaf samples of this batch. The higher the unfolding rate, generally, the better the unfolding degree of the tobacco leaves, the better the loosening effect, and the better the processing quality of the tobacco leaves.

[0047] The above specific embodiments are merely several alternative embodiments of the present invention. Based on the technical solution of the present invention and the relevant revelations of the above embodiments, those skilled in the art can make various alternative improvements and combinations to the above specific embodiments.

Claims

1. A method for identifying and calculating the unfolding rate of tobacco leaves after loose re-damping, characterized in that, Including: S1: Collect the transmissive image of the tobacco leaves after loose moisture regain through an industrial camera, and adjust the camera focal length, exposure time and light source angle. The minimum resolution of the industrial camera is greater than 20 million pixels, and the light source angle forms an angle of 30° - 60° with the tobacco leaf plane; S2: Perform leaf segmentation processing on the transmissive image, including: S21: Convert the original image into a grayscale image, and separate the tobacco leaf area and the background through binarization processing, where the binarization threshold is 0.8; S22: Use the Robert operator to extract the edges of the tobacco leaves, the edge detection threshold is 0.3, and remove the background residual interference through filling processing, corrosion processing and opening operation. The kernel size of the corrosion processing is 1 pixel, which is used to eliminate the contour lines with a diameter less than 1 pixel in the background; S23: Delete the connected regions with an area less than 1000 pixels, and retain the complete tobacco leaf area; S3: Based on the segmented image, calculate the dark area area A1 and the light area area A2 of the tobacco leaves in the transmissive state, and set the gray threshold to 0.5 to distinguish the dark and light areas; S4: Calculate the unfolding rate K according to the dark area area A1 and the light area area A2, K = A2 / (A1 + A2), and the unfolding rate is used to characterize the loose processing quality of the tobacco leaves.

2. The method for identifying and calculating the leaf unfolding rate after loose moisture regain according to claim 1, wherein, In S21, the rgb2gray function is used for gray conversion, and the im2bw function is used for binarization processing.

3. The method for identifying and calculating the leaf unfolding rate after loose moisture regain according to claim 1, wherein In S22, the edge function is used in combination with the Robert operator for edge detection, and the imfill function is used to fill the closed area, and the imerode function is used to corrode the background contour line.

4. The method for identifying and calculating the leaf unfolding rate after loose re-damping according to claim 1, wherein, In S23, the bwareaopen function is used for opening operation to delete the connected regions with an area less than 1000 pixels.

5. The method for identifying and calculating the unfolding rate of tobacco leaves after loose conditioning according to claim 1, wherein In S3, the regionprops function is used to calculate the areas of the dark area and the light area.

6. The method for identifying and calculating the unfolding rate of tobacco leaves after loose conditioning according to claim 1, wherein The value range of the unfolding rate K is 0 - 1.

Images