A method for detecting tomato seed quality based on spectral technology

Abstract

The present invention relates to the field of image processing technology, and in particular to a tomato seed quality detection method based on spectral technology, comprising: obtaining the color similarity of every two pixels according to the color difference of a plurality of pixels adjacent to every two pixels; obtaining the directional factor of each pixel according to the directionality of the color difference; obtaining the regional similarity of every two pixels according to the directional factor and the color similarity; and then obtaining the empty bag area evaluation of each pixel; obtaining the empty bag degree of tomato seeds according to the empty bag area evaluation of each pixel, thereby realizing tomato seed quality detection. The present invention avoids the error caused by other areas in the seed spectral image that have similar color performance to the empty bag area, and the obtained empty bag area result is more accurate, thereby ensuring the accuracy of tomato seed quality detection.

Description

Technical Field

[0001] The invention relates to the technical field of image processing, and in particular to a tomato seed quality detection method based on spectral technology. Background Art

[0002] Tomatoes belong to the Solanaceae family. They are not only delicious but also rich in various nutrients. Ensuring the quality of tomato seeds is of key importance for agricultural production and horticultural cultivation. High-quality tomato seeds can not only increase yields, but also ensure healthy plant growth, reduce disease risks, and improve the quality and market value of fruits. Therefore, it is necessary to test the quality of tomato seeds.

[0003] Existing technologies usually use spectral technology to detect the quality of tomato seeds. It can detect the internal and surface characteristics of the seeds without destroying them, and can obtain data in real time to quickly obtain test results. When using spectral technology to detect the fullness of tomato seeds, it is often interfered by nutrients such as water and protein in the tomato seeds on the spectral image, resulting in the inability to accurately extract the empty area on the edge of the seeds, causing errors in calculating the fullness of the seeds, and further leading to errors in the assessment of the quality of the tomato seeds. Summary of the invention

[0004] The invention provides a tomato seed quality detection method based on spectral technology, so as to solve the problem that when the existing spectral technology is used to detect the quality of tomato seeds, the result is inaccurate due to the interference of components such as water and protein.

[0005] The tomato seed quality detection method based on spectral technology of the present invention adopts the following technical scheme:

[0006] An embodiment of the present invention provides a method for detecting tomato seed quality based on spectral technology, the method comprising the following steps:

[0007] Acquire a seed spectrum image, wherein the seed spectrum image is an RGB image;

[0008] According to the distance between each pixel and other pixels in the seed spectrum image in the color space, the color difference between each pixel and other pixels is obtained; the reference point of each pixel is obtained, and all the reference points of each pixel constitute several reference sequences of each pixel; according to the color difference between each pixel and the corresponding reference points of other pixels in the same reference sequence, the color similarity between each pixel and other pixels is obtained; according to the directionality of the color difference of the pixels around each pixel, the directional factor of each pixel is obtained; according to the directional factor of each pixel and the color similarity with other pixels, the regional similarity between each pixel and other pixels is obtained;

[0009] Obtain the empty bag area evaluation of each pixel according to the color value of each pixel in the color space and the area similarity with other pixels;

[0010] The empty seed degree of tomato seeds is obtained according to the empty seed area evaluation of each pixel point; and the tomato seed quality detection is realized according to the empty seed degree of tomato seeds.

[0011] Furthermore, the obtaining of the reference point of each pixel, all the reference points of each pixel constituting a plurality of reference sequences of each pixel, includes:

[0012] A reference length is preset, and the four straight lines in the eight-neighborhood directions of each pixel are recorded as reference lines for each pixel. With each pixel as the center, pixels with a common reference length are selected on both sides of each reference line as reference points; all reference points of each pixel on each reference line constitute each reference sequence of each pixel.

[0013] Furthermore, obtaining the color similarity between each pixel and other pixels includes:

[0014] The color similarity G between the i-th pixel and the j-th pixel i,j The calculation method is:

[0015]

[0016] Where: T represents the number of reference sequences, P i,j,t It represents the reference color difference between the i-th pixel and the j-th pixel in the t-th reference sequence; exp() represents an exponential function with a natural constant as the base.

[0017] Furthermore, the specific method for obtaining the reference color difference is:

[0018] The reference color difference P between the i-th pixel and the j-th pixel in the t-th reference sequence i,j,t The calculation method is:

[0019]

[0020] Where: L represents the reference length, d i,j,t,k Indicates the color difference between the kth reference point in the tth reference sequence of the i-th pixel and the kth reference point in the tth reference sequence of the j-th pixel.

[0021] Furthermore, obtaining the directionality factor of each pixel includes:

[0022] The reference sequence that obtains the minimum value among the color change factors of all reference sequences of the i-th pixel point is recorded as the r-th minimum reference sequence; the reference sequence perpendicular to the straight line corresponding to the minimum reference sequence is recorded as the vertical reference sequence;

[0023] Directivity factor C of the i-th pixel i The calculation method is:

[0024] C i =S i ′ ,r ×exp(-S i,r ×|S i ′ ,r -S i,r |)

[0025] Where: S i ′ ,r represents the color change factor of the vertical reference sequence of the rth minimum reference sequence of the i-th pixel, S i,r It represents the color change factor of the rth minimum reference sequence of the i-th pixel; exp() represents an exponential function with a natural constant as the base; || represents the absolute value function.

[0026] Furthermore, the specific method for obtaining the color change factor is:

[0027] The average value of the color differences between every two pixels in the t-th reference sequence of the ith pixel is recorded as the color change factor of the t-th reference sequence of the ith pixel.

[0028] Furthermore, obtaining the regional similarity between each pixel and other pixels includes:

[0029] The regional similarity D between the i-th pixel and the j-th pixel i,j The calculation method is:

[0030] D i,j =Sigmoid(G i,j ×C i ×C j )

[0031] Where: G i,j Indicates the color similarity between the i-th pixel and the j-th pixel, C i represents the directional factor of the i-th pixel, C j Represents the directional factor of the j-th pixel; Sigmoid() represents the Sigmoid function.

[0032] Furthermore, the step of obtaining the empty packing degree of tomato seeds according to the empty packing area evaluation of each pixel point includes:

[0033] The empty bag area evaluation of each pixel is taken as the gray value of each pixel in the seed spectral image to obtain the gray image of the seed spectral image; the gray image of the seed spectral image is threshold segmented by the Otsu threshold segmentation method, and the area with a gray value greater than the threshold in the threshold segmentation result is recorded as the tomato seed empty bag area; the ratio of the number of pixels in the tomato seed empty bag area to the total number of pixels in the tomato seed is normalized using the Sigmoid function, and the normalized result is recorded as the empty bag degree of the tomato seed.

[0034] Furthermore, the specific method for obtaining the empty package area evaluation is:

[0035] The calculation method for the empty bag area evaluation of the i-th pixel is:

[0036]

[0037] Where: g represents the preset hyperparameter, n represents the number of pixels in the seed spectrum image, n-1 represents the number of pixels in the seed spectrum image except the i-th pixel, D i,j Represents the regional similarity between the i-th pixel and the j-th pixel, R j Represents the color value of the j-th pixel in the red channel of the RGB color space; Sigmoid[] represents the Sigmoid function.

[0038] Furthermore, the tomato seed quality detection according to the empty bag degree of the tomato seeds comprises:

[0039] A quality threshold is preset. When the empty bag degree of the current tomato seed is greater than the quality threshold, the current tomato seed is marked as a low-quality seed.

[0040] The beneficial effects of the technical solution of the present invention are as follows: the color similarity of each two pixels is obtained according to the color difference of the corresponding reference points in the reference sequence of each two pixels. The color difference of the two pixels alone is difficult to accurately reflect the color difference of the two pixels. The color similarity result of each two pixels obtained by taking the color difference corresponding to several adjacent pixels as a reference is more reliable; the directional factor of each pixel is obtained according to the directionality of the color difference of the pixels around each pixel, and the nutritional area and the empty bag area of ​​the seed spectral image are distinguished by the directional performance, so as to avoid the error caused by the nutritional area to the seed quality detection result; the empty bag area of ​​each pixel is obtained according to the regional similarity of each two pixels and the color value of each pixel in the color space. Domain evaluation, regional similarity reflects the probability that every two pixels belong to the empty bag area. According to the regional similarity between each pixel and all pixels except this pixel and the color value in the color space, all pixels except this pixel are used as reference to obtain the empty bag area evaluation of each pixel, and the result is more accurate; the empty bag area evaluation is used as the gray value of each pixel in the seed spectral image, and the gray value reflects the degree to which each pixel is a pixel in the empty bag area; the seed spectral image is thresholded by gray value to obtain the empty bag degree of tomato seeds, avoiding the error caused by other areas in the seed spectral image with similar color to the empty bag area, and the obtained empty bag area result is more accurate, which ensures the accuracy of tomato seed quality detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0042] Figure 1 The present invention is a flowchart of the steps of a tomato seed quality detection method based on spectral technology. DETAILED DESCRIPTION

[0043] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation method, structure, characteristics and effects of a tomato seed quality detection method based on spectral technology proposed by the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0044] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0045] The specific scheme of the tomato seed quality detection method based on spectral technology provided by the present invention is described in detail below with reference to the accompanying drawings.

[0046] See also Figure 1 , which shows a flowchart of a method for detecting tomato seed quality based on spectral technology provided by an embodiment of the present invention, the method comprising the following steps:

[0047] Step S001, obtaining a seed spectrum image.

[0048] The purpose of this embodiment is to detect the quality of tomato seeds through tomato seed spectral images, so firstly, tomato seed spectral images are collected. Specifically, tomato seed spectral images are collected through a multispectral seed phenotype imaging system, which are recorded as seed spectral images. The seed spectral images are RGB three-channel images.

[0049] Step S002, obtaining the color similarity between each pixel and other pixels; obtaining the directionality factor of each pixel; obtaining the area similarity between each pixel and other pixels.

[0050] It should be noted that when the tomato seeds are not full, empty capsules will appear near the edges of the seeds. Compared with normal full seeds, the internal structure and chemical composition of empty seeds are different. Different chemical compositions have different reflection characteristics for near-infrared light. The empty capsule areas of tomato seeds with incomplete grains appear as red bands at the head and tail of the seeds in the seed spectral image. To screen out the empty capsule areas of tomato seeds, the color similarity of every two pixels in the seed spectral image is first analyzed. To avoid errors caused by a small number of noise points with similar colors, the color similarity of every two pixels in the seed spectral image is obtained based on the color difference between several adjacent pixels of every two pixels in the seed spectral image. The smaller the color difference between adjacent pixels in the same direction of two pixels, the greater the color similarity of the two pixels.

[0051] Specifically, a reference length is preset, and the four straight lines in the eight-neighborhood direction of each pixel point are recorded as the reference straight lines of each pixel point. With each pixel point as the center, pixels of a common reference length are selected on both sides of each reference straight line and recorded as reference points; all reference points of each pixel point on each reference straight line constitute each reference sequence of each pixel point, and each reference straight line corresponds to a reference sequence; in particular, if the number of reference points that can be obtained on one side is less than half of the reference length, the maximum number of reference points that can be obtained is taken; this embodiment is described with a reference length of 10, and other embodiments can be set to other values, which are not specifically limited in this embodiment;

[0052] The Euclidean distance between any two pixels in the RGB color space is recorded as the color difference between the two pixels; the reference color difference P of the t-th reference sequence between the i-th pixel and the j-th pixel is i,j,t The calculation method is:

[0053]

[0054] Where: L represents the reference length, d i,j,t,k Indicates the color difference between the kth reference point in the tth reference sequence of the i-th pixel and the kth reference point in the tth reference sequence of the j-th pixel; in particular, if the number of reference points in the tth reference sequence of the i-th pixel is different from the number of reference points in the tth reference sequence of the j-th pixel, then L takes the minimum value of the number of reference points in the tth reference sequence of these two pixels;

[0055] The color similarity G between the i-th pixel and the j-th pixel i,j The calculation method is:

[0056]

[0057] Where: T represents the number of reference sequences, P i,j,t represents the reference color difference between the i-th pixel and the j-th pixel in the t-th reference sequence; exp() represents an exponential function with a natural constant as the base;

[0058] It should be further explained that, since the reflection characteristics of nutrients such as water, protein, and fat in tomato seeds are similar to those of the empty area, it is difficult to accurately determine the empty area of ​​tomato seeds only by color similarity; the pixel points of the empty area are located at the edge of the tomato seeds, which appear as a strip area in the same direction as the edge of the tomato seeds, while the nutrient area is located inside the tomato seeds, which appears as a clustered or randomly scattered small area. Therefore, on the seed spectral image, the color difference between adjacent pixels in the direction of the strip extension is small, while the color difference between adjacent pixels in the direction perpendicular to the strip extension is large; according to the directionality shown by the color difference of the pixels around each pixel, the directional factor of each pixel is obtained; the more the color change of the pixels around each pixel conforms to the directional characteristics shown by the empty area, the greater the directional factor of the pixel; combining the directional factor of each pixel and the color similarity of each two pixels, the regional similarity of each two pixels is obtained; the greater the directional factor of the two pixels and the greater the color similarity of the two pixels, the greater the regional similarity of the two pixels, and the more likely the two pixels belong to the empty area.

[0059] Specifically, the average value of the color difference between every two pixels in the t-th reference sequence of the i-th pixel is recorded as the color change factor of the t-th reference sequence of the i-th pixel; similarly, the color change factor of the reference sequence of each pixel in each direction is obtained;

[0060] The reference sequence with the minimum color change factor among all reference sequences of the i-th pixel is recorded as the r-th minimum reference sequence; the reference sequence perpendicular to the straight line corresponding to the minimum reference sequence is recorded as the vertical reference sequence; the directional factor C of the i-th pixel is recorded as i The calculation method is:

[0061] C i =S i ′ ,r ×exp(-S i,r ×|S i ′ ,r -S i,r |)

[0062] Where: S i ′ ,r represents the color change factor of the vertical reference sequence of the rth minimum reference sequence of the i-th pixel, S i,r represents the color change factor of the rth minimum reference sequence of the i-th pixel; exp() represents an exponential function with a natural constant as the base; || represents an absolute value function;

[0063] It should be noted that S i,r is the minimum value of the color change factor of all reference sequences of the i-th pixel point, then the direction of the reference straight line corresponding to the r-th minimum reference sequence represents the direction perpendicular to the direction of the pixel strip extension in the empty bag area, S i,r The smaller the value of S i ′ ,r The larger the value of , the more the color difference of the pixels around the i-th pixel is consistent with the directionality of the strip extension of the empty bag area, and the larger the directionality factor of the i-th pixel is; |S i ′ ,r -S i,r | indicates that the larger the value, the greater the difference in the color change factor between the rth minimum reference sequence of the i-th pixel and its vertical reference sequence. The color difference of the pixels around the i-th pixel is more consistent with the directionality of the strip extension of the empty bag area, and the greater the directionality factor of the i-th pixel.

[0064] Furthermore, the regional similarity D between the i-th pixel and the j-th pixeli,j The calculation method is:

[0065] D i,j =Sigmoid(G i,j ×C i ×C j )

[0066] Where: G i,j Indicates the color similarity between the i-th pixel and the j-th pixel, C i represents the directional factor of the i-th pixel, C j Represents the directional factor of the j-th pixel; Sigmoid() represents the Sigmoid function;

[0067] It should be noted that G i,j The larger the value of is, the more similar the colors of the i-th pixel and the j-th pixel are; i ×C j The larger the value of G is, the more the pixels around the i-th pixel and the j-th pixel conform to the empty packet area characteristics; i,j ×C i ×C j The larger the value of , the more likely it is that the i-th pixel and the j-th pixel belong to the empty bag area at the same time, and the greater the regional similarity between the i-th pixel and the j-th pixel;

[0068] Similarly, the regional similarity between every two pixels is obtained.

[0069] Step S003: Obtain the empty bag area evaluation of each pixel.

[0070] It should be noted that the regional similarity reflects the possibility that every two pixels belong to the empty bag area at the same time. The regional similarity between each pixel and all other pixels is compared to obtain the empty bag area evaluation of each pixel;

[0071] Specifically, the calculation method for evaluating the empty bag area of ​​the i-th pixel is:

[0072]

[0073] Where: g represents the preset hyperparameter, n represents the number of pixels in the seed spectrum image, n-1 represents the number of pixels in the seed spectrum image except the i-th pixel, D i,j Represents the regional similarity between the i-th pixel and the j-th pixel, R j Represents the color value of the red channel of the j-th pixel in the RGB color space; Sigmoid[] represents the Sigmoid function;

[0074] It should be noted that the hyperparameter g is set to The value of g is mapped to a certain range. This embodiment is described by taking g=255 as an example. Other embodiments may be set to other values, which are not specifically limited in this embodiment. i,j ×R j represents the product of the regional similarity between the i-th pixel and the j-th pixel and the color value of the red channel of the j-th pixel in the RGB color space, D i,j The larger the value of is, the greater the influence of the jth pixel on the empty bag area evaluation of the i-th pixel. i,j ×R j The larger it is, the greater the empty bag area evaluation of the i-th pixel point is; It represents the average of the product of the regional similarity between the i-th pixel and all pixels except the i-th pixel and the color value of the red channel of the j-th pixel in the RGB color space. The larger the value, the greater the empty bag area evaluation of the i-th pixel.

[0075] Similarly, the empty bag area evaluation of each pixel is obtained.

[0076] Step S004, obtaining the empty seed degree of the tomato seeds; and implementing tomato seed quality detection according to the empty seed degree of the tomato seeds.

[0077] It should be noted that the grayscale value of each pixel in the grayscale image of the seed spectrum image reflects the degree to which each pixel is a pixel in the empty bag area. The empty bag area in the seed spectrum image is obtained according to the grayscale value, thereby realizing tomato seed quality detection.

[0078] Specifically, the empty bag area evaluation of each pixel is used as the gray value of each pixel in the seed spectrum image to obtain the gray image of the seed spectrum image; the gray image of the seed spectrum image is threshold segmented by the Otsu threshold segmentation method, and the area with a gray value greater than the threshold in the threshold segmentation result is recorded as the empty bag area of ​​the tomato seeds; the number of pixels in the empty bag area of ​​the tomato seeds is obtained, the total number of pixels of the tomato seeds in the seed spectrum image is obtained, the ratio of the number of pixels in the empty bag area of ​​the tomato seeds to the total number of pixels of the tomato seeds is normalized by using the Sigmoid function, and the normalized result is recorded as the empty bag degree of the tomato seeds;

[0079] A quality threshold is preset. When the empty bag degree of tomato seeds is greater than the quality threshold, the seeds are marked as low-quality seeds, thereby realizing tomato seed quality detection.

[0080] It should be noted that the exp(-x) model used in this embodiment only indicates that the result of the negative correlation and constraint model output is in the interval (0,1], where x is the input of this model. It can be replaced by other models with the same purpose during specific implementation. This embodiment only takes the exp(-x) model as an example for description without specific limitation.

[0081] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A tomato seed quality detection method based on spectral technology, characterized in that: The method comprises the following steps: Acquire a seed spectrum image, wherein the seed spectrum image is an RGB image; According to the distance between each pixel and other pixels in the seed spectrum image in the color space, the color difference between each pixel and other pixels is obtained; the reference point of each pixel is obtained, and all the reference points of each pixel constitute several reference sequences of each pixel; according to the color difference between each pixel and the corresponding reference points of other pixels in the same reference sequence, the color similarity between each pixel and other pixels is obtained; according to the directionality of the color difference of the pixels around each pixel, the directional factor of each pixel is obtained; according to the directional factor of each pixel and the color similarity with other pixels, the regional similarity between each pixel and other pixels is obtained; Obtain the empty bag area evaluation of each pixel according to the color value of each pixel in the color space and the area similarity with other pixels; The empty packing degree of tomato seeds is obtained according to the empty packing area evaluation of each pixel point; and the quality detection of tomato seeds is realized according to the empty packing degree of tomato seeds; The obtaining of the color similarity between each pixel and other pixels includes: The color similarity G between the i-th pixel and the j-th pixel i,j The calculation method is: Where: T represents the number of reference sequences, P i,j,t represents the reference color difference between the i-th pixel and the j-th pixel in the t-th reference sequence; exp() represents an exponential function with a natural constant as the base; The obtaining of the directionality factor of each pixel point includes: The reference sequence that obtains the minimum value among the color change factors of all reference sequences of the i-th pixel point is recorded as the r-th minimum reference sequence; the reference sequence perpendicular to the straight line corresponding to the minimum reference sequence is recorded as the vertical reference sequence; Directivity factor C of the i-th pixel i The calculation method is: C i =S i ′ ,r ×exp(-S i,r ×|S i ′ ,r -S i,r |) Where: S i ′ ,r represents the color change factor of the vertical reference sequence of the rth minimum reference sequence of the i-th pixel, S i,r represents the color change factor of the rth minimum reference sequence of the i-th pixel; exp() represents an exponential function with a natural constant as the base; || represents an absolute value function; The obtaining of the regional similarity between each pixel and other pixels includes: The regional similarity D between the i-th pixel and the j-th pixel i,j The calculation method is: D i,j =Sigmoid(G i,j ×C i ×C j ) Where: G i,j Indicates the color similarity between the i-th pixel and the j-th pixel, C i represents the directional factor of the i-th pixel, C j Represents the directional factor of the j-th pixel; Sigmoid() represents the Sigmoid function; The specific method for obtaining the empty package area evaluation is: The calculation method for the empty bag area evaluation of the i-th pixel is: Where: g represents the preset hyperparameter, n represents the number of pixels in the seed spectrum image, n-1 represents the number of pixels in the seed spectrum image except the i-th pixel, D i,j Represents the regional similarity between the i-th pixel and the j-th pixel, R j Represents the color value of the j-th pixel in the red channel of the RGB color space; Sigmoid[] represents the Sigmoid function.

2. According to claim 1, a tomato seed quality detection method based on spectral technology is characterized in that: The step of obtaining the reference point of each pixel, wherein all the reference points of each pixel constitute a plurality of reference sequences of each pixel, includes: A reference length is preset, and the four straight lines in the eight-neighborhood directions of each pixel are recorded as reference lines for each pixel. With each pixel as the center, pixels with a common reference length are selected on both sides of each reference line as reference points; all reference points of each pixel on each reference line constitute each reference sequence of each pixel.

3. The tomato seed quality detection method based on spectral technology according to claim 1, characterized in that: The specific method for obtaining the reference color difference is: The reference color difference P of the tth reference sequence between the i-th pixel and the j-th pixel i,j,t The calculation method is: Where: L represents the reference length, d i,j,t,k Indicates the color difference between the kth reference point in the tth reference sequence of the i-th pixel and the kth reference point in the tth reference sequence of the j-th pixel.

4. The tomato seed quality detection method based on spectral technology according to claim 1, characterized in that: The specific method for obtaining the color change factor is: The average value of the color differences between every two pixels in the t-th reference sequence of the ith pixel is recorded as the color change factor of the t-th reference sequence of the ith pixel.

5. The tomato seed quality detection method based on spectral technology according to claim 1, characterized in that: The step of obtaining the empty packing degree of tomato seeds according to the empty packing area evaluation of each pixel point includes: The empty bag area evaluation of each pixel is taken as the gray value of each pixel in the seed spectral image to obtain the gray image of the seed spectral image; the gray image of the seed spectral image is threshold segmented by the Otsu threshold segmentation method, and the area with a gray value greater than the threshold in the threshold segmentation result is recorded as the tomato seed empty bag area; the ratio of the number of pixels in the tomato seed empty bag area to the total number of pixels in the tomato seed is normalized using the Sigmoid function, and the normalized result is recorded as the empty bag degree of the tomato seed.

6. The tomato seed quality detection method based on spectral technology according to claim 1, characterized in that: The tomato seed quality detection according to the empty bag degree of the tomato seeds comprises: A quality threshold is preset. When the empty bag degree of the current tomato seed is greater than the quality threshold, the current tomato seed is marked as a low-quality seed.

Images