A fresh cut flower grading system and judgment method
By designing a fresh cut flower grading system and feature extraction algorithm, the problem of poor feature extraction stability in the quality grading detection of fresh cut flower is solved, and automatic grading and assembly of large batches of fresh cut flowers is realized, ensuring measurement accuracy and grading accuracy.
Patent Information
- Application Number
- CN202010471681.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-05-29
- Publication Date
- 2024-12-06
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing technology has little research on the quality grading detection of fresh cut flowers, mainly because the abstraction of grading standards and the different types of fresh cut flowers lead to poor stability of feature extraction algorithms.
A fresh cut flower grading system including rollers, conveyors, components, bases, belts, graders, pushers and sealing rods was designed to extract the damaged area, color, flower bloom index, stem length and For characteristics such as curvature, a comprehensive evaluation model is established to make quality judgments.
Automatic grading and assembly of large batches of fresh cut flowers is realized, ensuring the accuracy and robustness of flower opening diameter measurement, laying the foundation for determining the storage time of fresh cut flowers, and improving the accuracy of grading.
Smart Images

Figure CN111495780B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of automatic flower grading, in particular to a fresh cut flower grading system and a judgment method. Background Art
[0002] In recent years, machine vision technology has been widely used in agricultural product inspection and grading. By extracting multiple shape features of potato images and using multivariate linear regression to predict the quality and shape of potatoes, good results have been achieved. However, there are few studies on machine vision in the grading of fresh cut flowers. The main reason is that the grading standards are abstract and the different shapes of fresh cut flowers lead to poor stability of feature extraction algorithms. Summary of the invention
[0003] In order to solve the above-mentioned deficiencies of the prior art, the object of the present invention is to provide a fresh cut flower grading system and a judgment method to overcome the defects in the prior art.
[0004] In order to achieve the above-mentioned purpose, the present invention provides a fresh cut flower grading system and judgment method, comprising a roller, a conveyor belt, a portioner, a base, a belt, a classifier, a branch pusher, a sealing rod and a bracket, the roller is rotatably connected to one end of the bracket, the conveyor belt is mounted in the middle of the bracket, one end of the portioner is fixedly connected to the other end of the bracket, the other end of the portioner is fixedly connected to the base, the belt is rotatably connected to the base, the classifier is fixedly connected to one side of the base, and a grading box is fixedly connected to the same side of the base at intervals, a plurality of branch pushers are mounted one by one on the base at the entrance of the grading box, a plurality of sealing rods are hinged at the entrance of the grading box, a camera is provided in the classifier, and the classifier is electrically connected to the branch pusher.
[0005] As a further explanation of the present invention, preferably, a plurality of long strip-shaped gear teeth are fixedly connected at intervals inside the drum, one end of the conveyor belt extends into the drum, and shift plates are fixedly connected to both sides of the end of the conveyor belt extending into the drum, and the interval between the shift plate and the outer end of the gear teeth is 1 to 3 mm.
[0006] As a further explanation of the present invention, preferably, the conveyor belt is inclined in the length direction and extends into the drum at the highest height. A guide rail is fixedly connected to the bracket above the conveyor belt. The guide rail is distributed in a V shape along the length direction of the conveyor belt, and the narrowed end of the guide rail points to the portioner.
[0007] As a further illustration of the present invention, preferably, the length direction of the distributor is inclined, a groove is formed in the middle of the distributor, and a plurality of rotating wheels are rotatably connected to both sides of the groove along the length direction, and the spacing between the rotating wheels on the same side is between 0.5 mm and 2 mm.
[0008] As a further explanation of the present invention, preferably, an L-shaped branch channel is opened on the top of the base, the short narrow portion on one side of the bending end of the branch channel is communicated with the divider, and a grading channel is opened on the base at the long narrow portion on the other side of the bending end of the branch channel. The length direction of the grading channel is the same as the length direction of the base, the classifier is erected on the grading channel, and the grading box and the branch pusher are both arranged on both sides of the grading channel.
[0009] As a further explanation of the present invention, preferably, a plurality of V-shaped clamping claws are fixedly connected to the belt at intervals, the width of the clamping claw is smaller than the outer diameter of the flower bud, and a hook is provided on the outer side of the end of the clamping claw in contact with the flower branch, and the hook is bent toward the opening direction of the clamping claw.
[0010] As a further illustration of the present invention, preferably, the base is inclined in the length direction and the inclination angle is between 2° and 15°.
[0011] The present invention also provides a method for determining the classification of fresh cut flowers, comprising the following steps:
[0012] Ⅰ. When the fresh cut flowers pass through the classifier, the belt stops rotating, and the industrial cameras on the top and side of the classifier take pictures, and the bud images of the fresh cut flowers are cut out through the mask;
[0013] Ⅱ. The damaged area and damage range of fresh cut flower buds were obtained by extracting yellow-brown and black colors, and the damage level was determined by the damage ratio;
[0014] III. Extract the red color to obtain the flower color of the fresh cut flower buds, and then convert it into grayscale value to determine the flower color grade;
[0015] IV. Determine the degree of bud opening by extracting the flower opening index;
[0016] V. Extract the length and curvature of the flower stems and establish the grade standards for length and curvature;
[0017] VI. By extracting the image features of flower bud damage, flower color, blooming index, stem length and stem inclination, and finding the relationship between the image features and each index, and integrating the grading standards of each image feature, a comprehensive evaluation model for the quality of fresh cut flowers is established.
[0018] As a further illustration of the present invention, preferably, when performing damage and flower color extraction, the split function is used to separate the three channels of the image, extract the R channel image, and perform binarization. When the threshold is 60, the black pixels in the flower buds can be extracted, and these pixels are clustered by K-means. When K=4, the black damaged area can be extracted; the cvtColor function is used to convert the color space to the HSV color space, extract the H channel, and use the K-means function to cluster the H channel image to extract the yellow-brown damaged area when K=3; and cvtColor is used to convert the image to the R channel. In the R channel, the gray value with the largest pixel gray value is extracted. The higher the gray value, the brighter the flower color.
[0019] As a further explanation of the present invention, preferably, a mathematical model is established with the number of rows of flower buds in the image as the independent variable and the number of pixels in each row of flower buds as the dependent variable. The measurement position is obtained by analyzing the difference between the number of pixels in each row of flower buds and the number of pixels in the fifth row thereafter, and the number of flower bud pixels at that position is used as the measured flower opening diameter.
[0020] The present invention has the following beneficial effects:
[0021] The invention designs a novel flower grading device to separate a large number of flowers into single branches, transport the flowers one by one to a grader for photographing and analysis, and finally differentiate and package them, thereby realizing the automatic grading and packaging of large quantities of fresh cut flowers. In addition, the invention relies on a newly designed algorithm to ensure good robustness and accuracy in positioning and measuring the flower opening diameter measurement position, thereby laying a foundation for extracting flowering index characteristics in the automatic grading of rose fresh cut flowers and determining the storage time of fresh flowers. BRIEF DESCRIPTION OF THE DRAWINGS
[0022] Figure 1 is a general assembly diagram of the device of the present invention;
[0023] Figure 2 yes Figure 1 A magnified view of middle;
[0024] Figure 3 is a distribution diagram of classification boxes of the present invention;
[0025] Figure 4 yes Figure 3 Enlarged view of B.
[0026] Description of reference numerals:
[0027] 1. Roller; 11. Gear teeth; 2. Conveyor belt; 21. Plate; 22. Guide rail; 3. Sharer; 31. Rotating wheel; 32. Branch collecting road; 4. Base; 41. Branching road; 42. Grading road; 43. Grading box; 44. Residual branch box; 5. Belt; 51. Clamp; 6. Classifier; 7. Branch pusher; 71. Articulated rod; 72. Push block; 8. Sealing rod; 9. Bracket. DETAILED DESCRIPTION
[0028] In order to further understand the structure, features and other purposes of the present invention, the preferred embodiments are described in detail with reference to the accompanying drawings as follows. The embodiments described in the drawings are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention.
[0029] A fresh cut flower grading system and judgment method, combined with Figure 1 , Figure 3 , including a roller 1, a conveyor belt 2, a portioner 3, a base 4, a belt 5, a classifier 6, a branch pusher 7, a sealing rod 8 and a bracket 9. The roller 1 is rotatably connected to one end of the bracket 9, the conveyor belt 2 is mounted in the middle of the bracket 9, one end of the portioner 3 is fixedly connected to the other end of the bracket 9, and the other end of the portioner 3 is fixedly connected to the base 4. The belt 5 is rotatably connected to the base 4, the classifier 6 is fixedly connected to the upper side of the base 4, and four grading boxes 43 (not all of which are shown in the figure) are fixedly connected on the same side of the base 4 at intervals. Four branch pushers 7 are mounted one by one on the base 4 at the entrance of the grading box 43, and four sealing rods 8 are hinged at the entrance of the grading box 43. The sealing rod 8 is driven by a motor to push the graded fresh cut flowers into the grading box 43.
[0030] Combination Figure 1 , Figure 2 The bracket 9 is a square bracket. The bracket 9 is located under the drum 1 and is rotatably connected to two rollers. The two rollers are driven by a motor. The drum 1 is placed on two long rollers so that the rollers drive the drum 1 to rotate. A plurality of long strips of gear teeth 11 are fixedly connected inside the drum 1 at intervals. The spacing between the gear teeth 11 is slightly larger than the width of the flower branch. One end of the conveyor belt 2 extends into the drum 1. The conveyor belt 2 extends into the drum 1. Both sides of the end of the drum 1 are fixedly connected with a paddle 21. The interval between the paddle 21 and the outer end of the gear tooth 11 is 1 to 3 mm. The large When a large amount of fresh cut flowers are needed, a large amount of fresh cut flowers can be placed at the bottom of the drum 1. At this time, under the action of gravity, the fresh cut flowers are embedded in the gaps of the gear teeth 11 and are transferred to the upper end of the drum 1 through the rotation of the drum 1. When the fresh cut flowers move to the paddle plate 21, the fresh cut flowers that are not embedded in the gaps of the gear teeth 11 are blocked, thereby realizing the function of automatically sorting out single fresh cut flowers, which is convenient for subsequent single-branch grading judgment; the single fresh cut flowers in the gaps of the gear teeth 11 bypass the paddle plate 21 and move to the upper end of the drum 1 until they fall onto the conveyor belt 2 under the action of gravity.
[0031] Combination Figure 1 , Figure 2The conveyor belt 2 is a wide transmission belt, which is driven to rotate by a pulley and a motor. The conveyor belt 2 is inclined in the length direction and extends into the drum 1 at the highest height. A guide rail 22 is fixedly connected to the bracket above the conveyor belt 2. The guide rail 22 is distributed in a V shape along the length direction of the conveyor belt 2, and the narrowed end of the guide rail 22 points to the portioner 3; a belt-driven conveyor belt 2 is set, and the inclined length direction is matched to enable a single branch of fresh cut flowers to move to the portioner 3 faster, so as to avoid the accumulation of fresh cut flowers on the conveyor belt 2 and affect the single branch separation effect of the drum 1; the guide rail 22 is set to allow a single branch of fresh cut flowers to be directional moved, so that the subsequent portioner 3 can perform reversing transportation of fresh cut flowers.
[0032] Combination Figure 1 , Figure 3 , the portioner 3 is a straight plate inclined in the length direction, and a groove is formed in the middle of the portioner 3. The groove width is wider than the width of the flower branch and smaller than the outer diameter of the flower bud; a number of rotating wheels 31 are connected to the two sides of the groove along the length direction, and the spacing between the rotating wheels 31 on the same side is between 0.5mm and 2mm to avoid mutual interference between the rotating wheels 31 and affecting the rotation; an inclined V-shaped branch collecting channel 32 is fixedly connected to the top of the portioner 3, and a groove is also formed at the bottom of the branch collecting channel 32 to play a transitional role. When a single branch of fresh cut flowers slides to the portioner 3, the flower branch passes through the groove under the action of gravity, and the flower bud falls on the rotating wheel 31 at this time. Relying on the action of gravity and the small friction of the rotating wheel 31, the fresh cut flowers fall vertically in the length direction to the base 4, so that the subsequent grader 6 can take pictures to judge the flower grading.
[0033] Combination Figure 3 , Figure 4The base 4 is a square plate-shaped bracket, the top surface of the base 4 is polished and smooth, the base 4 is inclined in the length direction and the inclination angle is between 2° and 15°, which can make the fresh cut flowers move more smoothly under the dual effects of gravity and the belt 5; an L-shaped branching path 41 is opened on the top of the base 4, the short narrow part on one side of the bent end of the branching path 41 is connected to the portioner 3, and a grading path 42 is opened on the base 4 at the long narrow part on the other side of the bent end of the branching path 41. The length direction of the grading path 42 is the same as that of the base 4, the classifier 6 is mounted on the grading path 42, and the grading box 43 and the branch pusher 7 are both arranged on both sides of the grading path 42. The belt 5 is an endless belt, which is driven to rotate by two turntables and a motor. A plurality of V-shaped clamps 51 are fixedly connected to the belt 5 at intervals, and the width of the clamps 51 is smaller than the outer diameter of the flower buds. When the fresh cut flowers move from the portioner 3 to the branching road 41, the belt 5 rotates to make the clamps 51 buckle the branches of the fresh cut flowers, and the clamps 51 are close to the flower buds. At this time, under the action of the single clamp 51 and the bent branching road 41, the single clamp 51 can clamp a single branch of fresh cut flowers, and the speed of the belt 5 is controlled to analyze and judge the single branch of fresh cut flowers for a sufficient time. A hook is provided on the outside of the end of the clamp 51 that contacts the flower branch, and the hook is bent in the direction of the opening of the clamp 51, which can ensure that the clamp 51 can still buckle the flower branch and push the fresh cut flowers to move when the fresh cut flowers move to the grading road 42.
[0034] Combination Figure 3 , Figure 4 The classifier 6 is provided with a camera, and the cameras in the classifier 6 are distributed on the top and both sides of the classifier 6. The classifier 6 is also provided with a controller, a driver and a CPU. The controller, the camera and the driver are electrically connected to the CPU, and the driver is electrically connected to the branch pusher 7, so that the fresh cut flowers passing through the classifier 6 can be stored separately according to the analysis and judgment results; the branch pusher 7 includes a hinged rod 71 and a push block 72, one end of the hinged rod 71 is hinged on the base 4, the hinged rod 71 is driven to rotate by a motor, one side of the push block 72 is hinged to the other end of the hinged rod 71, and the bottom of the push block 72 is slidably connected to the On the base 4, the branch pusher 7 can also be composed of a cylinder and a push plate. The push block 72 and the push plate move in the opposite direction of the grading box 73. The push block 72 is located in the opposite direction of the belt 5 and is provided with an infrared sensor. When the classifier 6 makes a judgment, when the fresh cut flowers move to the infrared sensor of the corresponding grade, the driver controls the motor on the corresponding branch pusher 7 and the motor on the sealing rod 8 to rotate. At this time, the push block 72 or the push plate pushes the fresh cut flowers into the grading box 43 to achieve automatic grading of the fresh cut flowers. The sealing rods on other grading boxes 43 do not move to support the flower buds so that the flower buds can move smoothly. The end of the base 4 is fixedly connected with a residual branch box 44. When the fresh cut flowers do not meet the grading requirements, all the branch pushers 7 and the sealing rods 8 do not move and move into the residual branch box 44 under the action of the belt 5, which is convenient for subsequent centralized processing.
[0035] The present invention also provides a method for determining the classification of fresh cut flowers, comprising the following steps:
[0036] Ⅰ. When the fresh cut flowers pass through the grader 6, the belt 5 stops rotating, and the industrial cameras on the top and side of the grader 6 take pictures. First, the cvtcolor function is used to convert the image into a grayscale image, and then GaussianBlur is used to perform Gaussian filtering to improve the image quality. Then, the R channel image is subtracted from the G channel image to obtain a bud image without the calyx, stems and leaves. Then, the THRESH_BINARY|THRESH_OTSU method in threshold is used for binarization, and the fill contour is extracted using the findContours function. The fill contour is used as a mask to filter the background, and the bud image is extracted. The bud image of the fresh cut flower extracted by the mask has the highest stability;
[0037] Ⅱ. Since the grayscale value of the damaged area of fresh cut flowers is similar to that of the shadow gullies of the petals, the extraction of damage will be easily affected, which is also the difficulty of damage extraction. The damage is divided into yellow-brown and black damage according to color;
[0038] At this time, the split function is used to separate the three channels of the image, extract the R channel image, and perform binarization. When the threshold is 60, the black pixels in the flower buds can be extracted, and these pixels are clustered by K-means. When K=4, the black damaged area can be extracted;
[0039] The color space was converted to the HSV color space using the cvtColor function, the H channel was extracted, and the H channel image was clustered using the K-means function when K=3 to extract the yellow-brown damage area;
[0040] III. The main color of fresh cut roses is red, so cvtColor is used to convert the image to the R channel. In the R channel, the gray value with the largest pixel gray value is extracted. The higher the gray value, the brighter the flower color.
[0041] IV. By analyzing the changing characteristics of the length of each row of flower buds, it is found that before the measurement position, the number of pixels in each row of flower buds will surge several times. This is caused by the transition from the inner petals to the outer petals of the flower buds in the image. The measurement position is just before the outermost petals, after the transition from the inner petals to the outermost petals. Therefore, the number of rows of flower buds in the image is taken as the independent variable, and the number of pixels in each row of flower buds is taken as the dependent variable to establish a mathematical model.
[0042] The measurement position is obtained by analyzing the difference between the number of flower bud pixels in each row and the number of flower bud pixels in the fifth row thereafter, and the number of flower bud pixels at this position is used as the measured flower opening diameter. The degree of flower bud opening is determined by extracting the flower opening index.
[0043] Ⅴ. By viewing the flower stem from the side Figure 2 After value conversion, skeletonization is performed, and the vertical morphological operation factor is used to perform an opening operation to erase the branches and leaves, leaving only the trunk, and extract the length of the branches;
[0044] The curvature of the flower stem can be assessed by simply establishing a vertical reference line and obtaining the gap between the reference line and the flower stem. The length and curvature of the flower stem can be extracted, and the grade standards for length and curvature can be established.
[0045] Ⅵ. By extracting the image features of flower bud damage, flower color, flower opening index, stem length and flower stem inclination, and finding the relationship between the image features and each index, and combining the grading standards of each image feature, a comprehensive evaluation model for the quality of fresh cut flowers is established; after obtaining the grade of each image feature through the above method, the comprehensive quality of fresh cut flowers is determined by the lowest grade standard among the five image features of fresh flowers.
[0046] Among them, under the same lighting conditions, through the above-mentioned device and calculation method, the grayscale value distribution range of the damaged area extracted from the clustered images of 80 rose fresh cut flowers is statistically analyzed, and the grayscale value range of the black damage B component is 35-49, the grayscale value range of the G component is 34-51, and the grayscale value range of the R component is 43-64; the grayscale value distribution range of the yellow-brown damage in the H channel is 6-26, and the damage is extracted to extract the damaged area of the fresh cut flower, and the number of pixels in the damaged area of the fresh cut flower is counted to obtain the damaged area, which is then divided by the total area of the flower buds in the top view, and the proportion of the damaged area is used to represent the degree of damage to the flower buds; as well as the data judgment of flower color, flower opening index, stem length and flower stem inclination, the data of graded judgment is relatively concentrated, indicating that it is basically the same as that of manual comparison, indicating that the classification accuracy of the algorithm is high.
[0047] Through manual comparison and manual grading judgment criteria, the corresponding range of the damage ratio of the algorithm is converted as follows:
[0048] First-degree injury: injury ratio <1.25%;
[0049] Second level injury: 1.25% ≤ injury ratio < 3.17%;
[0050] Level 3 injury: 3.17% ≤ injury ratio < 6.58%;
[0051] Level 4 injury: 6.58% ≤ injury ratio < 10.67%;
[0052] Discarded flowers: damage ratio ≥ 10.67%.
[0053] The corresponding range of the R channel grayscale value converted into color grade and algorithm is as follows:
[0054] First-level color: R channel gray value>174;
[0055] Secondary flower color: 161 < R-channel gray value ≤ 174;
[0056] Tertiary flower color: 135 < R-channel gray value ≤ 161;
[0057] Quaternary flower color: 113 < R-channel gray value ≤ 135;
[0058] Discarded fresh flowers: R-channel gray value ≤ 113;
[0059] The corresponding ranges of the converted flower opening index and the flower opening diameter value measured by the algorithm are as follows:
[0060] Primary flower opening index: The flower opening diameter measured by the algorithm < 0.54;
[0061] Secondary flower opening index: 0.54 ≤ The flower opening diameter measured by the algorithm < 2.24;
[0062] Tertiary flower opening index: 2.24 ≤ The flower opening diameter measured by the algorithm < 4.13;
[0063] Quaternary flower opening index: 4.13 ≤ The flower opening diameter measured by the algorithm < 6.59;
[0064] Discarded fresh flowers: The flower opening diameter measured by the algorithm ≥ 6.59.
[0065] The grading ranges of the flower stem length are as follows:
[0066] Primary fresh flowers: Curvature < 5°;
[0067] Secondary fresh flowers: 5° ≤ Curvature < 10°;
[0068] Tertiary fresh flowers: 10° ≤ Curvature < 25°;
[0069] Quaternary fresh flowers: 25° ≤ Curvature < 45°;
[0070] Discarded fresh flowers: Curvature ≥ 40cm.
[0071] The corresponding ranges of the converted curvature and the curvature measured by the algorithm are as follows:
[0072] Primary fresh flowers: Length ≥ 65cm;
[0073] Secondary fresh flowers: 55cm ≤ Length < 65cm;
[0074] Tertiary fresh flowers: 50cm ≤ Length < 55cm;
[0075] Quaternary fresh flowers: 40cm ≤ Length < 50cm;
[0076] Discarded fresh flowers: Length < 40cm.
[0077] In summary, the present invention designs a new type of flower grading equipment to separate a large number of flowers into single branches, and transports them to the grader 6 one by one for photographing and analysis, and finally differentiates and packages them, so as to realize the automatic grading and packaging of large quantities of fresh cut flowers. In addition, relying on the newly designed algorithm, it is ensured that the positioning and measurement of the flower opening diameter measurement position have good robustness and accuracy, laying a foundation for the extraction of flowering index characteristics in the automatic grading of rose fresh cut flowers and the determination of flower storage time.
[0078] It should be noted that the above invention content and specific implementation methods are intended to demonstrate the practical application of the technical solution provided by the present invention and should not be interpreted as limiting the scope of protection of the present invention. Those skilled in the art can make various modifications, equivalent substitutions or improvements within the spirit and principles of the present invention. The scope of protection of the present invention shall be subject to the attached claims.
Claims
1. A fresh cut flower grading system, characterized in that: The invention comprises a roller (1), a conveyor belt (2), a portioner (3), a base (4), a belt (5), a classifier (6), a branch pusher (7), a sealing rod (8) and a bracket (9), wherein the roller (1) is rotatably connected to one end of the bracket (9), the conveyor belt (2) is mounted in the middle of the bracket (9), one end of the portioner (3) is fixedly connected to the other end of the bracket (9), the other end of the portioner (3) is fixedly connected to the base (4), the belt (5) is rotatably connected to the base (4), the classifier (6) is fixedly connected to one side of the base (4), and the bottom A grading box (43) is fixedly connected to the same side of the seat (4), and a plurality of branch pushers (7) are mounted one by one on the base (4) at the entrance of the grading box (43). A plurality of sealing rods (8) are hinged at the entrance of the grading box (43). A camera is arranged inside the grading box (6), and the grading box (6) and the branch pusher (7) are electrically connected. After the grading box (6) takes a picture, the image is converted into a grayscale image using the cvtcolor function, and then Gaussian filtering is performed using GaussianBlur to improve the image quality. Subsequently, the R channel image is subtracted from the G channel image to obtain a flower bud image with the calyx, branch stem and leaves removed, and the image features of the yellow-brown and black damaged areas, flower color, flower opening index, stem length and flower stem tilt degree are extracted, and the relationship between the image features and each index is found, and a comprehensive evaluation model for the quality of fresh cut flowers is established by combining the grading standards of each image feature.
2. A fresh cut flower grading system according to claim 1, characterized in that: A plurality of long strip-shaped gear teeth (11) are fixedly connected at intervals inside the roller (1), one end of the conveyor belt (2) extends into the roller (1), and both sides of the end of the conveyor belt (2) extending into the roller (1) are fixedly connected with a shifting plate (21), and the interval between the shifting plate (21) and the outer end of the gear teeth (11) is 1 to 3 mm.
3. A fresh cut flower grading system according to claim 2, characterized in that: The conveyor belt (2) is inclined in the longitudinal direction and extends into the roller (1) at the highest height. A guide rail (22) is fixedly connected to the bracket (9) above the conveyor belt (2). The guide rail (22) is distributed in a V shape along the longitudinal direction of the conveyor belt (2), and the narrowed end of the guide rail (22) points to the portioner (3).
4. A fresh cut flower grading system according to claim 3, characterized in that: The portioner (3) is inclined in the length direction, and a groove is formed in the middle of the portioner (3). A plurality of rotating wheels (31) are rotatably connected to both sides of the groove along the length direction, and the spacing between the rotating wheels (31) on the same side is between 0.5 mm and 2 mm.
5. A fresh cut flower grading system according to claim 1, characterized in that: An L-shaped branching channel (41) is provided on the top of the base (4), a short narrow portion on one side of the bent end of the branching channel (41) is communicated with the divider (3), a grading channel (42) is provided on the base (4) at the long narrow portion on the other side of the bent end of the branching channel (41), the length direction of the grading channel (42) is the same as the length direction of the base (4), the classifier (6) is mounted on the grading channel (42), and the grading box (43) and the branch pusher (7) are both arranged on both sides of the grading channel (42).
6. A fresh cut flower grading system according to claim 5, characterized in that: A plurality of V-shaped clamping claws (51) are fixedly connected to the belt (5) at intervals. The width of the clamping claw (51) is smaller than the outer diameter of the flower bud. A curved hook is provided on the outer side of the end of the clamping claw (51) that contacts the flower branch. The curved hook is bent toward the opening direction of the clamping claw (51).
7. A fresh cut flower grading system according to claim 6, characterized in that: The base (4) is inclined in the longitudinal direction and the inclination angle is between 2° and 15°.
8. A method for judging the classification of fresh cut flowers, characterized in that: The following steps are involved: Ⅰ. When the fresh cut flowers pass through the classifier (6), the belt (5) stops rotating, and the industrial cameras on the top and side of the classifier (6) take pictures, and the bud images of the fresh cut flowers are cut out through the mask; Ⅱ. Use the split function to separate the three channels of the image, extract the R channel image, and perform binarization. When the threshold is 60, the black pixels in the flower buds can be extracted, and these pixels can be clustered by K-means. When K=4, the black damaged area can be extracted; use the cvtColor function to convert the color space to the HSV color space, extract the H channel, and use the K-means function to cluster the H channel image when K=3 to extract the yellow-brown damaged area; obtain the damaged area and damage range of the fresh cut flower buds by extracting the yellow-brown and black colors, and use the damage ratio to judge the damage level; Ⅲ. Use cvtColor to convert the image to the R channel. In the R channel, extract the gray value with the largest pixel gray value. Get the flower color of the fresh cut flower buds by extracting red. After converting it into gray value, judge the flower color grade. The higher the gray value, the brighter the flower color. IV. Determine the degree of bud opening by extracting the flower opening index; V. Extract the length and curvature of the flower stems and establish the grade standards for length and curvature; VI. By extracting the image features of flower bud damage, flower color, blooming index, stem length and stem inclination, and finding the relationship between the image features and each index, and integrating the grading standards of each image feature, a comprehensive evaluation model for the quality of fresh cut flowers is established.
9. A method for determining the classification of fresh cut flowers according to claim 8, characterized in that: A mathematical model was established with the number of rows of flower buds in the image as the independent variable and the number of pixels in each row of flower buds as the dependent variable. The measurement position was obtained by analyzing the difference between the number of pixels in each row of flower buds and the number of pixels in the fifth row. The number of flower bud pixels at that position was used as the measured flower opening diameter.
