A quality grading method for potted miniature roses based on machine vision

Abstract

The present invention discloses a method for grading the quality of potted miniature roses based on machine vision, which belongs to the technical field of quality grading of potted miniature roses. It aims to solve the problems in the existing method for grading the quality of potted miniature roses, such as heavy workload and high labor intensity for workers, low grading efficiency, and lack of limit fixation for rose pots, which makes them prone to shaking and affects the accuracy of quality grading. The method for grading the quality of potted miniature roses based on machine vision is based on a detection and grading device, and comprises the following steps: S1, collecting side views and top views of the potted miniature roses, and pre-processing the images by adopting a median filtering algorithm, a Gaussian sampling method, binarization and morphological processing; a clamping mechanism is provided on a limit ring plate, and a protective mechanism is provided on the end of the sliding rod away from the slide groove a; the present invention clamps and fixes the rose pots into the limit ring plate through the clamping mechanism, and automatically grades the potted miniature roses.

Description

Technical Field

[0001] The present invention belongs to the technical field of potted miniature rose quality grading, and in particular relates to a potted miniature rose quality grading method based on machine vision. Background Art

[0002] Miniature roses, belonging to the genus Rosa in the family Rosaceae, are deciduous or semi-deciduous evergreen shrubs. They are a relatively new species in the rose family. They are characterized by their short, globular form and numerous flower heads, earning them the nickname "Diamond Rose" for their unique qualities. They are primarily used as ornamental potted plants, lawn embellishments, and decorative patterns. Their small size, unique colors, and year-round blooms make them ideal for potted planting. In the past 20 years, miniature roses have experienced rapid growth and have become the best-selling miniature potted rose in the US and Japanese markets.

[0003] At present, potted miniature roses have been commercialized, packaged and sold. With the improvement of people's living standards, the quality requirements for potted miniature roses are getting higher and higher. Therefore, quality grading of potted miniature roses is required before packaging. However, the existing method of quality grading of potted miniature roses has the following shortcomings:

[0004] ① When producing and packaging potted miniature roses, in order to obtain better product ornamental value, the potted miniature roses are generally observed one by one manually, and then manually graded before packaging. The workers have a heavy workload, high labor intensity, and low grading efficiency, which affects the overall production efficiency.

[0005] ② When grading the quality of potted miniature roses, the rose pots lack limit fixation and are prone to shaking, affecting the accuracy of quality grading.

[0006] Therefore, a quality grading method for potted miniature roses based on machine vision is needed to solve the problems of existing quality grading methods for potted miniature roses, such as heavy workload for workers, high labor intensity, low grading efficiency, and lack of limit fixation of rose pots, which makes them prone to shaking and affects the accuracy of quality grading. Summary of the Invention

[0007] The purpose of the present invention is to provide a quality grading method for potted miniature roses based on machine vision to solve the problems raised in the above background technology.

[0008] To achieve the above objectives, the present invention provides the following technical solution: a method for grading the quality of potted miniature roses based on machine vision, the method being based on a detection and grading device and comprising the following steps:

[0009] S1. Collect side and top views of potted miniature roses and pre-process the images using median filtering algorithm, Gaussian sampling method, binarization, and morphological processing;

[0010] S2. Calculate the plant height characteristics using the green binary saliency map of the side view, calculate the number of flowers, uniformity, and pest and disease status using the red binary saliency map of the top view, and calculate the flower cover using the red and green binary saliency maps of the top view, extracting features such as height, number of flowers, uniformity, flower cover, and pest and disease status respectively;

[0011] S3, using the continuous projection method to screen the characteristic vectors of the samples;

[0012] S4. Use LS-SVM as the classifier of the filtered features, train it on the sample training set, obtain the optimal classifier model, and use its predicted category on the sample prediction set as the grading level to complete the grading of the quality of potted miniature roses.

[0013] The cam is fixed on the top of the support frame, and the cam is fixed on the top of the support frame, and the cam is fixed on the bottom of the support frame.

[0014] It is further worth mentioning that the clamping mechanism includes a threaded rod threadedly connected to the side wall of the limiting annular plate, a connector plate is fixed to the outer end of the threaded rod, and a rotating handle is fixed on the side of the connector plate.

[0015] It should be further explained that the protective mechanism includes a connecting rod fixed to the end of the sliding rod away from the sliding groove a, and a cover is fixed to the end of the connecting rod away from the sliding rod, and the cover is located directly above the placement plate.

[0016] As a preferred embodiment, the bracket seat is provided with sliding grooves b located on both sides of the sliding groove a, and the side surfaces of the sliding rod are fixed with limiting columns that are slidably adapted to the sliding grooves b.

[0017] As a preferred embodiment, a rubber sleeve is provided on the surface of the rotating handle, and the length of the rubber sleeve is equal to the length of the rotating handle.

[0018] As a preferred embodiment, the inner end of the threaded rod is provided with an arc surface, and the cross-section of the handle is rectangular.

[0019] As a preferred embodiment, the area of the cover shell is smaller than the area of the placement plate, and both ends of the cover shell are opened.

[0020] Compared with the prior art, the present invention provides a method for grading the quality of potted miniature roses based on machine vision, which has at least the following beneficial effects:

[0021] (1) The flower cover is calculated by the red binary saliency map and the green binary saliency map of the top view, and the features such as height, number of flowers, uniformity, flower cover, and pest and disease status are extracted and transmitted to the display screen for display; the feature vector of the sample is screened using the continuous projection method; LS-SVM is used as a classifier for the screened features and trained on the sample training set to obtain the optimal classifier model, and its predicted category on the sample prediction set is used as the grading level to complete the grading of the quality of potted miniature roses, realizing the automatic grading of potted miniature roses with an accuracy of 98%. The grading model is relatively stable, taking into account standardization, intelligence and practicality, and improving the efficiency of potted miniature rose grading.

[0022] (2) The rose pot is clamped and fixed to the limiting ring plate through a clamping mechanism to prevent the rose pot from shaking and deflecting due to external forces during detection, thereby improving detection accuracy.

[0023] (3) The rose potted plants under inspection are protected by a cover to prevent impurities in the external environment from entering the rose potted plants, avoid the rose potted plants from shaking due to wind and affecting image acquisition, and improve the accuracy of image acquisition. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 Schematic diagram of the method flow of the present invention;

[0025] Figure 2 It is a schematic diagram of the overall structure of the present invention;

[0026] Figure 3 Schematic diagram of the local structure of camera b of the present invention;

[0027] Figure 4 This is a schematic diagram of the partial structure of the limiting annular plate of the present invention;

[0028] Figure 5 It is a schematic diagram of the local structure of the limiting column of the present invention;

[0029] Figure 6 for Figure 2 A in the middle is an enlarged structural diagram;

[0030] Figure 7 for Figure 4 Enlarged structural diagram at point B in the middle.

[0031] In the figure: 1. Clamping mechanism; 101. Threaded rod; 102. Connector plate; 103. Turning handle; 104. Arc surface; 105. Rubber sleeve; 2. Protective mechanism; 201. Cover; 202. Connecting rod; 3. Bracket; 4. Bottom plate; 5. Analyzer; 6. Display screen; 7. Base; 8. Connecting rod a; 9. Mounting plate a; 10. Placement plate; 11. Limiting ring plate; 12. Connecting rod b; 13. Mounting plate b; 14. Camera a; 15. Slide a; 16. Sliding rod; 17. Limiting column; 18. Camera b; 19. Slide b. DETAILED DESCRIPTION

[0032] The present invention will be further described below with reference to the embodiments.

[0033] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the described 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.

[0034] The following examples are intended to illustrate the present invention but are not intended to limit the scope of protection of the present invention. The conditions in the examples may be further adjusted according to specific conditions. Simple improvements to the method of the present invention within the scope of the present invention are also within the scope of protection claimed in the present invention.

[0035] See also Figure 1-7 The present invention provides a method for grading the quality of potted miniature roses based on machine vision. The method is based on a detection and grading device and includes the following steps:

[0036] S1. Collect side and top views of potted miniature roses and pre-process the images using median filtering algorithm, Gaussian sampling method, binarization, and morphological processing;

[0037] S2. Calculate the plant height characteristics using the green binary saliency map of the side view, calculate the number of flowers, uniformity, and pest and disease status using the red binary saliency map of the top view, and calculate the flower cover using the red and green binary saliency maps of the top view, extracting features such as height, number of flowers, uniformity, flower cover, and pest and disease status respectively;

[0038] S3, using the continuous projection method to screen the characteristic vectors of the samples;

[0039] S4. Use LS-SVM as the classifier of the filtered features, train it on the sample training set, obtain the optimal classifier model, and use its predicted category on the sample prediction set as the grading level to complete the grading of the quality of potted miniature roses.

[0040] Further as Figure 2 、 Figure 3 and Figure 7As shown, it is worth mentioning in detail that a method for grading the quality of potted miniature roses based on machine vision is provided. The detection and grading device includes a bracket seat 3, a base 7 is fixed at the bottom end of the bracket seat 3, a bottom plate 4 is fixed at the bottom end of the base 7, an analyzer 5 located on one side of the bracket seat 3 is placed on the top side of the bottom plate 4, a display screen 6 is provided on the analyzer 5, a placement plate 10 is fixed on the side of the bracket seat 3, a limiting annular plate 11 is fixed on the top side of the placement plate 10, a clamping mechanism 1 is provided on the limiting annular plate 11, a slide groove a15 is provided on the side of the bracket seat 3, a sliding rod 16 is slidably connected in the slide groove a15, and the sliding rod 16 is away from the slide groove a15 A protective mechanism 2 is provided at one end of the sliding rod 16, a connecting rod b12 is fixed to the top of the connecting rod b12, a mounting plate b13 is fixed to the top of the connecting rod b12, a camera a14 is fixed to the mounting plate b13, a connecting rod a8 is fixed to the side of the bracket seat 3, a mounting plate a9 is fixed to the end of the connecting rod a8 away from the bracket seat 3, a camera b18 is fixed to the mounting plate a9, and the camera b18, the camera a14 and the analyzer 5 are electrically connected to each other; first, place the rose pot on the top side of the placement plate 10 and locate it in the limiting annular plate 11, and by rotating the input end of the clamping mechanism 1, the rose pot is clamped and fixed to the limiting annular plate 11 through the clamping mechanism 1 , avoid the rose pot being shaken and offset by external forces during detection, improve detection accuracy, and protect the rose pot being detected by the protective mechanism 2 to prevent impurities in the external environment from entering the rose pot, avoid the rose pot being shaken by the wind and affecting the image acquisition, improve the accuracy of image acquisition, start the analyzer 5 to work, camera b18 and camera a14 respectively collect the side view and top view of the potted miniature rose and transmit them to the analyzer 5, the processor in the analyzer 5 adopts the median filtering algorithm, Gaussian sampling method, binarization and morphological processing to pre-process the image; through the green binary display of the side view The height characteristics of the plants are calculated by using the red binary saliency map of the top view, the number of flowers, neatness and pest and disease status are calculated by using the red binary saliency map of the top view, and the flower cover is calculated by using the red binary saliency map and the green binary saliency map of the top view. The features such as height, number of flowers, neatness, flower cover, pest and disease status are extracted and transmitted to the display screen 6 for display; the feature vector of the sample is screened using the continuous projection method; LS-SVM is used as a classifier for the screened features and trained on the sample training set to obtain the optimal classifier model, and its predicted category on the sample prediction set is used as the grading level to complete the grading of the quality of potted miniature roses.

[0041] Further as Figure 2 、 Figure 4 and Figure 6As shown, it is worth mentioning that the clamping mechanism 1 includes a threaded rod 101 threadedly connected to the side wall of the limiting annular plate 11, and a joint plate 102 is fixed to the outer end of the threaded rod 101, and a turning handle 103 is fixed to the side of the joint plate 102; the rose pot is placed on the top side of the placement plate 10 and located in the limiting annular plate 11, and the turning handle 103 is rotated. The turning handle 103 drives the joint plate 102 to rotate, and the joint plate 102 drives the threaded rod 101 to tighten, clamping the rose pot to the limiting annular plate 11, avoiding the rose pot from shaking and deflecting due to external force during detection, thereby improving the detection accuracy.

[0042] Further as Figure 2 、 Figure 3 and Figure 5 As shown, it is worth mentioning that the protective mechanism 2 includes a connecting rod 202 fixed to the end of the sliding rod 16 away from the slide groove a15, and a cover shell 201 is fixed to the end of the connecting rod 202 away from the sliding rod 16, and the cover shell 201 is located directly above the placement plate 10; when placing the rose pot, pull up the cover shell 201, and the connecting rod 202 drives the sliding rod 16 to slide in the slide groove a15, which is convenient for placing the rose pot in the limiting annular plate 11, and then put down the cover shell 201. The rose pot under inspection is protected by the cover shell 201 to prevent impurities in the external environment from entering the rose pot, avoid the problem of the rose pot plant shaking due to wind force affecting image acquisition, and improve the accuracy of image acquisition.

[0043] This solution has the following working process: first, place the rose pot on the top side of the placement plate 10 and locate it in the limiting annular plate 11, turn the handle 103, the handle 103 drives the joint plate 102 to rotate, the joint plate 102 drives the threaded rod 101 to tighten, and the rose pot is clamped and fixed in the limiting annular plate 11. When placing the rose pot, pull up the cover 201, the connecting rod 202 drives the sliding rod 16 to slide in the slide groove a15, which is convenient for placing the rose pot in the limiting annular plate 11, and then put down the cover 201. The rose pot under inspection is protected by the cover 201, and the analyzer 5 is started to work. The specific model of the analyzer 5 is W345YU6. The camera b18 and the camera a14 respectively collect the side view and the top view of the potted miniature rose and transmit them to the analyzer 5. The analyzer 5 The processor inside adopts median filtering algorithm, Gaussian sampling method, binarization and morphological processing to pre-process the image; the plant height characteristics are calculated by the green binary saliency map of the side view, the number of flowers, neatness and pest and disease status are calculated by the red binary saliency map of the top view, and the flower cover is calculated by the red binary saliency map and the green binary saliency map of the top view. The height, number of flowers, neatness, flower cover, pest and disease status and other features are extracted and transmitted to the display screen 6 for display; the continuous projection method is used to screen the feature vector of the sample; LS-SVM is used as the classifier of the screened features, and it is trained on the sample training set to obtain the optimal classifier model, and its predicted category on the sample prediction set is used as the classification level to complete the classification of the quality of potted miniature roses.

[0044] According to the above working process, it can be seen that: automatic grading of potted miniature roses is achieved with an accuracy of 98%, the grading model is relatively stable, and takes into account standardization, intelligence and practicality, thereby improving the efficiency of grading potted miniature roses. The rose pot is clamped and fixed to the limiting annular plate 11 by the clamping mechanism 1 to avoid the rose pot being shaken and deflected by external forces during detection, thereby improving the detection accuracy. The rose pot being detected is protected by the cover 201 to prevent impurities in the external environment from entering the rose pot, thereby avoiding the problem of the rose pot being shaken by the wind and affecting the image acquisition, thereby improving the accuracy of image acquisition.

[0045] Further as Figure 5 and Figure 7 As shown, it is worth mentioning that the bracket seat 3 is provided with a slide groove b19 located on both sides of the slide groove a15, and a limit column 17 that slides with the slide groove b19 is fixed on the side of the sliding rod 16; during specific operation, when the sliding rod 16 slides in the slide groove a15, the limit column 17 slides in the slide groove b19 to assist in guiding.

[0046] Further as Figure 6As shown, it is worth mentioning that a rubber sleeve 105 is provided on the surface of the handle 103, and the length of the rubber sleeve 105 is equal to the length of the handle 103; during operation, the rubber sleeve 105 effectively increases the friction when holding and rotating the handle 103, saving effort in operation.

[0047] Further as Figure 6 As shown, it is worth noting that the inner end of the threaded rod 101 is provided with an arc surface 104, and the cross-section of the handle 103 is rectangular; during specific operation, the arc surface 104 effectively reduces the wear of the inner end of the threaded rod 101.

[0048] Further as Figure 3 As shown, it is worth noting that the area of the cover shell 201 is smaller than the area of the placement plate 10, and both ends of the cover shell 201 are opened.

[0049] In summary: the automatic grading of potted miniature roses is achieved with an accuracy of 98%. The grading model is relatively stable, taking into account standardization, intelligence and practicality, and improving the efficiency of potted miniature rose grading. The rose pot is clamped and fixed to the limiting annular plate 11 by the clamping mechanism 1, so as to avoid the rose pot being shaken and offset by external forces during detection, thereby improving the detection accuracy. The rose pot being detected is protected by the cover 201 to prevent impurities in the external environment from entering the rose pot, and to avoid the problem that the rose pot plants are shaken by wind and affect image acquisition, thereby improving the accuracy of image acquisition. When the sliding rod 16 slides in the slide groove a15, the limiting column 17 slides in the slide groove b19 to assist in guiding. The provided rubber sleeve 105 effectively increases the friction when holding and rotating the handle 103, and the operation is labor-saving. The provided arc surface 104 effectively reduces the wear of the inner end of the threaded rod 101.

[0050] Unless otherwise defined, the technical or scientific terms used in the present invention shall have the ordinary meaning understood by persons having ordinary skills in the field to which the present invention belongs. The words "include" or "comprise" and the like used in the present invention mean that the elements or objects preceding the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. The words "connect" or "connected" and the like are not limited to physical or mechanical connections, but may also include electrical connections, whether direct or indirect. The words "up", "down", "left", "right", etc. are only used to indicate relative position relationships. When the absolute position of the object being described changes, the relative position relationship may also change accordingly.

[0051] While 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 these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for grading the quality of potted miniature roses based on machine vision, characterized in that: The machine vision-based potted miniature rose quality grading method is based on a detection and grading device and includes the following steps: S1. Collect side and top views of potted miniature roses and pre-process the images using median filtering algorithm, Gaussian sampling method, binarization, and morphological processing; S2. Calculate the plant height characteristics using the green binary saliency map of the side view, calculate the number of flowers, uniformity, and pest and disease status using the red binary saliency map of the top view, and calculate the flower cover using the red and green binary saliency maps of the top view, extracting features such as height, number of flowers, uniformity, flower cover, and pest and disease status respectively; S3, using the continuous projection method to screen the characteristic vectors of the samples; S4. Use LS-SVM as the classifier of the filtered features, train it on the sample training set, obtain the optimal classifier model, and use its predicted category on the sample prediction set as the grading level to complete the grading of the quality of potted miniature roses.

2. The method for grading potted miniature roses based on machine vision according to claim 1, characterized in that: The detection and grading device comprises a bracket seat (3), a base (7) is fixed at the bottom end of the bracket seat (3), a bottom plate (4) is fixed at the bottom end of the base (7), an analyzer (5) located on one side of the bracket seat (3) is placed on the top side of the bottom plate (4), a display screen (6) is provided on the analyzer (5), a placement plate (10) is fixed on the side of the bracket seat (3), a limiting annular plate (11) is fixed on the top side of the placement plate (10), a clamping mechanism (1) is provided on the limiting annular plate (11), a sliding groove a (15) is provided on the side of the bracket seat (3), and a sliding connection is provided in the sliding groove a (15). A sliding rod (16) is connected, and a protective mechanism (2) is provided at one end of the sliding rod (16) away from the slide groove a (15), a connecting rod b (12) is fixed to the top of the sliding rod (16), a mounting plate b (13) is fixed to the top of the connecting rod b (12), a camera a (14) is fixed on the mounting plate b (13), a connecting rod a (8) is fixed on the side of the bracket seat (3), a mounting plate a (9) is fixed to the end of the connecting rod a (8) away from the bracket seat (3), a camera b (18) is fixed on the mounting plate a (9), and the camera b (18), camera a (14) and analyzer (5) are electrically connected to each other.

3. The method for grading quality of potted miniature roses based on machine vision according to claim 2, characterized in that: The clamping mechanism (1) comprises a threaded rod (101) threadedly connected to the side wall of the limiting annular plate (11); a joint plate (102) is fixed to the outer end of the threaded rod (101); and a rotating handle (103) is fixed to the side of the joint plate (102).

4. The method for grading quality of potted miniature roses based on machine vision according to claim 2, wherein: The protection mechanism (2) comprises a connecting rod (202) fixed to one end of the sliding rod (16) away from the sliding groove a (15); a cover (201) is fixed to one end of the connecting rod (202) away from the sliding rod (16); and the cover (201) is located directly above the placement plate (10).

5. The method for grading quality of potted miniature roses based on machine vision according to claim 4, characterized in that: The bracket seat (3) is provided with sliding grooves b (19) located on both sides of the sliding groove a (15), and a limiting column (17) is fixed on the side of the sliding rod (16) and is slidably adapted to the sliding groove b (19).

6. The method for grading quality of potted miniature roses based on machine vision according to claim 3, characterized in that: The surface of the rotating handle (103) is covered with a rubber sleeve (105), and the length of the rubber sleeve (105) is equal to that of the rotating handle (103).

7. The method for grading quality of potted miniature roses based on machine vision according to claim 3, characterized in that: The inner end of the threaded rod (101) is provided with an arc surface (104), and the cross section of the rotating handle (103) is rectangular.

8. The method for grading quality of potted miniature roses based on machine vision according to claim 4, characterized in that: The area of the cover shell (201) is smaller than the area of the placement plate (10), and both ends of the cover shell (201) are open.

Images