Method and system for counting in-vitro camellia oleifera pollen germination rate and germination length

Abstract

The invention relates to a method and system for counting the germination rate and germination length of in-vitro camellia oleifera pollen, and the method comprises the steps: S1, collecting camellia oleifera pollen, carrying out in-vitro culture, obtaining a microscopic image of a camellia oleifera pollen tube, and dividing the microscopic image into a training image and a to-be-detected image; s2, processing the training image in different labeling modes to obtain a detection label for counting the germination rate and a segmentation label for counting the germination length; s3, training by using the labeled training image, the detection label and the segmentation label to obtain an artificial intelligence model for detection and segmentation; s4, inputting the to-be-detected image into an artificial intelligence model for detection and segmentation to obtain a pollen grain counting result and a real length of a germination pollen tube; s5, according to the counting result of the pollen grains and the real length of the germination pollen tube, the germination rate and the germination length of the camellia oleifera pollen are measured and counted. The technical problems of tedious image processing, large workload and the like after germination of camellia oleifera pollen tubes are solved.

Description

Technical Field

[0001] The invention belongs to the field of combining agricultural biology with artificial intelligence, and in particular relates to a method and system for counting the germination rate and germination length of in vitro camellia oleifera pollen. Background Art

[0002] Camellia oleifera is one of the most important edible oil tree species in China, widely distributed in southern China. It has rich oil production, significant economic value, high nutritional and medicinal value, and is a pure natural green health food with antibacterial and anti-inflammatory, antioxidant, and blood sugar and lipid lowering effects. The domestic Camellia oleifera planting area exceeds 4.4 million hm2. 2 , with an annual output of about 600,000 tons of camellia oil. Camellia oleifera is one of the economic tree species with the largest cultivation area and the most widespread distribution in China.

[0003] In production, camellia oleifera generally has the phenomenon of more flowers and fewer fruits, and low fruit setting rate. The main reason for this phenomenon is that camellia oleifera has the characteristic of late-acting self-incompatibility (LSI), that is, although the self-pollination pollen tube can pass through the style, it eventually stays in the ovary and cannot reach the embryo sac to release sperm and complete double fertilization. Therefore, the self-pollination fruit setting rate of most varieties is low. In production, it is necessary to reasonably configure pollination trees or take artificial assisted pollination measures during the flowering period to obtain the expected yield and quality. In the early stage, the project team determined through plant microscopic observation, multi-omics and other technologies that LSI of Camellia oleifera is a process of programmed cell death (PCD) of self-pollinated pollen tubes, in which a variety of hormones such as abscisic acid (ABA), salicylic acid (SA), jasmonic acid (JA), indole acetic acid (IAA), gibberellic acid (GA3), and cytokinin (CTK) are involved. Therefore, in order to explore the specific role of plant hormones in the growth of Camellia oleifera pollen tubes, the project team carried out an in vitro culture experiment of Camellia oleifera pollen under a variety of exogenous hormones.

[0004] Among them, the germination rate and germination length of pollen tubes were used as reference indicators to determine the effects of different hormones on the growth of Camellia oleifera pollen tubes. In order to calculate the germination rate and germination length of Camellia oleifera pollen tubes, the growth state of pollen tubes under each treatment was observed and photographed using a microscope. For the uniformity of the data, at least 5 fields of view need to be collected for the same treatment, and the number of pollens counted in each field of view is not less than 30. The current experimental method is to use the Cell Counter tool of ImageJ software to obtain the germination rate data of Camellia oleifera pollen, use the broken line marking function of ImageJ software to obtain the pixel length of Camellia oleifera pollen tubes, and convert it to the actual length according to the scale to obtain the germination length data. Generally speaking, the field of view of the microscope is small, and there are many Camellia oleifera pollen grains and the pollen tube length is long, so the processing and analysis of images involves a lot of boring and repetitive labor, which greatly occupies the research time of scientific researchers and slows down the research progress. Therefore, there is an urgent need for an efficient and accurate method to count the germination rate and germination length of in vitro Camellia oleifera pollen. Summary of the invention

[0005] The present invention provides a method and system for statistically analyzing the germination rate and length of in vitro Camellia oleifera pollen, which solves the technical problems of cumbersome image processing and heavy workload after collecting Camellia oleifera pollen tube germination. By adopting a deep learning solution, artificial intelligence technology is used to automatically detect and segment pollen tube microscopic images, thereby obtaining the germination rate and germination length data after in vitro pollen culture.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0007] In a first aspect, a method for calculating the germination rate and germination length of in vitro camellia oleifera pollen is provided, comprising:

[0008] S1, collecting Camellia oleifera pollen and culturing it in vitro, obtaining microscopic images of Camellia oleifera pollen tubes, and dividing the microscopic images into training images and test images;

[0009] S2, processing the training images in different labeling methods to obtain detection labels for counting germination rates and segmentation labels for counting germination lengths;

[0010] S3, using the annotated training images, detection labels and segmentation labels to train an artificial intelligence model for detection and segmentation;

[0011] S4, inputting the image to be tested into an artificial intelligence model for detection and segmentation, obtaining the number of germinated pollen grains, the number of ungerminated pollen grains and the pixel length of germinated pollen tubes, obtaining the pollen grain counting result according to the number of germinated pollen grains and the number of ungerminated pollen grains, and calculating the actual length of the germinated pollen tube according to the pixel length of the germinated pollen tube;

[0012] S5. According to the pollen grain counting results and the actual length of the germinated pollen tube, the germination rate and germination length of Camellia oleifera pollen were determined and counted.

[0013] Further, step S1, collecting Camellia oleifera pollen and culturing in vitro, obtaining microscopic images of Camellia oleifera pollen tubes, and dividing the microscopic images into training images and test images, including:

[0014] S11, collecting Camellia oleifera pollen and storing it at low temperature;

[0015] S12, sowing the Camellia oleifera pollen in a pollen culture dish prepared with Camellia oleifera in vitro pollen suspension culture solution, fully mixing, and then placing in a constant temperature incubator in the dark for incubation;

[0016] S13, placing the current pollen culture dish under a microscope, adjusting the focal length of the microscope so that the pollen grains and pollen tubes in the current pollen culture dish can be clearly imaged under the microscope, and moving the current pollen culture dish to take pictures in multiple fields of view until microscopic images of all central areas of the current pollen culture dish are collected;

[0017] S14, replacing the pollen culture dish that has not been photographed, repeating step S13, and collecting the microscopic images of all pollen culture dishes;

[0018] S15, storing all the microscopic images of the pollen culture dishes in order, and randomly selecting and dividing the microscopic images into training images and test images.

[0019] Furthermore, in step S2, the training images are processed in different labeling methods to obtain detection labels for counting germination rates and segmentation labels for counting germination lengths, including:

[0020] S21, using image annotation software to select and annotate pollen grains in the training image using a rectangular frame tool, and setting label categories for ungerminated pollen grains and label categories for germinated pollen grains, respectively, to obtain detection labels for statistical germination rates;

[0021] S22, using image annotation software, using polygon tools to enclose and annotate the pollen tubes in the training image, and setting the label category normal for pollen tubes with clear images and not crossing other pollen tubes, the label category cross for pollen tubes with clear images and crossing other pollen tubes, and the label category error for pollen tubes with unclear images or beyond the focal length of the lens, to obtain segmentation labels.

[0022] Furthermore, in step S3, an artificial intelligence model for detection and segmentation is obtained by training using the annotated training images, detection labels, and segmentation labels, including:

[0023] S31, constructing a germination rate detection model for statistically analyzing pollen germination rate;

[0024] S32, using the annotated training images and detection labels to input into the germination rate detection model for training, and saving the best detection model weights for outputting the detection results of the pollen tubes;

[0025] S33, building an instance segmentation model for counting pollen tube length;

[0026] S34, using the annotated training images and segmentation labels to input into the instance segmentation model for training, and saving the best segmentation model weights for outputting the pollen tube segmentation results;

[0027] S35, randomly select some images to be tested, input them into the model trained in step S32 and step S34, and obtain new detection labels and new segmentation labels; after adjusting the detection labels and segmentation labels according to the new detection labels and new segmentation labels, continue to input them into the model in step S32 and step S34 for training;

[0028] S36, repeat step S35 until the germination rate detection model and the instance segmentation model reach the preset model performance, and obtain the artificial intelligence model for detection and segmentation.

[0029] Further, step S4, inputting the image to be tested into an artificial intelligence model for detection and segmentation, obtaining the number of germinated pollen grains, the number of ungerminated pollen grains and the pixel length of germinated pollen tubes, obtaining the pollen grain counting result according to the number of germinated pollen grains and the number of ungerminated pollen grains, and calculating the actual length of the germinated pollen tube according to the pixel length of the germinated pollen tube, including:

[0030] S41, inputting the image to be tested into an artificial intelligence model for detection and segmentation to obtain a target detection label and a target segmentation label of the image to be tested;

[0031] S42, screening the target detection label and the target segmentation label according to the preset label specification;

[0032] S43, counting the screened target detection labels to obtain label categories of the number of germinated pollen grains and the number of ungerminated pollen grains, and obtaining a pollen grain counting result according to the number of germinated pollen grains and the number of ungerminated pollen grains;

[0033] S44, selecting the target pollen tubes with the label categories of normal and cross in the filtered target segmentation labels, drawing a mask image of each target pollen tube using a polygon label, and processing the mask image using a thinning algorithm to obtain the pixel length of each target pollen tube;

[0034] S45, converting the pixel length according to the ratio between the image pixels and the real length of the image to be measured to obtain the real length of each target pollen tube.

[0035] In a second aspect, a system for counting the germination rate and germination length of in vitro camellia oleifera pollen is provided, comprising:

[0036] A microscopic image acquisition module is used to collect Camellia oleifera pollen and culture it in vitro, obtain microscopic images of Camellia oleifera pollen tubes, and divide the microscopic images into training images and images to be tested;

[0037] A microscopic image annotation processing module is used to process the training images in different annotation methods to obtain detection labels for statistical germination rates and segmentation labels for statistical germination lengths;

[0038] A model training module is used to train an artificial intelligence model for detection and segmentation using annotated training images, detection labels, and segmentation labels;

[0039] A calculation module based on an artificial intelligence model is used to input the image to be tested into the artificial intelligence model for detection and segmentation, obtain the number of germinated pollen grains, the number of ungerminated pollen grains and the pixel length of germinated pollen tubes, obtain the pollen grain counting result according to the number of germinated pollen grains and the number of ungerminated pollen grains, and calculate the real length of the germinated pollen tube according to the pixel length of the germinated pollen tube;

[0040] The germination rate and germination length statistics module is used to measure and count the germination rate and germination length of Camellia oleifera pollen based on the pollen grain counting results and the actual length of the germinated pollen tube.

[0041] Furthermore, the microscopic image acquisition module includes:

[0042] Camellia oleifera pollen collection unit, used for collecting Camellia oleifera pollen and preserving it at low temperature;

[0043] The in vitro culture unit is used to sow the oil-tea camellia pollen in a pollen culture dish prepared with an in vitro oil-tea camellia pollen suspension culture solution, mix thoroughly, and then place in a constant temperature incubator to culture in the dark;

[0044] A microscopic image capturing unit is used to place the current pollen culture dish under a microscope, adjust the focal length of the microscope so that the pollen grains and pollen tubes in the current pollen culture dish can be clearly imaged under the microscope, move the current pollen culture dish to capture multiple fields of view, until the microscopic images of all the central areas of the current pollen culture dish are collected; replace the pollen culture dish that has not been completed, and complete the collection of microscopic images of all pollen culture dishes;

[0045] The microscopic image division unit is used to store the microscopic images of all pollen culture dishes in order, and randomly select and divide the microscopic images into training images and images to be tested.

[0046] Furthermore, the microscopic image annotation processing module includes:

[0047] A pollen grain labeling processing unit is used to select and label the pollen grains in the training image using an image labeling software and a rectangular frame tool, and to set label categories for ungerminated pollen grains and label categories for germinated pollen grains, respectively, to obtain detection labels for statistical germination rates;

[0048] The pollen tube annotation processing unit is used to use the image annotation software to enclose and annotate the pollen tubes in the training image using a polygon tool, and set the label category normal for pollen tubes with clear imaging and not crossing other pollen tubes, the label category cross for pollen tubes with clear imaging and crossing other pollen tubes, and the label category error for pollen tubes with unclear imaging or beyond the focal length of the lens to obtain segmentation labels.

[0049] Furthermore, the model training module includes:

[0050] A germination rate detection model building unit, used to build a germination rate detection model for counting pollen germination rate;

[0051] A germination rate detection model training unit is used to input the labeled training images and detection labels into the germination rate detection model for training, and save the best detection model weights for outputting the pollen tube detection results;

[0052] An instance segmentation model building unit, used to build an instance segmentation model for counting pollen tube lengths;

[0053] An instance segmentation model training unit is used to input the labeled training images and segmentation labels into the instance segmentation model for training, and save the best segmentation model weights for outputting the pollen tube segmentation results;

[0054] The model optimization unit is used to randomly select some images to be tested, input them into the trained models in the germination rate detection model training unit and the instance segmentation model training unit, and obtain new detection labels and new segmentation labels; after adjusting the detection labels and segmentation labels according to the new detection labels and the new segmentation labels, continue to send them into the models in the germination rate detection model training unit and the instance segmentation model training unit for training; repeat the execution until the germination rate detection model and the instance segmentation model reach the preset model performance, and obtain the artificial intelligence model for detection and segmentation.

[0055] Furthermore, the computing module based on the artificial intelligence model includes:

[0056] A label acquisition unit, used to input the image to be tested into an artificial intelligence model for detection and segmentation, and obtain an object detection label and an object segmentation label of the image to be tested;

[0057] A label screening unit, used to screen target detection labels and target segmentation labels according to preset label specifications;

[0058] The pollen grain counting unit is used to count the selected target detection labels, obtain the label categories of the number of germinated pollen grains and the number of ungerminated pollen grains, and obtain the pollen grain counting result according to the number of germinated pollen grains and the number of ungerminated pollen grains;

[0059] The pollen tube pixel length statistical unit is used to select the target pollen tubes with the label categories of normal and cross in the filtered target segmentation labels, draw the mask image of each target pollen tube using the polygon label, and use the thinning algorithm to process the mask image to obtain the pixel length of each target pollen tube;

[0060] The pollen tube real length calculation unit is used to convert the pixel length according to the ratio between the image pixels and the real length of the image to be measured, so as to obtain the real length of each target pollen tube.

[0061] The beneficial effects achieved by the present invention are:

[0062] Camellia oleifera pollen is collected and cultured in vitro to obtain microscopic images of the pollen tubes of the camellia oleifera, and the microscopic images are divided into training images and images to be tested; the training images are processed in different annotation methods to obtain detection labels for counting germination rates and segmentation labels for counting germination lengths; an artificial intelligence model for detection and segmentation is trained using the annotated training images, detection labels and segmentation labels; the images to be tested are input into the artificial intelligence model for detection and segmentation to obtain the number of germinated pollen grains, the number of ungerminated pollen grains and the pixel length of germinated pollen tubes, the pollen grain counting result is obtained based on the number of germinated pollen grains and the number of ungerminated pollen grains, and the true length of the germinated pollen tubes is calculated based on the pixel length of the germinated pollen tubes; the germination rate and germination length of the camellia oleifera pollen are determined and counted based on the pollen grain counting result and the true length of the germinated pollen tubes; the technical problems of cumbersome image processing and heavy workload after collecting the germination of the camellia oleifera pollen tubes are solved. By adopting a deep learning solution and using artificial intelligence technology to automatically detect and segment pollen tube microscopic images, we can obtain the germination rate and germination length data after in vitro pollen culture. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 A flow chart of the method for counting the germination rate and germination length of in vitro camellia oleifera pollen according to the present invention;

[0064] Figure 2A schematic diagram of the rectangular frame marking and classification of the detection label of the present invention;

[0065] Figure 3 A schematic diagram of the contents of the detection label of the present invention;

[0066] Figure 4 A schematic diagram of the segmentation labeling and classification of the present invention;

[0067] Figure 5 A structural diagram of a system for counting the germination rate and germination length of in vitro camellia oleifera pollen according to the present invention;

[0068] Figure 6 is a structural diagram of a microscopic image acquisition module of the present invention;

[0069] Figure 7 This is a structural diagram of the model training module of the present invention;

[0070] Figure 8 This is a structural diagram of the computing module based on the artificial intelligence model of the present invention. DETAILED DESCRIPTION

[0071] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0072] like Figure 1 As shown, the embodiment of the present invention provides a method for counting the germination rate and germination length of in vitro Camellia oleifera pollen, comprising:

[0073] S1, collecting Camellia oleifera pollen and culturing it in vitro, obtaining microscopic images of Camellia oleifera pollen tubes, and dividing the microscopic images into training images and test images;

[0074] In this embodiment, the specific process of in vitro culturing and photographing Camellia oleifera pollen is as follows:

[0075] S11, collect Camellia oleifera pollen and store it at low temperature; collect pollen from Camellia oleifera ‘Huashuo’ during its flowering period, select the flower buds at the whitening stage, gently pull off the anthers with alcohol-sterilized pointed tweezers, spread them flat on sulfuric acid paper, dry them at room temperature 26℃ for 14 h, collect the pollen in a 1.5 mL centrifuge tube after the anthers are fully cracked, and store them in a -80℃ refrigerator;

[0076] S12, sowing the camellia oleifera pollen in a pollen culture dish prepared with an in vitro camellia oleifera pollen suspension culture solution, mixing thoroughly and placing in a constant temperature incubator for culture in the dark; preparing an in vitro camellia oleifera pollen suspension culture system, the components of which include 10% sucrose and 0.015% boric acid, with a pH of 5.8; taking out the pollen collected in step S11 in advance, and thawing it at room temperature for 2 hours; adding different amounts of 10mM abscisic acid (ABA) mother solution to the culture solution, pouring 10 mL of culture solution into each culture dish, so that the final concentrations of ABA treatment are: 0 μM, 10μM, 30μM, 50μM, 70μM, 90μM, respectively, and each treatment is repeated 3 times, and the camellia oleifera pollen is evenly shaken off in the culture solution of different treatment concentrations with a brush, mixing thoroughly and placing in a constant temperature incubator at 25°C for culture in the dark for 2 hours;

[0077] S13, placing the current pollen culture dish under a microscope, adjusting the focal length of the microscope so that the pollen grains and pollen tubes in the current pollen culture dish can be clearly imaged under the microscope, and moving the current pollen culture dish to take pictures in multiple fields of view until microscopic images of all central areas of the current pollen culture dish are collected;

[0078] After the pollen in the culture dish in the constant temperature incubator stops growing at low temperature, turn on the power of the microscope, use the OLYMPUSDP71 lens, adjust the ISO of the camera to "200", adjust the exposure time to the "auto" mode, place the pollen culture dish after in vitro culture in step S12 under the microscope, adjust the focal length of the microscope so that the pollen and pollen tubes can be clearly imaged under the microscope, move the current pollen culture dish, ensure that there are more than 30 pollen grains in each field of view, click the "shoot" button to shoot, and after shooting, move the current pollen culture dish to obtain a new field of view, and repeat the shooting steps until images of all central areas of the current culture dish are collected;

[0079] S14, replacing the pollen culture dish that has not been photographed, repeating step S13, and collecting the microscopic images of all pollen culture dishes;

[0080] S15, storing all the microscopic images of the pollen culture dishes in order, and randomly selecting and dividing the microscopic images into training images and test images;

[0081] According to the experimental conditions of the culture: oil tea variety, hormone type, hormone concentration, three dimensions are stored in folders, and the image files are renamed and classified according to the naming rules of "oil tea variety_hormone type_hormone concentration". For example, the above culture method can be named "'Asus'_ABA_70μM". After all the microscopic images of the pollen culture dishes are stored in order, they are randomly selected to be divided into training images for training the model and test images for testing after the model is trained.

[0082] S2, processing the training images in different annotation methods to obtain detection labels for counting germination rates and segmentation labels for counting germination lengths;

[0083] In this embodiment, the germination rate and germination length need to be marked in different ways:

[0084] S21, use the image annotation software (Labelme) to select and annotate the pollen grains in the training image using the rectangular box tool, change the label category of the ungerminated pollen grains to "0", change the label category of the germinated pollen grains to "1", export the detection annotation file in YOLO format, and obtain the detection label for statistical germination rate; Figure 2 As shown in , the label of the microscopic image should include: the label category of each pollen grain and the image resolution size; Figure 3 As shown, the ratio of the upper left position of the rectangular frame of each pollen grain to the aspect ratio of the whole image and the ratio of the width and height of the rectangular frame to the aspect ratio of the whole image;

[0085] S22, using the image annotation software (Labelme), the pollen tubes in the training image are enclosed and annotated using the polygon tool, and the label categories of the pollen tubes with clear imaging and not crossing other pollen tubes are set as normal, the label category of the pollen tubes with clear imaging and crossing other pollen tubes is set as cross, and the label category of the pollen tubes with unclear imaging or beyond the focal length of the lens is set as error, and the segmentation labels are obtained; Figure 4 As shown, the pollen tubes that grow out of the lens field of view also have their label categories changed to normal or cross according to the situation, and the instance segmentation annotation file in COCO format is exported, where the label of each microscopic image includes the image resolution size, the label category of the pollen tube, and the absolute coordinates of each point of the pollen tube polygon box.

[0086] S3, using the annotated training images, detection labels and segmentation labels to train an artificial intelligence model for detection and segmentation;

[0087] In this embodiment, the specific process of training the model is:

[0088] S31, using Python programming language as the basis, calling third-party Python libraries such as pytorch and ultralytics, to build an improved YOLOv8x detection model (i.e., germination rate detection model) for counting pollen germination rate;

[0089] S32, using the annotated training images and the detection labels in the YOLO format to input into the germination rate detection model for training, the specific training parameters are training for 500 cycles until the training loss value converges, and saving the best detection model weights for outputting the detection results of the pollen tubes;

[0090] S33, using Python programming language as the basis, calling third-party Python libraries such as pytorch and detectron2, to build an improved EVA instance segmentation model (i.e., instance segmentation model) for counting pollen tube length;

[0091] S34, using the annotated training images and the segmentation labels in COCO format to input into the instance segmentation model for training, the specific training parameters are training for 200,000 cycles until the training loss value converges, and saving the best segmentation model weights for outputting the pollen tube segmentation results;

[0092] S35, randomly select some images to be tested, input them into the models trained in step S32 and step S34, and obtain new detection labels and new segmentation labels; according to the new detection labels and new segmentation labels, use labelme software to manually adjust the detection labels and segmentation labels, and continue to input them into the models in step S32 and step S34 for training;

[0093] S36, repeating step S35, recording the final PR curve of each training process of the germination rate detection model, where P is the precision rate, R is the recall rate, , , where TP represents the number of correctly detected pollen tubes that have germinated, FP represents the number of incorrectly detected pollen tubes that have germinated, and FN represents the number of incorrectly detected pollen tubes that have not germinated. When the area enclosed by the PR curve is close to 1, it indicates that the germination rate detection model has excellent performance.

[0094] Record the final average intersection-over-union value of the instance segmentation model during each training process. The formula is:

[0095] ,

[0096] Among them, TP is correct and the pixel belongs to the current pollen tube, FN is wrong and the pixel does not belong to the current pollen tube, FP is wrong and the pixel belongs to the current pollen tube, k is the total number of pollen tubes of all categories in the current microscopic image, and the closer the average intersection-over-union ratio is to 1, the better the instance segmentation model has achieved.

[0097] S4, inputting the image to be tested into an artificial intelligence model for detection and segmentation, obtaining the number of germinated pollen grains, the number of ungerminated pollen grains and the pixel length of germinated pollen tubes, obtaining the pollen grain counting result according to the number of germinated pollen grains and the number of ungerminated pollen grains, and calculating the actual length of the germinated pollen tube according to the pixel length of the germinated pollen tube;

[0098] In this embodiment, the specific execution steps are as follows:

[0099] S41, inputting the image to be tested into an artificial intelligence model for detection and segmentation to obtain a target detection label and a target segmentation label of the image to be tested;

[0100] S42, screening the target detection label and the target segmentation label according to the preset label specification;

[0101] Specific screening includes clearing the rectangular boxes in the detection labels that are close to the edge of the image, clearing the pollen tube labels with the label "error" in the instance segmentation labels, the pollen tube labels whose polygon point coordinates are close to the edge of the image, and the pollen tube labels whose pixel length is less than 1 times the pixel diameter of the pollen image; thereby reducing the interference of non-compliant labels on the results;

[0102] S43, counting the screened target detection labels to obtain label categories of the number of germinated pollen grains and the number of ungerminated pollen grains, and obtaining a pollen grain counting result according to the number of germinated pollen grains and the number of ungerminated pollen grains;

[0103] S44, selecting the target pollen tubes with the label categories of normal and cross in the filtered target segmentation labels, drawing a mask image of each target pollen tube using a polygon label, and processing the mask image using a thinning algorithm to obtain the pixel length of each target pollen tube;

[0104] S45, converting the pixel length according to the ratio between the image pixels and the real length of the image to be measured to obtain the real length of each target pollen tube;

[0105] The formula for calculating the true length is:

[0106] ;

[0107] The ratio of pixel length to real length is calculated by measuring the pixel length of the micrometer by taking a micrometer photograph under the microscope.

[0108] In the above steps S43 and S45, the pollen grain counting results and the data of the actual length of the pollen tube are obtained. In order to facilitate intuitive viewing, third-party python libraries such as opencv-python, Pillow, and numpy can also be used to generate display images. Rectangular frames are drawn according to the detection labels and superimposed on the original image for display. The rectangular frames are divided into 2 groups according to the labels. In each group, they are sorted from left to right and from top to bottom according to the coordinates of the upper left corner. The rectangular frames labeled "germinated" are drawn in a special color (for example, red) and a counting mark is affixed, and the rectangular frames labeled "ungerminated" are drawn in a special color (for example, blue) and a counting mark is affixed, and the output is a picture.

[0109] S5. According to the pollen grain counting results and the actual length of the germinated pollen tube, the germination rate and germination length of Camellia oleifera pollen were determined and counted.

[0110] In this embodiment, the germination rate calculation formula is:

[0111] ;

[0112] The obtained germination rate and pollen tube germination length were screened, and the outliers were removed as the final results. The mean and variance were calculated to facilitate researchers' subsequent data analysis.

[0113] The beneficial effects of the embodiments of the present invention are:

[0114] Camellia oleifera pollen is collected and cultured in vitro to obtain microscopic images of the pollen tubes of the camellia oleifera, and the microscopic images are divided into training images and images to be tested; the training images are processed in different annotation methods to obtain detection labels for counting germination rates and segmentation labels for counting germination lengths; an artificial intelligence model for detection and segmentation is trained using the annotated training images, detection labels and segmentation labels; the images to be tested are input into the artificial intelligence model for detection and segmentation to obtain the number of germinated pollen grains, the number of ungerminated pollen grains and the pixel length of germinated pollen tubes, the pollen grain counting result is obtained based on the number of germinated pollen grains and the number of ungerminated pollen grains, and the true length of the germinated pollen tubes is calculated based on the pixel length of the germinated pollen tubes; the germination rate and germination length of the camellia oleifera pollen are determined and counted based on the pollen grain counting result and the true length of the germinated pollen tubes; the technical problems of cumbersome image processing and heavy workload after collecting the germination of the camellia oleifera pollen tubes are solved. By adopting a deep learning solution and using artificial intelligence technology to automatically detect and segment pollen tube microscopic images, we can obtain the germination rate and germination length data after in vitro pollen culture.

[0115] In combination with the method for counting the germination rate and germination length of in vitro camellia oleifera pollen described in the above embodiments, the system for counting the germination rate and germination length of in vitro camellia oleifera pollen is described below through an embodiment.

[0116] like Figure 5 As shown, the embodiment of the present invention provides a system for counting the germination rate and germination length of in vitro camellia oleifera pollen, comprising:

[0117] The microscopic image acquisition module 501 is used to collect Camellia oleifera pollen and culture it in vitro, obtain microscopic images of Camellia oleifera pollen tubes, and divide the microscopic images into training images and images to be tested;

[0118] The microscopic image annotation processing module 502 is used to process the training images in different annotation modes to obtain detection labels for counting germination rates and segmentation labels for counting germination lengths;

[0119] A model training module 503 is used to train an artificial intelligence model for detection and segmentation using the labeled training images, detection labels and segmentation labels;

[0120] The calculation module 504 based on the artificial intelligence model is used to input the image to be tested into the artificial intelligence model for detection and segmentation, obtain the number of germinated pollen grains, the number of ungerminated pollen grains and the pixel length of the germinated pollen tube, obtain the pollen grain counting result according to the number of germinated pollen grains and the number of ungerminated pollen grains, and calculate the real length of the germinated pollen tube according to the pixel length of the germinated pollen tube;

[0121] The germination rate and germination length statistics module 505 is used to measure and count the germination rate and germination length of Camellia oleifera pollen according to the pollen grain counting results and the actual length of the germinated pollen tube.

[0122] Combination Figure 5 The embodiment shown is preferably as Figure 6 As shown, in some embodiments of the present invention, the microscopic image acquisition module 501 includes:

[0123] Camellia oleifera pollen collection unit 601, used for collecting Camellia oleifera pollen and storing it at low temperature;

[0124] The in vitro culture unit 602 is used to spread the Camellia oleifera pollen in a pollen culture dish prepared with Camellia oleifera in vitro pollen suspension culture solution, mix thoroughly, and then place in a constant temperature incubator for culture in the dark;

[0125] The microscopic image capturing unit 603 is used to place the current pollen culture dish under a microscope, adjust the focal length of the microscope so that the pollen grains and pollen tubes in the current pollen culture dish can be clearly imaged under the microscope, move the current pollen culture dish to capture multiple fields of view, until the microscopic images of all the central areas of the current pollen culture dish are collected; replace the pollen culture dish that has not been completed, and complete the collection of microscopic images of all pollen culture dishes;

[0126] The microscopic image division unit 604 is used to store the microscopic images of all pollen culture dishes in order, and randomly select and divide the microscopic images into training images and images to be tested.

[0127] Combination Figure 5 In the embodiment shown, preferably, in some embodiments of the present invention, the microscopic image annotation processing module 502 includes:

[0128] A pollen grain labeling processing unit is used to select and label the pollen grains in the training image using an image labeling software and a rectangular frame tool, and to set label categories for ungerminated pollen grains and label categories for germinated pollen grains, respectively, to obtain detection labels for statistical germination rates;

[0129] The pollen tube annotation processing unit is used to use the image annotation software to enclose and annotate the pollen tubes in the training image using a polygon tool, and set the label category normal for pollen tubes with clear imaging and not crossing other pollen tubes, the label category cross for pollen tubes with clear imaging and crossing other pollen tubes, and the label category error for pollen tubes with unclear imaging or beyond the focal length of the lens to obtain segmentation labels.

[0130] Combination Figure 5 The embodiment shown is preferably as Figure 7 As shown, in some embodiments of the present invention, the model training module 503 includes:

[0131] A germination rate detection model building unit 701 is used to build a germination rate detection model for counting pollen germination rate;

[0132] A germination rate detection model training unit 702 is used to input the labeled training images and detection labels into the germination rate detection model for training, and save the best detection model weights for outputting the detection results of pollen tubes;

[0133] An instance segmentation model building unit 703 is used to build an instance segmentation model for counting pollen tube lengths;

[0134] An instance segmentation model training unit 704 is used to input the labeled training images and segmentation labels into the instance segmentation model for training, and save the best segmentation model weights for outputting the pollen tube segmentation results;

[0135] The model optimization unit 705 is used to randomly select part of the images to be tested, input them into the trained models in the germination rate detection model training unit and the instance segmentation model training unit, and obtain new detection labels and new segmentation labels; after adjusting the detection labels and segmentation labels according to the new detection labels and the new segmentation labels, continue to send them into the models in the germination rate detection model training unit and the instance segmentation model training unit for training; repeat the execution until the germination rate detection model and the instance segmentation model reach the preset model performance, and obtain the artificial intelligence model for detection and segmentation.

[0136] Combination Figure 5 The embodiment shown is preferably as Figure 8 As shown, in some embodiments of the present invention, the calculation module 504 based on the artificial intelligence model includes:

[0137] The label acquisition unit 801 is used to input the image to be tested into the artificial intelligence model for detection and segmentation to obtain the target detection label and target segmentation label of the image to be tested;

[0138] The label screening unit 802 is used to screen the target detection label and the target segmentation label according to the preset label specification;

[0139] The pollen grain counting unit 803 is used to count the selected target detection labels, obtain the label categories of the number of germinated pollen grains and the number of ungerminated pollen grains, and obtain the pollen grain counting result according to the number of germinated pollen grains and the number of ungerminated pollen grains;

[0140] The pollen tube pixel length counting unit 804 is used to select the target pollen tubes with the label categories of normal and cross in the filtered target segmentation labels, draw the mask image of each target pollen tube using the polygon label, and process the mask image using the thinning algorithm to obtain the pixel length of each target pollen tube;

[0141] The pollen tube real length calculation unit 805 is used to convert the pixel length according to the ratio between the image pixels and the real length of the image to be measured, so as to obtain the real length of each target pollen tube.

[0142] In summary, the beneficial effects of the system for counting the germination rate and germination length of in vitro camellia oleifera pollen according to the embodiment of the present invention are as follows:

[0143] The trained AI model uses established rules to automate the processing of image data, reducing human interference and improving data accuracy and reliability.

[0144] The existing methods are limited to manual identification and data acquisition, so image processing and analysis involve a lot of boring and repetitive work, which greatly takes up the research time of scientific researchers and slows down the research progress. The present invention can automatically obtain experimental data through a program after image acquisition, greatly improving the data analysis efficiency of scientific researchers engaged in related research.

[0145] The present invention uses less time to process a larger amount of data. Traditional methods may choose a smaller sample size when processing a large amount of data due to time and resource constraints, or may even select and screen image data that is difficult to process in order to reduce the workload, affecting the representativeness and depth of the statistical results. The present invention can quickly process a large amount of data, allowing researchers to perform analysis based on a larger sample size, thereby obtaining more comprehensive and in-depth statistical results.

[0146] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0147] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0148] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0149] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0150] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the scope of the claims of the present invention to be approved.

Claims

1. A method for calculating the germination rate and germination length of in vitro Camellia oleifera pollen, characterized in that: include: S1, collecting Camellia oleifera pollen and culturing it in vitro, obtaining a microscopic image of the Camellia oleifera pollen tube, and dividing the microscopic image into a training image and a test image; S2, processing the training images in different labeling methods to obtain detection labels for counting germination rates and segmentation labels for counting germination lengths; S3, using the labeled training image, the detection label and the segmentation label to train an artificial intelligence model for detection and segmentation; S4, inputting the image to be tested into the artificial intelligence model for detection and segmentation, obtaining the number of germinated pollen grains, the number of ungerminated pollen grains and the pixel length of germinated pollen tubes, obtaining a pollen grain counting result according to the number of germinated pollen grains and the number of ungerminated pollen grains, and calculating the actual length of the germinated pollen tube according to the pixel length of the germinated pollen tube; S5, determining and counting the germination rate and germination length of the Camellia oleifera pollen according to the pollen grain counting result and the true length of the germinated pollen tube.

2. The method for calculating the germination rate and germination length of in vitro camellia oleifera pollen according to claim 1, characterized in that: The step S1, collecting Camellia oleifera pollen and culturing it in vitro, obtaining a microscopic image of the Camellia oleifera pollen tube, and dividing the microscopic image into a training image and a test image, comprises: S11, collecting Camellia oleifera pollen and storing it at low temperature; S12, sowing the Camellia oleifera pollen in a pollen culture dish prepared with Camellia oleifera in vitro pollen suspension culture solution, mixing thoroughly, and then placing in a constant temperature incubator in the dark for culture; S13, placing the current pollen culture dish under a microscope, adjusting the focal length of the microscope so that the pollen grains and pollen tubes in the current pollen culture dish can be clearly imaged under the microscope, and moving the current pollen culture dish to take pictures in multiple fields of view until microscopic images of all central areas of the current pollen culture dish are collected; S14, replacing the pollen culture dish that has not been photographed, repeating step S13, and collecting the microscopic images of all pollen culture dishes; S15, storing all the microscopic images of the pollen culture dishes in order, and randomly selecting and dividing the microscopic images into training images and images to be tested.

3. The method for calculating the germination rate and germination length of in vitro camellia oleifera pollen according to claim 1, characterized in that: The step S2, processing the training images in different labeling methods to obtain detection labels for counting germination rates and segmentation labels for counting germination lengths, includes: S21, using image annotation software to select and annotate the pollen grains in the training image using a rectangular frame tool, and setting label categories for ungerminated pollen grains and label categories for germinated pollen grains, respectively, to obtain detection labels for statistically analyzing germination rates; S22, using the image annotation software to enclose and annotate the pollen tubes in the training image using a polygon tool, and setting the label category normal for pollen tubes that are clearly imaged and do not cross other pollen tubes, the label category cross for pollen tubes that are clearly imaged and cross other pollen tubes, and the label category error for pollen tubes that are not clearly imaged or are long beyond the focal length of the lens, to obtain segmentation labels.

4. The method for calculating the germination rate and germination length of in vitro camellia oleifera pollen according to claim 3, characterized in that: The step S3, using the labeled training image, the detection label and the segmentation label to train an artificial intelligence model for detection and segmentation, comprises: S31, constructing a germination rate detection model for statistically analyzing pollen germination rate; S32, using the labeled training image and the detection label to input into the germination rate detection model for training, and saving the best detection model weight for outputting the pollen tube detection result; S33, building an instance segmentation model for counting pollen tube length; S34, using the labeled training image and the segmentation label to input into the instance segmentation model for training, and saving the best segmentation model weight for outputting the pollen tube segmentation result; S35, randomly selecting part of the images to be tested, inputting them into the models trained in step S32 and step S34, and obtaining new detection labels and new segmentation labels; after adjusting the detection labels and the segmentation labels according to the new detection labels and the new segmentation labels, continue to input them into the models in step S32 and step S34 for training; S36, repeat step S35 until the germination rate detection model and the instance segmentation model reach preset model performance, and obtain an artificial intelligence model for detection and segmentation.

5. The method for calculating the germination rate and germination length of in vitro camellia oleifera pollen according to claim 4, characterized in that: The step S4, inputting the image to be tested into the artificial intelligence model for detection and segmentation, obtaining the number of germinated pollen grains, the number of ungerminated pollen grains and the pixel length of germinated pollen tubes, obtaining pollen grain counting results according to the number of germinated pollen grains and the number of ungerminated pollen grains, and calculating the actual length of the germinated pollen tubes according to the pixel length of the germinated pollen tubes, includes: S41, inputting the image to be tested into the artificial intelligence model for detection and segmentation to obtain a target detection label and a target segmentation label of the image to be tested; S42, screening the target detection label and the target segmentation label according to a preset label specification; S43, counting the selected target detection labels to obtain label categories of the number of germinated pollen grains and the number of ungerminated pollen grains, and obtaining a pollen grain counting result according to the number of germinated pollen grains and the number of ungerminated pollen grains; S44, selecting the target pollen tubes with the label categories of normal and cross in the filtered target segmentation labels, drawing a mask image of each target pollen tube using a polygon label, and processing the mask image using a thinning algorithm to obtain the pixel length of each target pollen tube; S45, converting the pixel length according to the ratio between the image pixels of the image to be measured and the real length to obtain the real length of each target pollen tube.

6. A system for counting the germination rate and germination length of in vitro Camellia oleifera pollen, characterized in that: include: A microscopic image acquisition module is used to collect Camellia oleifera pollen and culture it in vitro, obtain microscopic images of Camellia oleifera pollen tubes, and divide the microscopic images into training images and images to be tested; A microscopic image annotation processing module is used to process the training images in different annotation modes to obtain detection labels for counting germination rates and segmentation labels for counting germination lengths; A model training module, used to train an artificial intelligence model for detection and segmentation using the labeled training images, the detection labels and the segmentation labels; A calculation module based on an artificial intelligence model, used for inputting the image to be tested into the artificial intelligence model for detection and segmentation, obtaining the number of germinated pollen grains, the number of ungerminated pollen grains and the pixel length of germinated pollen tubes, obtaining a pollen grain counting result according to the number of germinated pollen grains and the number of ungerminated pollen grains, and calculating the true length of the germinated pollen tube according to the pixel length of the germinated pollen tube; The germination rate and germination length statistics module is used to measure and count the germination rate and germination length of the Camellia oleifera pollen according to the pollen grain counting result and the true length of the germinated pollen tube.

7. The system for counting the germination rate and germination length of in vitro camellia oleifera pollen according to claim 6, characterized in that: The microscopic image acquisition module comprises: Camellia oleifera pollen collection unit, used for collecting Camellia oleifera pollen and preserving it at low temperature; An in vitro culture unit is used to sow the Camellia oleifera pollen in a pollen culture dish prepared with Camellia oleifera in vitro pollen suspension culture solution, mix thoroughly and place in a constant temperature incubator for culture in the dark; A microscopic image capturing unit is used to place the current pollen culture dish under a microscope, adjust the focal length of the microscope so that the pollen grains and pollen tubes in the current pollen culture dish can be clearly imaged under the microscope, move the current pollen culture dish to capture multiple fields of view, until the microscopic images of all the central areas of the current pollen culture dish are captured; replace the pollen culture dish that has not been captured, and complete the capture of the microscopic images of all pollen culture dishes; The microscopic image division unit is used to store the microscopic images of all pollen culture dishes in order, and randomly select and divide the microscopic images into training images and images to be tested.

8. The system for counting the germination rate and germination length of in vitro camellia oleifera pollen according to claim 6, characterized in that: The microscopic image annotation processing module comprises: A pollen grain labeling processing unit is used to select and label the pollen grains in the training image using an image labeling software and a rectangular frame tool, and to set label categories of ungerminated pollen grains and label categories of germinated pollen grains respectively, so as to obtain detection labels for statistically analyzing germination rates; The pollen tube annotation processing unit is used to use the image annotation software to enclose and annotate the pollen tubes in the training image using a polygon tool, and set the label category normal for pollen tubes with clear imaging and not crossing other pollen tubes, the label category cross for pollen tubes with clear imaging and crossing other pollen tubes, and the label category error for pollen tubes with unclear imaging or long beyond the focal length of the lens to obtain segmentation labels.

9. The system for counting the germination rate and germination length of in vitro camellia oleifera pollen according to claim 8, characterized in that: The model training module includes: A germination rate detection model building unit, used to build a germination rate detection model for counting pollen germination rate; A germination rate detection model training unit, used to use the labeled training image and the detection label to input the germination rate detection model for training, and save the best detection model weight for outputting the detection result of the pollen tube; An instance segmentation model building unit, used to build an instance segmentation model for counting pollen tube lengths; An instance segmentation model training unit, used to input the labeled training image and the segmentation label into the instance segmentation model for training, and save the best segmentation model weights for outputting the pollen tube segmentation result; The model optimization unit is used to randomly select part of the images to be tested, input them into the trained models in the germination rate detection model training unit and the instance segmentation model training unit, and obtain new detection labels and new segmentation labels; after adjusting the detection labels and the segmentation labels according to the new detection labels and the new segmentation labels, continue to input them into the models in the germination rate detection model training unit and the instance segmentation model training unit for training; repeat the execution until the germination rate detection model and the instance segmentation model reach the preset model performance, and obtain the artificial intelligence model for detection and segmentation.

10. The system for counting the germination rate and germination length of in vitro camellia oleifera pollen according to claim 9, characterized in that: The computing module based on the artificial intelligence model includes: A label acquisition unit, used for inputting the image to be tested into the artificial intelligence model for detection and segmentation to obtain the target detection label and target segmentation label of the image to be tested; A label screening unit, used for screening the target detection label and the target segmentation label according to a preset label specification; A pollen grain counting unit is used to count the selected target detection labels to obtain label categories of the number of germinated pollen grains and the number of ungerminated pollen grains, and obtain a pollen grain counting result according to the number of germinated pollen grains and the number of ungerminated pollen grains; A pollen tube pixel length statistics unit is used to select target pollen tubes with label categories of normal and cross in the filtered target segmentation labels, draw a mask image of each target pollen tube using a polygon label, and process the mask image using a thinning algorithm to obtain the pixel length of each target pollen tube; The pollen tube real length calculation unit is used to convert the pixel length according to the ratio between the image pixels and the real length of the image to be measured, so as to obtain the real length of each target pollen tube.

Images