Quantitative method and system for plant leaf surface particulate matter

Abstract

The application provides a plant leaf surface particle quantification method and system, comprising the following steps: S1, preparing a plant leaf sample; S2, obtaining a scanning electron microscope image based on the prepared plant leaf sample; S3, segmenting the obtained scanning electron microscope image to obtain an image object; S4, selecting a training sample in the obtained image object, combining a parameter feature of the image object to construct and screen an optimal classification decision tree for automatically identifying a leaf surface particle, and automatically identifying and extracting information of the leaf surface particle in the scanning electron microscope image; and S5, based on the automatically identified and extracted leaf surface particle, performing quantity statistics and quantitative characterization of a shape feature on each particle.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing and the multidisciplinary field of ecological environment, in particular to a plant leaf surface particle quantification method and system, and more particularly to a plant leaf surface particle quantification method and system combining a scanning electron microscope image and machine learning. BACKGROUND

[0002] The prior art discloses identification of particles in the atmosphere, but there is no research on identification of particles on the leaves of vegetation.

[0003] Patent document CN113095109A (application number: 201911337953.6) discloses a crop leaf recognition model training method, recognition method and device, relating to the field of image processing. The crop leaf recognition model training method comprises: obtaining a training sample set, the training sample set comprising a plurality of training samples, for each training sample, respectively inputting a target crop image and a target crop label image corresponding to the target crop image into a crop leaf recognition model to obtain a recognition result; determining whether a preset training stop condition is met according to the recognition result of each target crop image and the target crop label image corresponding to the target crop image; if not, adjusting the model parameters of the crop leaf recognition model until the preset training stop condition is met, and obtaining a trained crop leaf recognition model. It can adapt to the complex environment of farmland, improve the recognition accuracy of the edge features of the crop leaf surface, and automatically and accurately recognize the target crop leaf surface. SUMMARY

[0004] In view of the deficiencies and defects in the prior art, the purpose of the present application is to provide an efficient and accurate plant leaf surface particle quantification method and system.

[0005] According to the plant leaf surface particle quantification method provided by the present application, the following steps are included:

[0006] Step S1: preparing a plant leaf sample;

[0007] Step S2: obtaining a scanning electron microscope image based on the prepared plant leaf sample;

[0008] Step S3: segmenting the obtained scanning electron microscope image to obtain an image object;

[0009] Step S4: selecting a training sample in the obtained image object, constructing and screening an optimal classification decision tree for automatically identifying leaf surface particles in combination with the parameter features of the image object, and automatically identifying and extracting information of the leaf surface particles in the scanning electron microscope image;

[0010] Step S5: based on the automatically identified and extracted leaf surface particles, performing quantitative characterization of the shape features and quantity statistics of each particle.

[0011] Preferably, the step S1 employs: selecting healthy plant leaves in natural environment, placing the plant leaves on a sterile operation table, cutting leaf samples on each plant leaf along the diagonal line of the leaf, avoiding leaf veins.

[0012] Preferably, the step S2 employs: placing the prepared plant leaf samples on an electron scanning electron microscope platform, selecting different image magnifications, sequentially imaging the leaf samples to generate scanning electron microscope images.

[0013] Preferably, the step S3 employs: the scanning electron microscope images are segmented by using eCognition Develope software, selecting a multi-resolution segmentation algorithm, setting corresponding segmentation parameters, and obtaining image objects.

[0014] Preferably, the parameter features of the image objects in the step S4 include: band mean, band brightness, band standard deviation, hue, color saturation, color intensity, elliptical fitting degree, rectangular fitting degree, image object area, circularity, compactness, shape index, boundary index, aspect ratio, spatial density, asymmetry, and gray level co-occurrence matrix.

[0015] According to the plant leaf surface particle quantification system provided by the application, the following technical effects are achieved:

[0016] Module M1: obtaining scanning electron microscope images based on plant leaf samples;

[0017] Module M2: segmenting the obtained scanning electron microscope images to obtain image objects;

[0018] Module M3: selecting training samples in the obtained image objects, constructing and screening an optimal classification decision tree for automatically identifying leaf surface particles in combination with the parameter features of the image objects, and automatically identifying and extracting information of the leaf surface particles in the scanning electron microscope images;

[0019] Module M4: based on the automatically identified and extracted leaf surface particles, performing quantity statistics and quantitative characterization of shape features on each particle.

[0020] Preferably, the step S1 employs: selecting healthy plant leaves in natural environment, placing the plant leaves on a sterile operation table, cutting leaf samples on each plant leaf along the diagonal line of the leaf, avoiding leaf veins.

[0021] Preferably, the step S2 employs: placing the prepared plant leaf samples on an electron scanning electron microscope platform, selecting different image magnifications, sequentially imaging the leaf samples to generate scanning electron microscope images.

[0022] Preferably, the module M2 adopts: a scanning electron microscope image is segmented by using eCognition Develope software, a multi-resolution segmentation algorithm is selected, corresponding segmentation parameters are set, and the scanning electron microscope image is segmented to obtain an image object.

[0023] Preferably, the shape index, the boundary index, the aspect ratio, the spatial density, the asymmetry, and the gray level co-occurrence matrix.

[0024] Compared with the prior art, the present application has the beneficial effects as follows: by adopting scanning electron microscope imaging of plant leaf blades, combining image processing technology and machine learning algorithms, automatic identification and information extraction of leaf surface particulates on plant leaf scanning electron microscope images are realized, the problems of low efficiency and single information in distinguishing particulates and quantifying the number and shape characteristics of particulates in traditional methods are solved, the identification accuracy and efficiency of plant leaf surface particulates are significantly improved, the evaluation index system of leaf surface particulate shape characteristics is enriched, and economic cost and time cost are obviously saved. BRIEF DESCRIPTION OF DRAWINGS

[0025] Other features, objects and advantages of the present application will become more apparent from the following detailed description of non-limiting embodiments, made with reference to the attached drawings:

[0026] Figure 1 The figure is a schematic diagram of image segmentation results under different classification parameters.

[0027] Figure 2 The figure is a scanning electron microscope image under different magnifications.

[0028] Figure 3 The figure is a schematic diagram of particulate recognition results.

[0029] Figure 4 The figure is a schematic diagram of quantification and characterization of particulate shape characteristics. DETAILED DESCRIPTION

[0030] The present application will be described in detail below with reference to specific embodiments. The following embodiments will help those skilled in the art to further understand the present application, but do not limit the present application in any form. It should be noted that, for those skilled in the art, without departing from the concept of the present application, a number of changes and improvements can be made. These all belong to the protection scope of the present application.

[0031] Example 1

[0032] The application provides a plant leaf granule quantification method and system, which comprises the following steps: combining plant leaf scanning electron microscope pictures and machine learning algorithms, using image information recognition and automatic extraction technology to recognize plant leaf granules and accurately characterize the quantity, position, size, shape and other characteristics of the plant leaf granules, and significantly improving the recognition efficiency and feature richness of the leaf granules.

[0033] The plant leaf granule quantification method comprises the following steps:

[0034] Step S1: preparing plant leaf samples;

[0035] Specifically, the step S1 adopts the following method: selecting healthy plant leaves in a natural environment, collecting the leaves and quickly bringing them back to the laboratory. The plant leaves are placed on a sterile operating table, and 5 leaf samples of 1cm 2 size are cut from each plant leaf along the diagonal line of the leaf, avoiding the leaf veins.

[0036] Step S2: obtaining scanning electron microscope images based on the prepared plant leaf samples;

[0037] Specifically, the step S2 adopts the following method: placing the prepared leaf samples on an electronic scanning electron microscope (Scanning Electron Microscope) platform, selecting different image magnifications, and sequentially imaging the leaf samples to generate scanning electron microscope images; wherein the magnifications include 200x, 500x, 1000x and 2000x.

[0038] Step S3: segmenting the obtained scanning electron microscope images to obtain image objects;

[0039] Specifically, the step S3 adopts the following method: importing the scanning electron microscope images into eCognition Develope software, selecting a multi-resolution segmentation (multi-resolution segmentation) algorithm, setting appropriate segmentation parameters, and performing image segmentation on the scanning electron microscope images to obtain image objects; wherein the segmentation parameters include size, color, shape, compactness and smoothness.

[0040] Step S4: selecting training samples in the obtained image objects, constructing and screening an optimal classification decision tree for automatically identifying leaf granules in combination with the parameter characteristics of the image objects, and automatically identifying and information extracting the leaf granules in the scanning electron microscope images; wherein the parameter characteristics of the image objects include band mean, band brightness, band standard deviation, hue, color saturation, color intensity, elliptical fitting degree, rectangular fitting degree, image object area, circularity, compactness, shape index, boundary index, aspect ratio, spatial density, asymmetry, gray level co-occurrence matrix and the like.

[0041] Step S5: Based on the automatically identified and extracted leaf surface particles, the number of each particle is counted and the shape characteristics are quantitatively characterized. The particle number and shape characteristics include: total number; density; area; size; diameter; weight; shape index; boundary length; fractal index, etc.

[0042] The plant leaf surface particle quantification system, such as Figures 1 to 4 as shown, comprises:

[0043] Module M1: Based on the plant leaf sample, a scanning electron microscope image is obtained;

[0044] Specifically, the module M1 uses: placing the prepared leaf sample on the platform of the electron scanning electron microscope (Scanning Electron Microscope), selecting different image magnifications, and sequentially imaging the leaf sample to generate a scanning electron microscope image; wherein the magnification includes: 200 times, 500 times, 1000 times, 2000 times.

[0045] The leaf sample includes: selecting healthy plant leaves in a natural environment, collecting the leaves and quickly bringing them back to the laboratory. Place the plant leaves on a sterile operating table, cut 5 leaf samples of 1 cm 2 size on each plant leaf along the diagonal line of the leaf, avoiding the leaf veins;

[0046] Module M2: The obtained scanning electron microscope image is segmented to obtain an image object;

[0047] Specifically, the module M2 uses: importing the scanning electron microscope image into eCognition Develope software, selecting a multi-resolution segmentation algorithm, setting appropriate segmentation parameters, and performing image segmentation on the scanning electron microscope image to obtain an image object; wherein the segmentation parameters include: size, color, shape, compactness, and smoothness.

[0048] Module M3: Select a training sample in the obtained image object, combine the parameter characteristics of the image object to construct and screen an optimal classification decision tree for automatically identifying leaf surface particles, and automatically identify and extract information from the leaf surface particles in the scanning electron microscope image; wherein the parameter characteristics of the image object include: band mean, band brightness, band standard deviation, hue, color saturation, color intensity, elliptical fitting degree, rectangular fitting degree, image object area, circularity, compactness, shape index, boundary index, aspect ratio, spatial density, asymmetry, and gray level co-occurrence matrix.

[0049] Module M4: based on the automatically identified and extracted leaf surface particles, the number of each particle is counted and the shape characteristics are quantitatively characterized. The particle number and shape characteristics include: total number; density; area; size; diameter; weight; shape index; boundary length; fractal index, etc.

[0050] Those skilled in the art know that, in addition to implementing the system, device and each module thereof provided by the present application in the form of pure computer readable program code, the same program can also be implemented in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers and embedded microcontrollers, etc. by logically programming the method steps. Therefore, the system, device and each module thereof provided by the present application can be considered as a hardware component, and the modules included therein for implementing various programs can also be considered as structures within the hardware component; the modules for implementing various functions can also be considered as both software programs for implementing methods and structures within hardware components.

[0051] The specific embodiments of the present application are described above. It needs to be understood that the present application is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essential content of the present application. The embodiments of the present application and the features in the embodiments can be arbitrarily combined with each other without conflict.

Claims

1. A method for quantifying particulate matter on plant leaves, characterized in that: include: Step S1: preparing plant leaf samples; Step S2: acquiring a scanning electron microscope image based on the prepared plant leaf sample; Step S3: segmenting the acquired SEM image to obtain image objects; Step S4: selecting training samples from the obtained image objects, constructing and screening the optimal classification decision tree for automatically identifying leaf surface particles based on the parameter features of the image objects, and automatically identifying and extracting information from leaf surface particles in the scanning electron microscope images; Step S5: Based on the automatically identified and extracted leaf surface particles, the number of particles is counted and the shape characteristics are quantitatively characterized for each particle; The parameter features of the image object in step S4 include: band mean, band brightness, band standard deviation, hue, color saturation, color intensity, ellipse fitting, rectangle fitting, image object area, roundness, compactness, shape index, boundary index, aspect ratio, spatial density, asymmetry and gray-level co-occurrence matrix.

2. The method for quantifying plant leaf particulate matter according to claim 1, characterized in that: The step S1 comprises the following steps: selecting healthy plant leaves in a natural environment, placing the plant leaves on a sterile operating table, and cutting leaf samples from each plant leaf along the diagonal line of the leaf while avoiding the veins.

3. The method for quantifying plant leaf particulate matter according to claim 1, characterized in that: The step S2 comprises placing the prepared plant leaf samples on a scanning electron microscope platform, selecting different image magnifications, and sequentially photographing and imaging the leaf samples to generate scanning electron microscope images.

4. The method for quantifying plant leaf particulate matter according to claim 1, characterized in that: The step S3 comprises: using eCognition Develope software to segment the scanning electron microscope image, selecting a multi-resolution segmentation algorithm, setting corresponding segmentation parameters, and performing image segmentation on the scanning electron microscope image to obtain an image object.

5. A plant leaf particle quantification system, characterized in that: include: Module M1: Acquisition of scanning electron microscope images based on plant leaf samples; Module M2: Segment the acquired SEM image to obtain image objects; Module M3: Select training samples from the obtained image objects, build and screen the optimal classification decision tree for automatic identification of leaf surface particles based on the parameter features of the image objects, and automatically identify and extract information about leaf surface particles in scanning electron microscope images; Module M4: Based on the automatically identified and extracted leaf surface particles, the number of particles and the quantitative characterization of their shape characteristics are performed; The parameter features of the image object in the module M3 include: band mean, band brightness, band standard deviation, hue, color saturation, color intensity, ellipse fit, rectangle fit, image object area, roundness, compactness, shape index, boundary index, aspect ratio, spatial density, asymmetry and gray-level co-occurrence matrix.

6. The plant leaf particle quantification system according to claim 5, characterized in that: Select healthy plant leaves in a natural environment, place the plant leaves on a sterile operating table, and cut leaf samples from each plant leaf along the diagonal line of the leaf, avoiding the veins.

7. The plant leaf particle quantification system according to claim 5, characterized in that: The module M1 uses: using an electronic scanning electron microscope platform, selecting different image magnifications, and sequentially photographing and imaging leaf samples to generate scanning electron microscope images.

8. The plant leaf particle quantification system according to claim 5, characterized in that: The module M2 uses: using eCognition Develope software to segment the scanning electron microscope image, selecting a multi-resolution segmentation algorithm, setting corresponding segmentation parameters, and performing image segmentation on the scanning electron microscope image to obtain image objects.

Images