Rhododendron species identification method and device, terminal equipment and readable medium

Abstract

The invention provides a rhododendron species identification method and device, terminal equipment and a readable medium, and is suitable for the technical field of data processing, and the method comprises the steps: obtaining a low-resolution spectral image of a to-be-identified rhododendron; performing spectral feature extraction on the low-resolution spectral image to match a plurality of high-resolution spectral features with the highest similarity, and calculating the weighted votes of the rhododendron species corresponding to the plurality of high-resolution spectral features; and determining the type with the highest weighted votes as the type of the rhododendron to be identified. According to the invention, based on the low-resolution spectral image of the to-be-identified azalea, through spectral feature extraction and matching of a plurality of high-resolution spectral features with the highest similarity, the azalea type identification is carried out, the identification efficiency is high, meanwhile, errors possibly brought by single high-resolution feature matching can be effectively avoided, and the identification accuracy is improved. The method fully considers the information of a plurality of similar features, and greatly improves the accuracy of rhododendron species determination.

Description

Technical Field

[0001] The present application belongs to the field of data processing technology, and in particular relates to a method, device, terminal equipment and readable medium for identifying azalea species. Background Art

[0002] The genus Rhododendron in the Ericaceae family contains many species. There are more than a thousand known species of rhododendrons in the world, and China has about 600 species, making it one of the countries with the richest rhododendron species in the world. Rhododendrons can be divided into different flower systems, such as spring azalea system, summer azalea system, western azalea system, eastern azalea system, and alpine azalea system. Different strains of azaleas have certain differences in flowering period, growth habits, medicinal value, flower shape and color. Therefore, research on the identification of azalea species has important industrial value and scientific significance.

[0003] At present, the identification method of azalea species mostly adopts molecular marker technology. This technology involves multiple complex experimental steps. The first is the time-consuming sequencing process, which requires professional equipment and technicians to operate. After sequencing, primer design is required, which requires a deep understanding of the gene sequence of azalea and the determination of suitable primers through repeated trials. Gel electrophoresis test is also a time-consuming step. Only a limited number of samples can be processed each time, and it is necessary to wait for the right time during the electrophoresis process to ensure clear band separation. When identifying multiple azaleas, it is necessary to screen out simple repetitive sequences with specific molecular marker potential from numerous DNA sequences. There are many species of azaleas and their genetic sequences are complex, which makes the screening process very cumbersome. For example, in an experiment involving multiple azalea samples, hundreds or thousands of DNA fragments may need to be analyzed one by one to determine which fragments can serve as effective molecular markers, which will undoubtedly greatly increase the time cost.

[0004] In addition, molecular marker technology requires high-precision experimental equipment, such as sequencers, PCR machines (for primer extension reactions) and electrophoresis equipment. The accuracy and stability of these equipment are crucial to the experimental results. For example, if the accuracy of the sequencer deviates, it may cause errors in the read gene sequence, thereby affecting the subsequent primer design and molecular marker screening. Moreover, these devices need regular maintenance and calibration, and any minor faults or improper parameter settings may have a negative impact on the experiment. Factors such as the temperature, humidity, and reagent purity of the experimental environment will also have a significant impact on the experimental results. For example, during the PCR reaction, slight fluctuations in temperature may affect the binding efficiency of primers and template DNA, thereby affecting the quality and quantity of the amplified products. During gel electrophoresis, the environmental humidity may affect the coagulation effect and electrophoresis speed of the gel, resulting in deviations in the migration distance of the DNA band, thereby affecting the judgment of the molecular marker. Furthermore, there may be a high degree of similarity in the gene sequences between azalea species, which will bring challenges to the accuracy of molecular markers. Because different species of azalea are related in the evolutionary process, some of their gene sequences are very similar, which makes it possible for some molecular markers to not accurately distinguish similar species. For example, the simple repeat sequences of some closely related species differ very little and may not show obvious differences on gel electrophoresis, leading to misjudgment. Some molecular markers discovered so far may have limitations in polymorphism. If a molecular marker shows the same or similar characteristics in multiple azaleas, it cannot be effectively used for species identification. For example, some simple repeat sequences are highly conserved in multiple azalea lines and do not have enough variation to distinguish different species, which reduces the accuracy of identification.

[0005] It can be seen that the existing molecular marker technology method for identifying rhododendron species has the problems of high time cost, being easily affected by experimental environmental parameters, and the accuracy and precision need to be improved. Summary of the invention

[0006] In view of this, the embodiments of the present application provide a method, apparatus, terminal device and readable medium for identifying azalea species, aiming to solve the problems that the existing method of identifying azalea species using molecular marker technology has high time cost, is easily affected by experimental environment parameters, and has accuracy and precision that needs to be improved.

[0007] A first aspect of an embodiment of the present application provides a method for identifying azalea species, comprising:

[0008] Obtaining a low-resolution spectral image of the azalea to be identified;

[0009] Extract spectral features from low-resolution spectral images to match several high-resolution spectral features with the highest similarity, and calculate the weighted votes of azalea species corresponding to several high-resolution spectral features;

[0010] The species with the highest number of weighted votes is determined as the species of the azalea to be identified.

[0011] A second aspect of an embodiment of the present application provides a device for identifying azalea species, comprising:

[0012] A low-resolution spectral image acquisition module is used to acquire a low-resolution spectral image of the azalea to be identified;

[0013] A high-resolution spectral feature matching module is used to extract spectral features from low-resolution spectral images to match several high-resolution spectral features with the highest similarity, and calculate the weighted votes of azalea species corresponding to the several high-resolution spectral features; and

[0014] The azalea species determination module is used to determine the species with the highest weighted votes as the species of the azalea to be identified.

[0015] A third aspect of an embodiment of the present application provides a terminal device, which includes a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the computer program, the steps of the azalea species identification method as described in any one of the first aspects above are implemented.

[0016] A fourth aspect of the embodiments of the present application provides a non-transitory computer-readable medium, comprising: a computer program stored thereon, wherein when the computer program is executed by a processor, the steps of the method for identifying azalea species as described in any one of the first aspects above are implemented.

[0017] Compared with the prior art, the embodiments of the present application have the following beneficial effects: based on the low-resolution spectral image of the azalea to be identified, the present application extracts spectral features and matches several high-resolution spectral features with the highest similarity, calculates the weighted votes of the azalea species corresponding to several high-resolution spectral features, so as to determine the species of the azalea to be identified, has high recognition efficiency, and can effectively avoid the errors that may be caused by matching a single high-resolution feature, fully considers the information of multiple similar features, and greatly improves the accuracy of azalea species determination. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0019] Figure 1This is a schematic diagram of the implementation process of a method for identifying azalea species provided in an embodiment of the present application;

[0020] Figure 2 It is a schematic diagram of the implementation process of another method for identifying azalea species provided in an embodiment of the present application;

[0021] Figure 3 It is a schematic diagram of an implementation flow of a similarity optimization method provided in an embodiment of the present application;

[0022] Figure 4 It is a schematic diagram of the implementation flow of another similarity optimization method provided in an embodiment of the present application;

[0023] Figure 5 is a schematic diagram of the structure of a device for identifying azalea species provided in an embodiment of the present application;

[0024] Figure 6 It is a schematic diagram of a terminal device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0025] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0026] In order to illustrate the technical solution described in this application, a specific embodiment is provided below for illustration.

[0027] Figure 1 The following is a flowchart of the method for identifying azalea species provided in the embodiment of the present application, which is described in detail as follows:

[0028] In the embodiment of the present application, the method for identifying the species of azalea comprises the following steps:

[0029] Step S101, obtaining a low-resolution spectral image of azalea to be identified.

[0030] In an optional embodiment of the present application, the spectral image can be acquired by a suitable spectral imaging device, such as a portable spectrometer, which is one of the commonly used devices for acquiring low-resolution spectral images. It is small and portable, and is convenient for use in field environments such as the wild. These devices are usually able to acquire spectral data within a certain wavelength range (such as visible light and near-infrared bands). The principle is to focus light onto a detector through an optical system, and the detector converts the optical signal into an electrical signal, and obtains spectral information after processing. For example, some handheld spectrometers can collect reflectance spectra by contacting or approaching the leaves, petals and other parts of azaleas, and their resolution may be between a few nanometers and tens of nanometers, which can meet the acquisition needs of low-resolution spectral images.

[0031] Step S102 , extracting spectral features from the low-resolution spectral image to match a number of high-resolution spectral features with the highest similarity, and calculating weighted votes of azalea species corresponding to the several high-resolution spectral features.

[0032] In the embodiment of the present application, the spectral characteristics of different types of azaleas, such as peak shape, relative absorption intensity, mean value of spectral bands, variance and texture characteristics, etc., will have certain differences. Spectral feature extraction of low-resolution spectral images includes extraction of spectral reflectance statistical characteristics (such as mean, variance, skewness and kurtosis), extraction of spectral curve characteristics (such as absorption peak and reflection peak position, absorption peak and reflection peak depth and width), extraction of principal component analysis characteristics, etc. Among them, calculating the mean spectral reflectance of each band can reflect the average level of spectral reflectance of the band; variance reflects the degree of discreteness of reflectance data in each band; skewness describes the asymmetry of reflectance distribution, and kurtosis describes the sharpness of distribution. These statistical features can help us understand the distribution form of spectral reflectance data, and have a certain auxiliary effect on distinguishing the spectral characteristics of different types of azaleas. By analyzing the spectral reflectance curve, find the position of the local maximum (reflection peak) and minimum (absorption peak). The position of these peaks is usually related to the characteristics of the absorption or reflection of light by the pigments (such as chlorophyll, anthocyanin, etc.) in azaleas. For example, chlorophyll has absorption peaks in the blue light (430-450nm) and red light (640-660nm) bands, and anthocyanins may have absorption peaks in the long-wave portion of visible light (550-700nm). Determining the locations of these peaks can provide important clues for identifying azalea species. The depth of the absorption peak can be calculated by the difference between the peak value and the adjacent valley value, while the depth of the reflection peak is the opposite. The width of the peak can be calculated by determining the boundary position on both sides of the peak using methods such as a certain slope threshold. These features can further characterize the shape of the spectral curve. The spectral curves of different types of azaleas have certain differences in the depth and width of the peak.

[0033] In addition, the principal component analysis feature extraction method can be: the spectral data of the low-resolution spectral image are arranged by band to form a matrix X (assuming that the number of image pixels is p and the number of bands is q, then the matrix X is p×q dimensions). First, standardize the data so that the mean of each column is 0 and the variance is 1. Calculate the covariance matrix For the eigenvalues ​​λ1≥λ2≥…≥λ q Decompose and obtain the eigenvalues ​​and corresponding eigenvectors υ1,υ2,…,υ q According to the cumulative contribution rate (such as selecting the principal component with a cumulative contribution rate of more than 85%), the first k principal components to be selected (k < q) are determined, and the low-resolution spectral feature vector F LR It can be expressed as F LR =X υ (where υ is a matrix composed of k selected eigenvectors.) Principal component analysis can reduce the data dimension while retaining the main spectral information features. These principal components can be used as important features of low-resolution spectral images for subsequent analysis.

[0034] In the embodiment of the present application, after extracting the spectral features of the low-resolution spectral image, the high-resolution spectral feature database is combined to match several high-resolution spectral features with the highest similarity. Specifically, the similarity between each low-resolution spectral feature vector and all high-resolution spectral feature vectors is calculated based on the Mahalanobis distance (for the low-resolution spectral feature vector F LR and high-resolution spectral features F HR Vector, Mahalanobis distance ), construct a similarity matrix M, where M ij Represents the similarity between the i-th low-resolution spectral feature and the j-th high-resolution spectral feature. According to the similarity matrix, for each low-resolution spectral feature, select several (for example, k) high-resolution spectral features with the highest similarity. The high-resolution spectral features corresponding to the first k maximum values ​​can be selected by sorting each row of the similarity matrix. The similarity between the low-resolution spectral feature and the high-resolution spectral feature is used as the weight w i (i=1,2…,k), that is, the higher the similarity, the greater the weight of the corresponding high-resolution spectral feature in the voting process. Suppose the azalea species corresponding to these k high-resolution spectral features are y1, y2,…, y k , for each azalea species c (assuming there are C azalea species), calculate its weighted votes V C :

[0035]

[0036] In the above formula (1), δ(y i ,c) is the indicator function, when yi = c, δ(y i ,c)=1, otherwise δ(y i ,c)=0. Thus, for each possible azalea species, its weighted vote is the sum of the similarities of all high-resolution spectral features of the same species.

[0037] Step S103: determining the species with the highest weighted votes as the species of azalea to be identified.

[0038] In the implementation mode of the present application, after calculating all types of weighted votes V1, V2, ..., V C After that, by comparing the size of these votes, the species with the highest weighted votes is determined as the species of the azalea to be identified. This is because this species has the largest "total weight" in the high-resolution spectral features similar to the low-resolution spectral features, which means that from the perspective of overall similarity, the spectral features of the azalea to be identified are most similar to the high-resolution spectral features of this species, so it is most likely to belong to this species. For example, if the calculated weighted votes for azalea species A are the highest, V A , then the species of the azalea to be identified is determined to be A. This method makes full use of the information of multiple similar high-resolution spectral features, avoids the errors that may be caused by relying solely on a single most similar feature for judgment, and thus improves the accuracy of species identification.

[0039] In a preferred embodiment of the present invention, Figure 2 As shown, in order to further improve the accuracy of azalea species identification, step S102 includes the following steps:

[0040] Step S201 , extracting spectral features from the low-resolution spectral image to obtain spectral features of multiple parts of the azalea to be identified.

[0041] In the embodiment of the present application, the low-resolution spectral image is divided into different regions according to the morphological structure of the azalea, corresponding to the leaves, petals, stems and other parts. This can be achieved by manual annotation or by using an image segmentation algorithm. For example, based on the threshold segmentation method, a threshold is set according to the difference in spectral reflectance of different parts, and the pixels in the image are divided into different categories, thereby separating the regions of leaves, petals and other parts.

[0042] Step S202 : Based on the multi-part spectral features, several high-resolution spectral features with the highest similarity are matched respectively, and weighted votes of azalea species corresponding to the several high-resolution spectral features are calculated respectively.

[0043] In an embodiment of the present application, for the spectral features of each part (leaves, petals, stems, etc.), the high-resolution spectral features of the specific parts are compared based on the similarity measurement method, and the similarity between the low-resolution spectral features of the azalea leaves to be identified and all the high-resolution spectral features (leaf parts) in the database is calculated. Assuming that there are m high-resolution spectral features in the database, a 1×m similarity vector (matrix) can be obtained by the above similarity measurement method. The same operation is performed on other parts such as petals and stems to obtain corresponding similarity vectors (matrices). According to the similarity vector (matrix), they are sorted. For example, if the Euclidean distance is used, the distance is sorted from small to large; if the cosine similarity is used, the similarity is sorted from large to small. The first k (pre-set number) high-resolution spectral features after sorting are selected as the features with the highest similarity. For other parts such as petals and stems, the k high-resolution spectral features with the highest similarity are also selected. Finally, the final azalea species can be determined by comprehensively considering the weighted votes of each part in different ways. For example, the weighted votes (V c-leaf ,V c-petal ,V c-stem ) to get the total weighted votes V c =V c-leaf +V c-petal +V c-stem Then select the species with the highest total weighted votes as the species of azalea to be identified. Or assign different weights according to the importance of each part and then perform comprehensive calculations.

[0044] In another preferred embodiment of the present application, by optimizing and adjusting the similarity, the matching accuracy between the low-resolution spectral features and the high-resolution spectral features is facilitated. Figure 3 As shown, the similarity is obtained by the following steps:

[0045] Step S301, obtaining local high-resolution spectral images and global low-resolution spectral images of azalea under several different growth environment factors.

[0046] In the present application embodiment, even if the spectral characteristics of the azalea of ​​the same kind under different growth environment factors such as growth stage, illumination conditions, temperature conditions, moisture conditions, soil conditions, altitude and topography and landforms, etc., all have certain differences, in order to ensure the recognition accuracy of the azalea species, it is necessary to collect the local high-resolution spectral images of the azalea under different growth growth environment factors as a sample, and carry out the matching training of low-resolution spectral characteristics and high-resolution spectral characteristics. Through the local high-resolution spectral images, the spectral differences between the same parts of different kinds of azaleas can be analyzed and compared. For example, the reflectivity platform height and slope of the blade of some kinds of azaleas in the near-infrared band may be different from other kinds, or there are differences in the absorption peak depth and position of petals at specific wavelengths. These subtle spectral differences help to set up a more accurate type identification model.

[0047] Step S302: extracting local high-resolution spectral features for identifying local azalea species based on the local high-resolution spectral image.

[0048] In the implementation mode of this application, Figure 4 As shown, the step of determining the local high-resolution spectral features for identifying the local azalea species based on the local high-resolution spectral image of the azalea comprises:

[0049] Step S401 , performing spectral feature division on each local high-resolution spectral image to obtain a plurality of initial local azalea species identification spectral features.

[0050] In this embodiment, spectral curve data in each local high-resolution spectral image can be extracted, such as using sub-interval band selection technology to divide the bands in each local high-resolution spectral image, and the multiple spectral intervals obtained by division are the initial local azalea species identification spectral characteristics.

[0051] Step S402, determining the position information of the local part of the azalea from the local high-resolution spectral image, and extracting the spectral features of the local high-resolution spectral image of the azalea based on the position information of the local part of the azalea to obtain the spectral features representing the local azalea species.

[0052] In this embodiment, band extraction can be performed on the local high-resolution spectral image through spectral data of the local part of the azalea reflected by the position information of the local part of the azalea, that is, band data in the local high-resolution spectral image that can be used to represent the local azalea species, that is, spectral characteristics characterizing the local azalea species, can be determined.

[0053] Step S403, when the band standard deviation of the initial local azalea species identification spectral feature and the local azalea species characterization spectral feature is less than a preset band standard deviation threshold, the initial local azalea species identification spectral feature is used as the local azalea species identification high-resolution spectral feature.

[0054] In this embodiment, the band standard deviation of the initial local azalea species identification spectral feature and the local azalea species characterization spectral feature is calculated to quantify the matching degree between the initial local azalea species identification spectral feature and the local azalea species characterization spectral feature, so as to determine whether the initial local azalea species identification spectral feature can be used as the high-resolution spectral feature for local azalea species identification in the spectral image. The calculation formula of the band standard deviation of the initial local azalea species identification spectral feature and the local azalea species characterization spectral feature is as follows:

[0055]

[0056] In the above formula (2), k represents the sample number, from 1 to n, n is the total number of samples, i represents the current band number, x ik represents the reflectance of the kth sample in band i, Represents the sample mean of band i.

[0057] Step S303: extracting low-resolution spectral features of various parts of the azalea according to the global low-resolution spectral image.

[0058] In an embodiment of the present application, the step of extracting low-resolution spectral features of various parts of azalea based on a global low-resolution spectral image includes: stitching multiple local high-resolution spectral images of azalea, and aligning the stitched images with the global low-resolution spectral image to determine the position information of various parts of the azalea in the global low-resolution spectral image; extracting low-resolution spectral features of various parts of the azalea in the global low-resolution spectral image based on the position information of various parts of the azalea in the global low-resolution spectral image.

[0059] In this embodiment, for multiple local high-resolution spectral images of azalea, it is first necessary to extract the feature points of the image. The scale-invariant feature transform (SIFT) algorithm or the accelerated robust feature (SURF) algorithm can be used. These algorithms can find stable feature points, such as edges and corners in images, at different image scales and rotation angles. Taking the SIFT algorithm as an example, it detects extreme points by constructing a Gaussian difference pyramid and assigns a direction to each feature point so that it has scale and rotation invariance. The extracted feature points are matched to find the corresponding feature points between different local images. A matching method based on feature descriptors is usually adopted, such as calculating the gradient histogram of the area around the feature point as a descriptor, and then determining the matching relationship by calculating the distance between the descriptors (such as the Euclidean distance). For example, in two local high-resolution spectral images, feature point pairs with similar gradient histogram descriptors are found, and these feature point pairs represent the overlapping parts between the images. According to the matched feature points, the transformation matrix between the images is calculated. If there are relations such as translation, rotation and scaling between the images, affine transformation or perspective transformation can be used to describe them. Solve the transformation matrix through the least square method and other methods to align the matching feature points as much as possible. After transforming the images, splice them together. The weighted average method can be used to merge the overlapping areas to avoid the appearance of stitching seams. For example, in the overlapping area, different weights are assigned according to the distance between the pixel points and the boundary of the two images, and the pixel values ​​are weighted and summed to obtain a stitched image with a natural transition.

[0060] Optionally, appropriate reference points are selected in the stitched high-resolution image and the global low-resolution image. These reference points can be obvious landmark features of azalea, such as larger flowers, groups of leaves of special shapes, etc. The reference points should be clearly identifiable in both resolution images, and the positions should be relatively fixed. The positions of the reference points can be determined by manual annotation or automatic detection. For automatic detection, the color, texture or shape features of the image can be used to identify the reference points. For example, based on the color and shape features of the flower, the center position of the flower is located in the image as a reference point through threshold segmentation and morphological operations. After the reference point is determined, the coordinate transformation relationship from the stitched high-resolution image to the global low-resolution image is calculated. This may involve transformations such as translation, scaling, and rotation. A similarity transformation model can be used to describe this relationship, and the transformation parameters are solved by minimizing the coordinate differences between the reference points. According to the transformation parameters obtained by the solution, the stitched high-resolution image is coordinate transformed to align it with the global low-resolution image. During the alignment process, an interpolation method (such as bilinear interpolation) can be used to process the transformed pixel positions to ensure the accuracy and continuity of the image.

[0061] Optionally, since the spliced ​​high-resolution image has been aligned with the global low-resolution image, a pixel position mapping relationship can be established between them. Through this mapping relationship, the position information of each azalea part in the spliced ​​high-resolution image is converted into the global low-resolution image. For example, for a specific leaf area in the spliced ​​high-resolution image, its coordinate range in the image can be converted into the corresponding coordinate range in the global low-resolution image through a mapping relationship, thereby determining the position of this leaf in the global image. Using a computer vision library (such as OpenCV) or a geographic information system (GIS) software, the position information of each part of the azalea in the global low-resolution image can be marked. The content of the marking can include part names (such as leaves, petals, stems, etc.), position coordinate ranges, etc. In order to ensure the accuracy of the position information, it can be verified by manual inspection or comparison with field survey data. For example, some parts of the azalea are marked in the field, and then checked with the position information determined in the image, and the inaccurate parts are adjusted and corrected.

[0062] Step S304, based on the spectral feature dimension, calculate the spectral feature similarity between the low-resolution spectral features of each part of the azalea and the high-resolution spectral features for local azalea species identification, and adjust the obtained similarity in combination with the growth environment factors.

[0063] In the embodiment of the present application, optionally, when comparing the low-resolution spectral features and the high-resolution spectral features of various parts of azalea, the following formula (3) can well reflect the similarity of their spectral feature distribution:

[0064]

[0065] In the above formula (3), x i ,y i , n are low-resolution spectral feature vector, high-resolution spectral feature vector and spectral feature dimension respectively. The value range of similarity s is between [-1,1]. The closer to 1, the more similar the low-resolution spectral feature and high-resolution spectral feature vectors are.

[0066] Furthermore, the similarity is adjusted according to the growth environment factors. Taking leaves as an example, the spectral feature similarity obtained by initial calculation is set as s leaf , the adjusted similarity s after considering growth environment factors such as light, temperature and soil leaf-adjusted It can be calculated by the following formula (4):

[0067] s leaf-adjusted =s leaf +w light-leaf ×Δ light-leaf +w temp-leaf ×Δ temp-leaf +wsoil-leaf ×Δ soil-leaf (4)

[0068] In the above formula (4), the influence weights of light factor, temperature factor and soil factor on the similarity of leaf spectral characteristics are w light-leaf ,w temp-leaf ,w soil-leaf ; Δ light-leaf ,Δ temp-leaf ,Δ soil-leaf They represent the adjustment amounts of light, temperature and soil factors on the similarity of leaf spectral characteristics, respectively. These adjustment amounts can be determined based on the relationship model between growth environment factors and spectral characteristics or experimental data. Similarly, similar adjustments are made to the similarity of petals and stems to obtain the similarity of spectral characteristics of each part after considering the growth environment factors.

[0069] In summary, the spectral characteristics of azaleas are significantly affected by the growth environment factors. By adjusting the similarity in combination with the growth environment factors, the interference of these external factors on the spectral characteristics can be more comprehensively considered. In addition, under different growth environments, the spectral characteristics of azaleas may change to a certain extent, making some originally similar species show differences in spectral characteristics, or originally different species appear more similar. Considering the growth environment factors to adjust the similarity can better deal with this situation.

[0070] Preferably, for a large area of ​​azalea area, an unmanned aerial vehicle equipped with a spectral imager can be used to obtain images. The unmanned aerial vehicle can quickly cover a large area and can shoot at different heights and angles. The spectral imager is installed on the unmanned aerial vehicle, and the reflected light from the azalea on the ground is collected through an optical lens, and the light is separated by wavelength through a spectroscopic system, and then the light intensity information of each wavelength is obtained by a detector array, and finally a spectral image is formed. The spectral image resolution obtained in this way is relatively low, but a large area of ​​spectral data can be provided, which is helpful for preliminary screening of azaleas in different growth areas. In the large area low-resolution spectral image collected, based on the azalea species identification method provided in the embodiment of the present application, all azalea species in the large area low-resolution spectral image can be uniformly identified.

[0071] Preferably, the high-resolution spectral feature also carries the growth environment factor information of the corresponding azalea species. The above-mentioned azalea species identification method also includes: displaying the species of azalea to be identified and the corresponding growth environment factor information in the low-resolution spectral image. In other words, since the high-resolution spectral feature corresponds to the growth environment factor, the azalea species identification method provided in the embodiment of the present application can synchronously determine the growth environment of the azalea after determining the species of the azalea.

[0072] Corresponding to the method of the above embodiment, Figure 5 The structural block diagram of the azalea species identification device provided in the embodiment of the present application is shown. For the convenience of explanation, only the part related to the embodiment of the present application is shown. Figure 5 The illustrated azalea species identification device may be an execution subject of the azalea species identification method provided in the aforementioned embodiment.

[0073] Reference Figure 5 , the azalea species identification device comprises:

[0074] A low-resolution spectral image acquisition module 510 is used to acquire a low-resolution spectral image of the azalea to be identified;

[0075] A high-resolution spectral feature matching module 520 is used to extract spectral features from the low-resolution spectral image to match a number of high-resolution spectral features with the highest similarity, and calculate the weighted votes of the azalea species corresponding to the several high-resolution spectral features; and

[0076] The azalea species determination module 530 is used to determine the species with the highest weighted votes as the azalea species to be identified.

[0077] The process of each module realizing its own function in the azalea species identification device provided in the embodiment of the present application can be specifically referred to the description of the aforementioned embodiment, which will not be repeated here.

[0078] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0079] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0080] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0081] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.

[0082] In addition, in the description of the present specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish descriptions, and cannot be understood as indicating or suggesting relative importance. It should also be understood that although the terms "first", "second", etc. are used to describe various elements in some embodiments of the present application in the text, these elements should not be limited by these terms. These terms are only used to distinguish one element from another element. For example, the first table can be named as the second table, and similarly, the second table can be named as the first table without departing from the scope of the various described embodiments. The first table and the second table are both tables, but they are not the same table.

[0083] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0084] The azalea species identification method provided in the embodiment of the present application can be applied to terminal devices such as mobile phones, tablet computers, wearable devices, vehicle-mounted devices, augmented reality (AR) / virtual reality (VR) devices, laptops, ultra-mobile personal computers (UMPC), netbooks, personal digital assistants (PDA), etc. The embodiment of the present application does not impose any restrictions on the specific type of the terminal device.

[0085] For example, the terminal device can be a station (STAION, ST) in a WLAN, a cellular phone, a cordless phone, a Session Initiation Protocol (SIP) phone, a Wireless Local Loop (WLL) station, a Personal Digital Assistant (PDA) device, a handheld device with wireless communication function, a computing device or other processing device connected to a wireless modem, a vehicle-mounted device, a vehicle networking terminal, a computer, a laptop computer, a handheld communication device, a handheld computing device, a satellite wireless device, a wireless modem card, a TV set-top box (settop box, STB), customer premises equipment (customer premises equipment, CPE) and / or other devices for communicating on a wireless system and a next-generation communication system, such as a mobile terminal in a 5G network or a mobile terminal in a future evolved Public Land Mobile Network (PLMN) network, etc.

[0086] As an example but not limitation, when the terminal device is a wearable device, the wearable device can also be a general term for wearable devices that are intelligently designed and developed using wearable technology for daily wear, such as glasses, gloves, watches, clothing and shoes. A wearable device is a portable device that is worn directly on the body or integrated into the user's clothes or accessories. Wearable devices are not just hardware devices, but also achieve powerful functions through software support, data interaction, and cloud interaction. Broadly speaking, wearable smart devices include those that are fully functional, large in size, and can achieve complete or partial functions without relying on smartphones, such as smart watches or smart glasses, as well as those that only focus on a certain type of application function and need to be used in conjunction with other devices such as smartphones, such as various smart bracelets and smart jewelry for vital sign monitoring.

[0087] Figure 6 Schematic diagram of the structure of a terminal device provided by an embodiment of the present application. Figure 6 As shown, the terminal device 7 of this embodiment includes: at least one processor 70 ( Figure 6 Only one is shown in the figure), a memory 71, wherein the memory 71 stores a computer program 72 that can be run on the processor 70. When the processor 70 executes the computer program 72, the steps in the above-mentioned azalea species identification method embodiments are implemented, such as Figure 1 Alternatively, when the processor 70 executes the computer program 72, the functions of each module / unit in the above-mentioned device embodiments are realized, for example Figure 5Functions of modules 510 to 530 are shown.

[0088] The terminal device 7 may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The terminal device may include, but is not limited to, a processor 70 and a memory 71. Those skilled in the art will appreciate that Figure 6 It is only an example of the terminal device 7 and does not constitute a limitation on the terminal device 7. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the terminal device may also include an input sending device, a network access device, a bus, etc.

[0089] The processor 70 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0090] In some embodiments, the memory 71 may be an internal storage unit of the terminal device 7, such as a hard disk or memory of the terminal device 7. The memory 71 may also be an external storage device of the terminal device 7, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the terminal device 7. Further, the memory 71 may also include both an internal storage unit and an external storage device of the terminal device 7. The memory 71 is used to store an operating system, an application program, a boot loader (BootLoader), data, and other programs, such as the program code of the computer program. The memory 71 may also be used to temporarily store data that has been sent or is to be sent.

[0091] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0092] An embodiment of the present application also provides a terminal device, which includes at least one memory, at least one processor, and a computer program stored in the at least one memory and executable on the at least one processor, wherein when the processor executes the computer program, the terminal device implements the steps in any of the above-mentioned method embodiments.

[0093] An embodiment of the present application further provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the above-mentioned method embodiments can be implemented.

[0094] An embodiment of the present application provides a computer program product. When the computer program product runs on a terminal device, the terminal device can implement the steps in the above-mentioned method embodiments when executing the computer program product.

[0095] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0096] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0097] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0098] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0099] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for identifying azalea species, characterized in that: include: Obtaining a low-resolution spectral image of the azalea to be identified; Extract spectral features from low-resolution spectral images to match several high-resolution spectral features with the highest similarity, and calculate the weighted votes of azalea species corresponding to several high-resolution spectral features; The species with the highest number of weighted votes is determined as the species of the azalea to be identified.

2. The method for identifying azalea species according to claim 1, characterized in that: The step of extracting spectral features from the low-resolution spectral image to match a plurality of high-resolution spectral features with the highest similarity, and calculating weighted votes of azalea species corresponding to the plurality of high-resolution spectral features, comprises: Extract spectral features from low-resolution spectral images to obtain spectral features of multiple parts of the azalea to be identified; Based on the multi-part spectral features, several high-resolution spectral features with the highest similarity are matched respectively, and weighted votes of azalea species corresponding to the several high-resolution spectral features are calculated respectively.

3. The method for identifying azalea species according to claim 1, characterized in that: The similarity is obtained by the following method: Obtain local high-resolution spectral images and global low-resolution spectral images of azalea under several different growth environment factors; According to the local high-resolution spectral image, the high-resolution spectral features for local azalea species identification are extracted; According to the global low-resolution spectral image, the low-resolution spectral features of each part of the azalea are extracted; Based on the spectral feature dimension, the spectral feature similarity between the low-resolution spectral features of each part of azalea and the high-resolution spectral features for local azalea species identification was calculated, and the obtained similarity was adjusted in combination with the growth environment factors.

4. The method for identifying azalea species according to claim 3, characterized in that: The step of determining the local high-resolution spectral features for identifying the local azalea species based on the local high-resolution spectral image of the azalea comprises: The spectral features of each local high-resolution spectral image are divided to obtain a plurality of initial local azalea species identification spectral features; Determine the position information of the local part of the azalea from the local high-resolution spectral image, extract the spectral features of the local high-resolution spectral image of the azalea based on the position information of the local part of the azalea, and obtain the spectral features representing the local azalea species; When the band standard deviation of the initial local azalea species identification spectral feature and the local azalea species characterization spectral feature is less than a preset band standard deviation threshold, the initial local azalea species identification spectral feature is used as the local azalea species identification high-resolution spectral feature.

5. The method for identifying azalea species according to claim 3, characterized in that: The step of extracting low-resolution spectral features of various parts of azalea according to the global low-resolution spectral image comprises: Splicing a plurality of local high-resolution spectral images of azalea, and aligning the spliced ​​images with the global low-resolution spectral image to determine the position information of each part of the azalea in the global low-resolution spectral image; According to the position information of each part of the azalea in the global low-resolution spectral image, the low-resolution spectral features of each part of the azalea in the global low-resolution spectral image are extracted.

6. The method for identifying azalea species according to claim 3, characterized in that: The growth environment factors include growth stage, light conditions, temperature conditions, moisture conditions, soil conditions, altitude and topography; the spectral characteristics include spectral reflectance statistical characteristics, spectral curve characteristics and principal component analysis characteristics.

7. The method for identifying azalea species according to claim 1, characterized in that: The high-resolution spectral features also carry information on growth environment factors corresponding to the azalea species; The azalea species identification method also includes: The type of azalea to be identified and the corresponding growth environment factor information are displayed in the low-resolution spectral image.

8. A device for identifying azalea species, characterized in that: include: A low-resolution spectral image acquisition module is used to acquire a low-resolution spectral image of the azalea to be identified; A high-resolution spectral feature matching module is used to extract spectral features from low-resolution spectral images to match several high-resolution spectral features with the highest similarity, and calculate the weighted votes of azalea species corresponding to several high-resolution spectral features; as well as The azalea species determination module is used to determine the species with the highest weighted votes as the species of the azalea to be identified.

9. A terminal device, characterized in that: The terminal device includes a memory and a processor. The memory stores a computer program that can be run on the processor. When the processor executes the computer program, the steps of the azalea species identification method according to any one of claims 1 to 7 are implemented.

10. A non-transitory computer-readable medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method for identifying azalea species according to any one of claims 1 to 7 are implemented.

Images